5 Changes between 0.9.8zf and 0.9.8zg [11 Jun 2015]
7 *) Malformed ECParameters causes infinite loop
9 When processing an ECParameters structure OpenSSL enters an infinite loop
10 if the curve specified is over a specially malformed binary polynomial
13 This can be used to perform denial of service against any
14 system which processes public keys, certificate requests or
15 certificates. This includes TLS clients and TLS servers with
16 client authentication enabled.
18 This issue was reported to OpenSSL by Joseph Barr-Pixton.
22 *) Exploitable out-of-bounds read in X509_cmp_time
24 X509_cmp_time does not properly check the length of the ASN1_TIME
25 string and can read a few bytes out of bounds. In addition,
26 X509_cmp_time accepts an arbitrary number of fractional seconds in the
29 An attacker can use this to craft malformed certificates and CRLs of
30 various sizes and potentially cause a segmentation fault, resulting in
31 a DoS on applications that verify certificates or CRLs. TLS clients
32 that verify CRLs are affected. TLS clients and servers with client
33 authentication enabled may be affected if they use custom verification
36 This issue was reported to OpenSSL by Robert Swiecki (Google), and
37 independently by Hanno Böck.
41 *) PKCS7 crash with missing EnvelopedContent
43 The PKCS#7 parsing code does not handle missing inner EncryptedContent
44 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
45 with missing content and trigger a NULL pointer dereference on parsing.
47 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
48 structures from untrusted sources are affected. OpenSSL clients and
49 servers are not affected.
51 This issue was reported to OpenSSL by Michal Zalewski (Google).
55 *) CMS verify infinite loop with unknown hash function
57 When verifying a signedData message the CMS code can enter an infinite loop
58 if presented with an unknown hash function OID. This can be used to perform
59 denial of service against any system which verifies signedData messages using
61 This issue was reported to OpenSSL by Johannes Bauer.
65 *) Race condition handling NewSessionTicket
67 If a NewSessionTicket is received by a multi-threaded client when attempting to
68 reuse a previous ticket then a race condition can occur potentially leading to
69 a double free of the ticket data.
73 Changes between 0.9.8ze and 0.9.8zf [19 Mar 2015]
75 *) Segmentation fault in ASN1_TYPE_cmp fix
77 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
78 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
79 certificate signature algorithm consistency this can be used to crash any
80 certificate verification operation and exploited in a DoS attack. Any
81 application which performs certificate verification is vulnerable including
82 OpenSSL clients and servers which enable client authentication.
86 *) ASN.1 structure reuse memory corruption fix
88 Reusing a structure in ASN.1 parsing may allow an attacker to cause
89 memory corruption via an invalid write. Such reuse is and has been
90 strongly discouraged and is believed to be rare.
92 Applications that parse structures containing CHOICE or ANY DEFINED BY
93 components may be affected. Certificate parsing (d2i_X509 and related
94 functions) are however not affected. OpenSSL clients and servers are
99 *) PKCS7 NULL pointer dereferences fix
101 The PKCS#7 parsing code does not handle missing outer ContentInfo
102 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
103 missing content and trigger a NULL pointer dereference on parsing.
105 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
106 otherwise parse PKCS#7 structures from untrusted sources are
107 affected. OpenSSL clients and servers are not affected.
109 This issue was reported to OpenSSL by Michal Zalewski (Google).
113 *) DoS via reachable assert in SSLv2 servers fix
115 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
116 servers that both support SSLv2 and enable export cipher suites by sending
117 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
119 This issue was discovered by Sean Burford (Google) and Emilia Käsper
120 (OpenSSL development team).
124 *) Use After Free following d2i_ECPrivatekey error fix
126 A malformed EC private key file consumed via the d2i_ECPrivateKey function
127 could cause a use after free condition. This, in turn, could cause a double
128 free in several private key parsing functions (such as d2i_PrivateKey
129 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
130 for applications that receive EC private keys from untrusted
131 sources. This scenario is considered rare.
133 This issue was discovered by the BoringSSL project and fixed in their
138 *) X509_to_X509_REQ NULL pointer deref fix
140 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
141 the certificate key is invalid. This function is rarely used in practice.
143 This issue was discovered by Brian Carpenter.
147 *) Removed the export and SSLv2 ciphers from the DEFAULT ciphers
150 Changes between 0.9.8zd and 0.9.8ze [15 Jan 2015]
152 *) Build fixes for the Windows and OpenVMS platforms
153 [Matt Caswell and Richard Levitte]
155 Changes between 0.9.8zc and 0.9.8zd [8 Jan 2015]
157 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
158 message can cause a segmentation fault in OpenSSL due to a NULL pointer
159 dereference. This could lead to a Denial Of Service attack. Thanks to
160 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
164 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
165 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
166 method would be set to NULL which could later result in a NULL pointer
167 dereference. Thanks to Frank Schmirler for reporting this issue.
171 *) Abort handshake if server key exchange message is omitted for ephemeral
174 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
175 reporting this issue.
179 *) Remove non-export ephemeral RSA code on client and server. This code
180 violated the TLS standard by allowing the use of temporary RSA keys in
181 non-export ciphersuites and could be used by a server to effectively
182 downgrade the RSA key length used to a value smaller than the server
183 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
184 INRIA or reporting this issue.
188 *) Fix various certificate fingerprint issues.
190 By using non-DER or invalid encodings outside the signed portion of a
191 certificate the fingerprint can be changed without breaking the signature.
192 Although no details of the signed portion of the certificate can be changed
193 this can cause problems with some applications: e.g. those using the
194 certificate fingerprint for blacklists.
196 1. Reject signatures with non zero unused bits.
198 If the BIT STRING containing the signature has non zero unused bits reject
199 the signature. All current signature algorithms require zero unused bits.
201 2. Check certificate algorithm consistency.
203 Check the AlgorithmIdentifier inside TBS matches the one in the
204 certificate signature. NB: this will result in signature failure
205 errors for some broken certificates.
207 Thanks to Konrad Kraszewski from Google for reporting this issue.
209 3. Check DSA/ECDSA signatures use DER.
211 Reencode DSA/ECDSA signatures and compare with the original received
212 signature. Return an error if there is a mismatch.
214 This will reject various cases including garbage after signature
215 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
216 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
217 (negative or with leading zeroes).
219 Further analysis was conducted and fixes were developed by Stephen Henson
220 of the OpenSSL core team.
225 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
226 results on some platforms, including x86_64. This bug occurs at random
227 with a very low probability, and is not known to be exploitable in any
228 way, though its exact impact is difficult to determine. Thanks to Pieter
229 Wuille (Blockstream) who reported this issue and also suggested an initial
230 fix. Further analysis was conducted by the OpenSSL development team and
231 Adam Langley of Google. The final fix was developed by Andy Polyakov of
232 the OpenSSL core team.
236 Changes between 0.9.8zb and 0.9.8zc [15 Oct 2014]
238 *) Session Ticket Memory Leak.
240 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
241 integrity of that ticket is first verified. In the event of a session
242 ticket integrity check failing, OpenSSL will fail to free memory
243 causing a memory leak. By sending a large number of invalid session
244 tickets an attacker could exploit this issue in a Denial Of Service
249 *) Build option no-ssl3 is incomplete.
251 When OpenSSL is configured with "no-ssl3" as a build option, servers
252 could accept and complete a SSL 3.0 handshake, and clients could be
253 configured to send them.
255 [Akamai and the OpenSSL team]
257 *) Add support for TLS_FALLBACK_SCSV.
258 Client applications doing fallback retries should call
259 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
261 [Adam Langley, Bodo Moeller]
263 *) Add additional DigestInfo checks.
265 Reencode DigestInto in DER and check against the original when
266 verifying RSA signature: this will reject any improperly encoded
267 DigestInfo structures.
269 Note: this is a precautionary measure and no attacks are currently known.
273 Changes between 0.9.8za and 0.9.8zb [6 Aug 2014]
275 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
276 to a denial of service attack. A malicious server can crash the client
277 with a null pointer dereference (read) by specifying an anonymous (EC)DH
278 ciphersuite and sending carefully crafted handshake messages.
280 Thanks to Felix Gröbert (Google) for discovering and researching this
285 *) By sending carefully crafted DTLS packets an attacker could cause openssl
286 to leak memory. This can be exploited through a Denial of Service attack.
287 Thanks to Adam Langley for discovering and researching this issue.
291 *) An attacker can force openssl to consume large amounts of memory whilst
292 processing DTLS handshake messages. This can be exploited through a
293 Denial of Service attack.
294 Thanks to Adam Langley for discovering and researching this issue.
298 *) An attacker can force an error condition which causes openssl to crash
299 whilst processing DTLS packets due to memory being freed twice. This
300 can be exploited through a Denial of Service attack.
301 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
306 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
307 X509_name_oneline, X509_name_print_ex et al. to leak some information
308 from the stack. Applications may be affected if they echo pretty printing
309 output to the attacker.
311 Thanks to Ivan Fratric (Google) for discovering this issue.
313 [Emilia Käsper, and Steve Henson]
315 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
316 for corner cases. (Certain input points at infinity could lead to
317 bogus results, with non-infinity inputs mapped to infinity too.)
320 Changes between 0.9.8y and 0.9.8za [5 Jun 2014]
322 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
323 handshake can force the use of weak keying material in OpenSSL
324 SSL/TLS clients and servers.
326 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
327 researching this issue. (CVE-2014-0224)
328 [KIKUCHI Masashi, Steve Henson]
330 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
331 OpenSSL DTLS client the code can be made to recurse eventually crashing
334 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
336 [Imre Rad, Steve Henson]
338 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
339 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
340 client or server. This is potentially exploitable to run arbitrary
341 code on a vulnerable client or server.
343 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
344 [Jüri Aedla, Steve Henson]
346 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
347 are subject to a denial of service attack.
349 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
350 this issue. (CVE-2014-3470)
351 [Felix Gröbert, Ivan Fratric, Steve Henson]
353 *) Fix for the attack described in the paper "Recovering OpenSSL
354 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
355 by Yuval Yarom and Naomi Benger. Details can be obtained from:
356 http://eprint.iacr.org/2014/140
358 Thanks to Yuval Yarom and Naomi Benger for discovering this
359 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
360 [Yuval Yarom and Naomi Benger]
362 Thanks to mancha for backporting the fix to the 0.9.8 branch.
364 *) Fix handling of warning-level alerts in SSL23 client mode so they
365 don't cause client-side termination (eg. on SNI unrecognized_name
366 warnings). Add client and server support for six additional alerts
367 per RFC 6066 and RFC 4279.
370 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
371 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
372 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
373 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
374 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
375 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
376 [Rob Stradling, Adam Langley]
378 Changes between 0.9.8x and 0.9.8y [5 Feb 2013]
380 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
382 This addresses the flaw in CBC record processing discovered by
383 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
384 at: http://www.isg.rhul.ac.uk/tls/
386 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
387 Security Group at Royal Holloway, University of London
388 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
389 Emilia Käsper for the initial patch.
391 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
393 *) Return an error when checking OCSP signatures when key is NULL.
394 This fixes a DoS attack. (CVE-2013-0166)
397 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
398 the right response is stapled. Also change SSL_get_certificate()
399 so it returns the certificate actually sent.
400 See http://rt.openssl.org/Ticket/Display.html?id=2836.
402 [Rob Stradling <rob.stradling@comodo.com>]
404 *) Fix possible deadlock when decoding public keys.
407 Changes between 0.9.8w and 0.9.8x [10 May 2012]
409 *) Sanity check record length before skipping explicit IV in DTLS
412 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
413 fuzzing as a service testing platform.
417 *) Initialise tkeylen properly when encrypting CMS messages.
418 Thanks to Solar Designer of Openwall for reporting this issue.
421 Changes between 0.9.8v and 0.9.8w [23 Apr 2012]
423 *) The fix for CVE-2012-2110 did not take into account that the
424 'len' argument to BUF_MEM_grow and BUF_MEM_grow_clean is an
425 int in OpenSSL 0.9.8, making it still vulnerable. Fix by
426 rejecting negative len parameter. (CVE-2012-2131)
427 [Tomas Hoger <thoger@redhat.com>]
429 Changes between 0.9.8u and 0.9.8v [19 Apr 2012]
431 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
432 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
433 in CRYPTO_realloc_clean.
435 Thanks to Tavis Ormandy, Google Security Team, for discovering this
436 issue and to Adam Langley <agl@chromium.org> for fixing it.
438 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
440 Changes between 0.9.8t and 0.9.8u [12 Mar 2012]
442 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
443 in CMS and PKCS7 code. When RSA decryption fails use a random key for
444 content decryption and always return the same error. Note: this attack
445 needs on average 2^20 messages so it only affects automated senders. The
446 old behaviour can be reenabled in the CMS code by setting the
447 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
448 an MMA defence is not necessary.
449 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
450 this issue. (CVE-2012-0884)
453 *) Fix CVE-2011-4619: make sure we really are receiving a
454 client hello before rejecting multiple SGC restarts. Thanks to
455 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
458 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
460 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
461 Thanks to Antonio Martin, Enterprise Secure Access Research and
462 Development, Cisco Systems, Inc. for discovering this bug and
463 preparing a fix. (CVE-2012-0050)
466 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
468 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
469 of the Vaudenay padding oracle attack on CBC mode encryption
470 which enables an efficient plaintext recovery attack against
471 the OpenSSL implementation of DTLS. Their attack exploits timing
472 differences arising during decryption processing. A research
473 paper describing this attack can be found at:
474 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
475 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
476 Security Group at Royal Holloway, University of London
477 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
478 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
479 for preparing the fix. (CVE-2011-4108)
480 [Robin Seggelmann, Michael Tuexen]
482 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
483 [Ben Laurie, Kasper <ekasper@google.com>]
485 *) Clear bytes used for block padding of SSL 3.0 records.
487 [Adam Langley (Google)]
489 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
490 Kadianakis <desnacked@gmail.com> for discovering this issue and
491 Adam Langley for preparing the fix. (CVE-2011-4619)
492 [Adam Langley (Google)]
494 *) Prevent malformed RFC3779 data triggering an assertion failure.
495 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
496 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
497 [Rob Austein <sra@hactrn.net>]
499 *) Fix ssl_ciph.c set-up race.
500 [Adam Langley (Google)]
502 *) Fix spurious failures in ecdsatest.c.
503 [Emilia Käsper (Google)]
505 *) Fix the BIO_f_buffer() implementation (which was mixing different
506 interpretations of the '..._len' fields).
507 [Adam Langley (Google)]
509 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
510 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
511 threads won't reuse the same blinding coefficients.
513 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
514 lock to call BN_BLINDING_invert_ex, and avoids one use of
515 BN_BLINDING_update for each BN_BLINDING structure (previously,
516 the last update always remained unused).
517 [Emilia Käsper (Google)]
519 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
520 for multi-threaded use of ECDH.
521 [Adam Langley (Google)]
523 *) Fix x509_name_ex_d2i memory leak on bad inputs.
526 *) Add protection against ECDSA timing attacks as mentioned in the paper
527 by Billy Bob Brumley and Nicola Tuveri, see:
529 http://eprint.iacr.org/2011/232.pdf
531 [Billy Bob Brumley and Nicola Tuveri]
533 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
535 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
536 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
538 *) Fix bug in string printing code: if *any* escaping is enabled we must
539 escape the escape character (backslash) or the resulting string is
543 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
545 *) Disable code workaround for ancient and obsolete Netscape browsers
546 and servers: an attacker can use it in a ciphersuite downgrade attack.
547 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
550 *) Fixed J-PAKE implementation error, originally discovered by
551 Sebastien Martini, further info and confirmation from Stefan
552 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
555 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
557 *) Fix extension code to avoid race conditions which can result in a buffer
558 overrun vulnerability: resumed sessions must not be modified as they can
559 be shared by multiple threads. CVE-2010-3864
562 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
565 *) Don't reencode certificate when calculating signature: cache and use
566 the original encoding instead. This makes signature verification of
567 some broken encodings work correctly.
570 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
571 is also one of the inputs.
572 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
574 *) Don't repeatedly append PBE algorithms to table if they already exist.
575 Sort table on each new add. This effectively makes the table read only
576 after all algorithms are added and subsequent calls to PKCS12_pbe_add
580 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
582 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
585 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
586 access or freeing data twice (CVE-2010-0742)
587 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
589 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
590 common in certificates and some applications which only call
591 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
595 Reduce copying into .apps and .test in makevms.com
596 Don't try to use blank CA certificate in CA.com
597 Allow use of C files from original directories in maketests.com
598 [Steven M. Schweda" <sms@antinode.info>]
600 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
602 *) When rejecting SSL/TLS records due to an incorrect version number, never
603 update s->server with a new major version number. As of
604 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
605 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
606 the previous behavior could result in a read attempt at NULL when
607 receiving specific incorrect SSL/TLS records once record payload
608 protection is active. (CVE-2010-0740)
609 [Bodo Moeller, Adam Langley <agl@chromium.org>]
611 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
612 could be crashed if the relevant tables were not present (e.g. chrooted).
613 [Tomas Hoger <thoger@redhat.com>]
615 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
617 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
618 [Martin Olsson, Neel Mehta]
620 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
621 accommodate for stack sorting, always a write lock!).
624 *) On some versions of WIN32 Heap32Next is very slow. This can cause
625 excessive delays in the RAND_poll(): over a minute. As a workaround
626 include a time check in the inner Heap32Next loop too.
629 *) The code that handled flushing of data in SSL/TLS originally used the
630 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
631 the problem outlined in PR#1949. The fix suggested there however can
632 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
633 of Apache). So instead simplify the code to flush unconditionally.
634 This should be fine since flushing with no data to flush is a no op.
637 *) Handle TLS versions 2.0 and later properly and correctly use the
638 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
639 off ancient servers have a habit of sticking around for a while...
642 *) Modify compression code so it frees up structures without using the
643 ex_data callbacks. This works around a problem where some applications
644 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
645 restarting) then use compression (e.g. SSL with compression) later.
646 This results in significant per-connection memory leaks and
647 has caused some security issues including CVE-2008-1678 and
651 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
652 change when encrypting or decrypting.
655 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
656 connect and renegotiate with servers which do not support RI.
657 Until RI is more widely deployed this option is enabled by default.
660 *) Add "missing" ssl ctrls to clear options and mode.
663 *) If client attempts to renegotiate and doesn't support RI respond with
664 a no_renegotiation alert as required by RFC5746. Some renegotiating
665 TLS clients will continue a connection gracefully when they receive
666 the alert. Unfortunately OpenSSL mishandled this alert and would hang
667 waiting for a server hello which it will never receive. Now we treat a
668 received no_renegotiation alert as a fatal error. This is because
669 applications requesting a renegotiation might well expect it to succeed
670 and would have no code in place to handle the server denying it so the
671 only safe thing to do is to terminate the connection.
674 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
675 peer supports secure renegotiation and 0 otherwise. Print out peer
676 renegotiation support in s_client/s_server.
679 *) Replace the highly broken and deprecated SPKAC certification method with
680 the updated NID creation version. This should correctly handle UTF8.
683 *) Implement RFC5746. Re-enable renegotiation but require the extension
684 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
685 turns out to be a bad idea. It has been replaced by
686 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
687 SSL_CTX_set_options(). This is really not recommended unless you
688 know what you are doing.
689 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
691 *) Fixes to stateless session resumption handling. Use initial_ctx when
692 issuing and attempting to decrypt tickets in case it has changed during
693 servername handling. Use a non-zero length session ID when attempting
694 stateless session resumption: this makes it possible to determine if
695 a resumption has occurred immediately after receiving server hello
696 (several places in OpenSSL subtly assume this) instead of later in
700 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
701 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
702 fixes for a few places where the return code is not checked
704 [Julia Lawall <julia@diku.dk>]
706 *) Add --strict-warnings option to Configure script to include devteam
707 warnings in other configurations.
710 *) Add support for --libdir option and LIBDIR variable in makefiles. This
711 makes it possible to install openssl libraries in locations which
712 have names other than "lib", for example "/usr/lib64" which some
714 [Steve Henson, based on patch from Jeremy Utley]
716 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
717 X690 8.9.12 and can produce some misleading textual output of OIDs.
718 [Steve Henson, reported by Dan Kaminsky]
720 *) Delete MD2 from algorithm tables. This follows the recommendation in
721 several standards that it is not used in new applications due to
722 several cryptographic weaknesses. For binary compatibility reasons
723 the MD2 API is still compiled in by default.
726 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
730 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
731 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
733 [Guenter <lists@gknw.net>]
735 *) Fix the server certificate chain building code to use X509_verify_cert(),
736 it used to have an ad-hoc builder which was unable to cope with anything
737 other than a simple chain.
738 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
740 *) Don't check self signed certificate signatures in X509_verify_cert()
741 by default (a flag can override this): it just wastes time without
742 adding any security. As a useful side effect self signed root CAs
743 with non-FIPS digests are now usable in FIPS mode.
746 *) In dtls1_process_out_of_seq_message() the check if the current message
747 is already buffered was missing. For every new message was memory
748 allocated, allowing an attacker to perform an denial of service attack
749 with sending out of seq handshake messages until there is no memory
750 left. Additionally every future messege was buffered, even if the
751 sequence number made no sense and would be part of another handshake.
752 So only messages with sequence numbers less than 10 in advance will be
753 buffered. (CVE-2009-1378)
754 [Robin Seggelmann, discovered by Daniel Mentz]
756 *) Records are buffered if they arrive with a future epoch to be
757 processed after finishing the corresponding handshake. There is
758 currently no limitation to this buffer allowing an attacker to perform
759 a DOS attack with sending records with future epochs until there is no
760 memory left. This patch adds the pqueue_size() function to detemine
761 the size of a buffer and limits the record buffer to 100 entries.
763 [Robin Seggelmann, discovered by Daniel Mentz]
765 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
766 parent structure is freed. (CVE-2009-1379)
769 *) Handle non-blocking I/O properly in SSL_shutdown() call.
770 [Darryl Miles <darryl-mailinglists@netbauds.net>]
773 [Ilya O. <vrghost@gmail.com>]
775 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
777 *) Disable renegotiation completely - this fixes a severe security
778 problem (CVE-2009-3555) at the cost of breaking all
779 renegotiation. Renegotiation can be re-enabled by setting
780 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
781 run-time. This is really not recommended unless you know what
785 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
787 *) Don't set val to NULL when freeing up structures, it is freed up by
788 underlying code. If sizeof(void *) > sizeof(long) this can result in
789 zeroing past the valid field. (CVE-2009-0789)
790 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
792 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
793 checked correctly. This would allow some invalid signed attributes to
794 appear to verify correctly. (CVE-2009-0591)
795 [Ivan Nestlerode <inestlerode@us.ibm.com>]
797 *) Reject UniversalString and BMPString types with invalid lengths. This
798 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
799 a legal length. (CVE-2009-0590)
802 *) Set S/MIME signing as the default purpose rather than setting it
803 unconditionally. This allows applications to override it at the store
807 *) Permit restricted recursion of ASN1 strings. This is needed in practice
808 to handle some structures.
811 *) Improve efficiency of mem_gets: don't search whole buffer each time
813 [Jeremy Shapiro <jnshapir@us.ibm.com>]
815 *) New -hex option for openssl rand.
818 *) Print out UTF8String and NumericString when parsing ASN1.
821 *) Support NumericString type for name components.
824 *) Allow CC in the environment to override the automatically chosen
825 compiler. Note that nothing is done to ensure flags work with the
829 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
831 *) Properly check EVP_VerifyFinal() and similar return values
833 [Ben Laurie, Bodo Moeller, Google Security Team]
835 *) Enable TLS extensions by default.
838 *) Allow the CHIL engine to be loaded, whether the application is
839 multithreaded or not. (This does not release the developer from the
840 obligation to set up the dynamic locking callbacks.)
841 [Sander Temme <sander@temme.net>]
843 *) Use correct exit code if there is an error in dgst command.
844 [Steve Henson; problem pointed out by Roland Dirlewanger]
846 *) Tweak Configure so that you need to say "experimental-jpake" to enable
847 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
850 *) Add experimental JPAKE support, including demo authentication in
851 s_client and s_server.
854 *) Set the comparison function in v3_addr_canonize().
855 [Rob Austein <sra@hactrn.net>]
857 *) Add support for XMPP STARTTLS in s_client.
858 [Philip Paeps <philip@freebsd.org>]
860 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
861 to ensure that even with this option, only ciphersuites in the
862 server's preference list will be accepted. (Note that the option
863 applies only when resuming a session, so the earlier behavior was
864 just about the algorithm choice for symmetric cryptography.)
867 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
869 *) Fix NULL pointer dereference if a DTLS server received
870 ChangeCipherSpec as first record (CVE-2009-1386).
873 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
874 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
877 *) The fix in 0.9.8c that supposedly got rid of unsafe
878 double-checked locking was incomplete for RSA blinding,
879 addressing just one layer of what turns out to have been
880 doubly unsafe triple-checked locking.
882 So now fix this for real by retiring the MONT_HELPER macro
883 in crypto/rsa/rsa_eay.c.
885 [Bodo Moeller; problem pointed out by Marius Schilder]
887 *) Various precautionary measures:
889 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
891 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
892 (NB: This would require knowledge of the secret session ticket key
893 to exploit, in which case you'd be SOL either way.)
895 - Change bn_nist.c so that it will properly handle input BIGNUMs
896 outside the expected range.
898 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
901 [Neel Mehta, Bodo Moeller]
903 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
904 the load fails. Useful for distros.
905 [Ben Laurie and the FreeBSD team]
907 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
910 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
913 *) Expand ENGINE to support engine supplied SSL client certificate functions.
915 This work was sponsored by Logica.
918 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
919 keystores. Support for SSL/TLS client authentication too.
920 Not compiled unless enable-capieng specified to Configure.
922 This work was sponsored by Logica.
925 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
926 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
927 attribute creation routines such as certifcate requests and PKCS#12
931 Changes between 0.9.8g and 0.9.8h [28 May 2008]
933 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
934 handshake which could lead to a cilent crash as found using the
935 Codenomicon TLS test suite (CVE-2008-1672)
936 [Steve Henson, Mark Cox]
938 *) Fix double free in TLS server name extensions which could lead to
939 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
942 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
944 Clear the error queue to ensure that error entries left from
945 older function calls do not interfere with the correct operation.
946 [Lutz Jaenicke, Erik de Castro Lopo]
948 *) Remove root CA certificates of commercial CAs:
950 The OpenSSL project does not recommend any specific CA and does not
951 have any policy with respect to including or excluding any CA.
952 Therefore it does not make any sense to ship an arbitrary selection
953 of root CA certificates with the OpenSSL software.
956 *) RSA OAEP patches to fix two separate invalid memory reads.
957 The first one involves inputs when 'lzero' is greater than
958 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
959 before the beginning of from). The second one involves inputs where
960 the 'db' section contains nothing but zeroes (there is a one-byte
961 invalid read after the end of 'db').
962 [Ivan Nestlerode <inestlerode@us.ibm.com>]
964 *) Partial backport from 0.9.9-dev:
966 Introduce bn_mul_mont (dedicated Montgomery multiplication
967 procedure) as a candidate for BIGNUM assembler implementation.
968 While 0.9.9-dev uses assembler for various architectures, only
969 x86_64 is available by default here in the 0.9.8 branch, and
970 32-bit x86 is available through a compile-time setting.
972 To try the 32-bit x86 assembler implementation, use Configure
973 option "enable-montasm" (which exists only for this backport).
975 As "enable-montasm" for 32-bit x86 disclaims code stability
976 anyway, in this constellation we activate additional code
977 backported from 0.9.9-dev for further performance improvements,
978 namely BN_from_montgomery_word. (To enable this otherwise,
979 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
981 [Andy Polyakov (backport partially by Bodo Moeller)]
983 *) Add TLS session ticket callback. This allows an application to set
984 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
985 values. This is useful for key rollover for example where several key
986 sets may exist with different names.
989 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
990 This was broken until now in 0.9.8 releases, such that the only way
991 a registered ENGINE could be used (assuming it initialises
992 successfully on the host) was to explicitly set it as the default
993 for the relevant algorithms. This is in contradiction with 0.9.7
994 behaviour and the documentation. With this fix, when an ENGINE is
995 registered into a given algorithm's table of implementations, the
996 'uptodate' flag is reset so that auto-discovery will be used next
997 time a new context for that algorithm attempts to select an
999 [Ian Lister (tweaked by Geoff Thorpe)]
1001 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
1002 implemention in the following ways:
1004 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
1007 Lack of BER streaming support means one pass streaming processing is
1008 only supported if data is detached: setting the streaming flag is
1009 ignored for embedded content.
1011 CMS support is disabled by default and must be explicitly enabled
1012 with the enable-cms configuration option.
1015 *) Update the GMP engine glue to do direct copies between BIGNUM and
1016 mpz_t when openssl and GMP use the same limb size. Otherwise the
1017 existing "conversion via a text string export" trick is still used.
1018 [Paul Sheer <paulsheer@gmail.com>]
1020 *) Zlib compression BIO. This is a filter BIO which compressed and
1021 uncompresses any data passed through it.
1024 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
1025 RFC3394 compatible AES key wrapping.
1028 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
1029 sets string data without copying. X509_ALGOR_set0() and
1030 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
1031 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
1032 from an X509_ATTRIBUTE structure optionally checking it occurs only
1033 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
1037 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
1038 to get the expected BN_FLG_CONSTTIME behavior.
1039 [Bodo Moeller (Google)]
1043 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
1044 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
1045 - added some more tests to do_tests.pl
1046 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
1047 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
1048 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
1049 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
1050 - various changes to netware.pl to enable gcc-cross builds on Win32
1052 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
1053 - various changes to fix missing prototype warnings
1054 - fixed x86nasm.pl to create correct asm files for NASM COFF output
1055 - added AES, WHIRLPOOL and CPUID assembler code to build files
1056 - added missing AES assembler make rules to mk1mf.pl
1057 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
1058 [Guenter Knauf <eflash@gmx.net>]
1060 *) Implement certificate status request TLS extension defined in RFC3546.
1061 A client can set the appropriate parameters and receive the encoded
1062 OCSP response via a callback. A server can query the supplied parameters
1063 and set the encoded OCSP response in the callback. Add simplified examples
1064 to s_client and s_server.
1067 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
1069 *) Fix various bugs:
1070 + Binary incompatibility of ssl_ctx_st structure
1071 + DTLS interoperation with non-compliant servers
1072 + Don't call get_session_cb() without proposed session
1073 + Fix ia64 assembler code
1074 [Andy Polyakov, Steve Henson]
1076 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
1078 *) DTLS Handshake overhaul. There were longstanding issues with
1079 OpenSSL DTLS implementation, which were making it impossible for
1080 RFC 4347 compliant client to communicate with OpenSSL server.
1081 Unfortunately just fixing these incompatibilities would "cut off"
1082 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
1083 server keeps tolerating non RFC compliant syntax. The opposite is
1084 not true, 0.9.8f client can not communicate with earlier server.
1085 This update even addresses CVE-2007-4995.
1088 *) Changes to avoid need for function casts in OpenSSL: some compilers
1089 (gcc 4.2 and later) reject their use.
1090 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
1093 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1094 RFC4507bis. The encrypted ticket format is an encrypted encoded
1095 SSL_SESSION structure, that way new session features are automatically
1098 If a client application caches session in an SSL_SESSION structure
1099 support is transparent because tickets are now stored in the encoded
1102 The SSL_CTX structure automatically generates keys for ticket
1103 protection in servers so again support should be possible
1104 with no application modification.
1106 If a client or server wishes to disable RFC4507 support then the option
1107 SSL_OP_NO_TICKET can be set.
1109 Add a TLS extension debugging callback to allow the contents of any client
1110 or server extensions to be examined.
1112 This work was sponsored by Google.
1115 *) Add initial support for TLS extensions, specifically for the server_name
1116 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1117 have new members for a host name. The SSL data structure has an
1118 additional member SSL_CTX *initial_ctx so that new sessions can be
1119 stored in that context to allow for session resumption, even after the
1120 SSL has been switched to a new SSL_CTX in reaction to a client's
1121 server_name extension.
1123 New functions (subject to change):
1125 SSL_get_servername()
1126 SSL_get_servername_type()
1129 New CTRL codes and macros (subject to change):
1131 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1132 - SSL_CTX_set_tlsext_servername_callback()
1133 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1134 - SSL_CTX_set_tlsext_servername_arg()
1135 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1137 openssl s_client has a new '-servername ...' option.
1139 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1140 '-key2 ...', '-servername_fatal' (subject to change). This allows
1141 testing the HostName extension for a specific single host name ('-cert'
1142 and '-key' remain fallbacks for handshakes without HostName
1143 negotiation). If the unrecogninzed_name alert has to be sent, this by
1144 default is a warning; it becomes fatal with the '-servername_fatal'
1147 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
1149 *) Add AES and SSE2 assembly language support to VC++ build.
1152 *) Mitigate attack on final subtraction in Montgomery reduction.
1155 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
1156 (which previously caused an internal error).
1159 *) Squeeze another 10% out of IGE mode when in != out.
1162 *) AES IGE mode speedup.
1163 [Dean Gaudet (Google)]
1165 *) Add the Korean symmetric 128-bit cipher SEED (see
1166 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
1167 add SEED ciphersuites from RFC 4162:
1169 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
1170 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
1171 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
1172 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
1174 To minimize changes between patchlevels in the OpenSSL 0.9.8
1175 series, SEED remains excluded from compilation unless OpenSSL
1176 is configured with 'enable-seed'.
1177 [KISA, Bodo Moeller]
1179 *) Mitigate branch prediction attacks, which can be practical if a
1180 single processor is shared, allowing a spy process to extract
1181 information. For detailed background information, see
1182 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
1183 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
1184 and Necessary Software Countermeasures"). The core of the change
1185 are new versions BN_div_no_branch() and
1186 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
1187 respectively, which are slower, but avoid the security-relevant
1188 conditional branches. These are automatically called by BN_div()
1189 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
1190 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
1191 remove a conditional branch.
1193 BN_FLG_CONSTTIME is the new name for the previous
1194 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
1195 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
1196 in the exponent causes BN_mod_exp_mont() to use the alternative
1197 implementation in BN_mod_exp_mont_consttime().) The old name
1198 remains as a deprecated alias.
1200 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
1201 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
1202 constant-time implementations for more than just exponentiation.
1203 Here too the old name is kept as a deprecated alias.
1205 BN_BLINDING_new() will now use BN_dup() for the modulus so that
1206 the BN_BLINDING structure gets an independent copy of the
1207 modulus. This means that the previous "BIGNUM *m" argument to
1208 BN_BLINDING_new() and to BN_BLINDING_create_param() now
1209 essentially becomes "const BIGNUM *m", although we can't actually
1210 change this in the header file before 0.9.9. It allows
1211 RSA_setup_blinding() to use BN_with_flags() on the modulus to
1212 enable BN_FLG_CONSTTIME.
1214 [Matthew D Wood (Intel Corp)]
1216 *) In the SSL/TLS server implementation, be strict about session ID
1217 context matching (which matters if an application uses a single
1218 external cache for different purposes). Previously,
1219 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
1220 set. This did ensure strict client verification, but meant that,
1221 with applications using a single external cache for quite
1222 different requirements, clients could circumvent ciphersuite
1223 restrictions for a given session ID context by starting a session
1224 in a different context.
1227 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
1228 a ciphersuite string such as "DEFAULT:RSA" cannot enable
1229 authentication-only ciphersuites.
1232 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
1233 not complete and could lead to a possible single byte overflow
1234 (CVE-2007-5135) [Ben Laurie]
1236 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
1238 *) Since AES128 and AES256 (and similarly Camellia128 and
1239 Camellia256) share a single mask bit in the logic of
1240 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
1241 kludge to work properly if AES128 is available and AES256 isn't
1242 (or if Camellia128 is available and Camellia256 isn't).
1245 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
1246 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
1247 When a point or a seed is encoded in a BIT STRING, we need to
1248 prevent the removal of trailing zero bits to get the proper DER
1249 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
1250 of a NamedBitList, for which trailing 0 bits need to be removed.)
1253 *) Have SSL/TLS server implementation tolerate "mismatched" record
1254 protocol version while receiving ClientHello even if the
1255 ClientHello is fragmented. (The server can't insist on the
1256 particular protocol version it has chosen before the ServerHello
1257 message has informed the client about his choice.)
1260 *) Add RFC 3779 support.
1261 [Rob Austein for ARIN, Ben Laurie]
1263 *) Load error codes if they are not already present instead of using a
1264 static variable. This allows them to be cleanly unloaded and reloaded.
1265 Improve header file function name parsing.
1268 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
1269 or CAPABILITY handshake as required by RFCs.
1272 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
1274 *) Introduce limits to prevent malicious keys being able to
1275 cause a denial of service. (CVE-2006-2940)
1276 [Steve Henson, Bodo Moeller]
1278 *) Fix ASN.1 parsing of certain invalid structures that can result
1279 in a denial of service. (CVE-2006-2937) [Steve Henson]
1281 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
1282 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
1284 *) Fix SSL client code which could crash if connecting to a
1285 malicious SSLv2 server. (CVE-2006-4343)
1286 [Tavis Ormandy and Will Drewry, Google Security Team]
1288 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
1289 match only those. Before that, "AES256-SHA" would be interpreted
1290 as a pattern and match "AES128-SHA" too (since AES128-SHA got
1291 the same strength classification in 0.9.7h) as we currently only
1292 have a single AES bit in the ciphersuite description bitmap.
1293 That change, however, also applied to ciphersuite strings such as
1294 "RC4-MD5" that intentionally matched multiple ciphersuites --
1295 namely, SSL 2.0 ciphersuites in addition to the more common ones
1296 from SSL 3.0/TLS 1.0.
1298 So we change the selection algorithm again: Naming an explicit
1299 ciphersuite selects this one ciphersuite, and any other similar
1300 ciphersuite (same bitmap) from *other* protocol versions.
1301 Thus, "RC4-MD5" again will properly select both the SSL 2.0
1302 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
1304 Since SSL 2.0 does not have any ciphersuites for which the
1305 128/256 bit distinction would be relevant, this works for now.
1306 The proper fix will be to use different bits for AES128 and
1307 AES256, which would have avoided the problems from the beginning;
1308 however, bits are scarce, so we can only do this in a new release
1309 (not just a patchlevel) when we can change the SSL_CIPHER
1310 definition to split the single 'unsigned long mask' bitmap into
1311 multiple values to extend the available space.
1315 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
1317 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
1318 (CVE-2006-4339) [Ben Laurie and Google Security Team]
1320 *) Add AES IGE and biIGE modes.
1323 *) Change the Unix randomness entropy gathering to use poll() when
1324 possible instead of select(), since the latter has some
1325 undesirable limitations.
1326 [Darryl Miles via Richard Levitte and Bodo Moeller]
1328 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
1329 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
1330 cannot be implicitly activated as part of, e.g., the "AES" alias.
1331 However, please upgrade to OpenSSL 0.9.9[-dev] for
1332 non-experimental use of the ECC ciphersuites to get TLS extension
1333 support, which is required for curve and point format negotiation
1334 to avoid potential handshake problems.
1337 *) Disable rogue ciphersuites:
1339 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
1340 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
1341 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
1343 The latter two were purportedly from
1344 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
1347 Also deactivate the remaining ciphersuites from
1348 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
1349 unofficial, and the ID has long expired.
1352 *) Fix RSA blinding Heisenbug (problems sometimes occured on
1353 dual-core machines) and other potential thread-safety issues.
1356 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
1357 versions), which is now available for royalty-free use
1358 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
1359 Also, add Camellia TLS ciphersuites from RFC 4132.
1361 To minimize changes between patchlevels in the OpenSSL 0.9.8
1362 series, Camellia remains excluded from compilation unless OpenSSL
1363 is configured with 'enable-camellia'.
1366 *) Disable the padding bug check when compression is in use. The padding
1367 bug check assumes the first packet is of even length, this is not
1368 necessarily true if compresssion is enabled and can result in false
1369 positives causing handshake failure. The actual bug test is ancient
1370 code so it is hoped that implementations will either have fixed it by
1371 now or any which still have the bug do not support compression.
1374 Changes between 0.9.8a and 0.9.8b [04 May 2006]
1376 *) When applying a cipher rule check to see if string match is an explicit
1377 cipher suite and only match that one cipher suite if it is.
1380 *) Link in manifests for VC++ if needed.
1381 [Austin Ziegler <halostatue@gmail.com>]
1383 *) Update support for ECC-based TLS ciphersuites according to
1384 draft-ietf-tls-ecc-12.txt with proposed changes (but without
1385 TLS extensions, which are supported starting with the 0.9.9
1386 branch, not in the OpenSSL 0.9.8 branch).
1389 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
1390 opaque EVP_CIPHER_CTX handling.
1393 *) Fixes and enhancements to zlib compression code. We now only use
1394 "zlib1.dll" and use the default __cdecl calling convention on Win32
1395 to conform with the standards mentioned here:
1396 http://www.zlib.net/DLL_FAQ.txt
1397 Static zlib linking now works on Windows and the new --with-zlib-include
1398 --with-zlib-lib options to Configure can be used to supply the location
1399 of the headers and library. Gracefully handle case where zlib library
1403 *) Several fixes and enhancements to the OID generation code. The old code
1404 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
1405 handle numbers larger than ULONG_MAX, truncated printing and had a
1406 non standard OBJ_obj2txt() behaviour.
1409 *) Add support for building of engines under engine/ as shared libraries
1410 under VC++ build system.
1413 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
1414 Hopefully, we will not see any false combination of paths any more.
1417 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
1419 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
1420 (part of SSL_OP_ALL). This option used to disable the
1421 countermeasure against man-in-the-middle protocol-version
1422 rollback in the SSL 2.0 server implementation, which is a bad
1423 idea. (CVE-2005-2969)
1425 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
1426 for Information Security, National Institute of Advanced Industrial
1427 Science and Technology [AIST], Japan)]
1429 *) Add two function to clear and return the verify parameter flags.
1432 *) Keep cipherlists sorted in the source instead of sorting them at
1433 runtime, thus removing the need for a lock.
1436 *) Avoid some small subgroup attacks in Diffie-Hellman.
1437 [Nick Mathewson and Ben Laurie]
1439 *) Add functions for well-known primes.
1442 *) Extended Windows CE support.
1443 [Satoshi Nakamura and Andy Polyakov]
1445 *) Initialize SSL_METHOD structures at compile time instead of during
1446 runtime, thus removing the need for a lock.
1449 *) Make PKCS7_decrypt() work even if no certificate is supplied by
1450 attempting to decrypt each encrypted key in turn. Add support to
1454 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
1456 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
1459 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
1462 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
1463 key into the same file any more.
1466 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
1469 *) Add -utf8 command line and config file option to 'ca'.
1470 [Stefan <stf@udoma.org]
1472 *) Removed the macro des_crypt(), as it seems to conflict with some
1473 libraries. Use DES_crypt().
1476 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
1477 involves renaming the source and generated shared-libs for
1478 both. The engines will accept the corrected or legacy ids
1479 ('ncipher' and '4758_cca' respectively) when binding. NB,
1480 this only applies when building 'shared'.
1481 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
1483 *) Add attribute functions to EVP_PKEY structure. Modify
1484 PKCS12_create() to recognize a CSP name attribute and
1485 use it. Make -CSP option work again in pkcs12 utility.
1488 *) Add new functionality to the bn blinding code:
1489 - automatic re-creation of the BN_BLINDING parameters after
1490 a fixed number of uses (currently 32)
1491 - add new function for parameter creation
1492 - introduce flags to control the update behaviour of the
1493 BN_BLINDING parameters
1494 - hide BN_BLINDING structure
1495 Add a second BN_BLINDING slot to the RSA structure to improve
1496 performance when a single RSA object is shared among several
1500 *) Add support for DTLS.
1501 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
1503 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
1504 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
1507 *) Remove buggy and incompletet DH cert support from
1508 ssl/ssl_rsa.c and ssl/s3_both.c
1511 *) Use SHA-1 instead of MD5 as the default digest algorithm for
1512 the apps/openssl applications.
1515 *) Compile clean with "-Wall -Wmissing-prototypes
1516 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
1517 DEBUG_SAFESTACK must also be set.
1520 *) Change ./Configure so that certain algorithms can be disabled by default.
1521 The new counterpiece to "no-xxx" is "enable-xxx".
1523 The patented RC5 and MDC2 algorithms will now be disabled unless
1524 "enable-rc5" and "enable-mdc2", respectively, are specified.
1526 (IDEA remains enabled despite being patented. This is because IDEA
1527 is frequently required for interoperability, and there is no license
1528 fee for non-commercial use. As before, "no-idea" can be used to
1529 avoid this algorithm.)
1533 *) Add processing of proxy certificates (see RFC 3820). This work was
1534 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
1535 EGEE (Enabling Grids for E-science in Europe).
1538 *) RC4 performance overhaul on modern architectures/implementations, such
1539 as Intel P4, IA-64 and AMD64.
1542 *) New utility extract-section.pl. This can be used specify an alternative
1543 section number in a pod file instead of having to treat each file as
1544 a separate case in Makefile. This can be done by adding two lines to the
1547 =for comment openssl_section:XXX
1549 The blank line is mandatory.
1553 *) New arguments -certform, -keyform and -pass for s_client and s_server
1554 to allow alternative format key and certificate files and passphrase
1558 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
1559 update associated structures and add various utility functions.
1561 Add new policy related verify parameters, include policy checking in
1562 standard verify code. Enhance 'smime' application with extra parameters
1563 to support policy checking and print out.
1566 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
1567 Nehemiah processors. These extensions support AES encryption in hardware
1568 as well as RNG (though RNG support is currently disabled).
1569 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
1571 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
1574 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
1575 [Andy Polyakov and a number of other people]
1577 *) Improved PowerPC platform support. Most notably BIGNUM assembler
1578 implementation contributed by IBM.
1579 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
1581 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
1582 exponent rather than 'unsigned long'. There is a corresponding change to
1583 the new 'rsa_keygen' element of the RSA_METHOD structure.
1584 [Jelte Jansen, Geoff Thorpe]
1586 *) Functionality for creating the initial serial number file is now
1587 moved from CA.pl to the 'ca' utility with a new option -create_serial.
1589 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
1590 number file to 1, which is bound to cause problems. To avoid
1591 the problems while respecting compatibility between different 0.9.7
1592 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
1593 CA.pl for serial number initialization. With the new release 0.9.8,
1594 we can fix the problem directly in the 'ca' utility.)
1597 *) Reduced header interdepencies by declaring more opaque objects in
1598 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
1599 give fewer recursive includes, which could break lazy source code - so
1600 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
1601 developers should define this symbol when building and using openssl to
1602 ensure they track the recommended behaviour, interfaces, [etc], but
1603 backwards-compatible behaviour prevails when this isn't defined.
1606 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
1609 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
1610 This will generate a random key of the appropriate length based on the
1611 cipher context. The EVP_CIPHER can provide its own random key generation
1612 routine to support keys of a specific form. This is used in the des and
1613 3des routines to generate a key of the correct parity. Update S/MIME
1614 code to use new functions and hence generate correct parity DES keys.
1615 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
1616 valid (weak or incorrect parity).
1619 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
1620 as looking them up. This is useful when the verified structure may contain
1621 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
1622 present unless the new PKCS7_NO_CRL flag is asserted.
1625 *) Extend ASN1 oid configuration module. It now additionally accepts the
1628 shortName = some long name, 1.2.3.4
1631 *) Reimplemented the BN_CTX implementation. There is now no more static
1632 limitation on the number of variables it can handle nor the depth of the
1633 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
1634 information can now expand as required, and rather than having a single
1635 static array of bignums, BN_CTX now uses a linked-list of such arrays
1636 allowing it to expand on demand whilst maintaining the usefulness of
1637 BN_CTX's "bundling".
1640 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
1641 to allow all RSA operations to function using a single BN_CTX.
1644 *) Preliminary support for certificate policy evaluation and checking. This
1645 is initially intended to pass the tests outlined in "Conformance Testing
1646 of Relying Party Client Certificate Path Processing Logic" v1.07.
1649 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
1650 remained unused and not that useful. A variety of other little bignum
1651 tweaks and fixes have also been made continuing on from the audit (see
1655 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
1656 associated ASN1, EVP and SSL functions and old ASN1 macros.
1659 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
1660 and this should never fail. So the return value from the use of
1661 BN_set_word() (which can fail due to needless expansion) is now deprecated;
1662 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
1665 *) BN_CTX_get() should return zero-valued bignums, providing the same
1666 initialised value as BN_new().
1667 [Geoff Thorpe, suggested by Ulf Möller]
1669 *) Support for inhibitAnyPolicy certificate extension.
1672 *) An audit of the BIGNUM code is underway, for which debugging code is
1673 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
1674 is considered valid when processing BIGNUMs, and causes execution to
1675 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
1676 further steps are taken to deliberately pollute unused data in BIGNUM
1677 structures to try and expose faulty code further on. For now, openssl will
1678 (in its default mode of operation) continue to tolerate the inconsistent
1679 forms that it has tolerated in the past, but authors and packagers should
1680 consider trying openssl and their own applications when compiled with
1681 these debugging symbols defined. It will help highlight potential bugs in
1682 their own code, and will improve the test coverage for OpenSSL itself. At
1683 some point, these tighter rules will become openssl's default to improve
1684 maintainability, though the assert()s and other overheads will remain only
1685 in debugging configurations. See bn.h for more details.
1686 [Geoff Thorpe, Nils Larsch, Ulf Möller]
1688 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
1689 that can only be obtained through BN_CTX_new() (which implicitly
1690 initialises it). The presence of this function only made it possible
1691 to overwrite an existing structure (and cause memory leaks).
1694 *) Because of the callback-based approach for implementing LHASH as a
1695 template type, lh_insert() adds opaque objects to hash-tables and
1696 lh_doall() or lh_doall_arg() are typically used with a destructor callback
1697 to clean up those corresponding objects before destroying the hash table
1698 (and losing the object pointers). So some over-zealous constifications in
1699 LHASH have been relaxed so that lh_insert() does not take (nor store) the
1700 objects as "const" and the lh_doall[_arg] callback wrappers are not
1701 prototyped to have "const" restrictions on the object pointers they are
1702 given (and so aren't required to cast them away any more).
1705 *) The tmdiff.h API was so ugly and minimal that our own timing utility
1706 (speed) prefers to use its own implementation. The two implementations
1707 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
1708 its object type properly exposed (MS_TM) instead of casting to/from "char
1709 *". This may still change yet if someone realises MS_TM and "ms_time_***"
1710 aren't necessarily the greatest nomenclatures - but this is what was used
1711 internally to the implementation so I've used that for now.
1714 *) Ensure that deprecated functions do not get compiled when
1715 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
1716 the self-tests were still using deprecated key-generation functions so
1717 these have been updated also.
1720 *) Reorganise PKCS#7 code to separate the digest location functionality
1721 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
1722 New function PKCS7_set_digest() to set the digest type for PKCS#7
1723 digestedData type. Add additional code to correctly generate the
1724 digestedData type and add support for this type in PKCS7 initialization
1728 *) New function PKCS7_set0_type_other() this initializes a PKCS7
1729 structure of type "other".
1732 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
1733 sure the loop does correctly stop and breaking ("division by zero")
1734 modulus operations are not performed. The (pre-generated) prime
1735 table crypto/bn/bn_prime.h was already correct, but it could not be
1736 re-generated on some platforms because of the "division by zero"
1737 situation in the script.
1738 [Ralf S. Engelschall]
1740 *) Update support for ECC-based TLS ciphersuites according to
1741 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
1742 SHA-1 now is only used for "small" curves (where the
1743 representation of a field element takes up to 24 bytes); for
1744 larger curves, the field element resulting from ECDH is directly
1745 used as premaster secret.
1746 [Douglas Stebila (Sun Microsystems Laboratories)]
1748 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
1749 curve secp160r1 to the tests.
1750 [Douglas Stebila (Sun Microsystems Laboratories)]
1752 *) Add the possibility to load symbols globally with DSO.
1753 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
1755 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
1756 control of the error stack.
1759 *) Add support for STORE in ENGINE.
1762 *) Add the STORE type. The intention is to provide a common interface
1763 to certificate and key stores, be they simple file-based stores, or
1764 HSM-type store, or LDAP stores, or...
1765 NOTE: The code is currently UNTESTED and isn't really used anywhere.
1768 *) Add a generic structure called OPENSSL_ITEM. This can be used to
1769 pass a list of arguments to any function as well as provide a way
1770 for a function to pass data back to the caller.
1773 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
1774 works like BUF_strdup() but can be used to duplicate a portion of
1775 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
1779 *) Add the function sk_find_ex() which works like sk_find(), but will
1780 return an index to an element even if an exact match couldn't be
1781 found. The index is guaranteed to point at the element where the
1782 searched-for key would be inserted to preserve sorting order.
1785 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
1786 takes an extra flags argument for optional functionality. Currently,
1787 the following flags are defined:
1789 OBJ_BSEARCH_VALUE_ON_NOMATCH
1790 This one gets OBJ_bsearch_ex() to return a pointer to the first
1791 element where the comparing function returns a negative or zero
1794 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
1795 This one gets OBJ_bsearch_ex() to return a pointer to the first
1796 element where the comparing function returns zero. This is useful
1797 if there are more than one element where the comparing function
1801 *) Make it possible to create self-signed certificates with 'openssl ca'
1802 in such a way that the self-signed certificate becomes part of the
1803 CA database and uses the same mechanisms for serial number generation
1804 as all other certificate signing. The new flag '-selfsign' enables
1805 this functionality. Adapt CA.sh and CA.pl.in.
1808 *) Add functionality to check the public key of a certificate request
1809 against a given private. This is useful to check that a certificate
1810 request can be signed by that key (self-signing).
1813 *) Make it possible to have multiple active certificates with the same
1814 subject in the CA index file. This is done only if the keyword
1815 'unique_subject' is set to 'no' in the main CA section (default
1816 if 'CA_default') of the configuration file. The value is saved
1817 with the database itself in a separate index attribute file,
1818 named like the index file with '.attr' appended to the name.
1821 *) Generate muti valued AVAs using '+' notation in config files for
1825 *) Support for nameConstraints certificate extension.
1828 *) Support for policyConstraints certificate extension.
1831 *) Support for policyMappings certificate extension.
1834 *) Make sure the default DSA_METHOD implementation only uses its
1835 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
1836 and change its own handlers to be NULL so as to remove unnecessary
1837 indirection. This lets alternative implementations fallback to the
1838 default implementation more easily.
1841 *) Support for directoryName in GeneralName related extensions
1845 *) Make it possible to link applications using Makefile.shared.
1846 Make that possible even when linking against static libraries!
1849 *) Support for single pass processing for S/MIME signing. This now
1850 means that S/MIME signing can be done from a pipe, in addition
1851 cleartext signing (multipart/signed type) is effectively streaming
1852 and the signed data does not need to be all held in memory.
1854 This is done with a new flag PKCS7_STREAM. When this flag is set
1855 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
1856 is done after the data is output (and digests calculated) in
1857 SMIME_write_PKCS7().
1860 *) Add full support for -rpath/-R, both in shared libraries and
1861 applications, at least on the platforms where it's known how
1865 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
1866 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
1867 will now compute a table of multiples of the generator that
1868 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
1869 faster (notably in the case of a single point multiplication,
1870 scalar * generator).
1871 [Nils Larsch, Bodo Moeller]
1873 *) IPv6 support for certificate extensions. The various extensions
1874 which use the IP:a.b.c.d can now take IPv6 addresses using the
1875 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
1879 *) Added an ENGINE that implements RSA by performing private key
1880 exponentiations with the GMP library. The conversions to and from
1881 GMP's mpz_t format aren't optimised nor are any montgomery forms
1882 cached, and on x86 it appears OpenSSL's own performance has caught up.
1883 However there are likely to be other architectures where GMP could
1884 provide a boost. This ENGINE is not built in by default, but it can be
1885 specified at Configure time and should be accompanied by the necessary
1886 linker additions, eg;
1887 ./config -DOPENSSL_USE_GMP -lgmp
1890 *) "openssl engine" will not display ENGINE/DSO load failure errors when
1891 testing availability of engines with "-t" - the old behaviour is
1892 produced by increasing the feature's verbosity with "-tt".
1895 *) ECDSA routines: under certain error conditions uninitialized BN objects
1896 could be freed. Solution: make sure initialization is performed early
1897 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
1901 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
1902 and DH_METHOD (eg. by ENGINE implementations) to override the normal
1903 software implementations. For DSA and DH, parameter generation can
1904 also be overriden by providing the appropriate method callbacks.
1907 *) Change the "progress" mechanism used in key-generation and
1908 primality testing to functions that take a new BN_GENCB pointer in
1909 place of callback/argument pairs. The new API functions have "_ex"
1910 postfixes and the older functions are reimplemented as wrappers for
1911 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
1912 declarations of the old functions to help (graceful) attempts to
1913 migrate to the new functions. Also, the new key-generation API
1914 functions operate on a caller-supplied key-structure and return
1915 success/failure rather than returning a key or NULL - this is to
1916 help make "keygen" another member function of RSA_METHOD etc.
1918 Example for using the new callback interface:
1920 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
1924 BN_GENCB_set(&my_cb, my_callback, my_arg);
1926 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
1927 /* For the meaning of a, b in calls to my_callback(), see the
1928 * documentation of the function that calls the callback.
1929 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
1930 * my_callback should return 1 if it wants BN_is_prime_ex()
1931 * to continue, or 0 to stop.
1936 *) Change the ZLIB compression method to be stateful, and make it
1937 available to TLS with the number defined in
1938 draft-ietf-tls-compression-04.txt.
1941 *) Add the ASN.1 structures and functions for CertificatePair, which
1942 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
1944 CertificatePair ::= SEQUENCE {
1945 forward [0] Certificate OPTIONAL,
1946 reverse [1] Certificate OPTIONAL,
1947 -- at least one of the pair shall be present -- }
1949 Also implement the PEM functions to read and write certificate
1950 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
1952 This needed to be defined, mostly for the sake of the LDAP
1953 attribute crossCertificatePair, but may prove useful elsewhere as
1957 *) Make it possible to inhibit symlinking of shared libraries in
1958 Makefile.shared, for Cygwin's sake.
1961 *) Extend the BIGNUM API by creating a function
1962 void BN_set_negative(BIGNUM *a, int neg);
1963 and a macro that behave like
1964 int BN_is_negative(const BIGNUM *a);
1966 to avoid the need to access 'a->neg' directly in applications.
1969 *) Implement fast modular reduction for pseudo-Mersenne primes
1970 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
1971 EC_GROUP_new_curve_GFp() will now automatically use this
1973 [Nils Larsch <nla@trustcenter.de>]
1975 *) Add new lock type (CRYPTO_LOCK_BN).
1978 *) Change the ENGINE framework to automatically load engines
1979 dynamically from specific directories unless they could be
1980 found to already be built in or loaded. Move all the
1981 current engines except for the cryptodev one to a new
1983 The engines in engines/ are built as shared libraries if
1984 the "shared" options was given to ./Configure or ./config.
1985 Otherwise, they are inserted in libcrypto.a.
1986 /usr/local/ssl/engines is the default directory for dynamic
1987 engines, but that can be overriden at configure time through
1988 the usual use of --prefix and/or --openssldir, and at run
1989 time with the environment variable OPENSSL_ENGINES.
1990 [Geoff Thorpe and Richard Levitte]
1992 *) Add Makefile.shared, a helper makefile to build shared
1993 libraries. Addapt Makefile.org.
1996 *) Add version info to Win32 DLLs.
1997 [Peter 'Luna' Runestig" <peter@runestig.com>]
1999 *) Add new 'medium level' PKCS#12 API. Certificates and keys
2000 can be added using this API to created arbitrary PKCS#12
2001 files while avoiding the low level API.
2003 New options to PKCS12_create(), key or cert can be NULL and
2004 will then be omitted from the output file. The encryption
2005 algorithm NIDs can be set to -1 for no encryption, the mac
2006 iteration count can be set to 0 to omit the mac.
2008 Enhance pkcs12 utility by making the -nokeys and -nocerts
2009 options work when creating a PKCS#12 file. New option -nomac
2010 to omit the mac, NONE can be set for an encryption algorithm.
2011 New code is modified to use the enhanced PKCS12_create()
2012 instead of the low level API.
2015 *) Extend ASN1 encoder to support indefinite length constructed
2016 encoding. This can output sequences tags and octet strings in
2017 this form. Modify pk7_asn1.c to support indefinite length
2018 encoding. This is experimental and needs additional code to
2019 be useful, such as an ASN1 bio and some enhanced streaming
2022 Extend template encode functionality so that tagging is passed
2023 down to the template encoder.
2026 *) Let 'openssl req' fail if an argument to '-newkey' is not
2027 recognized instead of using RSA as a default.
2030 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
2031 As these are not official, they are not included in "ALL";
2032 the "ECCdraft" ciphersuite group alias can be used to select them.
2033 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
2035 *) Add ECDH engine support.
2036 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
2038 *) Add ECDH in new directory crypto/ecdh/.
2039 [Douglas Stebila (Sun Microsystems Laboratories)]
2041 *) Let BN_rand_range() abort with an error after 100 iterations
2042 without success (which indicates a broken PRNG).
2045 *) Change BN_mod_sqrt() so that it verifies that the input value
2046 is really the square of the return value. (Previously,
2047 BN_mod_sqrt would show GIGO behaviour.)
2050 *) Add named elliptic curves over binary fields from X9.62, SECG,
2051 and WAP/WTLS; add OIDs that were still missing.
2053 [Sheueling Chang Shantz and Douglas Stebila
2054 (Sun Microsystems Laboratories)]
2056 *) Extend the EC library for elliptic curves over binary fields
2057 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
2060 EC_GF2m_simple_method
2064 EC_GROUP_new_curve_GF2m
2065 EC_GROUP_set_curve_GF2m
2066 EC_GROUP_get_curve_GF2m
2067 EC_POINT_set_affine_coordinates_GF2m
2068 EC_POINT_get_affine_coordinates_GF2m
2069 EC_POINT_set_compressed_coordinates_GF2m
2071 Point compression for binary fields is disabled by default for
2072 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
2075 As binary polynomials are represented as BIGNUMs, various members
2076 of the EC_GROUP and EC_POINT data structures can be shared
2077 between the implementations for prime fields and binary fields;
2078 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
2079 are essentially identical to their ..._GFp counterparts.
2080 (For simplicity, the '..._GFp' prefix has been dropped from
2081 various internal method names.)
2083 An internal 'field_div' method (similar to 'field_mul' and
2084 'field_sqr') has been added; this is used only for binary fields.
2086 [Sheueling Chang Shantz and Douglas Stebila
2087 (Sun Microsystems Laboratories)]
2089 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
2090 through methods ('mul', 'precompute_mult').
2092 The generic implementations (now internally called 'ec_wNAF_mul'
2093 and 'ec_wNAF_precomputed_mult') remain the default if these
2094 methods are undefined.
2096 [Sheueling Chang Shantz and Douglas Stebila
2097 (Sun Microsystems Laboratories)]
2099 *) New function EC_GROUP_get_degree, which is defined through
2100 EC_METHOD. For curves over prime fields, this returns the bit
2101 length of the modulus.
2103 [Sheueling Chang Shantz and Douglas Stebila
2104 (Sun Microsystems Laboratories)]
2106 *) New functions EC_GROUP_dup, EC_POINT_dup.
2107 (These simply call ..._new and ..._copy).
2109 [Sheueling Chang Shantz and Douglas Stebila
2110 (Sun Microsystems Laboratories)]
2112 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
2113 Polynomials are represented as BIGNUMs (where the sign bit is not
2114 used) in the following functions [macros]:
2117 BN_GF2m_sub [= BN_GF2m_add]
2118 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
2119 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
2120 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
2122 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
2123 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
2124 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
2125 BN_GF2m_cmp [= BN_ucmp]
2127 (Note that only the 'mod' functions are actually for fields GF(2^m).
2128 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
2130 For some functions, an the irreducible polynomial defining a
2131 field can be given as an 'unsigned int[]' with strictly
2132 decreasing elements giving the indices of those bits that are set;
2133 i.e., p[] represents the polynomial
2134 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
2136 p[0] > p[1] > ... > p[k] = 0.
2137 This applies to the following functions:
2142 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
2143 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
2145 BN_GF2m_mod_sqrt_arr
2146 BN_GF2m_mod_solve_quad_arr
2150 Conversion can be performed by the following functions:
2155 bntest.c has additional tests for binary polynomial arithmetic.
2157 Two implementations for BN_GF2m_mod_div() are available.
2158 The default algorithm simply uses BN_GF2m_mod_inv() and
2159 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
2160 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
2161 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
2163 [Sheueling Chang Shantz and Douglas Stebila
2164 (Sun Microsystems Laboratories)]
2166 *) Add new error code 'ERR_R_DISABLED' that can be used when some
2167 functionality is disabled at compile-time.
2168 [Douglas Stebila <douglas.stebila@sun.com>]
2170 *) Change default behaviour of 'openssl asn1parse' so that more
2171 information is visible when viewing, e.g., a certificate:
2173 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
2174 mode the content of non-printable OCTET STRINGs is output in a
2175 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
2176 avoid the appearance of a printable string.
2177 [Nils Larsch <nla@trustcenter.de>]
2179 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
2181 EC_GROUP_set_asn1_flag()
2182 EC_GROUP_get_asn1_flag()
2183 EC_GROUP_set_point_conversion_form()
2184 EC_GROUP_get_point_conversion_form()
2185 These control ASN1 encoding details:
2186 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
2187 has been set to OPENSSL_EC_NAMED_CURVE.
2188 - Points are encoded in uncompressed form by default; options for
2189 asn1_for are as for point2oct, namely
2190 POINT_CONVERSION_COMPRESSED
2191 POINT_CONVERSION_UNCOMPRESSED
2192 POINT_CONVERSION_HYBRID
2194 Also add 'seed' and 'seed_len' members to EC_GROUP with access
2197 EC_GROUP_get0_seed()
2198 EC_GROUP_get_seed_len()
2199 This is used only for ASN1 purposes (so far).
2200 [Nils Larsch <nla@trustcenter.de>]
2202 *) Add 'field_type' member to EC_METHOD, which holds the NID
2203 of the appropriate field type OID. The new function
2204 EC_METHOD_get_field_type() returns this value.
2205 [Nils Larsch <nla@trustcenter.de>]
2210 EC_POINT_point2hex()
2211 EC_POINT_hex2point()
2212 providing useful interfaces to EC_POINT_point2oct() and
2213 EC_POINT_oct2point().
2214 [Nils Larsch <nla@trustcenter.de>]
2216 *) Change internals of the EC library so that the functions
2217 EC_GROUP_set_generator()
2218 EC_GROUP_get_generator()
2219 EC_GROUP_get_order()
2220 EC_GROUP_get_cofactor()
2221 are implemented directly in crypto/ec/ec_lib.c and not dispatched
2222 to methods, which would lead to unnecessary code duplication when
2223 adding different types of curves.
2224 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
2226 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
2227 arithmetic, and such that modified wNAFs are generated
2228 (which avoid length expansion in many cases).
2231 *) Add a function EC_GROUP_check_discriminant() (defined via
2232 EC_METHOD) that verifies that the curve discriminant is non-zero.
2234 Add a function EC_GROUP_check() that makes some sanity tests
2235 on a EC_GROUP, its generator and order. This includes
2236 EC_GROUP_check_discriminant().
2237 [Nils Larsch <nla@trustcenter.de>]
2239 *) Add ECDSA in new directory crypto/ecdsa/.
2241 Add applications 'openssl ecparam' and 'openssl ecdsa'
2242 (these are based on 'openssl dsaparam' and 'openssl dsa').
2244 ECDSA support is also included in various other files across the
2245 library. Most notably,
2246 - 'openssl req' now has a '-newkey ecdsa:file' option;
2247 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
2248 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
2249 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
2250 them suitable for ECDSA where domain parameters must be
2251 extracted before the specific public key;
2252 - ECDSA engine support has been added.
2253 [Nils Larsch <nla@trustcenter.de>]
2255 *) Include some named elliptic curves, and add OIDs from X9.62,
2256 SECG, and WAP/WTLS. Each curve can be obtained from the new
2258 EC_GROUP_new_by_curve_name(),
2259 and the list of available named curves can be obtained with
2260 EC_get_builtin_curves().
2261 Also add a 'curve_name' member to EC_GROUP objects, which can be
2263 EC_GROUP_set_curve_name()
2264 EC_GROUP_get_curve_name()
2265 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
2267 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
2268 was actually never needed) and in BN_mul(). The removal in BN_mul()
2269 required a small change in bn_mul_part_recursive() and the addition
2270 of the functions bn_cmp_part_words(), bn_sub_part_words() and
2271 bn_add_part_words(), which do the same thing as bn_cmp_words(),
2272 bn_sub_words() and bn_add_words() except they take arrays with
2276 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
2278 *) Cleanse PEM buffers before freeing them since they may contain
2280 [Benjamin Bennett <ben@psc.edu>]
2282 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2283 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2284 authentication-only ciphersuites.
2287 *) Since AES128 and AES256 share a single mask bit in the logic of
2288 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2289 kludge to work properly if AES128 is available and AES256 isn't.
2292 *) Expand security boundary to match 1.1.1 module.
2295 *) Remove redundant features: hash file source, editing of test vectors
2296 modify fipsld to use external fips_premain.c signature.
2299 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
2300 run algorithm test programs.
2303 *) Make algorithm test programs more tolerant of whitespace.
2306 *) Have SSL/TLS server implementation tolerate "mismatched" record
2307 protocol version while receiving ClientHello even if the
2308 ClientHello is fragmented. (The server can't insist on the
2309 particular protocol version it has chosen before the ServerHello
2310 message has informed the client about his choice.)
2313 *) Load error codes if they are not already present instead of using a
2314 static variable. This allows them to be cleanly unloaded and reloaded.
2317 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
2319 *) Introduce limits to prevent malicious keys being able to
2320 cause a denial of service. (CVE-2006-2940)
2321 [Steve Henson, Bodo Moeller]
2323 *) Fix ASN.1 parsing of certain invalid structures that can result
2324 in a denial of service. (CVE-2006-2937) [Steve Henson]
2326 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2327 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2329 *) Fix SSL client code which could crash if connecting to a
2330 malicious SSLv2 server. (CVE-2006-4343)
2331 [Tavis Ormandy and Will Drewry, Google Security Team]
2333 *) Change ciphersuite string processing so that an explicit
2334 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
2335 will no longer include "AES128-SHA"), and any other similar
2336 ciphersuite (same bitmap) from *other* protocol versions (so that
2337 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
2338 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
2339 changes from 0.9.8b and 0.9.8d.
2342 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
2344 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2345 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2347 *) Change the Unix randomness entropy gathering to use poll() when
2348 possible instead of select(), since the latter has some
2349 undesirable limitations.
2350 [Darryl Miles via Richard Levitte and Bodo Moeller]
2352 *) Disable rogue ciphersuites:
2354 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2355 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2356 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2358 The latter two were purportedly from
2359 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2362 Also deactive the remaining ciphersuites from
2363 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2364 unofficial, and the ID has long expired.
2367 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2368 dual-core machines) and other potential thread-safety issues.
2371 Changes between 0.9.7i and 0.9.7j [04 May 2006]
2373 *) Adapt fipsld and the build system to link against the validated FIPS
2374 module in FIPS mode.
2377 *) Fixes for VC++ 2005 build under Windows.
2380 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
2381 from a Windows bash shell such as MSYS. It is autodetected from the
2382 "config" script when run from a VC++ environment. Modify standard VC++
2383 build to use fipscanister.o from the GNU make build.
2386 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
2388 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
2389 The value now differs depending on if you build for FIPS or not.
2390 BEWARE! A program linked with a shared FIPSed libcrypto can't be
2391 safely run with a non-FIPSed libcrypto, as it may crash because of
2392 the difference induced by this change.
2395 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
2397 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2398 (part of SSL_OP_ALL). This option used to disable the
2399 countermeasure against man-in-the-middle protocol-version
2400 rollback in the SSL 2.0 server implementation, which is a bad
2401 idea. (CVE-2005-2969)
2403 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2404 for Information Security, National Institute of Advanced Industrial
2405 Science and Technology [AIST], Japan)]
2407 *) Minimal support for X9.31 signatures and PSS padding modes. This is
2408 mainly for FIPS compliance and not fully integrated at this stage.
2411 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
2412 the exponentiation using a fixed-length exponent. (Otherwise,
2413 the information leaked through timing could expose the secret key
2414 after many signatures; cf. Bleichenbacher's attack on DSA with
2418 *) Make a new fixed-window mod_exp implementation the default for
2419 RSA, DSA, and DH private-key operations so that the sequence of
2420 squares and multiplies and the memory access pattern are
2421 independent of the particular secret key. This will mitigate
2422 cache-timing and potential related attacks.
2424 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
2425 and this is automatically used by BN_mod_exp_mont() if the new flag
2426 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
2427 will use this BN flag for private exponents unless the flag
2428 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
2429 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
2431 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
2433 *) Change the client implementation for SSLv23_method() and
2434 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
2435 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
2436 (Previously, the SSL 2.0 backwards compatible Client Hello
2437 message format would be used even with SSL_OP_NO_SSLv2.)
2440 *) Add support for smime-type MIME parameter in S/MIME messages which some
2444 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
2445 a threadsafe manner. Modify rsa code to use new function and add calls
2446 to dsa and dh code (which had race conditions before).
2449 *) Include the fixed error library code in the C error file definitions
2450 instead of fixing them up at runtime. This keeps the error code
2451 structures constant.
2454 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
2456 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
2459 *) Fixes for newer kerberos headers. NB: the casts are needed because
2460 the 'length' field is signed on one version and unsigned on another
2461 with no (?) obvious way to tell the difference, without these VC++
2462 complains. Also the "definition" of FAR (blank) is no longer included
2463 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
2464 some needed definitions.
2467 *) Undo Cygwin change.
2470 *) Added support for proxy certificates according to RFC 3820.
2471 Because they may be a security thread to unaware applications,
2472 they must be explicitely allowed in run-time. See
2473 docs/HOWTO/proxy_certificates.txt for further information.
2476 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
2478 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
2479 server and client random values. Previously
2480 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
2481 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
2483 This change has negligible security impact because:
2485 1. Server and client random values still have 24 bytes of pseudo random
2488 2. Server and client random values are sent in the clear in the initial
2491 3. The master secret is derived using the premaster secret (48 bytes in
2492 size for static RSA ciphersuites) as well as client server and random
2495 The OpenSSL team would like to thank the UK NISCC for bringing this issue
2498 [Stephen Henson, reported by UK NISCC]
2500 *) Use Windows randomness collection on Cygwin.
2503 *) Fix hang in EGD/PRNGD query when communication socket is closed
2504 prematurely by EGD/PRNGD.
2505 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
2507 *) Prompt for pass phrases when appropriate for PKCS12 input format.
2510 *) Back-port of selected performance improvements from development
2511 branch, as well as improved support for PowerPC platforms.
2514 *) Add lots of checks for memory allocation failure, error codes to indicate
2515 failure and freeing up memory if a failure occurs.
2516 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
2518 *) Add new -passin argument to dgst.
2521 *) Perform some character comparisons of different types in X509_NAME_cmp:
2522 this is needed for some certificates that reencode DNs into UTF8Strings
2523 (in violation of RFC3280) and can't or wont issue name rollover
2527 *) Make an explicit check during certificate validation to see that
2528 the CA setting in each certificate on the chain is correct. As a
2529 side effect always do the following basic checks on extensions,
2530 not just when there's an associated purpose to the check:
2532 - if there is an unhandled critical extension (unless the user
2533 has chosen to ignore this fault)
2534 - if the path length has been exceeded (if one is set at all)
2535 - that certain extensions fit the associated purpose (if one has
2539 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
2541 *) Avoid a race condition when CRLs are checked in a multi threaded
2542 environment. This would happen due to the reordering of the revoked
2543 entries during signature checking and serial number lookup. Now the
2544 encoding is cached and the serial number sort performed under a lock.
2545 Add new STACK function sk_is_sorted().
2548 *) Add Delta CRL to the extension code.
2551 *) Various fixes to s3_pkt.c so alerts are sent properly.
2552 [David Holmes <d.holmes@f5.com>]
2554 *) Reduce the chances of duplicate issuer name and serial numbers (in
2555 violation of RFC3280) using the OpenSSL certificate creation utilities.
2556 This is done by creating a random 64 bit value for the initial serial
2557 number when a serial number file is created or when a self signed
2558 certificate is created using 'openssl req -x509'. The initial serial
2559 number file is created using 'openssl x509 -next_serial' in CA.pl
2560 rather than being initialized to 1.
2563 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
2565 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
2566 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
2567 [Joe Orton, Steve Henson]
2569 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
2571 [Joe Orton, Steve Henson]
2573 *) Make it possible to have multiple active certificates with the same
2574 subject in the CA index file. This is done only if the keyword
2575 'unique_subject' is set to 'no' in the main CA section (default
2576 if 'CA_default') of the configuration file. The value is saved
2577 with the database itself in a separate index attribute file,
2578 named like the index file with '.attr' appended to the name.
2581 *) X509 verify fixes. Disable broken certificate workarounds when
2582 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
2583 keyUsage extension present. Don't accept CRLs with unhandled critical
2584 extensions: since verify currently doesn't process CRL extensions this
2585 rejects a CRL with *any* critical extensions. Add new verify error codes
2589 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
2590 A clarification of RFC2560 will require the use of OCTET STRINGs and
2591 some implementations cannot handle the current raw format. Since OpenSSL
2592 copies and compares OCSP nonces as opaque blobs without any attempt at
2593 parsing them this should not create any compatibility issues.
2596 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
2597 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
2598 this HMAC (and other) operations are several times slower than OpenSSL
2602 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
2603 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2605 *) Use the correct content when signing type "other".
2608 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
2610 *) Fix various bugs revealed by running the NISCC test suite:
2612 Stop out of bounds reads in the ASN1 code when presented with
2613 invalid tags (CVE-2003-0543 and CVE-2003-0544).
2615 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
2617 If verify callback ignores invalid public key errors don't try to check
2618 certificate signature with the NULL public key.
2622 *) New -ignore_err option in ocsp application to stop the server
2623 exiting on the first error in a request.
2626 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
2627 if the server requested one: as stated in TLS 1.0 and SSL 3.0
2631 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
2632 extra data after the compression methods not only for TLS 1.0
2633 but also for SSL 3.0 (as required by the specification).
2634 [Bodo Moeller; problem pointed out by Matthias Loepfe]
2636 *) Change X509_certificate_type() to mark the key as exported/exportable
2637 when it's 512 *bits* long, not 512 bytes.
2640 *) Change AES_cbc_encrypt() so it outputs exact multiple of
2641 blocks during encryption.
2644 *) Various fixes to base64 BIO and non blocking I/O. On write
2645 flushes were not handled properly if the BIO retried. On read
2646 data was not being buffered properly and had various logic bugs.
2647 This also affects blocking I/O when the data being decoded is a
2651 *) Various S/MIME bugfixes and compatibility changes:
2652 output correct application/pkcs7 MIME type if
2653 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
2654 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
2655 of files as .eml work). Correctly handle very long lines in MIME
2659 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
2661 *) Countermeasure against the Klima-Pokorny-Rosa extension of
2662 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
2663 a protocol version number mismatch like a decryption error
2664 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
2667 *) Turn on RSA blinding by default in the default implementation
2668 to avoid a timing attack. Applications that don't want it can call
2669 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
2670 They would be ill-advised to do so in most cases.
2671 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
2673 *) Change RSA blinding code so that it works when the PRNG is not
2674 seeded (in this case, the secret RSA exponent is abused as
2675 an unpredictable seed -- if it is not unpredictable, there
2676 is no point in blinding anyway). Make RSA blinding thread-safe
2677 by remembering the creator's thread ID in rsa->blinding and
2678 having all other threads use local one-time blinding factors
2679 (this requires more computation than sharing rsa->blinding, but
2680 avoids excessive locking; and if an RSA object is not shared
2681 between threads, blinding will still be very fast).
2684 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
2685 ENGINE as defaults for all supported algorithms irrespective of
2686 the 'flags' parameter. 'flags' is now honoured, so applications
2687 should make sure they are passing it correctly.
2690 *) Target "mingw" now allows native Windows code to be generated in
2691 the Cygwin environment as well as with the MinGW compiler.
2694 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
2696 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
2697 via timing by performing a MAC computation even if incorrrect
2698 block cipher padding has been found. This is a countermeasure
2699 against active attacks where the attacker has to distinguish
2700 between bad padding and a MAC verification error. (CVE-2003-0078)
2702 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
2703 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
2704 Martin Vuagnoux (EPFL, Ilion)]
2706 *) Make the no-err option work as intended. The intention with no-err
2707 is not to have the whole error stack handling routines removed from
2708 libcrypto, it's only intended to remove all the function name and
2709 reason texts, thereby removing some of the footprint that may not
2710 be interesting if those errors aren't displayed anyway.
2712 NOTE: it's still possible for any application or module to have it's
2713 own set of error texts inserted. The routines are there, just not
2714 used by default when no-err is given.
2717 *) Add support for FreeBSD on IA64.
2718 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
2720 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
2721 Kerberos function mit_des_cbc_cksum(). Before this change,
2722 the value returned by DES_cbc_cksum() was like the one from
2723 mit_des_cbc_cksum(), except the bytes were swapped.
2724 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
2726 *) Allow an application to disable the automatic SSL chain building.
2727 Before this a rather primitive chain build was always performed in
2728 ssl3_output_cert_chain(): an application had no way to send the
2729 correct chain if the automatic operation produced an incorrect result.
2731 Now the chain builder is disabled if either:
2733 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
2735 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
2737 The reasoning behind this is that an application would not want the
2738 auto chain building to take place if extra chain certificates are
2739 present and it might also want a means of sending no additional
2740 certificates (for example the chain has two certificates and the
2744 *) Add the possibility to build without the ENGINE framework.
2745 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2747 *) Under Win32 gmtime() can return NULL: check return value in
2748 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
2751 *) DSA routines: under certain error conditions uninitialized BN objects
2752 could be freed. Solution: make sure initialization is performed early
2753 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
2754 Nils Larsch <nla@trustcenter.de> via PR#459)
2757 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
2758 checked on reconnect on the client side, therefore session resumption
2759 could still fail with a "ssl session id is different" error. This
2760 behaviour is masked when SSL_OP_ALL is used due to
2761 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
2762 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2763 followup to PR #377.
2766 *) IA-32 assembler support enhancements: unified ELF targets, support
2767 for SCO/Caldera platforms, fix for Cygwin shared build.
2770 *) Add support for FreeBSD on sparc64. As a consequence, support for
2771 FreeBSD on non-x86 processors is separate from x86 processors on
2772 the config script, much like the NetBSD support.
2773 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
2775 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
2777 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
2780 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
2781 code (06) was taken as the first octet of the session ID and the last
2782 octet was ignored consequently. As a result SSLv2 client side session
2783 caching could not have worked due to the session ID mismatch between
2785 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2789 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
2790 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
2794 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
2795 seems that in spite of existing for more than a year, many application
2796 author have done nothing to provide the necessary callbacks, which
2797 means that this particular engine will not work properly anywhere.
2798 This is a very unfortunate situation which forces us, in the name
2799 of usability, to give the hw_ncipher.c a static lock, which is part
2801 NOTE: This is for the 0.9.7 series ONLY. This hack will never
2802 appear in 0.9.8 or later. We EXPECT application authors to have
2803 dealt properly with this when 0.9.8 is released (unless we actually
2804 make such changes in the libcrypto locking code that changes will
2805 have to be made anyway).
2808 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
2809 octets have been read, EOF or an error occurs. Without this change
2810 some truncated ASN1 structures will not produce an error.
2813 *) Disable Heimdal support, since it hasn't been fully implemented.
2814 Still give the possibility to force the use of Heimdal, but with
2815 warnings and a request that patches get sent to openssl-dev.
2818 *) Add the VC-CE target, introduce the WINCE sysname, and add
2819 INSTALL.WCE and appropriate conditionals to make it build.
2820 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2822 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
2823 cygssl-x.y.z.dll, where x, y and z are the major, minor and
2824 edit numbers of the version.
2825 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
2827 *) Introduce safe string copy and catenation functions
2828 (BUF_strlcpy() and BUF_strlcat()).
2829 [Ben Laurie (CHATS) and Richard Levitte]
2831 *) Avoid using fixed-size buffers for one-line DNs.
2832 [Ben Laurie (CHATS)]
2834 *) Add BUF_MEM_grow_clean() to avoid information leakage when
2835 resizing buffers containing secrets, and use where appropriate.
2836 [Ben Laurie (CHATS)]
2838 *) Avoid using fixed size buffers for configuration file location.
2839 [Ben Laurie (CHATS)]
2841 *) Avoid filename truncation for various CA files.
2842 [Ben Laurie (CHATS)]
2844 *) Use sizeof in preference to magic numbers.
2845 [Ben Laurie (CHATS)]
2847 *) Avoid filename truncation in cert requests.
2848 [Ben Laurie (CHATS)]
2850 *) Add assertions to check for (supposedly impossible) buffer
2852 [Ben Laurie (CHATS)]
2854 *) Don't cache truncated DNS entries in the local cache (this could
2855 potentially lead to a spoofing attack).
2856 [Ben Laurie (CHATS)]
2858 *) Fix various buffers to be large enough for hex/decimal
2859 representations in a platform independent manner.
2860 [Ben Laurie (CHATS)]
2862 *) Add CRYPTO_realloc_clean() to avoid information leakage when
2863 resizing buffers containing secrets, and use where appropriate.
2864 [Ben Laurie (CHATS)]
2866 *) Add BIO_indent() to avoid much slightly worrying code to do
2868 [Ben Laurie (CHATS)]
2870 *) Convert sprintf()/BIO_puts() to BIO_printf().
2871 [Ben Laurie (CHATS)]
2873 *) buffer_gets() could terminate with the buffer only half
2875 [Ben Laurie (CHATS)]
2877 *) Add assertions to prevent user-supplied crypto functions from
2878 overflowing internal buffers by having large block sizes, etc.
2879 [Ben Laurie (CHATS)]
2881 *) New OPENSSL_assert() macro (similar to assert(), but enabled
2883 [Ben Laurie (CHATS)]
2885 *) Eliminate unused copy of key in RC4.
2886 [Ben Laurie (CHATS)]
2888 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
2889 [Ben Laurie (CHATS)]
2891 *) Fix off-by-one error in EGD path.
2892 [Ben Laurie (CHATS)]
2894 *) If RANDFILE path is too long, ignore instead of truncating.
2895 [Ben Laurie (CHATS)]
2897 *) Eliminate unused and incorrectly sized X.509 structure
2899 [Ben Laurie (CHATS)]
2901 *) Eliminate unused and dangerous function knumber().
2902 [Ben Laurie (CHATS)]
2904 *) Eliminate unused and dangerous structure, KSSL_ERR.
2905 [Ben Laurie (CHATS)]
2907 *) Protect against overlong session ID context length in an encoded
2908 session object. Since these are local, this does not appear to be
2910 [Ben Laurie (CHATS)]
2912 *) Change from security patch (see 0.9.6e below) that did not affect
2913 the 0.9.6 release series:
2915 Remote buffer overflow in SSL3 protocol - an attacker could
2916 supply an oversized master key in Kerberos-enabled versions.
2918 [Ben Laurie (CHATS)]
2920 *) Change the SSL kerb5 codes to match RFC 2712.
2923 *) Make -nameopt work fully for req and add -reqopt switch.
2924 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
2926 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
2927 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
2929 *) Make sure tests can be performed even if the corresponding algorithms
2930 have been removed entirely. This was also the last step to make
2931 OpenSSL compilable with DJGPP under all reasonable conditions.
2932 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
2934 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
2935 to allow version independent disabling of normally unselected ciphers,
2936 which may be activated as a side-effect of selecting a single cipher.
2938 (E.g., cipher list string "RSA" enables ciphersuites that are left
2939 out of "ALL" because they do not provide symmetric encryption.
2940 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
2941 [Lutz Jaenicke, Bodo Moeller]
2943 *) Add appropriate support for separate platform-dependent build
2944 directories. The recommended way to make a platform-dependent
2945 build directory is the following (tested on Linux), maybe with
2948 # Place yourself outside of the OpenSSL source tree. In
2949 # this example, the environment variable OPENSSL_SOURCE
2950 # is assumed to contain the absolute OpenSSL source directory.
2951 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
2952 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
2953 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
2954 mkdir -p `dirname $F`
2955 ln -s $OPENSSL_SOURCE/$F $F
2958 To be absolutely sure not to disturb the source tree, a "make clean"
2959 is a good thing. If it isn't successfull, don't worry about it,
2960 it probably means the source directory is very clean.
2963 *) Make sure any ENGINE control commands make local copies of string
2964 pointers passed to them whenever necessary. Otherwise it is possible
2965 the caller may have overwritten (or deallocated) the original string
2966 data when a later ENGINE operation tries to use the stored values.
2967 [Götz Babin-Ebell <babinebell@trustcenter.de>]
2969 *) Improve diagnostics in file reading and command-line digests.
2970 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
2972 *) Add AES modes CFB and OFB to the object database. Correct an
2973 error in AES-CFB decryption.
2976 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
2977 allows existing EVP_CIPHER_CTX structures to be reused after
2978 calling EVP_*Final(). This behaviour is used by encryption
2979 BIOs and some applications. This has the side effect that
2980 applications must explicitly clean up cipher contexts with
2981 EVP_CIPHER_CTX_cleanup() or they will leak memory.
2984 *) Check the values of dna and dnb in bn_mul_recursive before calling
2985 bn_mul_comba (a non zero value means the a or b arrays do not contain
2986 n2 elements) and fallback to bn_mul_normal if either is not zero.
2989 *) Fix escaping of non-ASCII characters when using the -subj option
2990 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
2993 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
2994 form for "surname", serialNumber has no short form.
2995 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
2996 therefore remove "mail" short name for "internet 7".
2997 The OID for unique identifiers in X509 certificates is
2998 x500UniqueIdentifier, not uniqueIdentifier.
2999 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
3002 *) Add an "init" command to the ENGINE config module and auto initialize
3003 ENGINEs. Without any "init" command the ENGINE will be initialized
3004 after all ctrl commands have been executed on it. If init=1 the
3005 ENGINE is initailized at that point (ctrls before that point are run
3006 on the uninitialized ENGINE and after on the initialized one). If
3007 init=0 then the ENGINE will not be iniatialized at all.
3010 *) Fix the 'app_verify_callback' interface so that the user-defined
3011 argument is actually passed to the callback: In the
3012 SSL_CTX_set_cert_verify_callback() prototype, the callback
3013 declaration has been changed from
3016 int (*cb)(X509_STORE_CTX *,void *);
3017 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
3018 i=s->ctx->app_verify_callback(&ctx)
3019 has been changed into
3020 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
3022 To update applications using SSL_CTX_set_cert_verify_callback(),
3023 a dummy argument can be added to their callback functions.
3024 [D. K. Smetters <smetters@parc.xerox.com>]
3026 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
3027 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
3029 *) Add and OPENSSL_LOAD_CONF define which will cause
3030 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
3031 This allows older applications to transparently support certain
3032 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
3033 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
3034 load the config file and OPENSSL_add_all_algorithms_conf() which will
3035 always load it have also been added.
3038 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
3039 Adjust NIDs and EVP layer.
3040 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3042 *) Config modules support in openssl utility.
3044 Most commands now load modules from the config file,
3045 though in a few (such as version) this isn't done
3046 because it couldn't be used for anything.
3048 In the case of ca and req the config file used is
3049 the same as the utility itself: that is the -config
3050 command line option can be used to specify an
3054 *) Move default behaviour from OPENSSL_config(). If appname is NULL
3055 use "openssl_conf" if filename is NULL use default openssl config file.
3058 *) Add an argument to OPENSSL_config() to allow the use of an alternative
3059 config section name. Add a new flag to tolerate a missing config file
3060 and move code to CONF_modules_load_file().
3063 *) Support for crypto accelerator cards from Accelerated Encryption
3064 Processing, www.aep.ie. (Use engine 'aep')
3065 The support was copied from 0.9.6c [engine] and adapted/corrected
3066 to work with the new engine framework.
3067 [AEP Inc. and Richard Levitte]
3069 *) Support for SureWare crypto accelerator cards from Baltimore
3070 Technologies. (Use engine 'sureware')
3071 The support was copied from 0.9.6c [engine] and adapted
3072 to work with the new engine framework.
3075 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
3076 make the newer ENGINE framework commands for the CHIL engine work.
3077 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
3079 *) Make it possible to produce shared libraries on ReliantUNIX.
3080 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
3082 *) Add the configuration target debug-linux-ppro.
3083 Make 'openssl rsa' use the general key loading routines
3084 implemented in apps.c, and make those routines able to
3085 handle the key format FORMAT_NETSCAPE and the variant
3087 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3089 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
3090 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3092 *) Add -keyform to rsautl, and document -engine.
3093 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
3095 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
3096 BIO_R_NO_SUCH_FILE error code rather than the generic
3097 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
3100 *) Add new functions
3102 ERR_peek_last_error_line
3103 ERR_peek_last_error_line_data.
3104 These are similar to
3107 ERR_peek_error_line_data,
3108 but report on the latest error recorded rather than the first one
3109 still in the error queue.
3110 [Ben Laurie, Bodo Moeller]
3112 *) default_algorithms option in ENGINE config module. This allows things
3114 default_algorithms = ALL
3115 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
3118 *) Prelminary ENGINE config module.
3121 *) New experimental application configuration code.
3124 *) Change the AES code to follow the same name structure as all other
3125 symmetric ciphers, and behave the same way. Move everything to
3126 the directory crypto/aes, thereby obsoleting crypto/rijndael.
3127 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3129 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
3130 [Ben Laurie and Theo de Raadt]
3132 *) Add option to output public keys in req command.
3133 [Massimiliano Pala madwolf@openca.org]
3135 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
3136 (up to about 10% better than before for P-192 and P-224).
3139 *) New functions/macros
3141 SSL_CTX_set_msg_callback(ctx, cb)
3142 SSL_CTX_set_msg_callback_arg(ctx, arg)
3143 SSL_set_msg_callback(ssl, cb)
3144 SSL_set_msg_callback_arg(ssl, arg)
3146 to request calling a callback function
3148 void cb(int write_p, int version, int content_type,
3149 const void *buf, size_t len, SSL *ssl, void *arg)
3151 whenever a protocol message has been completely received
3152 (write_p == 0) or sent (write_p == 1). Here 'version' is the
3153 protocol version according to which the SSL library interprets
3154 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
3155 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
3156 the content type as defined in the SSL 3.0/TLS 1.0 protocol
3157 specification (change_cipher_spec(20), alert(21), handshake(22)).
3158 'buf' and 'len' point to the actual message, 'ssl' to the
3159 SSL object, and 'arg' is the application-defined value set by
3160 SSL[_CTX]_set_msg_callback_arg().
3162 'openssl s_client' and 'openssl s_server' have new '-msg' options
3163 to enable a callback that displays all protocol messages.
3166 *) Change the shared library support so shared libraries are built as
3167 soon as the corresponding static library is finished, and thereby get
3168 openssl and the test programs linked against the shared library.
3169 This still only happens when the keyword "shard" has been given to
3170 the configuration scripts.
3172 NOTE: shared library support is still an experimental thing, and
3173 backward binary compatibility is still not guaranteed.
3174 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
3176 *) Add support for Subject Information Access extension.
3177 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3179 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
3180 additional bytes when new memory had to be allocated, not just
3181 when reusing an existing buffer.
3184 *) New command line and configuration option 'utf8' for the req command.
3185 This allows field values to be specified as UTF8 strings.
3188 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
3189 runs for the former and machine-readable output for the latter.
3192 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
3193 of the e-mail address in the DN (i.e., it will go into a certificate
3194 extension only). The new configuration file option 'email_in_dn = no'
3195 has the same effect.
3196 [Massimiliano Pala madwolf@openca.org]
3198 *) Change all functions with names starting with des_ to be starting
3199 with DES_ instead. Add wrappers that are compatible with libdes,
3200 but are named _ossl_old_des_*. Finally, add macros that map the
3201 des_* symbols to the corresponding _ossl_old_des_* if libdes
3202 compatibility is desired. If OpenSSL 0.9.6c compatibility is
3203 desired, the des_* symbols will be mapped to DES_*, with one
3206 Since we provide two compatibility mappings, the user needs to
3207 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
3208 compatibility is desired. The default (i.e., when that macro
3209 isn't defined) is OpenSSL 0.9.6c compatibility.
3211 There are also macros that enable and disable the support of old
3212 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
3213 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
3214 are defined, the default will apply: to support the old des routines.
3216 In either case, one must include openssl/des.h to get the correct
3217 definitions. Do not try to just include openssl/des_old.h, that
3220 NOTE: This is a major break of an old API into a new one. Software
3221 authors are encouraged to switch to the DES_ style functions. Some
3222 time in the future, des_old.h and the libdes compatibility functions
3223 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
3224 default), and then completely removed.
3227 *) Test for certificates which contain unsupported critical extensions.
3228 If such a certificate is found during a verify operation it is
3229 rejected by default: this behaviour can be overridden by either
3230 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
3231 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
3232 X509_supported_extension() has also been added which returns 1 if a
3233 particular extension is supported.
3236 *) Modify the behaviour of EVP cipher functions in similar way to digests
3237 to retain compatibility with existing code.
3240 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
3241 compatibility with existing code. In particular the 'ctx' parameter does
3242 not have to be to be initialized before the call to EVP_DigestInit() and
3243 it is tidied up after a call to EVP_DigestFinal(). New function
3244 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
3245 EVP_MD_CTX_copy() changed to not require the destination to be
3246 initialized valid and new function EVP_MD_CTX_copy_ex() added which
3247 requires the destination to be valid.
3249 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
3250 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
3253 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
3254 so that complete 'Handshake' protocol structures are kept in memory
3255 instead of overwriting 'msg_type' and 'length' with 'body' data.
3258 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
3259 [Massimo Santin via Richard Levitte]
3261 *) Major restructuring to the underlying ENGINE code. This includes
3262 reduction of linker bloat, separation of pure "ENGINE" manipulation
3263 (initialisation, etc) from functionality dealing with implementations
3264 of specific crypto iterfaces. This change also introduces integrated
3265 support for symmetric ciphers and digest implementations - so ENGINEs
3266 can now accelerate these by providing EVP_CIPHER and EVP_MD
3267 implementations of their own. This is detailed in crypto/engine/README
3268 as it couldn't be adequately described here. However, there are a few
3269 API changes worth noting - some RSA, DSA, DH, and RAND functions that
3270 were changed in the original introduction of ENGINE code have now
3271 reverted back - the hooking from this code to ENGINE is now a good
3272 deal more passive and at run-time, operations deal directly with
3273 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
3274 dereferencing through an ENGINE pointer any more. Also, the ENGINE
3275 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
3276 they were not being used by the framework as there is no concept of a
3277 BIGNUM_METHOD and they could not be generalised to the new
3278 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
3279 ENGINE_cpy() has been removed as it cannot be consistently defined in
3283 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
3286 *) Change mkdef.pl to sort symbols that get the same entry number,
3287 and make sure the automatically generated functions ERR_load_*
3288 become part of libeay.num as well.
3291 *) New function SSL_renegotiate_pending(). This returns true once
3292 renegotiation has been requested (either SSL_renegotiate() call
3293 or HelloRequest/ClientHello receveived from the peer) and becomes
3294 false once a handshake has been completed.
3295 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
3296 sends a HelloRequest, but does not ensure that a handshake takes
3297 place. SSL_renegotiate_pending() is useful for checking if the
3298 client has followed the request.)
3301 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
3302 By default, clients may request session resumption even during
3303 renegotiation (if session ID contexts permit); with this option,
3304 session resumption is possible only in the first handshake.
3306 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
3307 more bits available for options that should not be part of
3308 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
3311 *) Add some demos for certificate and certificate request creation.
3314 *) Make maximum certificate chain size accepted from the peer application
3315 settable (SSL*_get/set_max_cert_list()), as proposed by
3316 "Douglas E. Engert" <deengert@anl.gov>.
3319 *) Add support for shared libraries for Unixware-7
3320 (Boyd Lynn Gerber <gerberb@zenez.com>).
3323 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
3324 be done prior to destruction. Use this to unload error strings from
3325 ENGINEs that load their own error strings. NB: This adds two new API
3326 functions to "get" and "set" this destroy handler in an ENGINE.
3329 *) Alter all existing ENGINE implementations (except "openssl" and
3330 "openbsd") to dynamically instantiate their own error strings. This
3331 makes them more flexible to be built both as statically-linked ENGINEs
3332 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
3333 Also, add stub code to each that makes building them as self-contained
3334 shared-libraries easier (see README.ENGINE).
3337 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
3338 implementations into applications that are completely implemented in
3339 self-contained shared-libraries. The "dynamic" ENGINE exposes control
3340 commands that can be used to configure what shared-library to load and
3341 to control aspects of the way it is handled. Also, made an update to
3342 the README.ENGINE file that brings its information up-to-date and
3343 provides some information and instructions on the "dynamic" ENGINE
3344 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
3347 *) Make it possible to unload ranges of ERR strings with a new
3348 "ERR_unload_strings" function.
3351 *) Add a copy() function to EVP_MD.
3354 *) Make EVP_MD routines take a context pointer instead of just the
3355 md_data void pointer.
3358 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
3359 that the digest can only process a single chunk of data
3360 (typically because it is provided by a piece of
3361 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
3362 is only going to provide a single chunk of data, and hence the
3363 framework needn't accumulate the data for oneshot drivers.
3366 *) As with "ERR", make it possible to replace the underlying "ex_data"
3367 functions. This change also alters the storage and management of global
3368 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
3369 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
3370 index counters. The API functions that use this state have been changed
3371 to take a "class_index" rather than pointers to the class's local STACK
3372 and counter, and there is now an API function to dynamically create new
3373 classes. This centralisation allows us to (a) plug a lot of the
3374 thread-safety problems that existed, and (b) makes it possible to clean
3375 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
3376 such data would previously have always leaked in application code and
3377 workarounds were in place to make the memory debugging turn a blind eye
3378 to it. Application code that doesn't use this new function will still
3379 leak as before, but their memory debugging output will announce it now
3380 rather than letting it slide.
3382 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
3383 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
3384 has a return value to indicate success or failure.
3387 *) Make it possible to replace the underlying "ERR" functions such that the
3388 global state (2 LHASH tables and 2 locks) is only used by the "default"
3389 implementation. This change also adds two functions to "get" and "set"
3390 the implementation prior to it being automatically set the first time
3391 any other ERR function takes place. Ie. an application can call "get",
3392 pass the return value to a module it has just loaded, and that module
3393 can call its own "set" function using that value. This means the
3394 module's "ERR" operations will use (and modify) the error state in the
3395 application and not in its own statically linked copy of OpenSSL code.
3398 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
3399 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
3400 the operation, and provides a more encapsulated way for external code
3401 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
3402 to use these functions rather than manually incrementing the counts.
3404 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
3407 *) Add EVP test program.
3410 *) Add symmetric cipher support to ENGINE. Expect the API to change!
3413 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
3414 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
3415 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
3416 These allow a CRL to be built without having to access X509_CRL fields
3417 directly. Modify 'ca' application to use new functions.
3420 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
3421 bug workarounds. Rollback attack detection is a security feature.
3422 The problem will only arise on OpenSSL servers when TLSv1 is not
3423 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
3424 Software authors not wanting to support TLSv1 will have special reasons
3425 for their choice and can explicitly enable this option.
3426 [Bodo Moeller, Lutz Jaenicke]
3428 *) Rationalise EVP so it can be extended: don't include a union of
3429 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
3430 (similar to those existing for EVP_CIPHER_CTX).
3435 EVP_MD_CTX_init(&md); /* new function call */
3436 EVP_DigestInit(&md, EVP_sha1());
3437 EVP_DigestUpdate(&md, in, len);
3438 EVP_DigestFinal(&md, out, NULL);
3439 EVP_MD_CTX_cleanup(&md); /* new function call */
3443 *) Make DES key schedule conform to the usual scheme, as well as
3444 correcting its structure. This means that calls to DES functions
3445 now have to pass a pointer to a des_key_schedule instead of a
3446 plain des_key_schedule (which was actually always a pointer
3449 des_key_schedule ks;
3451 des_set_key_checked(..., &ks);
3452 des_ncbc_encrypt(..., &ks, ...);
3454 (Note that a later change renames 'des_...' into 'DES_...'.)
3457 *) Initial reduction of linker bloat: the use of some functions, such as
3458 PEM causes large amounts of unused functions to be linked in due to
3459 poor organisation. For example pem_all.c contains every PEM function
3460 which has a knock on effect of linking in large amounts of (unused)
3461 ASN1 code. Grouping together similar functions and splitting unrelated
3462 functions prevents this.
3465 *) Cleanup of EVP macros.
3468 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
3469 correct _ecb suffix.
3472 *) Add initial OCSP responder support to ocsp application. The
3473 revocation information is handled using the text based index
3474 use by the ca application. The responder can either handle
3475 requests generated internally, supplied in files (for example
3476 via a CGI script) or using an internal minimal server.
3479 *) Add configuration choices to get zlib compression for TLS.
3482 *) Changes to Kerberos SSL for RFC 2712 compliance:
3483 1. Implemented real KerberosWrapper, instead of just using
3484 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
3485 2. Implemented optional authenticator field of KerberosWrapper.
3487 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
3488 and authenticator structs; see crypto/krb5/.
3490 Generalized Kerberos calls to support multiple Kerberos libraries.
3491 [Vern Staats <staatsvr@asc.hpc.mil>,
3492 Jeffrey Altman <jaltman@columbia.edu>
3493 via Richard Levitte]
3495 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
3496 already does with RSA. testdsa.h now has 'priv_key/pub_key'
3497 values for each of the key sizes rather than having just
3498 parameters (and 'speed' generating keys each time).
3501 *) Speed up EVP routines.
3504 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
3505 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
3506 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
3507 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
3509 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
3510 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
3511 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
3514 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
3516 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
3519 *) Added the OS2-EMX target.
3520 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
3522 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
3523 to support NCONF routines in extension code. New function CONF_set_nconf()
3524 to allow functions which take an NCONF to also handle the old LHASH
3525 structure: this means that the old CONF compatible routines can be
3526 retained (in particular wrt extensions) without having to duplicate the
3527 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
3530 *) Enhance the general user interface with mechanisms for inner control
3531 and with possibilities to have yes/no kind of prompts.
3534 *) Change all calls to low level digest routines in the library and
3535 applications to use EVP. Add missing calls to HMAC_cleanup() and
3536 don't assume HMAC_CTX can be copied using memcpy().
3537 [Verdon Walker <VWalker@novell.com>, Steve Henson]
3539 *) Add the possibility to control engines through control names but with
3540 arbitrary arguments instead of just a string.
3541 Change the key loaders to take a UI_METHOD instead of a callback
3542 function pointer. NOTE: this breaks binary compatibility with earlier
3543 versions of OpenSSL [engine].
3544 Adapt the nCipher code for these new conditions and add a card insertion
3548 *) Enhance the general user interface with mechanisms to better support
3549 dialog box interfaces, application-defined prompts, the possibility
3550 to use defaults (for example default passwords from somewhere else)
3551 and interrupts/cancellations.
3554 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
3555 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
3558 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
3559 tidy up some unnecessarily weird code in 'sk_new()').
3560 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
3562 *) Change the key loading routines for ENGINEs to use the same kind
3563 callback (pem_password_cb) as all other routines that need this
3567 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
3568 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
3569 than this minimum value is recommended.
3572 *) New random seeder for OpenVMS, using the system process statistics
3573 that are easily reachable.
3576 *) Windows apparently can't transparently handle global
3577 variables defined in DLLs. Initialisations such as:
3579 const ASN1_ITEM *it = &ASN1_INTEGER_it;
3581 wont compile. This is used by the any applications that need to
3582 declare their own ASN1 modules. This was fixed by adding the option
3583 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
3584 needed for static libraries under Win32.
3587 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
3588 setting of purpose and trust fields. New X509_STORE trust and
3589 purpose functions and tidy up setting in other SSL functions.
3592 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
3593 structure. These are inherited by X509_STORE_CTX when it is
3594 initialised. This allows various defaults to be set in the
3595 X509_STORE structure (such as flags for CRL checking and custom
3596 purpose or trust settings) for functions which only use X509_STORE_CTX
3597 internally such as S/MIME.
3599 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
3600 trust settings if they are not set in X509_STORE. This allows X509_STORE
3601 purposes and trust (in S/MIME for example) to override any set by default.
3603 Add command line options for CRL checking to smime, s_client and s_server
3607 *) Initial CRL based revocation checking. If the CRL checking flag(s)
3608 are set then the CRL is looked up in the X509_STORE structure and
3609 its validity and signature checked, then if the certificate is found
3610 in the CRL the verify fails with a revoked error.
3612 Various new CRL related callbacks added to X509_STORE_CTX structure.
3614 Command line options added to 'verify' application to support this.
3616 This needs some additional work, such as being able to handle multiple
3617 CRLs with different times, extension based lookup (rather than just
3618 by subject name) and ultimately more complete V2 CRL extension
3622 *) Add a general user interface API (crypto/ui/). This is designed
3623 to replace things like des_read_password and friends (backward
3624 compatibility functions using this new API are provided).
3625 The purpose is to remove prompting functions from the DES code
3626 section as well as provide for prompting through dialog boxes in
3627 a window system and the like.
3630 *) Add "ex_data" support to ENGINE so implementations can add state at a
3631 per-structure level rather than having to store it globally.
3634 *) Make it possible for ENGINE structures to be copied when retrieved by
3635 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
3636 This causes the "original" ENGINE structure to act like a template,
3637 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
3638 operational state can be localised to each ENGINE structure, despite the
3639 fact they all share the same "methods". New ENGINE structures returned in
3640 this case have no functional references and the return value is the single
3641 structural reference. This matches the single structural reference returned
3642 by ENGINE_by_id() normally, when it is incremented on the pre-existing
3646 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
3647 needs to match any other type at all we need to manually clear the
3651 *) Changes to the "openssl engine" utility to include;
3652 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
3653 about an ENGINE's available control commands.
3654 - executing control commands from command line arguments using the
3655 '-pre' and '-post' switches. '-post' is only used if '-t' is
3656 specified and the ENGINE is successfully initialised. The syntax for
3657 the individual commands are colon-separated, for example;
3658 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
3661 *) New dynamic control command support for ENGINEs. ENGINEs can now
3662 declare their own commands (numbers), names (strings), descriptions,
3663 and input types for run-time discovery by calling applications. A
3664 subset of these commands are implicitly classed as "executable"
3665 depending on their input type, and only these can be invoked through
3666 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
3667 can be based on user input, config files, etc). The distinction is
3668 that "executable" commands cannot return anything other than a boolean
3669 result and can only support numeric or string input, whereas some
3670 discoverable commands may only be for direct use through
3671 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
3672 pointers, or other custom uses. The "executable" commands are to
3673 support parameterisations of ENGINE behaviour that can be
3674 unambiguously defined by ENGINEs and used consistently across any
3675 OpenSSL-based application. Commands have been added to all the
3676 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
3677 control over shared-library paths without source code alterations.
3680 *) Changed all ENGINE implementations to dynamically allocate their
3681 ENGINEs rather than declaring them statically. Apart from this being
3682 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
3683 this also allows the implementations to compile without using the
3684 internal engine_int.h header.
3687 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
3688 'const' value. Any code that should be able to modify a RAND_METHOD
3689 should already have non-const pointers to it (ie. they should only
3690 modify their own ones).
3693 *) Made a variety of little tweaks to the ENGINE code.
3694 - "atalla" and "ubsec" string definitions were moved from header files
3695 to C code. "nuron" string definitions were placed in variables
3696 rather than hard-coded - allowing parameterisation of these values
3697 later on via ctrl() commands.
3698 - Removed unused "#if 0"'d code.
3699 - Fixed engine list iteration code so it uses ENGINE_free() to release
3700 structural references.
3701 - Constified the RAND_METHOD element of ENGINE structures.
3702 - Constified various get/set functions as appropriate and added
3703 missing functions (including a catch-all ENGINE_cpy that duplicates
3704 all ENGINE values onto a new ENGINE except reference counts/state).
3705 - Removed NULL parameter checks in get/set functions. Setting a method
3706 or function to NULL is a way of cancelling out a previously set
3707 value. Passing a NULL ENGINE parameter is just plain stupid anyway
3708 and doesn't justify the extra error symbols and code.
3709 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
3710 flags from engine_int.h to engine.h.
3711 - Changed prototypes for ENGINE handler functions (init(), finish(),
3712 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
3715 *) Implement binary inversion algorithm for BN_mod_inverse in addition
3716 to the algorithm using long division. The binary algorithm can be
3717 used only if the modulus is odd. On 32-bit systems, it is faster
3718 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
3719 roughly 5-15% for 256-bit moduli), so we use it only for moduli
3720 up to 450 bits. In 64-bit environments, the binary algorithm
3721 appears to be advantageous for much longer moduli; here we use it
3722 for moduli up to 2048 bits.
3725 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
3726 could not support the combine flag in choice fields.
3729 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
3730 extensions from a certificate request to the certificate.
3733 *) Allow multiple 'certopt' and 'nameopt' options to be separated
3734 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
3735 file: this allows the display of the certificate about to be
3736 signed to be customised, to allow certain fields to be included
3737 or excluded and extension details. The old system didn't display
3738 multicharacter strings properly, omitted fields not in the policy
3739 and couldn't display additional details such as extensions.
3742 *) Function EC_POINTs_mul for multiple scalar multiplication
3743 of an arbitrary number of elliptic curve points
3744 \sum scalars[i]*points[i],
3745 optionally including the generator defined for the EC_GROUP:
3746 scalar*generator + \sum scalars[i]*points[i].
3748 EC_POINT_mul is a simple wrapper function for the typical case
3749 that the point list has just one item (besides the optional
3753 *) First EC_METHODs for curves over GF(p):
3755 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
3756 operations and provides various method functions that can also
3757 operate with faster implementations of modular arithmetic.
3759 EC_GFp_mont_method() reuses most functions that are part of
3760 EC_GFp_simple_method, but uses Montgomery arithmetic.
3762 [Bodo Moeller; point addition and point doubling
3763 implementation directly derived from source code provided by
3764 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
3766 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
3767 crypto/ec/ec_lib.c):
3769 Curves are EC_GROUP objects (with an optional group generator)
3770 based on EC_METHODs that are built into the library.
3772 Points are EC_POINT objects based on EC_GROUP objects.
3774 Most of the framework would be able to handle curves over arbitrary
3775 finite fields, but as there are no obvious types for fields other
3776 than GF(p), some functions are limited to that for now.
3779 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
3780 that the file contains a complete HTTP response.
3783 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
3784 change the def and num file printf format specifier from "%-40sXXX"
3785 to "%-39s XXX". The latter will always guarantee a space after the
3786 field while the former will cause them to run together if the field
3787 is 40 of more characters long.
3790 *) Constify the cipher and digest 'method' functions and structures
3791 and modify related functions to take constant EVP_MD and EVP_CIPHER
3795 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
3796 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
3799 *) Modify EVP_Digest*() routines so they now return values. Although the
3800 internal software routines can never fail additional hardware versions
3804 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
3806 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
3807 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
3810 ERR_R_NESTED_ASN1_ERROR
3812 ERR_R_MISSING_ASN1_EOS
3813 were 4 .. 9, conflicting with
3814 ERR_LIB_RSA (= ERR_R_RSA_LIB)
3816 ERR_LIB_PEM (= ERR_R_PEM_LIB).
3817 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
3819 Add new error code 'ERR_R_INTERNAL_ERROR'.
3822 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
3826 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
3827 sets the subject name for a new request or supersedes the
3828 subject name in a given request. Formats that can be parsed are
3829 'CN=Some Name, OU=myOU, C=IT'
3831 'CN=Some Name/OU=myOU/C=IT'.
3833 Add options '-batch' and '-verbose' to 'openssl req'.
3834 [Massimiliano Pala <madwolf@hackmasters.net>]
3836 *) Introduce the possibility to access global variables through
3837 functions on platform were that's the best way to handle exporting
3838 global variables in shared libraries. To enable this functionality,
3839 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
3840 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
3841 is normally done by Configure or something similar).
3843 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
3844 in the source file (foo.c) like this:
3846 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
3847 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
3849 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
3850 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
3852 OPENSSL_DECLARE_GLOBAL(int,foo);
3853 #define foo OPENSSL_GLOBAL_REF(foo)
3854 OPENSSL_DECLARE_GLOBAL(double,bar);
3855 #define bar OPENSSL_GLOBAL_REF(bar)
3857 The #defines are very important, and therefore so is including the
3858 header file everywhere where the defined globals are used.
3860 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
3861 of ASN.1 items, but that structure is a bit different.
3863 The largest change is in util/mkdef.pl which has been enhanced with
3864 better and easier to understand logic to choose which symbols should
3865 go into the Windows .def files as well as a number of fixes and code
3866 cleanup (among others, algorithm keywords are now sorted
3867 lexicographically to avoid constant rewrites).
3870 *) In BN_div() keep a copy of the sign of 'num' before writing the
3871 result to 'rm' because if rm==num the value will be overwritten
3872 and produce the wrong result if 'num' is negative: this caused
3873 problems with BN_mod() and BN_nnmod().
3876 *) Function OCSP_request_verify(). This checks the signature on an
3877 OCSP request and verifies the signer certificate. The signer
3878 certificate is just checked for a generic purpose and OCSP request
3882 *) Add OCSP_check_validity() function to check the validity of OCSP
3883 responses. OCSP responses are prepared in real time and may only
3884 be a few seconds old. Simply checking that the current time lies
3885 between thisUpdate and nextUpdate max reject otherwise valid responses
3886 caused by either OCSP responder or client clock inaccuracy. Instead
3887 we allow thisUpdate and nextUpdate to fall within a certain period of
3888 the current time. The age of the response can also optionally be
3889 checked. Two new options -validity_period and -status_age added to
3893 *) If signature or public key algorithm is unrecognized print out its
3894 OID rather that just UNKNOWN.
3897 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
3898 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
3899 ID to be generated from the issuer certificate alone which can then be
3900 passed to OCSP_id_issuer_cmp().
3903 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
3904 ASN1 modules to export functions returning ASN1_ITEM pointers
3905 instead of the ASN1_ITEM structures themselves. This adds several
3906 new macros which allow the underlying ASN1 function/structure to
3907 be accessed transparently. As a result code should not use ASN1_ITEM
3908 references directly (such as &X509_it) but instead use the relevant
3909 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
3910 use of the new ASN1 code on platforms where exporting structures
3911 is problematical (for example in shared libraries) but exporting
3912 functions returning pointers to structures is not.
3915 *) Add support for overriding the generation of SSL/TLS session IDs.
3916 These callbacks can be registered either in an SSL_CTX or per SSL.
3917 The purpose of this is to allow applications to control, if they wish,
3918 the arbitrary values chosen for use as session IDs, particularly as it
3919 can be useful for session caching in multiple-server environments. A
3920 command-line switch for testing this (and any client code that wishes
3921 to use such a feature) has been added to "s_server".
3922 [Geoff Thorpe, Lutz Jaenicke]
3924 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
3925 of the form '#if defined(...) || defined(...) || ...' and
3926 '#if !defined(...) && !defined(...) && ...'. This also avoids
3927 the growing number of special cases it was previously handling.
3930 *) Make all configuration macros available for application by making
3931 sure they are available in opensslconf.h, by giving them names starting
3932 with "OPENSSL_" to avoid conflicts with other packages and by making
3933 sure e_os2.h will cover all platform-specific cases together with
3935 Additionally, it is now possible to define configuration/platform-
3936 specific names (called "system identities"). In the C code, these
3937 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
3938 macro with the name beginning with "OPENSSL_SYS_", which is determined
3939 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
3943 *) New option -set_serial to 'req' and 'x509' this allows the serial
3944 number to use to be specified on the command line. Previously self
3945 signed certificates were hard coded with serial number 0 and the
3946 CA options of 'x509' had to use a serial number in a file which was
3950 *) New options to 'ca' utility to support V2 CRL entry extensions.
3951 Currently CRL reason, invalidity date and hold instruction are
3952 supported. Add new CRL extensions to V3 code and some new objects.
3955 *) New function EVP_CIPHER_CTX_set_padding() this is used to
3956 disable standard block padding (aka PKCS#5 padding) in the EVP
3957 API, which was previously mandatory. This means that the data is
3958 not padded in any way and so the total length much be a multiple
3959 of the block size, otherwise an error occurs.
3962 *) Initial (incomplete) OCSP SSL support.
3965 *) New function OCSP_parse_url(). This splits up a URL into its host,
3966 port and path components: primarily to parse OCSP URLs. New -url
3967 option to ocsp utility.
3970 *) New nonce behavior. The return value of OCSP_check_nonce() now
3971 reflects the various checks performed. Applications can decide
3972 whether to tolerate certain situations such as an absent nonce
3973 in a response when one was present in a request: the ocsp application
3974 just prints out a warning. New function OCSP_add1_basic_nonce()
3975 this is to allow responders to include a nonce in a response even if
3976 the request is nonce-less.
3979 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
3980 skipped when using openssl x509 multiple times on a single input file,
3981 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
3984 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
3985 set string type: to handle setting ASN1_TIME structures. Fix ca
3986 utility to correctly initialize revocation date of CRLs.
3989 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
3990 the clients preferred ciphersuites and rather use its own preferences.
3991 Should help to work around M$ SGC (Server Gated Cryptography) bug in
3992 Internet Explorer by ensuring unchanged hash method during stepup.
3993 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
3996 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
3997 to aes and add a new 'exist' option to print out symbols that don't
4001 *) Additional options to ocsp utility to allow flags to be set and
4002 additional certificates supplied.
4005 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
4006 OCSP client a number of certificate to only verify the response
4010 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
4011 handle the new API. Currently only ECB, CBC modes supported. Add new
4014 Add TLS AES ciphersuites as described in RFC3268, "Advanced
4015 Encryption Standard (AES) Ciphersuites for Transport Layer
4016 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
4017 not enabled by default and were not part of the "ALL" ciphersuite
4018 alias because they were not yet official; they could be
4019 explicitly requested by specifying the "AESdraft" ciphersuite
4020 group alias. In the final release of OpenSSL 0.9.7, the group
4021 alias is called "AES" and is part of "ALL".)
4022 [Ben Laurie, Steve Henson, Bodo Moeller]
4024 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
4025 request to response.
4028 *) Functions for OCSP responders. OCSP_request_onereq_count(),
4029 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
4030 extract information from a certificate request. OCSP_response_create()
4031 creates a response and optionally adds a basic response structure.
4032 OCSP_basic_add1_status() adds a complete single response to a basic
4033 response and returns the OCSP_SINGLERESP structure just added (to allow
4034 extensions to be included for example). OCSP_basic_add1_cert() adds a
4035 certificate to a basic response and OCSP_basic_sign() signs a basic
4036 response with various flags. New helper functions ASN1_TIME_check()
4037 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
4038 (converts ASN1_TIME to GeneralizedTime).
4041 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
4042 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
4043 structure from a certificate. X509_pubkey_digest() digests the public_key
4044 contents: this is used in various key identifiers.
4047 *) Make sk_sort() tolerate a NULL argument.
4048 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
4050 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
4051 passed by the function are trusted implicitly. If any of them signed the
4052 response then it is assumed to be valid and is not verified.
4055 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
4056 to data. This was previously part of the PKCS7 ASN1 code. This
4057 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
4058 [Steve Henson, reported by Kenneth R. Robinette
4059 <support@securenetterm.com>]
4061 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
4062 routines: without these tracing memory leaks is very painful.
4063 Fix leaks in PKCS12 and PKCS7 routines.
4066 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
4067 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
4068 effectively meant GeneralizedTime would never be used. Now it
4069 is initialised to -1 but X509_time_adj() now has to check the value
4070 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
4071 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
4072 [Steve Henson, reported by Kenneth R. Robinette
4073 <support@securenetterm.com>]
4075 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
4076 result in a zero length in the ASN1_INTEGER structure which was
4077 not consistent with the structure when d2i_ASN1_INTEGER() was used
4078 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
4079 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
4080 where it did not print out a minus for negative ASN1_INTEGER.
4083 *) Add summary printout to ocsp utility. The various functions which
4084 convert status values to strings have been renamed to:
4085 OCSP_response_status_str(), OCSP_cert_status_str() and
4086 OCSP_crl_reason_str() and are no longer static. New options
4087 to verify nonce values and to disable verification. OCSP response
4088 printout format cleaned up.
4091 *) Add additional OCSP certificate checks. These are those specified
4092 in RFC2560. This consists of two separate checks: the CA of the
4093 certificate being checked must either be the OCSP signer certificate
4094 or the issuer of the OCSP signer certificate. In the latter case the
4095 OCSP signer certificate must contain the OCSP signing extended key
4096 usage. This check is performed by attempting to match the OCSP
4097 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
4098 in the OCSP_CERTID structures of the response.
4101 *) Initial OCSP certificate verification added to OCSP_basic_verify()
4102 and related routines. This uses the standard OpenSSL certificate
4103 verify routines to perform initial checks (just CA validity) and
4104 to obtain the certificate chain. Then additional checks will be
4105 performed on the chain. Currently the root CA is checked to see
4106 if it is explicitly trusted for OCSP signing. This is used to set
4107 a root CA as a global signing root: that is any certificate that
4108 chains to that CA is an acceptable OCSP signing certificate.
4111 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
4112 extensions from a separate configuration file.
4113 As when reading extensions from the main configuration file,
4114 the '-extensions ...' option may be used for specifying the
4116 [Massimiliano Pala <madwolf@comune.modena.it>]
4118 *) New OCSP utility. Allows OCSP requests to be generated or
4119 read. The request can be sent to a responder and the output
4120 parsed, outputed or printed in text form. Not complete yet:
4121 still needs to check the OCSP response validity.
4124 *) New subcommands for 'openssl ca':
4125 'openssl ca -status <serial>' prints the status of the cert with
4126 the given serial number (according to the index file).
4127 'openssl ca -updatedb' updates the expiry status of certificates
4129 [Massimiliano Pala <madwolf@comune.modena.it>]
4131 *) New '-newreq-nodes' command option to CA.pl. This is like
4132 '-newreq', but calls 'openssl req' with the '-nodes' option
4133 so that the resulting key is not encrypted.
4134 [Damien Miller <djm@mindrot.org>]
4136 *) New configuration for the GNU Hurd.
4137 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
4139 *) Initial code to implement OCSP basic response verify. This
4140 is currently incomplete. Currently just finds the signer's
4141 certificate and verifies the signature on the response.
4144 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
4145 value of OPENSSLDIR. This is available via the new '-d' option
4146 to 'openssl version', and is also included in 'openssl version -a'.
4149 *) Allowing defining memory allocation callbacks that will be given
4150 file name and line number information in additional arguments
4151 (a const char* and an int). The basic functionality remains, as
4152 well as the original possibility to just replace malloc(),
4153 realloc() and free() by functions that do not know about these
4154 additional arguments. To register and find out the current
4155 settings for extended allocation functions, the following
4156 functions are provided:
4158 CRYPTO_set_mem_ex_functions
4159 CRYPTO_set_locked_mem_ex_functions
4160 CRYPTO_get_mem_ex_functions
4161 CRYPTO_get_locked_mem_ex_functions
4163 These work the same way as CRYPTO_set_mem_functions and friends.
4164 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
4165 extended allocation function is enabled.
4166 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
4167 a conventional allocation function is enabled.
4168 [Richard Levitte, Bodo Moeller]
4170 *) Finish off removing the remaining LHASH function pointer casts.
4171 There should no longer be any prototype-casting required when using
4172 the LHASH abstraction, and any casts that remain are "bugs". See
4173 the callback types and macros at the head of lhash.h for details
4174 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
4177 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
4178 If /dev/[u]random devices are not available or do not return enough
4179 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
4181 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
4182 /etc/entropy will be queried once each in this sequence, quering stops
4183 when enough entropy was collected without querying more sockets.
4186 *) Change the Unix RAND_poll() variant to be able to poll several
4187 random devices, as specified by DEVRANDOM, until a sufficient amount
4188 of data has been collected. We spend at most 10 ms on each file
4189 (select timeout) and read in non-blocking mode. DEVRANDOM now
4190 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
4191 (previously it was just the string "/dev/urandom"), so on typical
4192 platforms the 10 ms delay will never occur.
4193 Also separate out the Unix variant to its own file, rand_unix.c.
4194 For VMS, there's a currently-empty rand_vms.c.
4197 *) Move OCSP client related routines to ocsp_cl.c. These
4198 provide utility functions which an application needing
4199 to issue a request to an OCSP responder and analyse the
4200 response will typically need: as opposed to those which an
4201 OCSP responder itself would need which will be added later.
4203 OCSP_request_sign() signs an OCSP request with an API similar
4204 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
4205 response. OCSP_response_get1_basic() extracts basic response
4206 from response. OCSP_resp_find_status(): finds and extracts status
4207 information from an OCSP_CERTID structure (which will be created
4208 when the request structure is built). These are built from lower
4209 level functions which work on OCSP_SINGLERESP structures but
4210 wont normally be used unless the application wishes to examine
4211 extensions in the OCSP response for example.
4213 Replace nonce routines with a pair of functions.
4214 OCSP_request_add1_nonce() adds a nonce value and optionally
4215 generates a random value. OCSP_check_nonce() checks the
4216 validity of the nonce in an OCSP response.
4219 *) Change function OCSP_request_add() to OCSP_request_add0_id().
4220 This doesn't copy the supplied OCSP_CERTID and avoids the
4221 need to free up the newly created id. Change return type
4222 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
4223 This can then be used to add extensions to the request.
4224 Deleted OCSP_request_new(), since most of its functionality
4225 is now in OCSP_REQUEST_new() (and the case insensitive name
4226 clash) apart from the ability to set the request name which
4227 will be added elsewhere.
4230 *) Update OCSP API. Remove obsolete extensions argument from
4231 various functions. Extensions are now handled using the new
4232 OCSP extension code. New simple OCSP HTTP function which
4233 can be used to send requests and parse the response.
4236 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
4237 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
4238 uses the special reorder version of SET OF to sort the attributes
4239 and reorder them to match the encoded order. This resolves a long
4240 standing problem: a verify on a PKCS7 structure just after signing
4241 it used to fail because the attribute order did not match the
4242 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
4243 it uses the received order. This is necessary to tolerate some broken
4244 software that does not order SET OF. This is handled by encoding
4245 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
4246 to produce the required SET OF.
4249 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
4250 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
4251 files to get correct declarations of the ASN.1 item variables.
4254 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
4255 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
4256 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
4257 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
4258 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
4259 ASN1_ITEM and no wrapper functions.
4262 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
4263 replace the old function pointer based I/O routines. Change most of
4264 the *_d2i_bio() and *_d2i_fp() functions to use these.
4267 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
4268 lines, recognice more "algorithms" that can be deselected, and make
4269 it complain about algorithm deselection that isn't recognised.
4272 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
4273 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
4274 to use new functions. Add NO_ASN1_OLD which can be set to remove
4275 some old style ASN1 functions: this can be used to determine if old
4276 code will still work when these eventually go away.
4279 *) New extension functions for OCSP structures, these follow the
4280 same conventions as certificates and CRLs.
4283 *) New function X509V3_add1_i2d(). This automatically encodes and
4284 adds an extension. Its behaviour can be customised with various
4285 flags to append, replace or delete. Various wrappers added for
4286 certifcates and CRLs.
4289 *) Fix to avoid calling the underlying ASN1 print routine when
4290 an extension cannot be parsed. Correct a typo in the
4291 OCSP_SERVICELOC extension. Tidy up print OCSP format.
4294 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
4295 entries for variables.
4298 *) Add functionality to apps/openssl.c for detecting locking
4299 problems: As the program is single-threaded, all we have
4300 to do is register a locking callback using an array for
4301 storing which locks are currently held by the program.
4304 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
4305 SSL_get_ex_data_X509_STORE_idx(), which is used in
4306 ssl_verify_cert_chain() and thus can be called at any time
4307 during TLS/SSL handshakes so that thread-safety is essential.
4308 Unfortunately, the ex_data design is not at all suited
4309 for multi-threaded use, so it probably should be abolished.
4312 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
4313 [Broadcom, tweaked and integrated by Geoff Thorpe]
4315 *) Move common extension printing code to new function
4316 X509V3_print_extensions(). Reorganise OCSP print routines and
4317 implement some needed OCSP ASN1 functions. Add OCSP extensions.
4320 *) New function X509_signature_print() to remove duplication in some
4324 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
4325 set (this was treated exactly the same as SET OF previously). This
4326 is used to reorder the STACK representing the structure to match the
4327 encoding. This will be used to get round a problem where a PKCS7
4328 structure which was signed could not be verified because the STACK
4329 order did not reflect the encoded order.
4332 *) Reimplement the OCSP ASN1 module using the new code.
4335 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
4336 for its ASN1 operations. The old style function pointers still exist
4337 for now but they will eventually go away.
4340 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
4341 completely replaces the old ASN1 functionality with a table driven
4342 encoder and decoder which interprets an ASN1_ITEM structure describing
4343 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
4344 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
4345 has also been converted to the new form.
4348 *) Change BN_mod_exp_recp so that negative moduli are tolerated
4349 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
4350 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
4351 for negative moduli.
4354 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
4355 of not touching the result's sign bit.
4358 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
4362 *) Changed the LHASH code to use prototypes for callbacks, and created
4363 macros to declare and implement thin (optionally static) functions
4364 that provide type-safety and avoid function pointer casting for the
4365 type-specific callbacks.
4368 *) Added Kerberos Cipher Suites to be used with TLS, as written in
4370 [Veers Staats <staatsvr@asc.hpc.mil>,
4371 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
4373 *) Reformat the FAQ so the different questions and answers can be divided
4374 in sections depending on the subject.
4377 *) Have the zlib compression code load ZLIB.DLL dynamically under
4381 *) New function BN_mod_sqrt for computing square roots modulo a prime
4382 (using the probabilistic Tonelli-Shanks algorithm unless
4383 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
4384 be handled deterministically).
4385 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
4387 *) Make BN_mod_inverse faster by explicitly handling small quotients
4388 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
4389 512 bits], about 30% for larger ones [1024 or 2048 bits].)
4392 *) New function BN_kronecker.
4395 *) Fix BN_gcd so that it works on negative inputs; the result is
4396 positive unless both parameters are zero.
4397 Previously something reasonably close to an infinite loop was
4398 possible because numbers could be growing instead of shrinking
4399 in the implementation of Euclid's algorithm.
4402 *) Fix BN_is_word() and BN_is_one() macros to take into account the
4403 sign of the number in question.
4405 Fix BN_is_word(a,w) to work correctly for w == 0.
4407 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
4408 because its test if the absolute value of 'a' equals 'w'.
4409 Note that BN_abs_is_word does *not* handle w == 0 reliably;
4410 it exists mostly for use in the implementations of BN_is_zero(),
4411 BN_is_one(), and BN_is_word().
4414 *) New function BN_swap.
4417 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
4418 the exponentiation functions are more likely to produce reasonable
4419 results on negative inputs.
4422 *) Change BN_mod_mul so that the result is always non-negative.
4423 Previously, it could be negative if one of the factors was negative;
4424 I don't think anyone really wanted that behaviour.
4427 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
4428 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
4429 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
4430 and add new functions:
4439 BN_mod_lshift1_quick
4443 These functions always generate non-negative results.
4445 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
4446 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
4448 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
4449 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
4450 be reduced modulo m.
4451 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
4454 The following entry accidentily appeared in the CHANGES file
4455 distributed with OpenSSL 0.9.7. The modifications described in
4456 it do *not* apply to OpenSSL 0.9.7.
4458 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4459 was actually never needed) and in BN_mul(). The removal in BN_mul()
4460 required a small change in bn_mul_part_recursive() and the addition
4461 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4462 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4463 bn_sub_words() and bn_add_words() except they take arrays with
4468 *) In 'openssl passwd', verify passwords read from the terminal
4469 unless the '-salt' option is used (which usually means that
4470 verification would just waste user's time since the resulting
4471 hash is going to be compared with some given password hash)
4472 or the new '-noverify' option is used.
4474 This is an incompatible change, but it does not affect
4475 non-interactive use of 'openssl passwd' (passwords on the command
4476 line, '-stdin' option, '-in ...' option) and thus should not
4480 *) Remove all references to RSAref, since there's no more need for it.
4483 *) Make DSO load along a path given through an environment variable
4484 (SHLIB_PATH) with shl_load().
4487 *) Constify the ENGINE code as a result of BIGNUM constification.
4488 Also constify the RSA code and most things related to it. In a
4489 few places, most notable in the depth of the ASN.1 code, ugly
4490 casts back to non-const were required (to be solved at a later
4494 *) Make it so the openssl application has all engines loaded by default.
4497 *) Constify the BIGNUM routines a little more.
4500 *) Add the following functions:
4502 ENGINE_load_cswift()
4504 ENGINE_load_atalla()
4506 ENGINE_load_builtin_engines()
4508 That way, an application can itself choose if external engines that
4509 are built-in in OpenSSL shall ever be used or not. The benefit is
4510 that applications won't have to be linked with libdl or other dso
4511 libraries unless it's really needed.
4513 Changed 'openssl engine' to load all engines on demand.
4514 Changed the engine header files to avoid the duplication of some
4515 declarations (they differed!).
4518 *) 'openssl engine' can now list capabilities.
4521 *) Better error reporting in 'openssl engine'.
4524 *) Never call load_dh_param(NULL) in s_server.
4527 *) Add engine application. It can currently list engines by name and
4528 identity, and test if they are actually available.
4531 *) Improve RPM specification file by forcing symbolic linking and making
4532 sure the installed documentation is also owned by root.root.
4533 [Damien Miller <djm@mindrot.org>]
4535 *) Give the OpenSSL applications more possibilities to make use of
4536 keys (public as well as private) handled by engines.
4539 *) Add OCSP code that comes from CertCo.
4542 *) Add VMS support for the Rijndael code.
4545 *) Added untested support for Nuron crypto accelerator.
4548 *) Add support for external cryptographic devices. This code was
4549 previously distributed separately as the "engine" branch.
4550 [Geoff Thorpe, Richard Levitte]
4552 *) Rework the filename-translation in the DSO code. It is now possible to
4553 have far greater control over how a "name" is turned into a filename
4554 depending on the operating environment and any oddities about the
4555 different shared library filenames on each system.
4558 *) Support threads on FreeBSD-elf in Configure.
4561 *) Fix for SHA1 assembly problem with MASM: it produces
4562 warnings about corrupt line number information when assembling
4563 with debugging information. This is caused by the overlapping
4565 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
4568 NCONF_get_number() has no error checking at all. As a replacement,
4569 NCONF_get_number_e() is defined (_e for "error checking") and is
4570 promoted strongly. The old NCONF_get_number is kept around for
4571 binary backward compatibility.
4572 Make it possible for methods to load from something other than a BIO,
4573 by providing a function pointer that is given a name instead of a BIO.
4574 For example, this could be used to load configuration data from an
4578 *) Fix for non blocking accept BIOs. Added new I/O special reason
4579 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
4580 with non blocking I/O was not possible because no retry code was
4581 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
4585 *) Added the beginnings of Rijndael support.
4588 *) Fix for bug in DirectoryString mask setting. Add support for
4589 X509_NAME_print_ex() in 'req' and X509_print_ex() function
4590 to allow certificate printing to more controllable, additional
4591 'certopt' option to 'x509' to allow new printing options to be
4595 *) Clean old EAY MD5 hack from e_os.h.
4598 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
4600 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4601 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4602 [Joe Orton, Steve Henson]
4604 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
4606 *) Fix additional bug revealed by the NISCC test suite:
4608 Stop bug triggering large recursion when presented with
4609 certain ASN.1 tags (CVE-2003-0851)
4612 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
4614 *) Fix various bugs revealed by running the NISCC test suite:
4616 Stop out of bounds reads in the ASN1 code when presented with
4617 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4619 If verify callback ignores invalid public key errors don't try to check
4620 certificate signature with the NULL public key.
4624 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4625 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4629 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4630 extra data after the compression methods not only for TLS 1.0
4631 but also for SSL 3.0 (as required by the specification).
4632 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4634 *) Change X509_certificate_type() to mark the key as exported/exportable
4635 when it's 512 *bits* long, not 512 bytes.
4638 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
4640 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4641 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4642 a protocol version number mismatch like a decryption error
4643 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4646 *) Turn on RSA blinding by default in the default implementation
4647 to avoid a timing attack. Applications that don't want it can call
4648 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4649 They would be ill-advised to do so in most cases.
4650 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4652 *) Change RSA blinding code so that it works when the PRNG is not
4653 seeded (in this case, the secret RSA exponent is abused as
4654 an unpredictable seed -- if it is not unpredictable, there
4655 is no point in blinding anyway). Make RSA blinding thread-safe
4656 by remembering the creator's thread ID in rsa->blinding and
4657 having all other threads use local one-time blinding factors
4658 (this requires more computation than sharing rsa->blinding, but
4659 avoids excessive locking; and if an RSA object is not shared
4660 between threads, blinding will still be very fast).
4663 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
4665 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4666 via timing by performing a MAC computation even if incorrrect
4667 block cipher padding has been found. This is a countermeasure
4668 against active attacks where the attacker has to distinguish
4669 between bad padding and a MAC verification error. (CVE-2003-0078)
4671 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4672 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4673 Martin Vuagnoux (EPFL, Ilion)]
4675 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
4677 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
4678 memory from it's contents. This is done with a counter that will
4679 place alternating values in each byte. This can be used to solve
4680 two issues: 1) the removal of calls to memset() by highly optimizing
4681 compilers, and 2) cleansing with other values than 0, since those can
4682 be read through on certain media, for example a swap space on disk.
4685 *) Bugfix: client side session caching did not work with external caching,
4686 because the session->cipher setting was not restored when reloading
4687 from the external cache. This problem was masked, when
4688 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
4689 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
4692 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
4693 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
4694 [Zeev Lieber <zeev-l@yahoo.com>]
4696 *) Undo an undocumented change introduced in 0.9.6e which caused
4697 repeated calls to OpenSSL_add_all_ciphers() and
4698 OpenSSL_add_all_digests() to be ignored, even after calling
4702 *) Change the default configuration reader to deal with last line not
4703 being properly terminated.
4706 *) Change X509_NAME_cmp() so it applies the special rules on handling
4707 DN values that are of type PrintableString, as well as RDNs of type
4708 emailAddress where the value has the type ia5String.
4709 [stefank@valicert.com via Richard Levitte]
4711 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
4712 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
4713 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
4714 the bitwise-OR of the two for use by the majority of applications
4715 wanting this behaviour, and update the docs. The documented
4716 behaviour and actual behaviour were inconsistent and had been
4717 changing anyway, so this is more a bug-fix than a behavioural
4719 [Geoff Thorpe, diagnosed by Nadav Har'El]
4721 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
4722 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
4725 *) Fix initialization code race conditions in
4726 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
4727 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
4728 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
4729 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
4730 ssl2_get_cipher_by_char(),
4731 ssl3_get_cipher_by_char().
4732 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
4734 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
4735 the cached sessions are flushed, as the remove_cb() might use ex_data
4736 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
4737 (see [openssl.org #212]).
4738 [Geoff Thorpe, Lutz Jaenicke]
4740 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
4741 length, instead of the encoding length to d2i_ASN1_OBJECT.
4744 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
4746 *) [In 0.9.6g-engine release:]
4747 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
4748 [Lynn Gazis <lgazis@rainbow.com>]
4750 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
4752 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
4753 and get fix the header length calculation.
4754 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
4755 Alon Kantor <alonk@checkpoint.com> (and others),
4758 *) Use proper error handling instead of 'assertions' in buffer
4759 overflow checks added in 0.9.6e. This prevents DoS (the
4760 assertions could call abort()).
4761 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
4763 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
4765 *) Add various sanity checks to asn1_get_length() to reject
4766 the ASN1 length bytes if they exceed sizeof(long), will appear
4767 negative or the content length exceeds the length of the
4769 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4771 *) Fix cipher selection routines: ciphers without encryption had no flags
4772 for the cipher strength set and where therefore not handled correctly
4773 by the selection routines (PR #130).
4776 *) Fix EVP_dsa_sha macro.
4780 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
4781 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
4782 that was added in OpenSSL 0.9.6d.
4784 As the countermeasure turned out to be incompatible with some
4785 broken SSL implementations, the new option is part of SSL_OP_ALL.
4786 SSL_OP_ALL is usually employed when compatibility with weird SSL
4787 implementations is desired (e.g. '-bugs' option to 's_client' and
4788 's_server'), so the new option is automatically set in many
4792 *) Changes in security patch:
4794 Changes marked "(CHATS)" were sponsored by the Defense Advanced
4795 Research Projects Agency (DARPA) and Air Force Research Laboratory,
4796 Air Force Materiel Command, USAF, under agreement number
4799 *) Add various sanity checks to asn1_get_length() to reject
4800 the ASN1 length bytes if they exceed sizeof(long), will appear
4801 negative or the content length exceeds the length of the
4802 supplied buffer. (CVE-2002-0659)
4803 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4805 *) Assertions for various potential buffer overflows, not known to
4807 [Ben Laurie (CHATS)]
4809 *) Various temporary buffers to hold ASCII versions of integers were
4810 too small for 64 bit platforms. (CVE-2002-0655)
4811 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
4813 *) Remote buffer overflow in SSL3 protocol - an attacker could
4814 supply an oversized session ID to a client. (CVE-2002-0656)
4815 [Ben Laurie (CHATS)]
4817 *) Remote buffer overflow in SSL2 protocol - an attacker could
4818 supply an oversized client master key. (CVE-2002-0656)
4819 [Ben Laurie (CHATS)]
4821 Changes between 0.9.6c and 0.9.6d [9 May 2002]
4823 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
4824 encoded as NULL) with id-dsa-with-sha1.
4825 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
4827 *) Check various X509_...() return values in apps/req.c.
4828 [Nils Larsch <nla@trustcenter.de>]
4830 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
4831 an end-of-file condition would erronously be flagged, when the CRLF
4832 was just at the end of a processed block. The bug was discovered when
4833 processing data through a buffering memory BIO handing the data to a
4834 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
4835 <ptsekov@syntrex.com> and Nedelcho Stanev.
4838 *) Implement a countermeasure against a vulnerability recently found
4839 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
4840 before application data chunks to avoid the use of known IVs
4841 with data potentially chosen by the attacker.
4844 *) Fix length checks in ssl3_get_client_hello().
4847 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
4848 to prevent ssl3_read_internal() from incorrectly assuming that
4849 ssl3_read_bytes() found application data while handshake
4850 processing was enabled when in fact s->s3->in_read_app_data was
4851 merely automatically cleared during the initial handshake.
4852 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
4854 *) Fix object definitions for Private and Enterprise: they were not
4855 recognized in their shortname (=lowercase) representation. Extend
4856 obj_dat.pl to issue an error when using undefined keywords instead
4857 of silently ignoring the problem (Svenning Sorensen
4858 <sss@sss.dnsalias.net>).
4861 *) Fix DH_generate_parameters() so that it works for 'non-standard'
4862 generators, i.e. generators other than 2 and 5. (Previously, the
4863 code did not properly initialise the 'add' and 'rem' values to
4864 BN_generate_prime().)
4866 In the new general case, we do not insist that 'generator' is
4867 actually a primitive root: This requirement is rather pointless;
4868 a generator of the order-q subgroup is just as good, if not
4872 *) Map new X509 verification errors to alerts. Discovered and submitted by
4873 Tom Wu <tom@arcot.com>.
4876 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
4877 returning non-zero before the data has been completely received
4878 when using non-blocking I/O.
4879 [Bodo Moeller; problem pointed out by John Hughes]
4881 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
4882 [Ben Laurie, Lutz Jaenicke]
4884 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
4885 Yoram Zahavi <YoramZ@gilian.com>).
4888 *) Add information about CygWin 1.3 and on, and preserve proper
4889 configuration for the versions before that.
4890 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4892 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
4893 check whether we deal with a copy of a session and do not delete from
4894 the cache in this case. Problem reported by "Izhar Shoshani Levi"
4895 <izhar@checkpoint.com>.
4898 *) Do not store session data into the internal session cache, if it
4899 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
4900 flag is set). Proposed by Aslam <aslam@funk.com>.
4903 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
4907 *) [In 0.9.6d-engine release:]
4908 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4909 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4911 *) Add the configuration target linux-s390x.
4912 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
4914 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
4915 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
4916 variable as an indication that a ClientHello message has been
4917 received. As the flag value will be lost between multiple
4918 invocations of ssl3_accept when using non-blocking I/O, the
4919 function may not be aware that a handshake has actually taken
4920 place, thus preventing a new session from being added to the
4923 To avoid this problem, we now set s->new_session to 2 instead of
4924 using a local variable.
4925 [Lutz Jaenicke, Bodo Moeller]
4927 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
4928 if the SSL_R_LENGTH_MISMATCH error is detected.
4929 [Geoff Thorpe, Bodo Moeller]
4931 *) New 'shared_ldflag' column in Configure platform table.
4934 *) Fix EVP_CIPHER_mode macro.
4935 ["Dan S. Camper" <dan@bti.net>]
4937 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
4938 type, we must throw them away by setting rr->length to 0.
4939 [D P Chang <dpc@qualys.com>]
4941 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
4943 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
4944 <Dominikus.Scherkl@biodata.com>. (The previous implementation
4945 worked incorrectly for those cases where range = 10..._2 and
4946 3*range is two bits longer than range.)
4949 *) Only add signing time to PKCS7 structures if it is not already
4953 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
4954 OBJ_ld_ce should be OBJ_id_ce.
4955 Also some ip-pda OIDs in crypto/objects/objects.txt were
4956 incorrect (cf. RFC 3039).
4957 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
4959 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
4960 returns early because it has nothing to do.
4961 [Andy Schneider <andy.schneider@bjss.co.uk>]
4963 *) [In 0.9.6c-engine release:]
4964 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
4965 [Andy Schneider <andy.schneider@bjss.co.uk>]
4967 *) [In 0.9.6c-engine release:]
4968 Add support for Cryptographic Appliance's keyserver technology.
4969 (Use engine 'keyclient')
4970 [Cryptographic Appliances and Geoff Thorpe]
4972 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
4973 is called via tools/c89.sh because arguments have to be
4974 rearranged (all '-L' options must appear before the first object
4976 [Richard Shapiro <rshapiro@abinitio.com>]
4978 *) [In 0.9.6c-engine release:]
4979 Add support for Broadcom crypto accelerator cards, backported
4981 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
4983 *) [In 0.9.6c-engine release:]
4984 Add support for SureWare crypto accelerator cards from
4985 Baltimore Technologies. (Use engine 'sureware')
4986 [Baltimore Technologies and Mark Cox]
4988 *) [In 0.9.6c-engine release:]
4989 Add support for crypto accelerator cards from Accelerated
4990 Encryption Processing, www.aep.ie. (Use engine 'aep')
4991 [AEP Inc. and Mark Cox]
4993 *) Add a configuration entry for gcc on UnixWare.
4994 [Gary Benson <gbenson@redhat.com>]
4996 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
4997 messages are stored in a single piece (fixed-length part and
4998 variable-length part combined) and fix various bugs found on the way.
5001 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
5002 instead. BIO_gethostbyname() does not know what timeouts are
5003 appropriate, so entries would stay in cache even when they have
5005 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
5007 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
5008 faced with a pathologically small ClientHello fragment that does
5009 not contain client_version: Instead of aborting with an error,
5010 simply choose the highest available protocol version (i.e.,
5011 TLS 1.0 unless it is disabled). In practice, ClientHello
5012 messages are never sent like this, but this change gives us
5013 strictly correct behaviour at least for TLS.
5016 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
5017 never resets s->method to s->ctx->method when called from within
5018 one of the SSL handshake functions.
5019 [Bodo Moeller; problem pointed out by Niko Baric]
5021 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
5022 (sent using the client's version number) if client_version is
5023 smaller than the protocol version in use. Also change
5024 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
5025 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
5026 the client will at least see that alert.
5029 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
5033 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
5034 client receives HelloRequest while in a handshake.
5035 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
5037 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
5038 should end in 'break', not 'goto end' which circuments various
5039 cleanups done in state SSL_ST_OK. But session related stuff
5040 must be disabled for SSL_ST_OK in the case that we just sent a
5043 Also avoid some overhead by not calling ssl_init_wbio_buffer()
5044 before just sending a HelloRequest.
5045 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
5047 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
5048 reveal whether illegal block cipher padding was found or a MAC
5049 verification error occured. (Neither SSLerr() codes nor alerts
5050 are directly visible to potential attackers, but the information
5051 may leak via logfiles.)
5053 Similar changes are not required for the SSL 2.0 implementation
5054 because the number of padding bytes is sent in clear for SSL 2.0,
5055 and the extra bytes are just ignored. However ssl/s2_pkt.c
5056 failed to verify that the purported number of padding bytes is in
5060 *) Add OpenUNIX-8 support including shared libraries
5061 (Boyd Lynn Gerber <gerberb@zenez.com>).
5064 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
5065 'wristwatch attack' using huge encoding parameters (cf.
5066 James H. Manger's CRYPTO 2001 paper). Note that the
5067 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
5068 encoding parameters and hence was not vulnerable.
5071 *) BN_sqr() bug fix.
5072 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
5074 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
5075 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
5076 followed by modular reduction.
5077 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
5079 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
5080 equivalent based on BN_pseudo_rand() instead of BN_rand().
5083 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
5084 This function was broken, as the check for a new client hello message
5085 to handle SGC did not allow these large messages.
5086 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
5089 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
5092 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
5093 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
5096 *) Rework the configuration and shared library support for Tru64 Unix.
5097 The configuration part makes use of modern compiler features and
5098 still retains old compiler behavior for those that run older versions
5099 of the OS. The shared library support part includes a variant that
5100 uses the RPATH feature, and is available through the special
5101 configuration target "alpha-cc-rpath", which will never be selected
5103 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
5105 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
5106 with the same message size as in ssl3_get_certificate_request().
5107 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
5108 messages might inadvertently be reject as too long.
5109 [Petr Lampa <lampa@fee.vutbr.cz>]
5111 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
5114 *) Modified SSL library such that the verify_callback that has been set
5115 specificly for an SSL object with SSL_set_verify() is actually being
5116 used. Before the change, a verify_callback set with this function was
5117 ignored and the verify_callback() set in the SSL_CTX at the time of
5118 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
5119 to allow the necessary settings.
5122 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
5123 explicitly to NULL, as at least on Solaris 8 this seems not always to be
5124 done automatically (in contradiction to the requirements of the C
5125 standard). This made problems when used from OpenSSH.
5128 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
5129 dh->length and always used
5131 BN_rand_range(priv_key, dh->p).
5133 BN_rand_range() is not necessary for Diffie-Hellman, and this
5134 specific range makes Diffie-Hellman unnecessarily inefficient if
5135 dh->length (recommended exponent length) is much smaller than the
5136 length of dh->p. We could use BN_rand_range() if the order of
5137 the subgroup was stored in the DH structure, but we only have
5142 BN_rand(priv_key, l, ...)
5144 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
5150 RSA_eay_public_encrypt
5151 RSA_eay_private_decrypt
5152 RSA_eay_private_encrypt (signing)
5153 RSA_eay_public_decrypt (signature verification)
5155 (default implementations for RSA_public_encrypt,
5156 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
5157 always reject numbers >= n.
5160 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
5161 to synchronize access to 'locking_thread'. This is necessary on
5162 systems where access to 'locking_thread' (an 'unsigned long'
5163 variable) is not atomic.
5166 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
5167 *before* setting the 'crypto_lock_rand' flag. The previous code had
5168 a race condition if 0 is a valid thread ID.
5169 [Travis Vitek <vitek@roguewave.com>]
5171 *) Add support for shared libraries under Irix.
5172 [Albert Chin-A-Young <china@thewrittenword.com>]
5174 *) Add configuration option to build on Linux on both big-endian and
5176 [Ralf Baechle <ralf@uni-koblenz.de>]
5178 *) Add the possibility to create shared libraries on HP-UX.
5181 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
5183 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
5184 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
5185 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
5186 PRNG state recovery was possible based on the output of
5187 one PRNG request appropriately sized to gain knowledge on
5188 'md' followed by enough consecutive 1-byte PRNG requests
5189 to traverse all of 'state'.
5191 1. When updating 'md_local' (the current thread's copy of 'md')
5192 during PRNG output generation, hash all of the previous
5193 'md_local' value, not just the half used for PRNG output.
5195 2. Make the number of bytes from 'state' included into the hash
5196 independent from the number of PRNG bytes requested.
5198 The first measure alone would be sufficient to avoid
5199 Markku-Juhani's attack. (Actually it had never occurred
5200 to me that the half of 'md_local' used for chaining was the
5201 half from which PRNG output bytes were taken -- I had always
5202 assumed that the secret half would be used.) The second
5203 measure makes sure that additional data from 'state' is never
5204 mixed into 'md_local' in small portions; this heuristically
5205 further strengthens the PRNG.
5208 *) Fix crypto/bn/asm/mips3.s.
5211 *) When only the key is given to "enc", the IV is undefined. Print out
5212 an error message in this case.
5215 *) Handle special case when X509_NAME is empty in X509 printing routines.
5218 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
5219 positive and less than q.
5222 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
5223 used: it isn't thread safe and the add_lock_callback should handle
5225 [Paul Rose <Paul.Rose@bridge.com>]
5227 *) Verify that incoming data obeys the block size in
5228 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
5232 [Ulf Möller, Bodo Möller]
5234 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
5235 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
5236 when fixing the server behaviour for backwards-compatible 'client
5237 hello' messages. (Note that the attack is impractical against
5238 SSL 3.0 and TLS 1.0 anyway because length and version checking
5239 means that the probability of guessing a valid ciphertext is
5240 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
5243 Before 0.9.5, the countermeasure (hide the error by generating a
5244 random 'decryption result') did not work properly because
5245 ERR_clear_error() was missing, meaning that SSL_get_error() would
5246 detect the supposedly ignored error.
5248 Both problems are now fixed.
5251 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
5252 (previously it was 1024).
5255 *) Fix for compatibility mode trust settings: ignore trust settings
5256 unless some valid trust or reject settings are present.
5259 *) Fix for blowfish EVP: its a variable length cipher.
5262 *) Fix various bugs related to DSA S/MIME verification. Handle missing
5263 parameters in DSA public key structures and return an error in the
5264 DSA routines if parameters are absent.
5267 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
5268 in the current directory if neither $RANDFILE nor $HOME was set.
5269 RAND_file_name() in 0.9.6a returned NULL in this case. This has
5270 caused some confusion to Windows users who haven't defined $HOME.
5271 Thus RAND_file_name() is changed again: e_os.h can define a
5272 DEFAULT_HOME, which will be used if $HOME is not set.
5273 For Windows, we use "C:"; on other platforms, we still require
5274 environment variables.
5276 *) Move 'if (!initialized) RAND_poll()' into regions protected by
5277 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
5278 having multiple threads call RAND_poll() concurrently.
5281 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
5282 combination of a flag and a thread ID variable.
5283 Otherwise while one thread is in ssleay_rand_bytes (which sets the
5284 flag), *other* threads can enter ssleay_add_bytes without obeying
5285 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
5286 that they do not hold after the first thread unsets add_do_not_lock).
5289 *) Change bctest again: '-x' expressions are not available in all
5293 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
5295 *) Fix a couple of memory leaks in PKCS7_dataDecode()
5296 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
5298 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
5299 the default extension for executables, if any. Also, make the perl
5300 scripts that use symlink() to test if it really exists and use "cp"
5301 if it doesn't. All this made OpenSSL compilable and installable in
5305 *) Fix for asn1_GetSequence() for indefinite length constructed data.
5306 If SEQUENCE is length is indefinite just set c->slen to the total
5307 amount of data available.
5308 [Steve Henson, reported by shige@FreeBSD.org]
5309 [This change does not apply to 0.9.7.]
5311 *) Change bctest to avoid here-documents inside command substitution
5312 (workaround for FreeBSD /bin/sh bug).
5313 For compatibility with Ultrix, avoid shell functions (introduced
5314 in the bctest version that searches along $PATH).
5317 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
5318 with des_encrypt() defined on some operating systems, like Solaris
5322 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
5323 On the Importance of Eliminating Errors in Cryptographic
5324 Computations, J. Cryptology 14 (2001) 2, 101-119,
5325 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
5328 *) MIPS assembler BIGNUM division bug fix.
5331 *) Disabled incorrect Alpha assembler code.
5334 *) Fix PKCS#7 decode routines so they correctly update the length
5335 after reading an EOC for the EXPLICIT tag.
5337 [This change does not apply to 0.9.7.]
5339 *) Fix bug in PKCS#12 key generation routines. This was triggered
5340 if a 3DES key was generated with a 0 initial byte. Include
5341 PKCS12_BROKEN_KEYGEN compilation option to retain the old
5342 (but broken) behaviour.
5345 *) Enhance bctest to search for a working bc along $PATH and print
5347 [Tim Rice <tim@multitalents.net> via Richard Levitte]
5349 *) Fix memory leaks in err.c: free err_data string if necessary;
5350 don't write to the wrong index in ERR_set_error_data.
5353 *) Implement ssl23_peek (analogous to ssl23_read), which previously
5357 *) Replace rdtsc with _emit statements for VC++ version 5.
5358 [Jeremy Cooper <jeremy@baymoo.org>]
5360 *) Make it possible to reuse SSLv2 sessions.
5363 *) In copy_email() check for >= 0 as a return value for
5364 X509_NAME_get_index_by_NID() since 0 is a valid index.
5365 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
5367 *) Avoid coredump with unsupported or invalid public keys by checking if
5368 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
5369 PKCS7_verify() fails with non detached data.
5372 *) Don't use getenv in library functions when run as setuid/setgid.
5373 New function OPENSSL_issetugid().
5376 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
5377 due to incorrect handling of multi-threading:
5379 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
5381 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
5383 3. Count how many times MemCheck_off() has been called so that
5384 nested use can be treated correctly. This also avoids
5385 inband-signalling in the previous code (which relied on the
5386 assumption that thread ID 0 is impossible).
5389 *) Add "-rand" option also to s_client and s_server.
5392 *) Fix CPU detection on Irix 6.x.
5393 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
5394 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
5396 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
5399 [This change does not apply to 0.9.7.]
5401 *) Use the cached encoding of an X509_NAME structure rather than
5402 copying it. This is apparently the reason for the libsafe "errors"
5403 but the code is actually correct.
5406 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
5407 Bleichenbacher's DSA attack.
5408 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
5409 to be set and top=0 forces the highest bit to be set; top=-1 is new
5410 and leaves the highest bit random.
5411 [Ulf Moeller, Bodo Moeller]
5413 *) In the NCONF_...-based implementations for CONF_... queries
5414 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
5415 a temporary CONF structure with the data component set to NULL
5416 (which gives segmentation faults in lh_retrieve).
5417 Instead, use NULL for the CONF pointer in CONF_get_string and
5418 CONF_get_number (which may use environment variables) and directly
5419 return NULL from CONF_get_section.
5422 *) Fix potential buffer overrun for EBCDIC.
5425 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
5426 keyUsage if basicConstraints absent for a CA.
5429 *) Make SMIME_write_PKCS7() write mail header values with a format that
5430 is more generally accepted (no spaces before the semicolon), since
5431 some programs can't parse those values properly otherwise. Also make
5432 sure BIO's that break lines after each write do not create invalid
5436 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
5437 macros previously used would not encode an empty SEQUENCE OF
5438 and break the signature.
5440 [This change does not apply to 0.9.7.]
5442 *) Zero the premaster secret after deriving the master secret in
5446 *) Add some EVP_add_digest_alias registrations (as found in
5447 OpenSSL_add_all_digests()) to SSL_library_init()
5448 aka OpenSSL_add_ssl_algorithms(). This provides improved
5449 compatibility with peers using X.509 certificates
5450 with unconventional AlgorithmIdentifier OIDs.
5453 *) Fix for Irix with NO_ASM.
5454 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
5456 *) ./config script fixes.
5457 [Ulf Moeller, Richard Levitte]
5459 *) Fix 'openssl passwd -1'.
5462 *) Change PKCS12_key_gen_asc() so it can cope with non null
5463 terminated strings whose length is passed in the passlen
5464 parameter, for example from PEM callbacks. This was done
5465 by adding an extra length parameter to asc2uni().
5466 [Steve Henson, reported by <oddissey@samsung.co.kr>]
5468 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
5469 call failed, free the DSA structure.
5472 *) Fix to uni2asc() to cope with zero length Unicode strings.
5473 These are present in some PKCS#12 files.
5476 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
5477 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
5478 when writing a 32767 byte record.
5479 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
5481 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
5482 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
5484 (RSA objects have a reference count access to which is protected
5485 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
5486 so they are meant to be shared between threads.)
5487 [Bodo Moeller, Geoff Thorpe; original patch submitted by
5488 "Reddie, Steven" <Steven.Reddie@ca.com>]
5490 *) Fix a deadlock in CRYPTO_mem_leaks().
5493 *) Use better test patterns in bntest.
5496 *) rand_win.c fix for Borland C.
5499 *) BN_rshift bugfix for n == 0.
5502 *) Add a 'bctest' script that checks for some known 'bc' bugs
5503 so that 'make test' does not abort just because 'bc' is broken.
5506 *) Store verify_result within SSL_SESSION also for client side to
5507 avoid potential security hole. (Re-used sessions on the client side
5508 always resulted in verify_result==X509_V_OK, not using the original
5509 result of the server certificate verification.)
5512 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
5513 SSL3_RT_APPLICATION_DATA, return 0.
5514 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
5518 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
5519 releases, have been re-implemented by renaming the previous
5520 implementations of ssl2_read and ssl3_read to ssl2_read_internal
5521 and ssl3_read_internal, respectively, and adding 'peek' parameters
5522 to them. The new ssl[23]_{read,peek} functions are calls to
5523 ssl[23]_read_internal with the 'peek' flag set appropriately.
5524 A 'peek' parameter has also been added to ssl3_read_bytes, which
5525 does the actual work for ssl3_read_internal.
5528 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
5529 the method-specific "init()" handler. Also clean up ex_data after
5530 calling the method-specific "finish()" handler. Previously, this was
5531 happening the other way round.
5534 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
5535 The previous value, 12, was not always sufficient for BN_mod_exp().
5538 *) Make sure that shared libraries get the internal name engine with
5539 the full version number and not just 0. This should mark the
5540 shared libraries as not backward compatible. Of course, this should
5541 be changed again when we can guarantee backward binary compatibility.
5544 *) Fix typo in get_cert_by_subject() in by_dir.c
5545 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
5547 *) Rework the system to generate shared libraries:
5549 - Make note of the expected extension for the shared libraries and
5550 if there is a need for symbolic links from for example libcrypto.so.0
5551 to libcrypto.so.0.9.7. There is extended info in Configure for
5554 - Make as few rebuilds of the shared libraries as possible.
5556 - Still avoid linking the OpenSSL programs with the shared libraries.
5558 - When installing, install the shared libraries separately from the
5562 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
5564 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
5565 and not in SSL_clear because the latter is also used by the
5566 accept/connect functions; previously, the settings made by
5567 SSL_set_read_ahead would be lost during the handshake.
5568 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
5570 *) Correct util/mkdef.pl to be selective about disabled algorithms.
5571 Previously, it would create entries for disableed algorithms no
5575 *) Added several new manual pages for SSL_* function.
5578 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
5580 *) In ssl23_get_client_hello, generate an error message when faced
5581 with an initial SSL 3.0/TLS record that is too small to contain the
5582 first two bytes of the ClientHello message, i.e. client_version.
5583 (Note that this is a pathologic case that probably has never happened
5584 in real life.) The previous approach was to use the version number
5585 from the record header as a substitute; but our protocol choice
5586 should not depend on that one because it is not authenticated
5587 by the Finished messages.
5590 *) More robust randomness gathering functions for Windows.
5591 [Jeffrey Altman <jaltman@columbia.edu>]
5593 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
5594 not set then we don't setup the error code for issuer check errors
5595 to avoid possibly overwriting other errors which the callback does
5596 handle. If an application does set the flag then we assume it knows
5597 what it is doing and can handle the new informational codes
5601 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
5602 a general "ANY" type, as such it should be able to decode anything
5603 including tagged types. However it didn't check the class so it would
5604 wrongly interpret tagged types in the same way as their universal
5605 counterpart and unknown types were just rejected. Changed so that the
5606 tagged and unknown types are handled in the same way as a SEQUENCE:
5607 that is the encoding is stored intact. There is also a new type
5608 "V_ASN1_OTHER" which is used when the class is not universal, in this
5609 case we have no idea what the actual type is so we just lump them all
5613 *) On VMS, stdout may very well lead to a file that is written to
5614 in a record-oriented fashion. That means that every write() will
5615 write a separate record, which will be read separately by the
5616 programs trying to read from it. This can be very confusing.
5618 The solution is to put a BIO filter in the way that will buffer
5619 text until a linefeed is reached, and then write everything a
5620 line at a time, so every record written will be an actual line,
5621 not chunks of lines and not (usually doesn't happen, but I've
5622 seen it once) several lines in one record. BIO_f_linebuffer() is
5625 Currently, it's a VMS-only method, because that's where it has
5626 been tested well enough.
5629 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
5630 it can return incorrect results.
5631 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
5632 but it was in 0.9.6-beta[12].)
5635 *) Disable the check for content being present when verifying detached
5636 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
5637 include zero length content when signing messages.
5640 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
5641 BIO_ctrl (for BIO pairs).
5644 *) Add DSO method for VMS.
5647 *) Bug fix: Montgomery multiplication could produce results with the
5651 *) Add RPM specification openssl.spec and modify it to build three
5652 packages. The default package contains applications, application
5653 documentation and run-time libraries. The devel package contains
5654 include files, static libraries and function documentation. The
5655 doc package contains the contents of the doc directory. The original
5656 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
5659 *) Add a large number of documentation files for many SSL routines.
5660 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
5662 *) Add a configuration entry for Sony News 4.
5663 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
5665 *) Don't set the two most significant bits to one when generating a
5666 random number < q in the DSA library.
5669 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
5670 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
5671 the underlying transport is blocking) if a handshake took place.
5672 (The default behaviour is needed by applications such as s_client
5673 and s_server that use select() to determine when to use SSL_read;
5674 but for applications that know in advance when to expect data, it
5675 just makes things more complicated.)
5678 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
5682 *) Add a few more EBCDIC conditionals that make `req' and `x509'
5683 work better on such systems.
5684 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5686 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
5687 Update PKCS12_parse() so it copies the friendlyName and the
5688 keyid to the certificates aux info.
5691 *) Fix bug in PKCS7_verify() which caused an infinite loop
5692 if there was more than one signature.
5693 [Sven Uszpelkat <su@celocom.de>]
5695 *) Major change in util/mkdef.pl to include extra information
5696 about each symbol, as well as presentig variables as well
5697 as functions. This change means that there's n more need
5698 to rebuild the .num files when some algorithms are excluded.
5701 *) Allow the verify time to be set by an application,
5702 rather than always using the current time.
5705 *) Phase 2 verify code reorganisation. The certificate
5706 verify code now looks up an issuer certificate by a
5707 number of criteria: subject name, authority key id
5708 and key usage. It also verifies self signed certificates
5709 by the same criteria. The main comparison function is
5710 X509_check_issued() which performs these checks.
5712 Lot of changes were necessary in order to support this
5713 without completely rewriting the lookup code.
5715 Authority and subject key identifier are now cached.
5717 The LHASH 'certs' is X509_STORE has now been replaced
5718 by a STACK_OF(X509_OBJECT). This is mainly because an
5719 LHASH can't store or retrieve multiple objects with
5720 the same hash value.
5722 As a result various functions (which were all internal
5723 use only) have changed to handle the new X509_STORE
5724 structure. This will break anything that messed round
5725 with X509_STORE internally.
5727 The functions X509_STORE_add_cert() now checks for an
5728 exact match, rather than just subject name.
5730 The X509_STORE API doesn't directly support the retrieval
5731 of multiple certificates matching a given criteria, however
5732 this can be worked round by performing a lookup first
5733 (which will fill the cache with candidate certificates)
5734 and then examining the cache for matches. This is probably
5735 the best we can do without throwing out X509_LOOKUP
5736 entirely (maybe later...).
5738 The X509_VERIFY_CTX structure has been enhanced considerably.
5740 All certificate lookup operations now go via a get_issuer()
5741 callback. Although this currently uses an X509_STORE it
5742 can be replaced by custom lookups. This is a simple way
5743 to bypass the X509_STORE hackery necessary to make this
5744 work and makes it possible to use more efficient techniques
5745 in future. A very simple version which uses a simple
5746 STACK for its trusted certificate store is also provided
5747 using X509_STORE_CTX_trusted_stack().
5749 The verify_cb() and verify() callbacks now have equivalents
5750 in the X509_STORE_CTX structure.
5752 X509_STORE_CTX also has a 'flags' field which can be used
5753 to customise the verify behaviour.
5756 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
5757 excludes S/MIME capabilities.
5760 *) When a certificate request is read in keep a copy of the
5761 original encoding of the signed data and use it when outputing
5762 again. Signatures then use the original encoding rather than
5763 a decoded, encoded version which may cause problems if the
5764 request is improperly encoded.
5767 *) For consistency with other BIO_puts implementations, call
5768 buffer_write(b, ...) directly in buffer_puts instead of calling
5771 In BIO_puts, increment b->num_write as in BIO_write.
5772 [Peter.Sylvester@EdelWeb.fr]
5774 *) Fix BN_mul_word for the case where the word is 0. (We have to use
5775 BN_zero, we may not return a BIGNUM with an array consisting of
5779 *) Avoid calling abort() from within the library when problems are
5780 detected, except if preprocessor symbols have been defined
5781 (such as REF_CHECK, BN_DEBUG etc.).
5784 *) New openssl application 'rsautl'. This utility can be
5785 used for low level RSA operations. DER public key
5786 BIO/fp routines also added.
5789 *) New Configure entry and patches for compiling on QNX 4.
5790 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
5792 *) A demo state-machine implementation was sponsored by
5793 Nuron (http://www.nuron.com/) and is now available in
5794 demos/state_machine.
5797 *) New options added to the 'dgst' utility for signature
5798 generation and verification.
5801 *) Unrecognized PKCS#7 content types are now handled via a
5802 catch all ASN1_TYPE structure. This allows unsupported
5803 types to be stored as a "blob" and an application can
5804 encode and decode it manually.
5807 *) Fix various signed/unsigned issues to make a_strex.c
5809 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
5811 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
5812 length if passed a buffer. ASN1_INTEGER_to_BN failed
5813 if passed a NULL BN and its argument was negative.
5814 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
5816 *) Modification to PKCS#7 encoding routines to output definite
5817 length encoding. Since currently the whole structures are in
5818 memory there's not real point in using indefinite length
5819 constructed encoding. However if OpenSSL is compiled with
5820 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
5823 *) Added BIO_vprintf() and BIO_vsnprintf().
5826 *) Added more prefixes to parse for in the the strings written
5827 through a logging bio, to cover all the levels that are available
5828 through syslog. The prefixes are now:
5830 PANIC, EMERG, EMR => LOG_EMERG
5831 ALERT, ALR => LOG_ALERT
5832 CRIT, CRI => LOG_CRIT
5833 ERROR, ERR => LOG_ERR
5834 WARNING, WARN, WAR => LOG_WARNING
5835 NOTICE, NOTE, NOT => LOG_NOTICE
5836 INFO, INF => LOG_INFO
5837 DEBUG, DBG => LOG_DEBUG
5839 and as before, if none of those prefixes are present at the
5840 beginning of the string, LOG_ERR is chosen.
5842 On Win32, the LOG_* levels are mapped according to this:
5844 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
5845 LOG_WARNING => EVENTLOG_WARNING_TYPE
5846 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
5850 *) Made it possible to reconfigure with just the configuration
5851 argument "reconf" or "reconfigure". The command line arguments
5852 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
5853 and are retrieved from there when reconfiguring.
5857 [Assar Westerlund <assar@sics.se>, Richard Levitte]
5859 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
5862 *) The obj_dat.pl script was messing up the sorting of object
5863 names. The reason was that it compared the quoted version
5864 of strings as a result "OCSP" > "OCSP Signing" because
5865 " > SPACE. Changed script to store unquoted versions of
5866 names and add quotes on output. It was also omitting some
5867 names from the lookup table if they were given a default
5868 value (that is if SN is missing it is given the same
5869 value as LN and vice versa), these are now added on the
5870 grounds that if an object has a name we should be able to
5871 look it up. Finally added warning output when duplicate
5872 short or long names are found.
5875 *) Changes needed for Tandem NSK.
5876 [Scott Uroff <scott@xypro.com>]
5878 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
5879 RSA_padding_check_SSLv23(), special padding was never detected
5880 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
5881 version rollback attacks was not effective.
5883 In s23_clnt.c, don't use special rollback-attack detection padding
5884 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
5885 client; similarly, in s23_srvr.c, don't do the rollback check if
5886 SSL 2.0 is the only protocol enabled in the server.
5889 *) Make it possible to get hexdumps of unprintable data with 'openssl
5890 asn1parse'. By implication, the functions ASN1_parse_dump() and
5891 BIO_dump_indent() are added.
5894 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
5895 these print out strings and name structures based on various
5896 flags including RFC2253 support and proper handling of
5897 multibyte characters. Added options to the 'x509' utility
5898 to allow the various flags to be set.
5901 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
5902 Also change the functions X509_cmp_current_time() and
5903 X509_gmtime_adj() work with an ASN1_TIME structure,
5904 this will enable certificates using GeneralizedTime in validity
5905 dates to be checked.
5908 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
5909 negative public key encodings) on by default,
5910 NO_NEG_PUBKEY_BUG can be set to disable it.
5913 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
5914 content octets. An i2c_ASN1_OBJECT is unnecessary because
5915 the encoding can be trivially obtained from the structure.
5918 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
5919 not read locks (CRYPTO_r_[un]lock).
5922 *) A first attempt at creating official support for shared
5923 libraries through configuration. I've kept it so the
5924 default is static libraries only, and the OpenSSL programs
5925 are always statically linked for now, but there are
5926 preparations for dynamic linking in place.
5927 This has been tested on Linux and Tru64.
5930 *) Randomness polling function for Win9x, as described in:
5931 Peter Gutmann, Software Generation of Practically Strong
5935 *) Fix so PRNG is seeded in req if using an already existing
5939 *) New options to smime application. -inform and -outform
5940 allow alternative formats for the S/MIME message including
5941 PEM and DER. The -content option allows the content to be
5942 specified separately. This should allow things like Netscape
5943 form signing output easier to verify.
5946 *) Fix the ASN1 encoding of tags using the 'long form'.
5949 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
5950 STRING types. These convert content octets to and from the
5951 underlying type. The actual tag and length octets are
5952 already assumed to have been read in and checked. These
5953 are needed because all other string types have virtually
5954 identical handling apart from the tag. By having versions
5955 of the ASN1 functions that just operate on content octets
5956 IMPLICIT tagging can be handled properly. It also allows
5957 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
5958 and ASN1_INTEGER are identical apart from the tag.
5961 *) Change the handling of OID objects as follows:
5963 - New object identifiers are inserted in objects.txt, following
5964 the syntax given in objects.README.
5965 - objects.pl is used to process obj_mac.num and create a new
5967 - obj_dat.pl is used to create a new obj_dat.h, using the data in
5970 This is currently kind of a hack, and the perl code in objects.pl
5971 isn't very elegant, but it works as I intended. The simplest way
5972 to check that it worked correctly is to look in obj_dat.h and
5973 check the array nid_objs and make sure the objects haven't moved
5974 around (this is important!). Additions are OK, as well as
5975 consistent name changes.
5978 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
5981 *) Addition of the command line parameter '-rand file' to 'openssl req'.
5982 The given file adds to whatever has already been seeded into the
5983 random pool through the RANDFILE configuration file option or
5984 environment variable, or the default random state file.
5987 *) mkstack.pl now sorts each macro group into lexical order.
5988 Previously the output order depended on the order the files
5989 appeared in the directory, resulting in needless rewriting
5993 *) Patches to make OpenSSL compile under Win32 again. Mostly
5994 work arounds for the VC++ problem that it treats func() as
5995 func(void). Also stripped out the parts of mkdef.pl that
5996 added extra typesafe functions: these no longer exist.
5999 *) Reorganisation of the stack code. The macros are now all
6000 collected in safestack.h . Each macro is defined in terms of
6001 a "stack macro" of the form SKM_<name>(type, a, b). The
6002 DEBUG_SAFESTACK is now handled in terms of function casts,
6003 this has the advantage of retaining type safety without the
6004 use of additional functions. If DEBUG_SAFESTACK is not defined
6005 then the non typesafe macros are used instead. Also modified the
6006 mkstack.pl script to handle the new form. Needs testing to see
6007 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
6008 the default if no major problems. Similar behaviour for ASN1_SET_OF
6009 and PKCS12_STACK_OF.
6012 *) When some versions of IIS use the 'NET' form of private key the
6013 key derivation algorithm is different. Normally MD5(password) is
6014 used as a 128 bit RC4 key. In the modified case
6015 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
6016 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
6017 as the old Netscape_RSA functions except they have an additional
6018 'sgckey' parameter which uses the modified algorithm. Also added
6019 an -sgckey command line option to the rsa utility. Thanks to
6020 Adrian Peck <bertie@ncipher.com> for posting details of the modified
6021 algorithm to openssl-dev.
6024 *) The evp_local.h macros were using 'c.##kname' which resulted in
6025 invalid expansion on some systems (SCO 5.0.5 for example).
6026 Corrected to 'c.kname'.
6027 [Phillip Porch <root@theporch.com>]
6029 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
6030 a STACK of email addresses from a certificate or request, these look
6031 in the subject name and the subject alternative name extensions and
6032 omit any duplicate addresses.
6035 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
6036 This makes DSA verification about 2 % faster.
6039 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
6040 (meaning that now 2^5 values will be precomputed, which is only 4 KB
6041 plus overhead for 1024 bit moduli).
6042 This makes exponentiations about 0.5 % faster for 1024 bit
6043 exponents (as measured by "openssl speed rsa2048").
6046 *) Rename memory handling macros to avoid conflicts with other
6048 Malloc => OPENSSL_malloc
6049 Malloc_locked => OPENSSL_malloc_locked
6050 Realloc => OPENSSL_realloc
6051 Free => OPENSSL_free
6054 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
6055 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
6058 *) CygWin32 support.
6059 [John Jarvie <jjarvie@newsguy.com>]
6061 *) The type-safe stack code has been rejigged. It is now only compiled
6062 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
6063 by default all type-specific stack functions are "#define"d back to
6064 standard stack functions. This results in more streamlined output
6065 but retains the type-safety checking possibilities of the original
6069 *) The STACK code has been cleaned up, and certain type declarations
6070 that didn't make a lot of sense have been brought in line. This has
6071 also involved a cleanup of sorts in safestack.h to more correctly
6072 map type-safe stack functions onto their plain stack counterparts.
6073 This work has also resulted in a variety of "const"ifications of
6074 lots of the code, especially "_cmp" operations which should normally
6075 be prototyped with "const" parameters anyway.
6078 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
6079 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
6080 (The PRNG state consists of two parts, the large pool 'state' and 'md',
6081 where all of 'md' is used each time the PRNG is used, but 'state'
6082 is used only indexed by a cyclic counter. As entropy may not be
6083 well distributed from the beginning, 'md' is important as a
6084 chaining variable. However, the output function chains only half
6085 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
6086 all of 'md', and seeding with STATE_SIZE dummy bytes will result
6087 in all of 'state' being rewritten, with the new values depending
6088 on virtually all of 'md'. This overcomes the 80 bit limitation.)
6091 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
6092 the handshake is continued after ssl_verify_cert_chain();
6093 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
6094 can lead to 'unexplainable' connection aborts later.
6095 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
6097 *) Major EVP API cipher revision.
6098 Add hooks for extra EVP features. This allows various cipher
6099 parameters to be set in the EVP interface. Support added for variable
6100 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
6101 setting of RC2 and RC5 parameters.
6103 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
6106 Remove lots of duplicated code from the EVP library. For example *every*
6107 cipher init() function handles the 'iv' in the same way according to the
6108 cipher mode. They also all do nothing if the 'key' parameter is NULL and
6109 for CFB and OFB modes they zero ctx->num.
6111 New functionality allows removal of S/MIME code RC2 hack.
6113 Most of the routines have the same form and so can be declared in terms
6116 By shifting this to the top level EVP_CipherInit() it can be removed from
6117 all individual ciphers. If the cipher wants to handle IVs or keys
6118 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
6121 Change lots of functions like EVP_EncryptUpdate() to now return a
6122 value: although software versions of the algorithms cannot fail
6123 any installed hardware versions can.
6126 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
6127 this option is set, tolerate broken clients that send the negotiated
6128 protocol version number instead of the requested protocol version
6132 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
6133 i.e. non-zero for export ciphersuites, zero otherwise.
6134 Previous versions had this flag inverted, inconsistent with
6135 rsa_tmp_cb (..._TMP_RSA_CB).
6136 [Bodo Moeller; problem reported by Amit Chopra]
6138 *) Add missing DSA library text string. Work around for some IIS
6139 key files with invalid SEQUENCE encoding.
6142 *) Add a document (doc/standards.txt) that list all kinds of standards
6143 and so on that are implemented in OpenSSL.
6146 *) Enhance c_rehash script. Old version would mishandle certificates
6147 with the same subject name hash and wouldn't handle CRLs at all.
6148 Added -fingerprint option to crl utility, to support new c_rehash
6152 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
6155 *) Fix for SSL server purpose checking. Server checking was
6156 rejecting certificates which had extended key usage present
6157 but no ssl client purpose.
6158 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
6160 *) Make PKCS#12 code work with no password. The PKCS#12 spec
6161 is a little unclear about how a blank password is handled.
6162 Since the password in encoded as a BMPString with terminating
6163 double NULL a zero length password would end up as just the
6164 double NULL. However no password at all is different and is
6165 handled differently in the PKCS#12 key generation code. NS
6166 treats a blank password as zero length. MSIE treats it as no
6167 password on export: but it will try both on import. We now do
6168 the same: PKCS12_parse() tries zero length and no password if
6169 the password is set to "" or NULL (NULL is now a valid password:
6170 it wasn't before) as does the pkcs12 application.
6173 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
6174 perror when PEM_read_bio_X509_REQ fails, the error message must
6175 be obtained from the error queue.
6178 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
6179 it in ERR_remove_state if appropriate, and change ERR_get_state
6180 accordingly to avoid race conditions (this is necessary because
6181 thread_hash is no longer constant once set).
6184 *) Bugfix for linux-elf makefile.one.
6187 *) RSA_get_default_method() will now cause a default
6188 RSA_METHOD to be chosen if one doesn't exist already.
6189 Previously this was only set during a call to RSA_new()
6190 or RSA_new_method(NULL) meaning it was possible for
6191 RSA_get_default_method() to return NULL.
6194 *) Added native name translation to the existing DSO code
6195 that will convert (if the flag to do so is set) filenames
6196 that are sufficiently small and have no path information
6197 into a canonical native form. Eg. "blah" converted to
6198 "libblah.so" or "blah.dll" etc.
6201 *) New function ERR_error_string_n(e, buf, len) which is like
6202 ERR_error_string(e, buf), but writes at most 'len' bytes
6203 including the 0 terminator. For ERR_error_string_n, 'buf'
6205 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
6207 *) CONF library reworked to become more general. A new CONF
6208 configuration file reader "class" is implemented as well as a
6209 new functions (NCONF_*, for "New CONF") to handle it. The now
6210 old CONF_* functions are still there, but are reimplemented to
6211 work in terms of the new functions. Also, a set of functions
6212 to handle the internal storage of the configuration data is
6213 provided to make it easier to write new configuration file
6214 reader "classes" (I can definitely see something reading a
6215 configuration file in XML format, for example), called _CONF_*,
6216 or "the configuration storage API"...
6218 The new configuration file reading functions are:
6220 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
6221 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
6223 NCONF_default, NCONF_WIN32
6225 NCONF_dump_fp, NCONF_dump_bio
6227 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
6228 NCONF_new creates a new CONF object. This works in the same way
6229 as other interfaces in OpenSSL, like the BIO interface.
6230 NCONF_dump_* dump the internal storage of the configuration file,
6231 which is useful for debugging. All other functions take the same
6232 arguments as the old CONF_* functions wth the exception of the
6233 first that must be a `CONF *' instead of a `LHASH *'.
6235 To make it easer to use the new classes with the old CONF_* functions,
6236 the function CONF_set_default_method is provided.
6239 *) Add '-tls1' option to 'openssl ciphers', which was already
6240 mentioned in the documentation but had not been implemented.
6241 (This option is not yet really useful because even the additional
6242 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
6245 *) Initial DSO code added into libcrypto for letting OpenSSL (and
6246 OpenSSL-based applications) load shared libraries and bind to
6247 them in a portable way.
6248 [Geoff Thorpe, with contributions from Richard Levitte]
6250 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
6252 *) Make sure _lrotl and _lrotr are only used with MSVC.
6254 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
6255 (the default implementation of RAND_status).
6257 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
6258 to '-clrext' (= clear extensions), as intended and documented.
6259 [Bodo Moeller; inconsistency pointed out by Michael Attili
6262 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
6263 was larger than the MD block size.
6264 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
6266 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
6267 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
6268 using the passed key: if the passed key was a private key the result
6269 of X509_print(), for example, would be to print out all the private key
6273 *) des_quad_cksum() byte order bug fix.
6274 [Ulf Möller, using the problem description in krb4-0.9.7, where
6275 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
6277 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
6279 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
6281 *) For easily testing in shell scripts whether some command
6282 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
6283 returns with exit code 0 iff no command of the given name is available.
6284 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
6285 the output goes to stdout and nothing is printed to stderr.
6286 Additional arguments are always ignored.
6288 Since for each cipher there is a command of the same name,
6289 the 'no-cipher' compilation switches can be tested this way.
6291 ('openssl no-XXX' is not able to detect pseudo-commands such
6292 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
6295 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
6298 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
6299 is set; it will be thrown away anyway because each handshake creates
6301 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
6302 to parameters -- in previous versions (since OpenSSL 0.9.3) the
6303 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
6304 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
6307 *) New s_client option -ign_eof: EOF at stdin is ignored, and
6308 'Q' and 'R' lose their special meanings (quit/renegotiate).
6309 This is part of what -quiet does; unlike -quiet, -ign_eof
6310 does not suppress any output.
6313 *) Add compatibility options to the purpose and trust code. The
6314 purpose X509_PURPOSE_ANY is "any purpose" which automatically
6315 accepts a certificate or CA, this was the previous behaviour,
6316 with all the associated security issues.
6318 X509_TRUST_COMPAT is the old trust behaviour: only and
6319 automatically trust self signed roots in certificate store. A
6320 new trust setting X509_TRUST_DEFAULT is used to specify that
6321 a purpose has no associated trust setting and it should instead
6322 use the value in the default purpose.
6325 *) Fix the PKCS#8 DSA private key code so it decodes keys again
6326 and fix a memory leak.
6329 *) In util/mkerr.pl (which implements 'make errors'), preserve
6330 reason strings from the previous version of the .c file, as
6331 the default to have only downcase letters (and digits) in
6332 automatically generated reasons codes is not always appropriate.
6335 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
6336 using strerror. Previously, ERR_reason_error_string() returned
6337 library names as reason strings for SYSerr; but SYSerr is a special
6338 case where small numbers are errno values, not library numbers.
6341 *) Add '-dsaparam' option to 'openssl dhparam' application. This
6342 converts DSA parameters into DH parameters. (When creating parameters,
6343 DSA_generate_parameters is used.)
6346 *) Include 'length' (recommended exponent length) in C code generated
6347 by 'openssl dhparam -C'.
6350 *) The second argument to set_label in perlasm was already being used
6351 so couldn't be used as a "file scope" flag. Moved to third argument
6355 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
6356 instead of RAND_bytes for encryption IVs and salts.
6359 *) Include RAND_status() into RAND_METHOD instead of implementing
6360 it only for md_rand.c Otherwise replacing the PRNG by calling
6361 RAND_set_rand_method would be impossible.
6364 *) Don't let DSA_generate_key() enter an infinite loop if the random
6365 number generation fails.
6368 *) New 'rand' application for creating pseudo-random output.
6371 *) Added configuration support for Linux/IA64
6372 [Rolf Haberrecker <rolf@suse.de>]
6374 *) Assembler module support for Mingw32.
6377 *) Shared library support for HPUX (in shlib/).
6378 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
6380 *) Shared library support for Solaris gcc.
6381 [Lutz Behnke <behnke@trustcenter.de>]
6383 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
6385 *) PKCS7_encrypt() was adding text MIME headers twice because they
6386 were added manually and by SMIME_crlf_copy().
6389 *) In bntest.c don't call BN_rand with zero bits argument.
6390 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
6392 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
6393 case was implemented. This caused BN_div_recp() to fail occasionally.
6396 *) Add an optional second argument to the set_label() in the perl
6397 assembly language builder. If this argument exists and is set
6398 to 1 it signals that the assembler should use a symbol whose
6399 scope is the entire file, not just the current function. This
6400 is needed with MASM which uses the format label:: for this scope.
6401 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
6403 *) Change the ASN1 types so they are typedefs by default. Before
6404 almost all types were #define'd to ASN1_STRING which was causing
6405 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
6409 *) Change names of new functions to the new get1/get0 naming
6410 convention: After 'get1', the caller owns a reference count
6411 and has to call ..._free; 'get0' returns a pointer to some
6412 data structure without incrementing reference counters.
6413 (Some of the existing 'get' functions increment a reference
6414 counter, some don't.)
6415 Similarly, 'set1' and 'add1' functions increase reference
6416 counters or duplicate objects.
6419 *) Allow for the possibility of temp RSA key generation failure:
6420 the code used to assume it always worked and crashed on failure.
6423 *) Fix potential buffer overrun problem in BIO_printf().
6424 [Ulf Möller, using public domain code by Patrick Powell; problem
6425 pointed out by David Sacerdote <das33@cornell.edu>]
6427 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
6428 RAND_egd() and RAND_status(). In the command line application,
6429 the EGD socket can be specified like a seed file using RANDFILE
6433 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
6434 Some CAs (e.g. Verisign) distribute certificates in this form.
6437 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
6438 list to exclude them. This means that no special compilation option
6439 is needed to use anonymous DH: it just needs to be included in the
6443 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
6444 EVP_MD_type. The old functionality is available in a new macro called
6445 EVP_MD_md(). Change code that uses it and update docs.
6448 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
6449 where the 'void *' argument is replaced by a function pointer argument.
6450 Previously 'void *' was abused to point to functions, which works on
6451 many platforms, but is not correct. As these functions are usually
6452 called by macros defined in OpenSSL header files, most source code
6453 should work without changes.
6456 *) <openssl/opensslconf.h> (which is created by Configure) now contains
6457 sections with information on -D... compiler switches used for
6458 compiling the library so that applications can see them. To enable
6459 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
6460 must be defined. E.g.,
6461 #define OPENSSL_ALGORITHM_DEFINES
6462 #include <openssl/opensslconf.h>
6463 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
6464 [Richard Levitte, Ulf and Bodo Möller]
6466 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
6470 *) Change the 'other' type in certificate aux info to a STACK_OF
6471 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
6472 the required ASN1 format: arbitrary types determined by an OID.
6475 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
6476 argument to 'req'. This is not because the function is newer or
6477 better than others it just uses the work 'NEW' in the certificate
6478 request header lines. Some software needs this.
6481 *) Reorganise password command line arguments: now passwords can be
6482 obtained from various sources. Delete the PEM_cb function and make
6483 it the default behaviour: i.e. if the callback is NULL and the
6484 usrdata argument is not NULL interpret it as a null terminated pass
6485 phrase. If usrdata and the callback are NULL then the pass phrase
6486 is prompted for as usual.
6489 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
6490 the support is automatically enabled. The resulting binaries will
6491 autodetect the card and use it if present.
6492 [Ben Laurie and Compaq Inc.]
6494 *) Work around for Netscape hang bug. This sends certificate request
6495 and server done in one record. Since this is perfectly legal in the
6496 SSL/TLS protocol it isn't a "bug" option and is on by default. See
6497 the bugs/SSLv3 entry for more info.
6500 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
6503 *) Add -rand argument to smime and pkcs12 applications and read/write
6507 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
6510 *) Add command line password options to the remaining applications.
6513 *) Bug fix for BN_div_recp() for numerators with an even number of
6517 *) More tests in bntest.c, and changed test_bn output.
6520 *) ./config recognizes MacOS X now.
6523 *) Bug fix for BN_div() when the first words of num and divsor are
6524 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
6527 *) Add support for various broken PKCS#8 formats, and command line
6528 options to produce them.
6531 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
6532 get temporary BIGNUMs from a BN_CTX.
6535 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
6539 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
6540 include a #define from the old name to the new. The original intent
6541 was that statically linked binaries could for example just call
6542 SSLeay_add_all_ciphers() to just add ciphers to the table and not
6543 link with digests. This never worked becayse SSLeay_add_all_digests()
6544 and SSLeay_add_all_ciphers() were in the same source file so calling
6545 one would link with the other. They are now in separate source files.
6548 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
6551 *) Use a less unusual form of the Miller-Rabin primality test (it used
6552 a binary algorithm for exponentiation integrated into the Miller-Rabin
6553 loop, our standard modexp algorithms are faster).
6556 *) Support for the EBCDIC character set completed.
6557 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
6559 *) Source code cleanups: use const where appropriate, eliminate casts,
6560 use void * instead of char * in lhash.
6563 *) Bugfix: ssl3_send_server_key_exchange was not restartable
6564 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
6565 this the server could overwrite ephemeral keys that the client
6569 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
6570 using 50 iterations of the Rabin-Miller test.
6572 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
6573 iterations of the Rabin-Miller test as required by the appendix
6574 to FIPS PUB 186[-1]) instead of DSA_is_prime.
6575 As BN_is_prime_fasttest includes trial division, DSA parameter
6576 generation becomes much faster.
6578 This implies a change for the callback functions in DSA_is_prime
6579 and DSA_generate_parameters: The callback function is called once
6580 for each positive witness in the Rabin-Miller test, not just
6581 occasionally in the inner loop; and the parameters to the
6582 callback function now provide an iteration count for the outer
6583 loop rather than for the current invocation of the inner loop.
6584 DSA_generate_parameters additionally can call the callback
6585 function with an 'iteration count' of -1, meaning that a
6586 candidate has passed the trial division test (when q is generated
6587 from an application-provided seed, trial division is skipped).
6590 *) New function BN_is_prime_fasttest that optionally does trial
6591 division before starting the Rabin-Miller test and has
6592 an additional BN_CTX * argument (whereas BN_is_prime always
6593 has to allocate at least one BN_CTX).
6594 'callback(1, -1, cb_arg)' is called when a number has passed the
6595 trial division stage.
6598 *) Fix for bug in CRL encoding. The validity dates weren't being handled
6602 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
6605 *) New function BN_pseudo_rand().
6608 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
6609 bignum version of BN_from_montgomery() with the working code from
6610 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
6614 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
6615 made it impossible to use the same SSL_SESSION data structure in
6616 SSL2 clients in multiple threads.
6619 *) The return value of RAND_load_file() no longer counts bytes obtained
6620 by stat(). RAND_load_file(..., -1) is new and uses the complete file
6621 to seed the PRNG (previously an explicit byte count was required).
6622 [Ulf Möller, Bodo Möller]
6624 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
6625 used (char *) instead of (void *) and had casts all over the place.
6628 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
6631 *) Retain source code compatibility for BN_prime_checks macro:
6632 BN_is_prime(..., BN_prime_checks, ...) now uses
6633 BN_prime_checks_for_size to determine the appropriate number of
6634 Rabin-Miller iterations.
6637 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
6638 DH_CHECK_P_NOT_SAFE_PRIME.
6639 (Check if this is true? OpenPGP calls them "strong".)
6642 *) Merge the functionality of "dh" and "gendh" programs into a new program
6643 "dhparam". The old programs are retained for now but will handle DH keys
6644 (instead of parameters) in future.
6647 *) Make the ciphers, s_server and s_client programs check the return values
6648 when a new cipher list is set.
6651 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
6652 ciphers. Before when the 56bit ciphers were enabled the sorting was
6655 The syntax for the cipher sorting has been extended to support sorting by
6656 cipher-strength (using the strength_bits hard coded in the tables).
6657 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
6659 Fix a bug in the cipher-command parser: when supplying a cipher command
6660 string with an "undefined" symbol (neither command nor alphanumeric
6661 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
6662 an error is flagged.
6664 Due to the strength-sorting extension, the code of the
6665 ssl_create_cipher_list() function was completely rearranged. I hope that
6666 the readability was also increased :-)
6667 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6669 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
6670 for the first serial number and places 2 in the serial number file. This
6671 avoids problems when the root CA is created with serial number zero and
6672 the first user certificate has the same issuer name and serial number
6676 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
6677 the new code. Add documentation for this stuff.
6680 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
6681 X509_*() to X509at_*() on the grounds that they don't handle X509
6682 structures and behave in an analagous way to the X509v3 functions:
6683 they shouldn't be called directly but wrapper functions should be used
6686 So we also now have some wrapper functions that call the X509at functions
6687 when passed certificate requests. (TO DO: similar things can be done with
6688 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
6689 things. Some of these need some d2i or i2d and print functionality
6690 because they handle more complex structures.)
6693 *) Add missing #ifndefs that caused missing symbols when building libssl
6694 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
6695 NO_RSA in ssl/s2*.c.
6696 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
6698 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
6699 has a return value which indicates the quality of the random data
6700 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
6701 error queue. New function RAND_pseudo_bytes() generates output that is
6702 guaranteed to be unique but not unpredictable. RAND_add is like
6703 RAND_seed, but takes an extra argument for an entropy estimate
6704 (RAND_seed always assumes full entropy).
6707 *) Do more iterations of Rabin-Miller probable prime test (specifically,
6708 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
6709 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
6710 in crypto/bn/bn_prime.c for the complete table). This guarantees a
6711 false-positive rate of at most 2^-80 for random input.
6714 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
6717 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
6718 in the 0.9.5 release), this returns the chain
6719 from an X509_CTX structure with a dup of the stack and all
6720 the X509 reference counts upped: so the stack will exist
6721 after X509_CTX_cleanup() has been called. Modify pkcs12.c
6724 Also make SSL_SESSION_print() print out the verify return
6728 *) Add manpage for the pkcs12 command. Also change the default
6729 behaviour so MAC iteration counts are used unless the new
6730 -nomaciter option is used. This improves file security and
6731 only older versions of MSIE (4.0 for example) need it.
6734 *) Honor the no-xxx Configure options when creating .DEF files.
6737 *) Add PKCS#10 attributes to field table: challengePassword,
6738 unstructuredName and unstructuredAddress. These are taken from
6739 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
6740 international characters are used.
6742 More changes to X509_ATTRIBUTE code: allow the setting of types
6743 based on strings. Remove the 'loc' parameter when adding
6744 attributes because these will be a SET OF encoding which is sorted
6748 *) Initial changes to the 'req' utility to allow request generation
6749 automation. This will allow an application to just generate a template
6750 file containing all the field values and have req construct the
6753 Initial support for X509_ATTRIBUTE handling. Stacks of these are
6754 used all over the place including certificate requests and PKCS#7
6755 structures. They are currently handled manually where necessary with
6756 some primitive wrappers for PKCS#7. The new functions behave in a
6757 manner analogous to the X509 extension functions: they allow
6758 attributes to be looked up by NID and added.
6760 Later something similar to the X509V3 code would be desirable to
6761 automatically handle the encoding, decoding and printing of the
6762 more complex types. The string types like challengePassword can
6763 be handled by the string table functions.
6765 Also modified the multi byte string table handling. Now there is
6766 a 'global mask' which masks out certain types. The table itself
6767 can use the flag STABLE_NO_MASK to ignore the mask setting: this
6768 is useful when for example there is only one permissible type
6769 (as in countryName) and using the mask might result in no valid
6773 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
6774 SSL_get_peer_finished to allow applications to obtain the latest
6775 Finished messages sent to the peer or expected from the peer,
6776 respectively. (SSL_get_peer_finished is usually the Finished message
6777 actually received from the peer, otherwise the protocol will be aborted.)
6779 As the Finished message are message digests of the complete handshake
6780 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
6781 be used for external authentication procedures when the authentication
6782 provided by SSL/TLS is not desired or is not enough.
6785 *) Enhanced support for Alpha Linux is added. Now ./config checks if
6786 the host supports BWX extension and if Compaq C is present on the
6787 $PATH. Just exploiting of the BWX extension results in 20-30%
6788 performance kick for some algorithms, e.g. DES and RC4 to mention
6789 a couple. Compaq C in turn generates ~20% faster code for MD5 and
6793 *) Add support for MS "fast SGC". This is arguably a violation of the
6794 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
6795 weak crypto and after checking the certificate is SGC a second one
6796 with strong crypto. MS SGC stops the first handshake after receiving
6797 the server certificate message and sends a second client hello. Since
6798 a server will typically do all the time consuming operations before
6799 expecting any further messages from the client (server key exchange
6800 is the most expensive) there is little difference between the two.
6802 To get OpenSSL to support MS SGC we have to permit a second client
6803 hello message after we have sent server done. In addition we have to
6804 reset the MAC if we do get this second client hello.
6807 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
6808 if a DER encoded private key is RSA or DSA traditional format. Changed
6809 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
6810 format DER encoded private key. Newer code should use PKCS#8 format which
6811 has the key type encoded in the ASN1 structure. Added DER private key
6812 support to pkcs8 application.
6815 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
6816 ciphersuites has been selected (as required by the SSL 3/TLS 1
6817 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
6818 is set, we interpret this as a request to violate the specification
6819 (the worst that can happen is a handshake failure, and 'correct'
6820 behaviour would result in a handshake failure anyway).
6823 *) In SSL_CTX_add_session, take into account that there might be multiple
6824 SSL_SESSION structures with the same session ID (e.g. when two threads
6825 concurrently obtain them from an external cache).
6826 The internal cache can handle only one SSL_SESSION with a given ID,
6827 so if there's a conflict, we now throw out the old one to achieve
6831 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
6832 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
6833 some routines that use cipher OIDs: some ciphers do not have OIDs
6834 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
6838 *) Simplify the trust setting structure and code. Now we just have
6839 two sequences of OIDs for trusted and rejected settings. These will
6840 typically have values the same as the extended key usage extension
6841 and any application specific purposes.
6843 The trust checking code now has a default behaviour: it will just
6844 check for an object with the same NID as the passed id. Functions can
6845 be provided to override either the default behaviour or the behaviour
6846 for a given id. SSL client, server and email already have functions
6847 in place for compatibility: they check the NID and also return "trusted"
6848 if the certificate is self signed.
6851 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
6852 traditional format into an EVP_PKEY structure.
6855 *) Add a password callback function PEM_cb() which either prompts for
6856 a password if usr_data is NULL or otherwise assumes it is a null
6857 terminated password. Allow passwords to be passed on command line
6858 environment or config files in a few more utilities.
6861 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
6862 keys. Add some short names for PKCS#8 PBE algorithms and allow them
6863 to be specified on the command line for the pkcs8 and pkcs12 utilities.
6864 Update documentation.
6867 *) Support for ASN1 "NULL" type. This could be handled before by using
6868 ASN1_TYPE but there wasn't any function that would try to read a NULL
6869 and produce an error if it couldn't. For compatibility we also have
6870 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
6871 don't allocate anything because they don't need to.
6874 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
6876 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
6878 *) Rebuild of the memory allocation routines used by OpenSSL code and
6879 possibly others as well. The purpose is to make an interface that
6880 provide hooks so anyone can build a separate set of allocation and
6881 deallocation routines to be used by OpenSSL, for example memory
6882 pool implementations, or something else, which was previously hard
6883 since Malloc(), Realloc() and Free() were defined as macros having
6884 the values malloc, realloc and free, respectively (except for Win32
6885 compilations). The same is provided for memory debugging code.
6886 OpenSSL already comes with functionality to find memory leaks, but
6887 this gives people a chance to debug other memory problems.
6889 With these changes, a new set of functions and macros have appeared:
6891 CRYPTO_set_mem_debug_functions() [F]
6892 CRYPTO_get_mem_debug_functions() [F]
6893 CRYPTO_dbg_set_options() [F]
6894 CRYPTO_dbg_get_options() [F]
6895 CRYPTO_malloc_debug_init() [M]
6897 The memory debug functions are NULL by default, unless the library
6898 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
6899 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
6900 gives the standard debugging functions that come with OpenSSL) or
6901 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
6902 provided by the library user) must be used. When the standard
6903 debugging functions are used, CRYPTO_dbg_set_options can be used to
6904 request additional information:
6905 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
6906 the CRYPTO_MDEBUG_xxx macro when compiling the library.
6908 Also, things like CRYPTO_set_mem_functions will always give the
6909 expected result (the new set of functions is used for allocation
6910 and deallocation) at all times, regardless of platform and compiler
6913 To finish it up, some functions that were never use in any other
6914 way than through macros have a new API and new semantic:
6917 CRYPTO_dbg_realloc()
6920 All macros of value have retained their old syntax.
6921 [Richard Levitte and Bodo Moeller]
6923 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
6924 ordering of SMIMECapabilities wasn't in "strength order" and there
6925 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
6929 *) Some ASN1 types with illegal zero length encoding (INTEGER,
6930 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
6931 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
6933 *) Merge in my S/MIME library for OpenSSL. This provides a simple
6934 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
6935 functionality to handle multipart/signed properly) and a utility
6936 called 'smime' to call all this stuff. This is based on code I
6937 originally wrote for Celo who have kindly allowed it to be
6938 included in OpenSSL.
6941 *) Add variants des_set_key_checked and des_set_key_unchecked of
6942 des_set_key (aka des_key_sched). Global variable des_check_key
6943 decides which of these is called by des_set_key; this way
6944 des_check_key behaves as it always did, but applications and
6945 the library itself, which was buggy for des_check_key == 1,
6946 have a cleaner way to pick the version they need.
6949 *) New function PKCS12_newpass() which changes the password of a
6953 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
6954 dynamic mix. In both cases the ids can be used as an index into the
6955 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
6956 functions so they accept a list of the field values and the
6957 application doesn't need to directly manipulate the X509_TRUST
6961 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
6965 *) Modify the way the V3 extension code looks up extensions. This now
6966 works in a similar way to the object code: we have some "standard"
6967 extensions in a static table which is searched with OBJ_bsearch()
6968 and the application can add dynamic ones if needed. The file
6969 crypto/x509v3/ext_dat.h now has the info: this file needs to be
6970 updated whenever a new extension is added to the core code and kept
6971 in ext_nid order. There is a simple program 'tabtest.c' which checks
6972 this. New extensions are not added too often so this file can readily
6973 be maintained manually.
6975 There are two big advantages in doing things this way. The extensions
6976 can be looked up immediately and no longer need to be "added" using
6977 X509V3_add_standard_extensions(): this function now does nothing.
6978 [Side note: I get *lots* of email saying the extension code doesn't
6979 work because people forget to call this function]
6980 Also no dynamic allocation is done unless new extensions are added:
6981 so if we don't add custom extensions there is no need to call
6982 X509V3_EXT_cleanup().
6985 *) Modify enc utility's salting as follows: make salting the default. Add a
6986 magic header, so unsalted files fail gracefully instead of just decrypting
6987 to garbage. This is because not salting is a big security hole, so people
6988 should be discouraged from doing it.
6991 *) Fixes and enhancements to the 'x509' utility. It allowed a message
6992 digest to be passed on the command line but it only used this
6993 parameter when signing a certificate. Modified so all relevant
6994 operations are affected by the digest parameter including the
6995 -fingerprint and -x509toreq options. Also -x509toreq choked if a
6996 DSA key was used because it didn't fix the digest.
6999 *) Initial certificate chain verify code. Currently tests the untrusted
7000 certificates for consistency with the verify purpose (which is set
7001 when the X509_STORE_CTX structure is set up) and checks the pathlength.
7003 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
7004 this is because it will reject chains with invalid extensions whereas
7005 every previous version of OpenSSL and SSLeay made no checks at all.
7007 Trust code: checks the root CA for the relevant trust settings. Trust
7008 settings have an initial value consistent with the verify purpose: e.g.
7009 if the verify purpose is for SSL client use it expects the CA to be
7010 trusted for SSL client use. However the default value can be changed to
7011 permit custom trust settings: one example of this would be to only trust
7012 certificates from a specific "secure" set of CAs.
7014 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
7015 which should be used for version portability: especially since the
7016 verify structure is likely to change more often now.
7018 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
7019 to set them. If not set then assume SSL clients will verify SSL servers
7022 Two new options to the verify program: -untrusted allows a set of
7023 untrusted certificates to be passed in and -purpose which sets the
7024 intended purpose of the certificate. If a purpose is set then the
7025 new chain verify code is used to check extension consistency.
7028 *) Support for the authority information access extension.
7031 *) Modify RSA and DSA PEM read routines to transparently handle
7032 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
7033 public keys in a format compatible with certificate
7034 SubjectPublicKeyInfo structures. Unfortunately there were already
7035 functions called *_PublicKey_* which used various odd formats so
7036 these are retained for compatibility: however the DSA variants were
7037 never in a public release so they have been deleted. Changed dsa/rsa
7038 utilities to handle the new format: note no releases ever handled public
7039 keys so we should be OK.
7041 The primary motivation for this change is to avoid the same fiasco
7042 that dogs private keys: there are several incompatible private key
7043 formats some of which are standard and some OpenSSL specific and
7044 require various evil hacks to allow partial transparent handling and
7045 even then it doesn't work with DER formats. Given the option anything
7046 other than PKCS#8 should be dumped: but the other formats have to
7047 stay in the name of compatibility.
7049 With public keys and the benefit of hindsight one standard format
7050 is used which works with EVP_PKEY, RSA or DSA structures: though
7051 it clearly returns an error if you try to read the wrong kind of key.
7053 Added a -pubkey option to the 'x509' utility to output the public key.
7054 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
7055 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
7056 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
7057 that do the same as the EVP_PKEY_assign_*() except they up the
7058 reference count of the added key (they don't "swallow" the
7062 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
7063 CRLs would fail if the file contained no certificates or no CRLs:
7064 added a new function to read in both types and return the number
7065 read: this means that if none are read it will be an error. The
7066 DER versions of the certificate and CRL reader would always fail
7067 because it isn't possible to mix certificates and CRLs in DER format
7068 without choking one or the other routine. Changed this to just read
7069 a certificate: this is the best we can do. Also modified the code
7070 in apps/verify.c to take notice of return codes: it was previously
7071 attempting to read in certificates from NULL pointers and ignoring
7072 any errors: this is one reason why the cert and CRL reader seemed
7073 to work. It doesn't check return codes from the default certificate
7074 routines: these may well fail if the certificates aren't installed.
7077 *) Code to support otherName option in GeneralName.
7080 *) First update to verify code. Change the verify utility
7081 so it warns if it is passed a self signed certificate:
7082 for consistency with the normal behaviour. X509_verify
7083 has been modified to it will now verify a self signed
7084 certificate if *exactly* the same certificate appears
7085 in the store: it was previously impossible to trust a
7086 single self signed certificate. This means that:
7087 openssl verify ss.pem
7088 now gives a warning about a self signed certificate but
7089 openssl verify -CAfile ss.pem ss.pem
7093 *) For servers, store verify_result in SSL_SESSION data structure
7094 (and add it to external session representation).
7095 This is needed when client certificate verifications fails,
7096 but an application-provided verification callback (set by
7097 SSL_CTX_set_cert_verify_callback) allows accepting the session
7098 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
7099 but returns 1): When the session is reused, we have to set
7100 ssl->verify_result to the appropriate error code to avoid
7102 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
7104 *) Fix a bug in the new PKCS#7 code: it didn't consider the
7105 case in PKCS7_dataInit() where the signed PKCS7 structure
7106 didn't contain any existing data because it was being created.
7107 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
7109 *) Add a salt to the key derivation routines in enc.c. This
7110 forms the first 8 bytes of the encrypted file. Also add a
7111 -S option to allow a salt to be input on the command line.
7114 *) New function X509_cmp(). Oddly enough there wasn't a function
7115 to compare two certificates. We do this by working out the SHA1
7116 hash and comparing that. X509_cmp() will be needed by the trust
7120 *) SSL_get1_session() is like SSL_get_session(), but increments
7121 the reference count in the SSL_SESSION returned.
7122 [Geoff Thorpe <geoff@eu.c2.net>]
7124 *) Fix for 'req': it was adding a null to request attributes.
7125 Also change the X509_LOOKUP and X509_INFO code to handle
7126 certificate auxiliary information.
7129 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
7133 *) Add the possibility to add extra information to the memory leak
7134 detecting output, to form tracebacks, showing from where each
7135 allocation was originated: CRYPTO_push_info("constant string") adds
7136 the string plus current file name and line number to a per-thread
7137 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
7138 is like calling CYRPTO_pop_info() until the stack is empty.
7139 Also updated memory leak detection code to be multi-thread-safe.
7142 *) Add options -text and -noout to pkcs7 utility and delete the
7143 encryption options which never did anything. Update docs.
7146 *) Add options to some of the utilities to allow the pass phrase
7147 to be included on either the command line (not recommended on
7148 OSes like Unix) or read from the environment. Update the
7149 manpages and fix a few bugs.
7152 *) Add a few manpages for some of the openssl commands.
7155 *) Fix the -revoke option in ca. It was freeing up memory twice,
7156 leaking and not finding already revoked certificates.
7159 *) Extensive changes to support certificate auxiliary information.
7160 This involves the use of X509_CERT_AUX structure and X509_AUX
7161 functions. An X509_AUX function such as PEM_read_X509_AUX()
7162 can still read in a certificate file in the usual way but it
7163 will also read in any additional "auxiliary information". By
7164 doing things this way a fair degree of compatibility can be
7165 retained: existing certificates can have this information added
7166 using the new 'x509' options.
7168 Current auxiliary information includes an "alias" and some trust
7169 settings. The trust settings will ultimately be used in enhanced
7170 certificate chain verification routines: currently a certificate
7171 can only be trusted if it is self signed and then it is trusted
7175 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
7176 The problem was that one of the replacement routines had not been working
7177 since SSLeay releases. For now the offending routine has been replaced
7178 with non-optimised assembler. Even so, this now gives around 95%
7179 performance improvement for 1024 bit RSA signs.
7182 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
7183 handling. Most clients have the effective key size in bits equal to
7184 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
7185 A few however don't do this and instead use the size of the decrypted key
7186 to determine the RC2 key length and the AlgorithmIdentifier to determine
7187 the effective key length. In this case the effective key length can still
7188 be 40 bits but the key length can be 168 bits for example. This is fixed
7189 by manually forcing an RC2 key into the EVP_PKEY structure because the
7190 EVP code can't currently handle unusual RC2 key sizes: it always assumes
7191 the key length and effective key length are equal.
7194 *) Add a bunch of functions that should simplify the creation of
7195 X509_NAME structures. Now you should be able to do:
7196 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
7197 and have it automatically work out the correct field type and fill in
7198 the structures. The more adventurous can try:
7199 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
7200 and it will (hopefully) work out the correct multibyte encoding.
7203 *) Change the 'req' utility to use the new field handling and multibyte
7204 copy routines. Before the DN field creation was handled in an ad hoc
7205 way in req, ca, and x509 which was rather broken and didn't support
7206 BMPStrings or UTF8Strings. Since some software doesn't implement
7207 BMPStrings or UTF8Strings yet, they can be enabled using the config file
7208 using the dirstring_type option. See the new comment in the default
7209 openssl.cnf for more info.
7212 *) Make crypto/rand/md_rand.c more robust:
7213 - Assure unique random numbers after fork().
7214 - Make sure that concurrent threads access the global counter and
7215 md serializably so that we never lose entropy in them
7216 or use exactly the same state in multiple threads.
7217 Access to the large state is not always serializable because
7218 the additional locking could be a performance killer, and
7219 md should be large enough anyway.
7222 *) New file apps/app_rand.c with commonly needed functionality
7223 for handling the random seed file.
7225 Use the random seed file in some applications that previously did not:
7227 dsaparam -genkey (which also ignored its '-rand' option),
7230 x509 (when signing).
7231 Except on systems with /dev/urandom, it is crucial to have a random
7232 seed file at least for key creation, DSA signing, and for DH exchanges;
7233 for RSA signatures we could do without one.
7235 gendh and gendsa (unlike genrsa) used to read only the first byte
7236 of each file listed in the '-rand' option. The function as previously
7237 found in genrsa is now in app_rand.c and is used by all programs
7238 that support '-rand'.
7241 *) In RAND_write_file, use mode 0600 for creating files;
7242 don't just chmod when it may be too late.
7245 *) Report an error from X509_STORE_load_locations
7246 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
7249 *) New function ASN1_mbstring_copy() this copies a string in either
7250 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
7251 into an ASN1_STRING type. A mask of permissible types is passed
7252 and it chooses the "minimal" type to use or an error if not type
7256 *) Add function equivalents to the various macros in asn1.h. The old
7257 macros are retained with an M_ prefix. Code inside the library can
7258 use the M_ macros. External code (including the openssl utility)
7259 should *NOT* in order to be "shared library friendly".
7262 *) Add various functions that can check a certificate's extensions
7263 to see if it usable for various purposes such as SSL client,
7264 server or S/MIME and CAs of these types. This is currently
7265 VERY EXPERIMENTAL but will ultimately be used for certificate chain
7266 verification. Also added a -purpose flag to x509 utility to
7267 print out all the purposes.
7270 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
7274 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
7275 for, obtain and decode and extension and obtain its critical flag.
7276 This allows all the necessary extension code to be handled in a
7277 single function call.
7280 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
7281 platforms. See crypto/rc4/rc4_enc.c for further details.
7284 *) New -noout option to asn1parse. This causes no output to be produced
7285 its main use is when combined with -strparse and -out to extract data
7286 from a file (which may not be in ASN.1 format).
7289 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
7290 when producing the local key id.
7291 [Richard Levitte <levitte@stacken.kth.se>]
7293 *) New option -dhparam in s_server. This allows a DH parameter file to be
7294 stated explicitly. If it is not stated then it tries the first server
7295 certificate file. The previous behaviour hard coded the filename
7299 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
7300 a public key to be input or output. For example:
7301 openssl rsa -in key.pem -pubout -out pubkey.pem
7302 Also added necessary DSA public key functions to handle this.
7305 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
7306 in the message. This was handled by allowing
7307 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
7308 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
7310 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
7311 to the end of the strings whereas this didn't. This would cause problems
7312 if strings read with d2i_ASN1_bytes() were later modified.
7313 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
7315 *) Fix for base64 decode bug. When a base64 bio reads only one line of
7316 data and it contains EOF it will end up returning an error. This is
7317 caused by input 46 bytes long. The cause is due to the way base64
7318 BIOs find the start of base64 encoded data. They do this by trying a
7319 trial decode on each line until they find one that works. When they
7320 do a flag is set and it starts again knowing it can pass all the
7321 data directly through the decoder. Unfortunately it doesn't reset
7322 the context it uses. This means that if EOF is reached an attempt
7323 is made to pass two EOFs through the context and this causes the
7324 resulting error. This can also cause other problems as well. As is
7325 usual with these problems it takes *ages* to find and the fix is
7326 trivial: move one line.
7327 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
7329 *) Ugly workaround to get s_client and s_server working under Windows. The
7330 old code wouldn't work because it needed to select() on sockets and the
7331 tty (for keypresses and to see if data could be written). Win32 only
7332 supports select() on sockets so we select() with a 1s timeout on the
7333 sockets and then see if any characters are waiting to be read, if none
7334 are present then we retry, we also assume we can always write data to
7335 the tty. This isn't nice because the code then blocks until we've
7336 received a complete line of data and it is effectively polling the
7337 keyboard at 1s intervals: however it's quite a bit better than not
7338 working at all :-) A dedicated Windows application might handle this
7339 with an event loop for example.
7342 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
7343 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
7344 will be called when RSA_sign() and RSA_verify() are used. This is useful
7345 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
7346 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
7347 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
7348 This necessitated the support of an extra signature type NID_md5_sha1
7349 for SSL signatures and modifications to the SSL library to use it instead
7350 of calling RSA_public_decrypt() and RSA_private_encrypt().
7353 *) Add new -verify -CAfile and -CApath options to the crl program, these
7354 will lookup a CRL issuers certificate and verify the signature in a
7355 similar way to the verify program. Tidy up the crl program so it
7356 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
7357 less strict. It will now permit CRL extensions even if it is not
7358 a V2 CRL: this will allow it to tolerate some broken CRLs.
7361 *) Initialize all non-automatic variables each time one of the openssl
7362 sub-programs is started (this is necessary as they may be started
7363 multiple times from the "OpenSSL>" prompt).
7364 [Lennart Bang, Bodo Moeller]
7366 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
7367 removing all other RSA functionality (this is what NO_RSA does). This
7368 is so (for example) those in the US can disable those operations covered
7369 by the RSA patent while allowing storage and parsing of RSA keys and RSA
7373 *) Non-copying interface to BIO pairs.
7374 (still largely untested)
7377 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
7378 ASCII string. This was handled independently in various places before.
7381 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
7382 UTF8 strings a character at a time.
7385 *) Use client_version from client hello to select the protocol
7386 (s23_srvr.c) and for RSA client key exchange verification
7387 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
7390 *) Add various utility functions to handle SPKACs, these were previously
7391 handled by poking round in the structure internals. Added new function
7392 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
7393 print, verify and generate SPKACs. Based on an original idea from
7394 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
7397 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
7400 *) Allow the config file extension section to be overwritten on the
7401 command line. Based on an original idea from Massimiliano Pala
7402 <madwolf@comune.modena.it>. The new option is called -extensions
7403 and can be applied to ca, req and x509. Also -reqexts to override
7404 the request extensions in req and -crlexts to override the crl extensions
7408 *) Add new feature to the SPKAC handling in ca. Now you can include
7409 the same field multiple times by preceding it by "XXXX." for example:
7412 this is the same syntax as used in the req config file.
7415 *) Allow certificate extensions to be added to certificate requests. These
7416 are specified in a 'req_extensions' option of the req section of the
7417 config file. They can be printed out with the -text option to req but
7418 are otherwise ignored at present.
7421 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
7422 data read consists of only the final block it would not decrypted because
7423 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
7424 A misplaced 'break' also meant the decrypted final block might not be
7425 copied until the next read.
7428 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
7429 a few extra parameters to the DH structure: these will be useful if
7430 for example we want the value of 'q' or implement X9.42 DH.
7433 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
7434 provides hooks that allow the default DSA functions or functions on a
7435 "per key" basis to be replaced. This allows hardware acceleration and
7436 hardware key storage to be handled without major modification to the
7437 library. Also added low level modexp hooks and CRYPTO_EX structure and
7438 associated functions.
7441 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
7442 as "read only": it can't be written to and the buffer it points to will
7443 not be freed. Reading from a read only BIO is much more efficient than
7444 a normal memory BIO. This was added because there are several times when
7445 an area of memory needs to be read from a BIO. The previous method was
7446 to create a memory BIO and write the data to it, this results in two
7447 copies of the data and an O(n^2) reading algorithm. There is a new
7448 function BIO_new_mem_buf() which creates a read only memory BIO from
7449 an area of memory. Also modified the PKCS#7 routines to use read only
7453 *) Bugfix: ssl23_get_client_hello did not work properly when called in
7454 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
7455 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
7456 but a retry condition occured while trying to read the rest.
7459 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
7460 NID_pkcs7_encrypted by default: this was wrong since this should almost
7461 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
7462 the encrypted data type: this is a more sensible place to put it and it
7463 allows the PKCS#12 code to be tidied up that duplicated this
7467 *) Changed obj_dat.pl script so it takes its input and output files on
7468 the command line. This should avoid shell escape redirection problems
7472 *) Initial support for certificate extension requests, these are included
7473 in things like Xenroll certificate requests. Included functions to allow
7474 extensions to be obtained and added.
7477 *) -crlf option to s_client and s_server for sending newlines as
7478 CRLF (as required by many protocols).
7481 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
7483 *) Install libRSAglue.a when OpenSSL is built with RSAref.
7484 [Ralf S. Engelschall]
7486 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
7487 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
7489 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
7493 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
7494 DH parameters/keys (q is lost during that conversion, but the resulting
7495 DH parameters contain its length).
7497 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
7498 much faster than DH_generate_parameters (which creates parameters
7499 where p = 2*q + 1), and also the smaller q makes DH computations
7500 much more efficient (160-bit exponentiation instead of 1024-bit
7501 exponentiation); so this provides a convenient way to support DHE
7502 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
7503 utter importance to use
7504 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
7506 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
7507 when such DH parameters are used, because otherwise small subgroup
7508 attacks may become possible!
7511 *) Avoid memory leak in i2d_DHparams.
7514 *) Allow the -k option to be used more than once in the enc program:
7515 this allows the same encrypted message to be read by multiple recipients.
7518 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
7519 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
7520 it will always use the numerical form of the OID, even if it has a short
7524 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
7525 method only got called if p,q,dmp1,dmq1,iqmp components were present,
7526 otherwise bn_mod_exp was called. In the case of hardware keys for example
7527 no private key components need be present and it might store extra data
7528 in the RSA structure, which cannot be accessed from bn_mod_exp.
7529 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
7530 private key operations.
7533 *) Added support for SPARC Linux.
7536 *) pem_password_cb function type incompatibly changed from
7537 typedef int pem_password_cb(char *buf, int size, int rwflag);
7539 ....(char *buf, int size, int rwflag, void *userdata);
7540 so that applications can pass data to their callbacks:
7541 The PEM[_ASN1]_{read,write}... functions and macros now take an
7542 additional void * argument, which is just handed through whenever
7543 the password callback is called.
7544 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
7546 New function SSL_CTX_set_default_passwd_cb_userdata.
7548 Compatibility note: As many C implementations push function arguments
7549 onto the stack in reverse order, the new library version is likely to
7550 interoperate with programs that have been compiled with the old
7551 pem_password_cb definition (PEM_whatever takes some data that
7552 happens to be on the stack as its last argument, and the callback
7553 just ignores this garbage); but there is no guarantee whatsoever that
7556 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
7557 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
7558 problems not only on Windows, but also on some Unix platforms.
7559 To avoid problematic command lines, these definitions are now in an
7560 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
7561 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
7564 *) MIPS III/IV assembler module is reimplemented.
7567 *) More DES library cleanups: remove references to srand/rand and
7568 delete an unused file.
7571 *) Add support for the the free Netwide assembler (NASM) under Win32,
7572 since not many people have MASM (ml) and it can be hard to obtain.
7573 This is currently experimental but it seems to work OK and pass all
7574 the tests. Check out INSTALL.W32 for info.
7577 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
7578 without temporary keys kept an extra copy of the server key,
7579 and connections with temporary keys did not free everything in case
7583 *) New function RSA_check_key and new openssl rsa option -check
7584 for verifying the consistency of RSA keys.
7585 [Ulf Moeller, Bodo Moeller]
7587 *) Various changes to make Win32 compile work:
7588 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
7589 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
7590 comparison" warnings.
7591 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
7594 *) Add a debugging option to PKCS#5 v2 key generation function: when
7595 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
7596 derived keys are printed to stderr.
7599 *) Copy the flags in ASN1_STRING_dup().
7600 [Roman E. Pavlov <pre@mo.msk.ru>]
7602 *) The x509 application mishandled signing requests containing DSA
7603 keys when the signing key was also DSA and the parameters didn't match.
7605 It was supposed to omit the parameters when they matched the signing key:
7606 the verifying software was then supposed to automatically use the CA's
7607 parameters if they were absent from the end user certificate.
7609 Omitting parameters is no longer recommended. The test was also
7610 the wrong way round! This was probably due to unusual behaviour in
7611 EVP_cmp_parameters() which returns 1 if the parameters match.
7612 This meant that parameters were omitted when they *didn't* match and
7613 the certificate was useless. Certificates signed with 'ca' didn't have
7615 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
7617 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
7618 The interface is as follows:
7619 Applications can use
7620 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
7621 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
7622 "off" is now the default.
7623 The library internally uses
7624 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
7625 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
7626 to disable memory-checking temporarily.
7628 Some inconsistent states that previously were possible (and were
7629 even the default) are now avoided.
7631 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
7632 with each memory chunk allocated; this is occasionally more helpful
7633 than just having a counter.
7635 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
7637 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
7641 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
7642 which largely parallels "options", but is for changing API behaviour,
7643 whereas "options" are about protocol behaviour.
7644 Initial "mode" flags are:
7646 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
7647 a single record has been written.
7648 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
7649 retries use the same buffer location.
7650 (But all of the contents must be
7654 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
7657 *) Fix problems with no-hmac etc.
7658 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
7660 *) New functions RSA_get_default_method(), RSA_set_method() and
7661 RSA_get_method(). These allows replacement of RSA_METHODs without having
7662 to mess around with the internals of an RSA structure.
7665 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
7666 Also really enable memory leak checks in openssl.c and in some
7668 [Chad C. Mulligan, Bodo Moeller]
7670 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
7671 up the length of negative integers. This has now been simplified to just
7672 store the length when it is first determined and use it later, rather
7673 than trying to keep track of where data is copied and updating it to
7675 [Steve Henson, reported by Brien Wheeler
7676 <bwheeler@authentica-security.com>]
7678 *) Add a new function PKCS7_signatureVerify. This allows the verification
7679 of a PKCS#7 signature but with the signing certificate passed to the
7680 function itself. This contrasts with PKCS7_dataVerify which assumes the
7681 certificate is present in the PKCS#7 structure. This isn't always the
7682 case: certificates can be omitted from a PKCS#7 structure and be
7683 distributed by "out of band" means (such as a certificate database).
7686 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
7687 function prototypes in pem.h, also change util/mkdef.pl to add the
7688 necessary function names.
7691 *) mk1mf.pl (used by Windows builds) did not properly read the
7692 options set by Configure in the top level Makefile, and Configure
7693 was not even able to write more than one option correctly.
7694 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
7697 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
7698 file to be loaded from a BIO or FILE pointer. The BIO version will
7699 for example allow memory BIOs to contain config info.
7702 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
7703 Whoever hopes to achieve shared-library compatibility across versions
7704 must use this, not the compile-time macro.
7705 (Exercise 0.9.4: Which is the minimum library version required by
7707 Note: All this applies only to multi-threaded programs, others don't
7711 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
7712 through a BIO pair triggered the default case, i.e.
7713 SSLerr(...,SSL_R_UNKNOWN_STATE).
7716 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
7717 can use the SSL library even if none of the specific BIOs is
7721 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
7722 for the encoded length.
7723 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
7725 *) Add initial documentation of the X509V3 functions.
7728 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
7729 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
7730 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
7731 secure PKCS#8 private key format with a high iteration count.
7734 *) Fix determination of Perl interpreter: A perl or perl5
7735 _directory_ in $PATH was also accepted as the interpreter.
7736 [Ralf S. Engelschall]
7738 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
7739 wrong with it but it was very old and did things like calling
7740 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
7744 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
7745 to use the new extension code.
7748 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
7749 with macros. This should make it easier to change their form, add extra
7750 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
7754 *) Add to configuration table a new entry that can specify an alternative
7755 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
7756 according to Mark Crispin <MRC@Panda.COM>.
7760 *) DES CBC did not update the IV. Weird.
7763 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
7764 Changing the behaviour of the former might break existing programs --
7765 where IV updating is needed, des_ncbc_encrypt can be used.
7768 *) When bntest is run from "make test" it drives bc to check its
7769 calculations, as well as internally checking them. If an internal check
7770 fails, it needs to cause bc to give a non-zero result or make test carries
7771 on without noticing the failure. Fixed.
7774 *) DES library cleanups.
7777 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
7778 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
7779 ciphers. NOTE: although the key derivation function has been verified
7780 against some published test vectors it has not been extensively tested
7781 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
7785 *) Instead of "mkdir -p", which is not fully portable, use new
7786 Perl script "util/mkdir-p.pl".
7789 *) Rewrite the way password based encryption (PBE) is handled. It used to
7790 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
7791 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
7792 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
7793 the 'parameter' field of the AlgorithmIdentifier is passed to the
7794 underlying key generation function so it must do its own ASN1 parsing.
7795 This has also changed the EVP_PBE_CipherInit() function which now has a
7796 'parameter' argument instead of literal salt and iteration count values
7797 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
7800 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
7801 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
7802 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
7803 KEY" because this clashed with PKCS#8 unencrypted string. Since this
7804 value was just used as a "magic string" and not used directly its
7805 value doesn't matter.
7808 *) Introduce some semblance of const correctness to BN. Shame C doesn't
7812 *) "linux-sparc64" configuration (ultrapenguin).
7813 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
7814 "linux-sparc" configuration.
7815 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
7817 *) config now generates no-xxx options for missing ciphers.
7820 *) Support the EBCDIC character set (work in progress).
7821 File ebcdic.c not yet included because it has a different license.
7822 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7824 *) Support BS2000/OSD-POSIX.
7825 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7827 *) Make callbacks for key generation use void * instead of char *.
7830 *) Make S/MIME samples compile (not yet tested).
7833 *) Additional typesafe stacks.
7836 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
7840 Changes between 0.9.3 and 0.9.3a [29 May 1999]
7842 *) New configuration variant "sco5-gcc".
7844 *) Updated some demos.
7845 [Sean O Riordain, Wade Scholine]
7847 *) Add missing BIO_free at exit of pkcs12 application.
7850 *) Fix memory leak in conf.c.
7853 *) Updates for Win32 to assembler version of MD5.
7856 *) Set #! path to perl in apps/der_chop to where we found it
7857 instead of using a fixed path.
7860 *) SHA library changes for irix64-mips4-cc.
7863 *) Improvements for VMS support.
7867 Changes between 0.9.2b and 0.9.3 [24 May 1999]
7869 *) Bignum library bug fix. IRIX 6 passes "make test" now!
7870 This also avoids the problems with SC4.2 and unpatched SC5.
7871 [Andy Polyakov <appro@fy.chalmers.se>]
7873 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
7874 These are required because of the typesafe stack would otherwise break
7875 existing code. If old code used a structure member which used to be STACK
7876 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
7877 sk_num or sk_value it would produce an error because the num, data members
7878 are not present in STACK_OF. Now it just produces a warning. sk_set
7879 replaces the old method of assigning a value to sk_value
7880 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
7881 that does this will no longer work (and should use sk_set instead) but
7882 this could be regarded as a "questionable" behaviour anyway.
7885 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
7886 correctly handle encrypted S/MIME data.
7889 *) Change type of various DES function arguments from des_cblock
7890 (which means, in function argument declarations, pointer to char)
7891 to des_cblock * (meaning pointer to array with 8 char elements),
7892 which allows the compiler to do more typechecking; it was like
7893 that back in SSLeay, but with lots of ugly casts.
7895 Introduce new type const_des_cblock.
7898 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
7899 problems: find RecipientInfo structure that matches recipient certificate
7900 and initialise the ASN1 structures properly based on passed cipher.
7903 *) Belatedly make the BN tests actually check the results.
7906 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
7907 to and from BNs: it was completely broken. New compilation option
7908 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
7909 key elements as negative integers.
7912 *) Reorganize and speed up MD5.
7913 [Andy Polyakov <appro@fy.chalmers.se>]
7916 [Richard Levitte <richard@levitte.org>]
7918 *) New option -out to asn1parse to allow the parsed structure to be
7919 output to a file. This is most useful when combined with the -strparse
7920 option to examine the output of things like OCTET STRINGS.
7923 *) Make SSL library a little more fool-proof by not requiring any longer
7924 that SSL_set_{accept,connect}_state be called before
7925 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
7926 in many applications because usually everything *appeared* to work as
7927 intended anyway -- now it really works as intended).
7930 *) Move openssl.cnf out of lib/.
7933 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
7934 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
7935 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
7936 [Ralf S. Engelschall]
7938 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
7939 handle PKCS#7 enveloped data properly.
7940 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
7942 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
7943 copying pointers. The cert_st handling is changed by this in
7944 various ways (and thus what used to be known as ctx->default_cert
7945 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
7946 any longer when s->cert does not give us what we need).
7947 ssl_cert_instantiate becomes obsolete by this change.
7948 As soon as we've got the new code right (possibly it already is?),
7949 we have solved a couple of bugs of the earlier code where s->cert
7950 was used as if it could not have been shared with other SSL structures.
7952 Note that using the SSL API in certain dirty ways now will result
7953 in different behaviour than observed with earlier library versions:
7954 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
7955 does not influence s as it used to.
7957 In order to clean up things more thoroughly, inside SSL_SESSION
7958 we don't use CERT any longer, but a new structure SESS_CERT
7959 that holds per-session data (if available); currently, this is
7960 the peer's certificate chain and, for clients, the server's certificate
7961 and temporary key. CERT holds only those values that can have
7962 meaningful defaults in an SSL_CTX.
7965 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
7966 from the internal representation. Various PKCS#7 fixes: remove some
7967 evil casts and set the enc_dig_alg field properly based on the signing
7971 *) Allow PKCS#12 password to be set from the command line or the
7972 environment. Let 'ca' get its config file name from the environment
7973 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
7977 *) Allow certificate policies extension to use an IA5STRING for the
7978 organization field. This is contrary to the PKIX definition but
7979 VeriSign uses it and IE5 only recognises this form. Document 'x509'
7983 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
7984 without disallowing inline assembler and the like for non-pedantic builds.
7987 *) Support Borland C++ builder.
7988 [Janez Jere <jj@void.si>, modified by Ulf Möller]
7993 *) SHA-1 cleanups and performance enhancements.
7994 [Andy Polyakov <appro@fy.chalmers.se>]
7996 *) Sparc v8plus assembler for the bignum library.
7997 [Andy Polyakov <appro@fy.chalmers.se>]
7999 *) Accept any -xxx and +xxx compiler options in Configure.
8002 *) Update HPUX configuration.
8005 *) Add missing sk_<type>_unshift() function to safestack.h
8006 [Ralf S. Engelschall]
8008 *) New function SSL_CTX_use_certificate_chain_file that sets the
8009 "extra_cert"s in addition to the certificate. (This makes sense
8010 only for "PEM" format files, as chains as a whole are not
8014 *) Support verify_depth from the SSL API.
8015 x509_vfy.c had what can be considered an off-by-one-error:
8016 Its depth (which was not part of the external interface)
8017 was actually counting the number of certificates in a chain;
8018 now it really counts the depth.
8021 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
8022 instead of X509err, which often resulted in confusing error
8023 messages since the error codes are not globally unique
8024 (e.g. an alleged error in ssl3_accept when a certificate
8025 didn't match the private key).
8027 *) New function SSL_CTX_set_session_id_context that allows to set a default
8028 value (so that you don't need SSL_set_session_id_context for each
8029 connection using the SSL_CTX).
8032 *) OAEP decoding bug fix.
8035 *) Support INSTALL_PREFIX for package builders, as proposed by
8039 *) New Configure options "threads" and "no-threads". For systems
8040 where the proper compiler options are known (currently Solaris
8041 and Linux), "threads" is the default.
8044 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
8047 *) Install various scripts to $(OPENSSLDIR)/misc, not to
8048 $(INSTALLTOP)/bin -- they shouldn't clutter directories
8049 such as /usr/local/bin.
8052 *) "make linux-shared" to build shared libraries.
8053 [Niels Poppe <niels@netbox.org>]
8055 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
8058 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
8059 extension adding in x509 utility.
8062 *) Remove NOPROTO sections and error code comments.
8065 *) Partial rewrite of the DEF file generator to now parse the ANSI
8069 *) New Configure options --prefix=DIR and --openssldir=DIR.
8072 *) Complete rewrite of the error code script(s). It is all now handled
8073 by one script at the top level which handles error code gathering,
8074 header rewriting and C source file generation. It should be much better
8075 than the old method: it now uses a modified version of Ulf's parser to
8076 read the ANSI prototypes in all header files (thus the old K&R definitions
8077 aren't needed for error creation any more) and do a better job of
8078 translating function codes into names. The old 'ASN1 error code imbedded
8079 in a comment' is no longer necessary and it doesn't use .err files which
8080 have now been deleted. Also the error code call doesn't have to appear all
8081 on one line (which resulted in some large lines...).
8084 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
8087 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
8088 0 (which usually indicates a closed connection), but continue reading.
8091 *) Fix some race conditions.
8094 *) Add support for CRL distribution points extension. Add Certificate
8095 Policies and CRL distribution points documentation.
8098 *) Move the autogenerated header file parts to crypto/opensslconf.h.
8101 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
8102 8 of keying material. Merlin has also confirmed interop with this fix
8103 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
8104 [Merlin Hughes <merlin@baltimore.ie>]
8106 *) Fix lots of warnings.
8107 [Richard Levitte <levitte@stacken.kth.se>]
8109 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
8110 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
8111 [Richard Levitte <levitte@stacken.kth.se>]
8113 *) Fix problems with sizeof(long) == 8.
8114 [Andy Polyakov <appro@fy.chalmers.se>]
8116 *) Change functions to ANSI C.
8119 *) Fix typos in error codes.
8120 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
8122 *) Remove defunct assembler files from Configure.
8125 *) SPARC v8 assembler BIGNUM implementation.
8126 [Andy Polyakov <appro@fy.chalmers.se>]
8128 *) Support for Certificate Policies extension: both print and set.
8129 Various additions to support the r2i method this uses.
8132 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
8133 return a const string when you are expecting an allocated buffer.
8136 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
8137 types DirectoryString and DisplayText.
8140 *) Add code to allow r2i extensions to access the configuration database,
8141 add an LHASH database driver and add several ctx helper functions.
8144 *) Fix an evil bug in bn_expand2() which caused various BN functions to
8145 fail when they extended the size of a BIGNUM.
8148 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
8149 support typesafe stack.
8152 *) Fix typo in SSL_[gs]et_options().
8153 [Nils Frostberg <nils@medcom.se>]
8155 *) Delete various functions and files that belonged to the (now obsolete)
8156 old X509V3 handling code.
8159 *) New Configure option "rsaref".
8162 *) Don't auto-generate pem.h.
8165 *) Introduce type-safe ASN.1 SETs.
8168 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
8169 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
8171 *) Introduce type-safe STACKs. This will almost certainly break lots of code
8172 that links with OpenSSL (well at least cause lots of warnings), but fear
8173 not: the conversion is trivial, and it eliminates loads of evil casts. A
8174 few STACKed things have been converted already. Feel free to convert more.
8175 In the fullness of time, I'll do away with the STACK type altogether.
8178 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
8179 specified in <certfile> by updating the entry in the index.txt file.
8180 This way one no longer has to edit the index.txt file manually for
8181 revoking a certificate. The -revoke option does the gory details now.
8182 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
8184 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
8185 `-text' option at all and this way the `-noout -text' combination was
8186 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
8187 [Ralf S. Engelschall]
8189 *) Make sure a corresponding plain text error message exists for the
8190 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
8191 verify callback function determined that a certificate was revoked.
8192 [Ralf S. Engelschall]
8194 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
8195 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
8196 all available cipers including rc5, which was forgotten until now.
8197 In order to let the testing shell script know which algorithms
8198 are available, a new (up to now undocumented) command
8199 "openssl list-cipher-commands" is used.
8202 *) Bugfix: s_client occasionally would sleep in select() when
8203 it should have checked SSL_pending() first.
8206 *) New functions DSA_do_sign and DSA_do_verify to provide access to
8207 the raw DSA values prior to ASN.1 encoding.
8210 *) Tweaks to Configure
8211 [Niels Poppe <niels@netbox.org>]
8213 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
8217 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
8220 *) New config option to avoid instructions that are illegal on the 80386.
8221 The default code is faster, but requires at least a 486.
8224 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
8225 SSL2_SERVER_VERSION (not used at all) macros, which are now the
8226 same as SSL2_VERSION anyway.
8229 *) New "-showcerts" option for s_client.
8232 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
8233 application. Various cleanups and fixes.
8236 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
8237 modify error routines to work internally. Add error codes and PBE init
8238 to library startup routines.
8241 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
8242 packing functions to asn1 and evp. Changed function names and error
8243 codes along the way.
8246 *) PKCS12 integration: and so it begins... First of several patches to
8247 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
8248 objects to objects.h
8251 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
8252 and display support for Thawte strong extranet extension.
8255 *) Add LinuxPPC support.
8256 [Jeff Dubrule <igor@pobox.org>]
8258 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
8259 bn_div_words in alpha.s.
8260 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
8262 *) Make sure the RSA OAEP test is skipped under -DRSAref because
8263 OAEP isn't supported when OpenSSL is built with RSAref.
8264 [Ulf Moeller <ulf@fitug.de>]
8266 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
8267 so they no longer are missing under -DNOPROTO.
8268 [Soren S. Jorvang <soren@t.dk>]
8271 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
8273 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
8274 doesn't work when the session is reused. Coming soon!
8277 *) Fix a security hole, that allows sessions to be reused in the wrong
8278 context thus bypassing client cert protection! All software that uses
8279 client certs and session caches in multiple contexts NEEDS PATCHING to
8280 allow session reuse! A fuller solution is in the works.
8281 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
8283 *) Some more source tree cleanups (removed obsolete files
8284 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
8285 permission on "config" script to be executable) and a fix for the INSTALL
8287 [Ulf Moeller <ulf@fitug.de>]
8289 *) Remove some legacy and erroneous uses of malloc, free instead of
8291 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
8293 *) Make rsa_oaep_test return non-zero on error.
8294 [Ulf Moeller <ulf@fitug.de>]
8296 *) Add support for native Solaris shared libraries. Configure
8297 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
8298 if someone would make that last step automatic.
8299 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
8301 *) ctx_size was not built with the right compiler during "make links". Fixed.
8304 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
8305 except NULL ciphers". This means the default cipher list will no longer
8306 enable NULL ciphers. They need to be specifically enabled e.g. with
8307 the string "DEFAULT:eNULL".
8310 *) Fix to RSA private encryption routines: if p < q then it would
8311 occasionally produce an invalid result. This will only happen with
8312 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
8315 *) Be less restrictive and allow also `perl util/perlpath.pl
8316 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
8317 because this way one can also use an interpreter named `perl5' (which is
8318 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
8319 installed as `perl').
8320 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8322 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
8323 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8325 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
8326 advapi32.lib to Win32 build and change the pem test comparision
8327 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
8328 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
8329 and crypto/des/ede_cbcm_enc.c.
8332 *) DES quad checksum was broken on big-endian architectures. Fixed.
8335 *) Comment out two functions in bio.h that aren't implemented. Fix up the
8336 Win32 test batch file so it (might) work again. The Win32 test batch file
8337 is horrible: I feel ill....
8340 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
8341 in e_os.h. Audit of header files to check ANSI and non ANSI
8342 sections: 10 functions were absent from non ANSI section and not exported
8343 from Windows DLLs. Fixed up libeay.num for new functions.
8346 *) Make `openssl version' output lines consistent.
8347 [Ralf S. Engelschall]
8349 *) Fix Win32 symbol export lists for BIO functions: Added
8350 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
8351 to ms/libeay{16,32}.def.
8352 [Ralf S. Engelschall]
8354 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
8355 fine under Unix and passes some trivial tests I've now added. But the
8356 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
8357 added to make sure no one expects that this stuff really works in the
8358 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
8359 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
8361 [Ralf S. Engelschall]
8363 *) Fix the generation of two part addresses in perl.
8364 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
8366 *) Add config entry for Linux on MIPS.
8367 [John Tobey <jtobey@channel1.com>]
8369 *) Make links whenever Configure is run, unless we are on Windoze.
8372 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
8373 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
8377 *) Add a useful kludge to allow package maintainers to specify compiler and
8378 other platforms details on the command line without having to patch the
8379 Configure script everytime: One now can use ``perl Configure
8380 <id>:<details>'', i.e. platform ids are allowed to have details appended
8381 to them (seperated by colons). This is treated as there would be a static
8382 pre-configured entry in Configure's %table under key <id> with value
8383 <details> and ``perl Configure <id>'' is called. So, when you want to
8384 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
8385 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
8386 now, which overrides the FreeBSD-elf entry on-the-fly.
8387 [Ralf S. Engelschall]
8389 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
8392 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
8393 on the `perl Configure ...' command line. This way one can compile
8394 OpenSSL libraries with Position Independent Code (PIC) which is needed
8395 for linking it into DSOs.
8396 [Ralf S. Engelschall]
8398 *) Remarkably, export ciphers were totally broken and no-one had noticed!
8402 *) Cleaned up the LICENSE document: The official contact for any license
8403 questions now is the OpenSSL core team under openssl-core@openssl.org.
8404 And add a paragraph about the dual-license situation to make sure people
8405 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
8406 to the OpenSSL toolkit.
8407 [Ralf S. Engelschall]
8409 *) General source tree makefile cleanups: Made `making xxx in yyy...'
8410 display consistent in the source tree and replaced `/bin/rm' by `rm'.
8411 Additonally cleaned up the `make links' target: Remove unnecessary
8412 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
8413 to speed processing and no longer clutter the display with confusing
8414 stuff. Instead only the actually done links are displayed.
8415 [Ralf S. Engelschall]
8417 *) Permit null encryption ciphersuites, used for authentication only. It used
8418 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
8419 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
8423 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
8424 signed attributes when verifying signatures (this would break them),
8425 the detached data encoding was wrong and public keys obtained using
8426 X509_get_pubkey() weren't freed.
8429 *) Add text documentation for the BUFFER functions. Also added a work around
8430 to a Win95 console bug. This was triggered by the password read stuff: the
8431 last character typed gets carried over to the next fread(). If you were
8432 generating a new cert request using 'req' for example then the last
8433 character of the passphrase would be CR which would then enter the first
8437 *) Added the new `Includes OpenSSL Cryptography Software' button as
8438 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
8439 button and can be used by applications based on OpenSSL to show the
8440 relationship to the OpenSSL project.
8441 [Ralf S. Engelschall]
8443 *) Remove confusing variables in function signatures in files
8444 ssl/ssl_lib.c and ssl/ssl.h.
8445 [Lennart Bong <lob@kulthea.stacken.kth.se>]
8447 *) Don't install bss_file.c under PREFIX/include/
8448 [Lennart Bong <lob@kulthea.stacken.kth.se>]
8450 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
8451 functions that return function pointers and has support for NT specific
8452 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
8453 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
8454 unsigned to signed types: this was killing the Win32 compile.
8457 *) Add new certificate file to stack functions,
8458 SSL_add_dir_cert_subjects_to_stack() and
8459 SSL_add_file_cert_subjects_to_stack(). These largely supplant
8460 SSL_load_client_CA_file(), and can be used to add multiple certs easily
8461 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
8462 This means that Apache-SSL and similar packages don't have to mess around
8463 to add as many CAs as they want to the preferred list.
8466 *) Experiment with doxygen documentation. Currently only partially applied to
8468 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
8469 openssl.doxy as the configuration file.
8472 *) Get rid of remaining C++-style comments which strict C compilers hate.
8473 [Ralf S. Engelschall, pointed out by Carlos Amengual]
8475 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
8476 compiled in by default: it has problems with large keys.
8479 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
8480 DH private keys and/or callback functions which directly correspond to
8481 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
8482 is needed for applications which have to configure certificates on a
8483 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
8485 For the RSA certificate situation is makes no difference, but
8486 for the DSA certificate situation this fixes the "no shared cipher"
8487 problem where the OpenSSL cipher selection procedure failed because the
8488 temporary keys were not overtaken from the context and the API provided
8489 no way to reconfigure them.
8490 The new functions now let applications reconfigure the stuff and they
8491 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
8492 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
8493 non-public-API function ssl_cert_instantiate() is used as a helper
8494 function and also to reduce code redundancy inside ssl_rsa.c.
8495 [Ralf S. Engelschall]
8497 *) Move s_server -dcert and -dkey options out of the undocumented feature
8498 area because they are useful for the DSA situation and should be
8499 recognized by the users.
8500 [Ralf S. Engelschall]
8502 *) Fix the cipher decision scheme for export ciphers: the export bits are
8503 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
8504 SSL_EXP_MASK. So, the original variable has to be used instead of the
8505 already masked variable.
8506 [Richard Levitte <levitte@stacken.kth.se>]
8508 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
8509 [Richard Levitte <levitte@stacken.kth.se>]
8511 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
8512 from `int' to `unsigned int' because it's a length and initialized by
8513 EVP_DigestFinal() which expects an `unsigned int *'.
8514 [Richard Levitte <levitte@stacken.kth.se>]
8516 *) Don't hard-code path to Perl interpreter on shebang line of Configure
8517 script. Instead use the usual Shell->Perl transition trick.
8518 [Ralf S. Engelschall]
8520 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
8521 (in addition to RSA certificates) to match the behaviour of `openssl dsa
8522 -noout -modulus' as it's already the case for `openssl rsa -noout
8523 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
8524 currently the public key is printed (a decision which was already done by
8525 `openssl dsa -modulus' in the past) which serves a similar purpose.
8526 Additionally the NO_RSA no longer completely removes the whole -modulus
8527 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
8529 [Ralf S. Engelschall]
8531 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
8532 BIO. See the source (crypto/evp/bio_ok.c) for more info.
8533 [Arne Ansper <arne@ats.cyber.ee>]
8535 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
8536 to be added. Now both 'req' and 'ca' can use new objects defined in the
8540 *) Add cool BIO that does syslog (or event log on NT).
8541 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
8543 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
8544 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
8545 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
8546 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
8549 *) Add preliminary config info for new extension code.
8552 *) Make RSA_NO_PADDING really use no padding.
8553 [Ulf Moeller <ulf@fitug.de>]
8555 *) Generate errors when private/public key check is done.
8558 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
8559 for some CRL extensions and new objects added.
8562 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
8563 key usage extension and fuller support for authority key id.
8566 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
8567 padding method for RSA, which is recommended for new applications in PKCS
8568 #1 v2.0 (RFC 2437, October 1998).
8569 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
8570 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
8571 against Bleichbacher's attack on RSA.
8572 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
8575 *) Updates to the new SSL compression code
8576 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8578 *) Fix so that the version number in the master secret, when passed
8579 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
8580 (because the server will not accept higher), that the version number
8581 is 0x03,0x01, not 0x03,0x00
8582 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8584 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
8585 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
8586 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
8589 *) Support for RAW extensions where an arbitrary extension can be
8590 created by including its DER encoding. See apps/openssl.cnf for
8594 *) Make sure latest Perl versions don't interpret some generated C array
8595 code as Perl array code in the crypto/err/err_genc.pl script.
8596 [Lars Weber <3weber@informatik.uni-hamburg.de>]
8598 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
8599 not many people have the assembler. Various Win32 compilation fixes and
8600 update to the INSTALL.W32 file with (hopefully) more accurate Win32
8604 *) Modify configure script 'Configure' to automatically create crypto/date.h
8605 file under Win32 and also build pem.h from pem.org. New script
8606 util/mkfiles.pl to create the MINFO file on environments that can't do a
8607 'make files': perl util/mkfiles.pl >MINFO should work.
8610 *) Major rework of DES function declarations, in the pursuit of correctness
8611 and purity. As a result, many evil casts evaporated, and some weirdness,
8612 too. You may find this causes warnings in your code. Zapping your evil
8613 casts will probably fix them. Mostly.
8616 *) Fix for a typo in asn1.h. Bug fix to object creation script
8617 obj_dat.pl. It considered a zero in an object definition to mean
8618 "end of object": none of the objects in objects.h have any zeros
8619 so it wasn't spotted.
8620 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
8622 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
8623 Masking (CBCM). In the absence of test vectors, the best I have been able
8624 to do is check that the decrypt undoes the encrypt, so far. Send me test
8625 vectors if you have them.
8628 *) Correct calculation of key length for export ciphers (too much space was
8629 allocated for null ciphers). This has not been tested!
8632 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
8633 message is now correct (it understands "crypto" and "ssl" on its
8634 command line). There is also now an "update" option. This will update
8635 the util/ssleay.num and util/libeay.num files with any new functions.
8637 perl util/mkdef.pl crypto ssl update
8638 it will update them.
8641 *) Overhauled the Perl interface (perl/*):
8642 - ported BN stuff to OpenSSL's different BN library
8643 - made the perl/ source tree CVS-aware
8644 - renamed the package from SSLeay to OpenSSL (the files still contain
8645 their history because I've copied them in the repository)
8646 - removed obsolete files (the test scripts will be replaced
8647 by better Test::Harness variants in the future)
8648 [Ralf S. Engelschall]
8650 *) First cut for a very conservative source tree cleanup:
8651 1. merge various obsolete readme texts into doc/ssleay.txt
8652 where we collect the old documents and readme texts.
8653 2. remove the first part of files where I'm already sure that we no
8654 longer need them because of three reasons: either they are just temporary
8655 files which were left by Eric or they are preserved original files where
8656 I've verified that the diff is also available in the CVS via "cvs diff
8657 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
8658 the crypto/md/ stuff).
8659 [Ralf S. Engelschall]
8661 *) More extension code. Incomplete support for subject and issuer alt
8662 name, issuer and authority key id. Change the i2v function parameters
8663 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
8664 what that's for :-) Fix to ASN1 macro which messed up
8665 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
8668 *) Preliminary support for ENUMERATED type. This is largely copied from the
8672 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
8673 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8675 *) Make sure `make rehash' target really finds the `openssl' program.
8676 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8678 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
8679 like to hear about it if this slows down other processors.
8682 *) Add CygWin32 platform information to Configure script.
8683 [Alan Batie <batie@aahz.jf.intel.com>]
8685 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
8686 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
8688 *) New program nseq to manipulate netscape certificate sequences
8691 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
8695 *) Fixes to BN code. Previously the default was to define BN_RECURSION
8696 but the BN code had some problems that would cause failures when
8697 doing certificate verification and some other functions.
8698 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8700 *) Add ASN1 and PEM code to support netscape certificate sequences.
8703 *) Add ASN1 and PEM code to support netscape certificate sequences.
8706 *) Add several PKIX and private extended key usage OIDs.
8709 *) Modify the 'ca' program to handle the new extension code. Modify
8710 openssl.cnf for new extension format, add comments.
8713 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
8714 and add a sample to openssl.cnf so req -x509 now adds appropriate
8718 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
8719 error code, add initial support to X509_print() and x509 application.
8722 *) Takes a deep breath and start addding X509 V3 extension support code. Add
8723 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
8724 stuff is currently isolated and isn't even compiled yet.
8727 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
8728 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
8729 Removed the versions check from X509 routines when loading extensions:
8730 this allows certain broken certificates that don't set the version
8731 properly to be processed.
8734 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
8735 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
8736 can still be regenerated with "make depend".
8739 *) Spelling mistake in C version of CAST-128.
8740 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
8742 *) Changes to the error generation code. The perl script err-code.pl
8743 now reads in the old error codes and retains the old numbers, only
8744 adding new ones if necessary. It also only changes the .err files if new
8745 codes are added. The makefiles have been modified to only insert errors
8746 when needed (to avoid needlessly modifying header files). This is done
8747 by only inserting errors if the .err file is newer than the auto generated
8748 C file. To rebuild all the error codes from scratch (the old behaviour)
8749 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
8750 or delete all the .err files.
8753 *) CAST-128 was incorrectly implemented for short keys. The C version has
8754 been fixed, but is untested. The assembler versions are also fixed, but
8755 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
8756 to regenerate it if needed.
8757 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
8758 Hagino <itojun@kame.net>]
8760 *) File was opened incorrectly in randfile.c.
8761 [Ulf Möller <ulf@fitug.de>]
8763 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
8764 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
8765 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
8766 al: it's just almost always a UTCTime. Note this patch adds new error
8767 codes so do a "make errors" if there are problems.
8770 *) Correct Linux 1 recognition in config.
8771 [Ulf Möller <ulf@fitug.de>]
8773 *) Remove pointless MD5 hash when using DSA keys in ca.
8774 [Anonymous <nobody@replay.com>]
8776 *) Generate an error if given an empty string as a cert directory. Also
8777 generate an error if handed NULL (previously returned 0 to indicate an
8778 error, but didn't set one).
8779 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
8781 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
8784 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
8785 parameters. This was causing a warning which killed off the Win32 compile.
8788 *) Remove C++ style comments from crypto/bn/bn_local.h.
8789 [Neil Costigan <neil.costigan@celocom.com>]
8791 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
8792 based on a text string, looking up short and long names and finally
8793 "dot" format. The "dot" format stuff didn't work. Added new function
8794 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
8795 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
8796 OID is not part of the table.
8799 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
8800 X509_LOOKUP_by_alias().
8803 *) Sort openssl functions by name.
8806 *) Get the gendsa program working (hopefully) and add it to app list. Remove
8807 encryption from sample DSA keys (in case anyone is interested the password
8811 *) Make _all_ *_free functions accept a NULL pointer.
8812 [Frans Heymans <fheymans@isaserver.be>]
8814 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
8816 [Anonymous <nobody@replay.com>]
8818 *) s_server should send the CAfile as acceptable CAs, not its own cert.
8819 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
8821 *) Don't blow it for numeric -newkey arguments to apps/req.
8822 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
8824 *) Temp key "for export" tests were wrong in s3_srvr.c.
8825 [Anonymous <nobody@replay.com>]
8827 *) Add prototype for temp key callback functions
8828 SSL_CTX_set_tmp_{rsa,dh}_callback().
8831 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
8832 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
8835 *) X509_name_add_entry() freed the wrong thing after an error.
8836 [Arne Ansper <arne@ats.cyber.ee>]
8838 *) rsa_eay.c would attempt to free a NULL context.
8839 [Arne Ansper <arne@ats.cyber.ee>]
8841 *) BIO_s_socket() had a broken should_retry() on Windoze.
8842 [Arne Ansper <arne@ats.cyber.ee>]
8844 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
8845 [Arne Ansper <arne@ats.cyber.ee>]
8847 *) Make sure the already existing X509_STORE->depth variable is initialized
8848 in X509_STORE_new(), but document the fact that this variable is still
8849 unused in the certificate verification process.
8850 [Ralf S. Engelschall]
8852 *) Fix the various library and apps files to free up pkeys obtained from
8853 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
8856 *) Fix reference counting in X509_PUBKEY_get(). This makes
8857 demos/maurice/example2.c work, amongst others, probably.
8858 [Steve Henson and Ben Laurie]
8860 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
8861 `openssl' and second, the shortcut symlinks for the `openssl <command>'
8862 are no longer created. This way we have a single and consistent command
8863 line interface `openssl <command>', similar to `cvs <command>'.
8864 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
8866 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
8867 BIT STRING wrapper always have zero unused bits.
8870 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
8873 *) Make the top-level INSTALL documentation easier to understand.
8876 *) Makefiles updated to exit if an error occurs in a sub-directory
8877 make (including if user presses ^C) [Paul Sutton]
8879 *) Make Montgomery context stuff explicit in RSA data structure.
8882 *) Fix build order of pem and err to allow for generated pem.h.
8885 *) Fix renumbering bug in X509_NAME_delete_entry().
8888 *) Enhanced the err-ins.pl script so it makes the error library number
8889 global and can add a library name. This is needed for external ASN1 and
8890 other error libraries.
8893 *) Fixed sk_insert which never worked properly.
8896 *) Fix ASN1 macros so they can handle indefinite length construted
8897 EXPLICIT tags. Some non standard certificates use these: they can now
8901 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
8902 into a single doc/ssleay.txt bundle. This way the information is still
8903 preserved but no longer messes up this directory. Now it's new room for
8904 the new set of documenation files.
8905 [Ralf S. Engelschall]
8907 *) SETs were incorrectly DER encoded. This was a major pain, because they
8908 shared code with SEQUENCEs, which aren't coded the same. This means that
8909 almost everything to do with SETs or SEQUENCEs has either changed name or
8910 number of arguments.
8911 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
8913 *) Fix test data to work with the above.
8916 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
8917 was already fixed by Eric for 0.9.1 it seems.
8918 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
8920 *) Autodetect FreeBSD3.
8923 *) Fix various bugs in Configure. This affects the following platforms:
8927 unixware-2.0-pentium
8931 *) Eliminate generated files from CVS. Reorder tests to regenerate files
8932 before they are needed.
8935 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
8939 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
8941 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
8942 changed SSLeay to OpenSSL in version strings.
8943 [Ralf S. Engelschall]
8945 *) Some fixups to the top-level documents.
8948 *) Fixed the nasty bug where rsaref.h was not found under compile-time
8949 because the symlink to include/ was missing.
8950 [Ralf S. Engelschall]
8952 *) Incorporated the popular no-RSA/DSA-only patches
8953 which allow to compile a RSA-free SSLeay.
8954 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
8956 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
8957 when "ssleay" is still not found.
8958 [Ralf S. Engelschall]
8960 *) Added more platforms to Configure: Cray T3E, HPUX 11,
8961 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
8963 *) Updated the README file.
8964 [Ralf S. Engelschall]
8966 *) Added various .cvsignore files in the CVS repository subdirs
8967 to make a "cvs update" really silent.
8968 [Ralf S. Engelschall]
8970 *) Recompiled the error-definition header files and added
8971 missing symbols to the Win32 linker tables.
8972 [Ralf S. Engelschall]
8974 *) Cleaned up the top-level documents;
8975 o new files: CHANGES and LICENSE
8976 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
8977 o merged COPYRIGHT into LICENSE
8978 o removed obsolete TODO file
8979 o renamed MICROSOFT to INSTALL.W32
8980 [Ralf S. Engelschall]
8982 *) Removed dummy files from the 0.9.1b source tree:
8983 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
8984 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
8985 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
8986 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
8987 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
8988 [Ralf S. Engelschall]
8990 *) Added various platform portability fixes.
8993 *) The Genesis of the OpenSSL rpject:
8994 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
8995 Young and Tim J. Hudson created while they were working for C2Net until
8997 [The OpenSSL Project]
9000 Changes between 0.9.0b and 0.9.1b [not released]
9002 *) Updated a few CA certificates under certs/
9005 *) Changed some BIGNUM api stuff.
9008 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
9009 DGUX x86, Linux Alpha, etc.
9012 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
9013 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
9017 *) Add -strparse option to asn1pars program which parses nested
9019 [Dr Stephen Henson <shenson@bigfoot.com>]
9021 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
9024 *) DSA fix for "ca" program.
9027 *) Added "-genkey" option to "dsaparam" program.
9030 *) Added RIPE MD160 (rmd160) message digest.
9033 *) Added -a (all) option to "ssleay version" command.
9036 *) Added PLATFORM define which is the id given to Configure.
9039 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
9042 *) Extended the ASN.1 parser routines.
9045 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
9048 *) Added a BN_CTX to the BN library.
9051 *) Fixed the weak key values in DES library
9054 *) Changed API in EVP library for cipher aliases.
9057 *) Added support for RC2/64bit cipher.
9060 *) Converted the lhash library to the crypto/mem.c functions.
9063 *) Added more recognized ASN.1 object ids.
9066 *) Added more RSA padding checks for SSL/TLS.
9069 *) Added BIO proxy/filter functionality.
9072 *) Added extra_certs to SSL_CTX which can be used
9073 send extra CA certificates to the client in the CA cert chain sending
9074 process. It can be configured with SSL_CTX_add_extra_chain_cert().
9077 *) Now Fortezza is denied in the authentication phase because
9078 this is key exchange mechanism is not supported by SSLeay at all.
9081 *) Additional PKCS1 checks.
9084 *) Support the string "TLSv1" for all TLS v1 ciphers.
9087 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
9088 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
9091 *) Fixed a few memory leaks.
9094 *) Fixed various code and comment typos.
9097 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
9098 bytes sent in the client random.
9099 [Edward Bishop <ebishop@spyglass.com>]