5 Changes between 0.9.8ze and 0.9.8zf [19 Mar 2015]
7 *) Segmentation fault in ASN1_TYPE_cmp fix
9 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
10 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
11 certificate signature algorithm consistency this can be used to crash any
12 certificate verification operation and exploited in a DoS attack. Any
13 application which performs certificate verification is vulnerable including
14 OpenSSL clients and servers which enable client authentication.
18 *) ASN.1 structure reuse memory corruption fix
20 Reusing a structure in ASN.1 parsing may allow an attacker to cause
21 memory corruption via an invalid write. Such reuse is and has been
22 strongly discouraged and is believed to be rare.
24 Applications that parse structures containing CHOICE or ANY DEFINED BY
25 components may be affected. Certificate parsing (d2i_X509 and related
26 functions) are however not affected. OpenSSL clients and servers are
31 *) PKCS7 NULL pointer dereferences fix
33 The PKCS#7 parsing code does not handle missing outer ContentInfo
34 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
35 missing content and trigger a NULL pointer dereference on parsing.
37 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
38 otherwise parse PKCS#7 structures from untrusted sources are
39 affected. OpenSSL clients and servers are not affected.
41 This issue was reported to OpenSSL by Michal Zalewski (Google).
45 *) DoS via reachable assert in SSLv2 servers fix
47 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
48 servers that both support SSLv2 and enable export cipher suites by sending
49 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
51 This issue was discovered by Sean Burford (Google) and Emilia Käsper
52 (OpenSSL development team).
56 *) Use After Free following d2i_ECPrivatekey error fix
58 A malformed EC private key file consumed via the d2i_ECPrivateKey function
59 could cause a use after free condition. This, in turn, could cause a double
60 free in several private key parsing functions (such as d2i_PrivateKey
61 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
62 for applications that receive EC private keys from untrusted
63 sources. This scenario is considered rare.
65 This issue was discovered by the BoringSSL project and fixed in their
70 *) X509_to_X509_REQ NULL pointer deref fix
72 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
73 the certificate key is invalid. This function is rarely used in practice.
75 This issue was discovered by Brian Carpenter.
79 *) Removed the export and SSLv2 ciphers from the DEFAULT ciphers
82 Changes between 0.9.8zd and 0.9.8ze [15 Jan 2015]
84 *) Build fixes for the Windows and OpenVMS platforms
85 [Matt Caswell and Richard Levitte]
87 Changes between 0.9.8zc and 0.9.8zd [8 Jan 2015]
89 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
90 message can cause a segmentation fault in OpenSSL due to a NULL pointer
91 dereference. This could lead to a Denial Of Service attack. Thanks to
92 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
96 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
97 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
98 method would be set to NULL which could later result in a NULL pointer
99 dereference. Thanks to Frank Schmirler for reporting this issue.
103 *) Abort handshake if server key exchange message is omitted for ephemeral
106 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
107 reporting this issue.
111 *) Remove non-export ephemeral RSA code on client and server. This code
112 violated the TLS standard by allowing the use of temporary RSA keys in
113 non-export ciphersuites and could be used by a server to effectively
114 downgrade the RSA key length used to a value smaller than the server
115 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
116 INRIA or reporting this issue.
120 *) Fix various certificate fingerprint issues.
122 By using non-DER or invalid encodings outside the signed portion of a
123 certificate the fingerprint can be changed without breaking the signature.
124 Although no details of the signed portion of the certificate can be changed
125 this can cause problems with some applications: e.g. those using the
126 certificate fingerprint for blacklists.
128 1. Reject signatures with non zero unused bits.
130 If the BIT STRING containing the signature has non zero unused bits reject
131 the signature. All current signature algorithms require zero unused bits.
133 2. Check certificate algorithm consistency.
135 Check the AlgorithmIdentifier inside TBS matches the one in the
136 certificate signature. NB: this will result in signature failure
137 errors for some broken certificates.
139 Thanks to Konrad Kraszewski from Google for reporting this issue.
141 3. Check DSA/ECDSA signatures use DER.
143 Reencode DSA/ECDSA signatures and compare with the original received
144 signature. Return an error if there is a mismatch.
146 This will reject various cases including garbage after signature
147 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
148 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
149 (negative or with leading zeroes).
151 Further analysis was conducted and fixes were developed by Stephen Henson
152 of the OpenSSL core team.
157 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
158 results on some platforms, including x86_64. This bug occurs at random
159 with a very low probability, and is not known to be exploitable in any
160 way, though its exact impact is difficult to determine. Thanks to Pieter
161 Wuille (Blockstream) who reported this issue and also suggested an initial
162 fix. Further analysis was conducted by the OpenSSL development team and
163 Adam Langley of Google. The final fix was developed by Andy Polyakov of
164 the OpenSSL core team.
168 Changes between 0.9.8zb and 0.9.8zc [15 Oct 2014]
170 *) Session Ticket Memory Leak.
172 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
173 integrity of that ticket is first verified. In the event of a session
174 ticket integrity check failing, OpenSSL will fail to free memory
175 causing a memory leak. By sending a large number of invalid session
176 tickets an attacker could exploit this issue in a Denial Of Service
181 *) Build option no-ssl3 is incomplete.
183 When OpenSSL is configured with "no-ssl3" as a build option, servers
184 could accept and complete a SSL 3.0 handshake, and clients could be
185 configured to send them.
187 [Akamai and the OpenSSL team]
189 *) Add support for TLS_FALLBACK_SCSV.
190 Client applications doing fallback retries should call
191 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
193 [Adam Langley, Bodo Moeller]
195 *) Add additional DigestInfo checks.
197 Reencode DigestInto in DER and check against the original when
198 verifying RSA signature: this will reject any improperly encoded
199 DigestInfo structures.
201 Note: this is a precautionary measure and no attacks are currently known.
205 Changes between 0.9.8za and 0.9.8zb [6 Aug 2014]
207 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
208 to a denial of service attack. A malicious server can crash the client
209 with a null pointer dereference (read) by specifying an anonymous (EC)DH
210 ciphersuite and sending carefully crafted handshake messages.
212 Thanks to Felix Gröbert (Google) for discovering and researching this
217 *) By sending carefully crafted DTLS packets an attacker could cause openssl
218 to leak memory. This can be exploited through a Denial of Service attack.
219 Thanks to Adam Langley for discovering and researching this issue.
223 *) An attacker can force openssl to consume large amounts of memory whilst
224 processing DTLS handshake messages. This can be exploited through a
225 Denial of Service attack.
226 Thanks to Adam Langley for discovering and researching this issue.
230 *) An attacker can force an error condition which causes openssl to crash
231 whilst processing DTLS packets due to memory being freed twice. This
232 can be exploited through a Denial of Service attack.
233 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
238 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
239 X509_name_oneline, X509_name_print_ex et al. to leak some information
240 from the stack. Applications may be affected if they echo pretty printing
241 output to the attacker.
243 Thanks to Ivan Fratric (Google) for discovering this issue.
245 [Emilia Käsper, and Steve Henson]
247 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
248 for corner cases. (Certain input points at infinity could lead to
249 bogus results, with non-infinity inputs mapped to infinity too.)
252 Changes between 0.9.8y and 0.9.8za [5 Jun 2014]
254 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
255 handshake can force the use of weak keying material in OpenSSL
256 SSL/TLS clients and servers.
258 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
259 researching this issue. (CVE-2014-0224)
260 [KIKUCHI Masashi, Steve Henson]
262 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
263 OpenSSL DTLS client the code can be made to recurse eventually crashing
266 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
268 [Imre Rad, Steve Henson]
270 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
271 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
272 client or server. This is potentially exploitable to run arbitrary
273 code on a vulnerable client or server.
275 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
276 [Jüri Aedla, Steve Henson]
278 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
279 are subject to a denial of service attack.
281 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
282 this issue. (CVE-2014-3470)
283 [Felix Gröbert, Ivan Fratric, Steve Henson]
285 *) Fix for the attack described in the paper "Recovering OpenSSL
286 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
287 by Yuval Yarom and Naomi Benger. Details can be obtained from:
288 http://eprint.iacr.org/2014/140
290 Thanks to Yuval Yarom and Naomi Benger for discovering this
291 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
292 [Yuval Yarom and Naomi Benger]
294 Thanks to mancha for backporting the fix to the 0.9.8 branch.
296 *) Fix handling of warning-level alerts in SSL23 client mode so they
297 don't cause client-side termination (eg. on SNI unrecognized_name
298 warnings). Add client and server support for six additional alerts
299 per RFC 6066 and RFC 4279.
302 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
303 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
304 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
305 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
306 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
307 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
308 [Rob Stradling, Adam Langley]
310 Changes between 0.9.8x and 0.9.8y [5 Feb 2013]
312 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
314 This addresses the flaw in CBC record processing discovered by
315 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
316 at: http://www.isg.rhul.ac.uk/tls/
318 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
319 Security Group at Royal Holloway, University of London
320 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
321 Emilia Käsper for the initial patch.
323 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
325 *) Return an error when checking OCSP signatures when key is NULL.
326 This fixes a DoS attack. (CVE-2013-0166)
329 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
330 the right response is stapled. Also change SSL_get_certificate()
331 so it returns the certificate actually sent.
332 See http://rt.openssl.org/Ticket/Display.html?id=2836.
334 [Rob Stradling <rob.stradling@comodo.com>]
336 *) Fix possible deadlock when decoding public keys.
339 Changes between 0.9.8w and 0.9.8x [10 May 2012]
341 *) Sanity check record length before skipping explicit IV in DTLS
344 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
345 fuzzing as a service testing platform.
349 *) Initialise tkeylen properly when encrypting CMS messages.
350 Thanks to Solar Designer of Openwall for reporting this issue.
353 Changes between 0.9.8v and 0.9.8w [23 Apr 2012]
355 *) The fix for CVE-2012-2110 did not take into account that the
356 'len' argument to BUF_MEM_grow and BUF_MEM_grow_clean is an
357 int in OpenSSL 0.9.8, making it still vulnerable. Fix by
358 rejecting negative len parameter. (CVE-2012-2131)
359 [Tomas Hoger <thoger@redhat.com>]
361 Changes between 0.9.8u and 0.9.8v [19 Apr 2012]
363 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
364 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
365 in CRYPTO_realloc_clean.
367 Thanks to Tavis Ormandy, Google Security Team, for discovering this
368 issue and to Adam Langley <agl@chromium.org> for fixing it.
370 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
372 Changes between 0.9.8t and 0.9.8u [12 Mar 2012]
374 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
375 in CMS and PKCS7 code. When RSA decryption fails use a random key for
376 content decryption and always return the same error. Note: this attack
377 needs on average 2^20 messages so it only affects automated senders. The
378 old behaviour can be reenabled in the CMS code by setting the
379 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
380 an MMA defence is not necessary.
381 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
382 this issue. (CVE-2012-0884)
385 *) Fix CVE-2011-4619: make sure we really are receiving a
386 client hello before rejecting multiple SGC restarts. Thanks to
387 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
390 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
392 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
393 Thanks to Antonio Martin, Enterprise Secure Access Research and
394 Development, Cisco Systems, Inc. for discovering this bug and
395 preparing a fix. (CVE-2012-0050)
398 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
400 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
401 of the Vaudenay padding oracle attack on CBC mode encryption
402 which enables an efficient plaintext recovery attack against
403 the OpenSSL implementation of DTLS. Their attack exploits timing
404 differences arising during decryption processing. A research
405 paper describing this attack can be found at:
406 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
407 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
408 Security Group at Royal Holloway, University of London
409 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
410 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
411 for preparing the fix. (CVE-2011-4108)
412 [Robin Seggelmann, Michael Tuexen]
414 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
415 [Ben Laurie, Kasper <ekasper@google.com>]
417 *) Clear bytes used for block padding of SSL 3.0 records.
419 [Adam Langley (Google)]
421 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
422 Kadianakis <desnacked@gmail.com> for discovering this issue and
423 Adam Langley for preparing the fix. (CVE-2011-4619)
424 [Adam Langley (Google)]
426 *) Prevent malformed RFC3779 data triggering an assertion failure.
427 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
428 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
429 [Rob Austein <sra@hactrn.net>]
431 *) Fix ssl_ciph.c set-up race.
432 [Adam Langley (Google)]
434 *) Fix spurious failures in ecdsatest.c.
435 [Emilia Käsper (Google)]
437 *) Fix the BIO_f_buffer() implementation (which was mixing different
438 interpretations of the '..._len' fields).
439 [Adam Langley (Google)]
441 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
442 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
443 threads won't reuse the same blinding coefficients.
445 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
446 lock to call BN_BLINDING_invert_ex, and avoids one use of
447 BN_BLINDING_update for each BN_BLINDING structure (previously,
448 the last update always remained unused).
449 [Emilia Käsper (Google)]
451 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
452 for multi-threaded use of ECDH.
453 [Adam Langley (Google)]
455 *) Fix x509_name_ex_d2i memory leak on bad inputs.
458 *) Add protection against ECDSA timing attacks as mentioned in the paper
459 by Billy Bob Brumley and Nicola Tuveri, see:
461 http://eprint.iacr.org/2011/232.pdf
463 [Billy Bob Brumley and Nicola Tuveri]
465 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
467 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
468 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
470 *) Fix bug in string printing code: if *any* escaping is enabled we must
471 escape the escape character (backslash) or the resulting string is
475 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
477 *) Disable code workaround for ancient and obsolete Netscape browsers
478 and servers: an attacker can use it in a ciphersuite downgrade attack.
479 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
482 *) Fixed J-PAKE implementation error, originally discovered by
483 Sebastien Martini, further info and confirmation from Stefan
484 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
487 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
489 *) Fix extension code to avoid race conditions which can result in a buffer
490 overrun vulnerability: resumed sessions must not be modified as they can
491 be shared by multiple threads. CVE-2010-3864
494 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
497 *) Don't reencode certificate when calculating signature: cache and use
498 the original encoding instead. This makes signature verification of
499 some broken encodings work correctly.
502 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
503 is also one of the inputs.
504 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
506 *) Don't repeatedly append PBE algorithms to table if they already exist.
507 Sort table on each new add. This effectively makes the table read only
508 after all algorithms are added and subsequent calls to PKCS12_pbe_add
512 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
514 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
517 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
518 access or freeing data twice (CVE-2010-0742)
519 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
521 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
522 common in certificates and some applications which only call
523 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
527 Reduce copying into .apps and .test in makevms.com
528 Don't try to use blank CA certificate in CA.com
529 Allow use of C files from original directories in maketests.com
530 [Steven M. Schweda" <sms@antinode.info>]
532 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
534 *) When rejecting SSL/TLS records due to an incorrect version number, never
535 update s->server with a new major version number. As of
536 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
537 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
538 the previous behavior could result in a read attempt at NULL when
539 receiving specific incorrect SSL/TLS records once record payload
540 protection is active. (CVE-2010-0740)
541 [Bodo Moeller, Adam Langley <agl@chromium.org>]
543 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
544 could be crashed if the relevant tables were not present (e.g. chrooted).
545 [Tomas Hoger <thoger@redhat.com>]
547 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
549 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
550 [Martin Olsson, Neel Mehta]
552 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
553 accommodate for stack sorting, always a write lock!).
556 *) On some versions of WIN32 Heap32Next is very slow. This can cause
557 excessive delays in the RAND_poll(): over a minute. As a workaround
558 include a time check in the inner Heap32Next loop too.
561 *) The code that handled flushing of data in SSL/TLS originally used the
562 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
563 the problem outlined in PR#1949. The fix suggested there however can
564 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
565 of Apache). So instead simplify the code to flush unconditionally.
566 This should be fine since flushing with no data to flush is a no op.
569 *) Handle TLS versions 2.0 and later properly and correctly use the
570 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
571 off ancient servers have a habit of sticking around for a while...
574 *) Modify compression code so it frees up structures without using the
575 ex_data callbacks. This works around a problem where some applications
576 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
577 restarting) then use compression (e.g. SSL with compression) later.
578 This results in significant per-connection memory leaks and
579 has caused some security issues including CVE-2008-1678 and
583 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
584 change when encrypting or decrypting.
587 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
588 connect and renegotiate with servers which do not support RI.
589 Until RI is more widely deployed this option is enabled by default.
592 *) Add "missing" ssl ctrls to clear options and mode.
595 *) If client attempts to renegotiate and doesn't support RI respond with
596 a no_renegotiation alert as required by RFC5746. Some renegotiating
597 TLS clients will continue a connection gracefully when they receive
598 the alert. Unfortunately OpenSSL mishandled this alert and would hang
599 waiting for a server hello which it will never receive. Now we treat a
600 received no_renegotiation alert as a fatal error. This is because
601 applications requesting a renegotiation might well expect it to succeed
602 and would have no code in place to handle the server denying it so the
603 only safe thing to do is to terminate the connection.
606 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
607 peer supports secure renegotiation and 0 otherwise. Print out peer
608 renegotiation support in s_client/s_server.
611 *) Replace the highly broken and deprecated SPKAC certification method with
612 the updated NID creation version. This should correctly handle UTF8.
615 *) Implement RFC5746. Re-enable renegotiation but require the extension
616 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
617 turns out to be a bad idea. It has been replaced by
618 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
619 SSL_CTX_set_options(). This is really not recommended unless you
620 know what you are doing.
621 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
623 *) Fixes to stateless session resumption handling. Use initial_ctx when
624 issuing and attempting to decrypt tickets in case it has changed during
625 servername handling. Use a non-zero length session ID when attempting
626 stateless session resumption: this makes it possible to determine if
627 a resumption has occurred immediately after receiving server hello
628 (several places in OpenSSL subtly assume this) instead of later in
632 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
633 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
634 fixes for a few places where the return code is not checked
636 [Julia Lawall <julia@diku.dk>]
638 *) Add --strict-warnings option to Configure script to include devteam
639 warnings in other configurations.
642 *) Add support for --libdir option and LIBDIR variable in makefiles. This
643 makes it possible to install openssl libraries in locations which
644 have names other than "lib", for example "/usr/lib64" which some
646 [Steve Henson, based on patch from Jeremy Utley]
648 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
649 X690 8.9.12 and can produce some misleading textual output of OIDs.
650 [Steve Henson, reported by Dan Kaminsky]
652 *) Delete MD2 from algorithm tables. This follows the recommendation in
653 several standards that it is not used in new applications due to
654 several cryptographic weaknesses. For binary compatibility reasons
655 the MD2 API is still compiled in by default.
658 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
662 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
663 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
665 [Guenter <lists@gknw.net>]
667 *) Fix the server certificate chain building code to use X509_verify_cert(),
668 it used to have an ad-hoc builder which was unable to cope with anything
669 other than a simple chain.
670 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
672 *) Don't check self signed certificate signatures in X509_verify_cert()
673 by default (a flag can override this): it just wastes time without
674 adding any security. As a useful side effect self signed root CAs
675 with non-FIPS digests are now usable in FIPS mode.
678 *) In dtls1_process_out_of_seq_message() the check if the current message
679 is already buffered was missing. For every new message was memory
680 allocated, allowing an attacker to perform an denial of service attack
681 with sending out of seq handshake messages until there is no memory
682 left. Additionally every future messege was buffered, even if the
683 sequence number made no sense and would be part of another handshake.
684 So only messages with sequence numbers less than 10 in advance will be
685 buffered. (CVE-2009-1378)
686 [Robin Seggelmann, discovered by Daniel Mentz]
688 *) Records are buffered if they arrive with a future epoch to be
689 processed after finishing the corresponding handshake. There is
690 currently no limitation to this buffer allowing an attacker to perform
691 a DOS attack with sending records with future epochs until there is no
692 memory left. This patch adds the pqueue_size() function to detemine
693 the size of a buffer and limits the record buffer to 100 entries.
695 [Robin Seggelmann, discovered by Daniel Mentz]
697 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
698 parent structure is freed. (CVE-2009-1379)
701 *) Handle non-blocking I/O properly in SSL_shutdown() call.
702 [Darryl Miles <darryl-mailinglists@netbauds.net>]
705 [Ilya O. <vrghost@gmail.com>]
707 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
709 *) Disable renegotiation completely - this fixes a severe security
710 problem (CVE-2009-3555) at the cost of breaking all
711 renegotiation. Renegotiation can be re-enabled by setting
712 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
713 run-time. This is really not recommended unless you know what
717 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
719 *) Don't set val to NULL when freeing up structures, it is freed up by
720 underlying code. If sizeof(void *) > sizeof(long) this can result in
721 zeroing past the valid field. (CVE-2009-0789)
722 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
724 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
725 checked correctly. This would allow some invalid signed attributes to
726 appear to verify correctly. (CVE-2009-0591)
727 [Ivan Nestlerode <inestlerode@us.ibm.com>]
729 *) Reject UniversalString and BMPString types with invalid lengths. This
730 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
731 a legal length. (CVE-2009-0590)
734 *) Set S/MIME signing as the default purpose rather than setting it
735 unconditionally. This allows applications to override it at the store
739 *) Permit restricted recursion of ASN1 strings. This is needed in practice
740 to handle some structures.
743 *) Improve efficiency of mem_gets: don't search whole buffer each time
745 [Jeremy Shapiro <jnshapir@us.ibm.com>]
747 *) New -hex option for openssl rand.
750 *) Print out UTF8String and NumericString when parsing ASN1.
753 *) Support NumericString type for name components.
756 *) Allow CC in the environment to override the automatically chosen
757 compiler. Note that nothing is done to ensure flags work with the
761 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
763 *) Properly check EVP_VerifyFinal() and similar return values
765 [Ben Laurie, Bodo Moeller, Google Security Team]
767 *) Enable TLS extensions by default.
770 *) Allow the CHIL engine to be loaded, whether the application is
771 multithreaded or not. (This does not release the developer from the
772 obligation to set up the dynamic locking callbacks.)
773 [Sander Temme <sander@temme.net>]
775 *) Use correct exit code if there is an error in dgst command.
776 [Steve Henson; problem pointed out by Roland Dirlewanger]
778 *) Tweak Configure so that you need to say "experimental-jpake" to enable
779 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
782 *) Add experimental JPAKE support, including demo authentication in
783 s_client and s_server.
786 *) Set the comparison function in v3_addr_canonize().
787 [Rob Austein <sra@hactrn.net>]
789 *) Add support for XMPP STARTTLS in s_client.
790 [Philip Paeps <philip@freebsd.org>]
792 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
793 to ensure that even with this option, only ciphersuites in the
794 server's preference list will be accepted. (Note that the option
795 applies only when resuming a session, so the earlier behavior was
796 just about the algorithm choice for symmetric cryptography.)
799 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
801 *) Fix NULL pointer dereference if a DTLS server received
802 ChangeCipherSpec as first record (CVE-2009-1386).
805 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
806 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
809 *) The fix in 0.9.8c that supposedly got rid of unsafe
810 double-checked locking was incomplete for RSA blinding,
811 addressing just one layer of what turns out to have been
812 doubly unsafe triple-checked locking.
814 So now fix this for real by retiring the MONT_HELPER macro
815 in crypto/rsa/rsa_eay.c.
817 [Bodo Moeller; problem pointed out by Marius Schilder]
819 *) Various precautionary measures:
821 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
823 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
824 (NB: This would require knowledge of the secret session ticket key
825 to exploit, in which case you'd be SOL either way.)
827 - Change bn_nist.c so that it will properly handle input BIGNUMs
828 outside the expected range.
830 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
833 [Neel Mehta, Bodo Moeller]
835 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
836 the load fails. Useful for distros.
837 [Ben Laurie and the FreeBSD team]
839 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
842 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
845 *) Expand ENGINE to support engine supplied SSL client certificate functions.
847 This work was sponsored by Logica.
850 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
851 keystores. Support for SSL/TLS client authentication too.
852 Not compiled unless enable-capieng specified to Configure.
854 This work was sponsored by Logica.
857 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
858 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
859 attribute creation routines such as certifcate requests and PKCS#12
863 Changes between 0.9.8g and 0.9.8h [28 May 2008]
865 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
866 handshake which could lead to a cilent crash as found using the
867 Codenomicon TLS test suite (CVE-2008-1672)
868 [Steve Henson, Mark Cox]
870 *) Fix double free in TLS server name extensions which could lead to
871 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
874 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
876 Clear the error queue to ensure that error entries left from
877 older function calls do not interfere with the correct operation.
878 [Lutz Jaenicke, Erik de Castro Lopo]
880 *) Remove root CA certificates of commercial CAs:
882 The OpenSSL project does not recommend any specific CA and does not
883 have any policy with respect to including or excluding any CA.
884 Therefore it does not make any sense to ship an arbitrary selection
885 of root CA certificates with the OpenSSL software.
888 *) RSA OAEP patches to fix two separate invalid memory reads.
889 The first one involves inputs when 'lzero' is greater than
890 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
891 before the beginning of from). The second one involves inputs where
892 the 'db' section contains nothing but zeroes (there is a one-byte
893 invalid read after the end of 'db').
894 [Ivan Nestlerode <inestlerode@us.ibm.com>]
896 *) Partial backport from 0.9.9-dev:
898 Introduce bn_mul_mont (dedicated Montgomery multiplication
899 procedure) as a candidate for BIGNUM assembler implementation.
900 While 0.9.9-dev uses assembler for various architectures, only
901 x86_64 is available by default here in the 0.9.8 branch, and
902 32-bit x86 is available through a compile-time setting.
904 To try the 32-bit x86 assembler implementation, use Configure
905 option "enable-montasm" (which exists only for this backport).
907 As "enable-montasm" for 32-bit x86 disclaims code stability
908 anyway, in this constellation we activate additional code
909 backported from 0.9.9-dev for further performance improvements,
910 namely BN_from_montgomery_word. (To enable this otherwise,
911 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
913 [Andy Polyakov (backport partially by Bodo Moeller)]
915 *) Add TLS session ticket callback. This allows an application to set
916 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
917 values. This is useful for key rollover for example where several key
918 sets may exist with different names.
921 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
922 This was broken until now in 0.9.8 releases, such that the only way
923 a registered ENGINE could be used (assuming it initialises
924 successfully on the host) was to explicitly set it as the default
925 for the relevant algorithms. This is in contradiction with 0.9.7
926 behaviour and the documentation. With this fix, when an ENGINE is
927 registered into a given algorithm's table of implementations, the
928 'uptodate' flag is reset so that auto-discovery will be used next
929 time a new context for that algorithm attempts to select an
931 [Ian Lister (tweaked by Geoff Thorpe)]
933 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
934 implemention in the following ways:
936 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
939 Lack of BER streaming support means one pass streaming processing is
940 only supported if data is detached: setting the streaming flag is
941 ignored for embedded content.
943 CMS support is disabled by default and must be explicitly enabled
944 with the enable-cms configuration option.
947 *) Update the GMP engine glue to do direct copies between BIGNUM and
948 mpz_t when openssl and GMP use the same limb size. Otherwise the
949 existing "conversion via a text string export" trick is still used.
950 [Paul Sheer <paulsheer@gmail.com>]
952 *) Zlib compression BIO. This is a filter BIO which compressed and
953 uncompresses any data passed through it.
956 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
957 RFC3394 compatible AES key wrapping.
960 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
961 sets string data without copying. X509_ALGOR_set0() and
962 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
963 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
964 from an X509_ATTRIBUTE structure optionally checking it occurs only
965 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
969 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
970 to get the expected BN_FLG_CONSTTIME behavior.
971 [Bodo Moeller (Google)]
975 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
976 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
977 - added some more tests to do_tests.pl
978 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
979 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
980 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
981 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
982 - various changes to netware.pl to enable gcc-cross builds on Win32
984 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
985 - various changes to fix missing prototype warnings
986 - fixed x86nasm.pl to create correct asm files for NASM COFF output
987 - added AES, WHIRLPOOL and CPUID assembler code to build files
988 - added missing AES assembler make rules to mk1mf.pl
989 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
990 [Guenter Knauf <eflash@gmx.net>]
992 *) Implement certificate status request TLS extension defined in RFC3546.
993 A client can set the appropriate parameters and receive the encoded
994 OCSP response via a callback. A server can query the supplied parameters
995 and set the encoded OCSP response in the callback. Add simplified examples
996 to s_client and s_server.
999 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
1001 *) Fix various bugs:
1002 + Binary incompatibility of ssl_ctx_st structure
1003 + DTLS interoperation with non-compliant servers
1004 + Don't call get_session_cb() without proposed session
1005 + Fix ia64 assembler code
1006 [Andy Polyakov, Steve Henson]
1008 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
1010 *) DTLS Handshake overhaul. There were longstanding issues with
1011 OpenSSL DTLS implementation, which were making it impossible for
1012 RFC 4347 compliant client to communicate with OpenSSL server.
1013 Unfortunately just fixing these incompatibilities would "cut off"
1014 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
1015 server keeps tolerating non RFC compliant syntax. The opposite is
1016 not true, 0.9.8f client can not communicate with earlier server.
1017 This update even addresses CVE-2007-4995.
1020 *) Changes to avoid need for function casts in OpenSSL: some compilers
1021 (gcc 4.2 and later) reject their use.
1022 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
1025 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1026 RFC4507bis. The encrypted ticket format is an encrypted encoded
1027 SSL_SESSION structure, that way new session features are automatically
1030 If a client application caches session in an SSL_SESSION structure
1031 support is transparent because tickets are now stored in the encoded
1034 The SSL_CTX structure automatically generates keys for ticket
1035 protection in servers so again support should be possible
1036 with no application modification.
1038 If a client or server wishes to disable RFC4507 support then the option
1039 SSL_OP_NO_TICKET can be set.
1041 Add a TLS extension debugging callback to allow the contents of any client
1042 or server extensions to be examined.
1044 This work was sponsored by Google.
1047 *) Add initial support for TLS extensions, specifically for the server_name
1048 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1049 have new members for a host name. The SSL data structure has an
1050 additional member SSL_CTX *initial_ctx so that new sessions can be
1051 stored in that context to allow for session resumption, even after the
1052 SSL has been switched to a new SSL_CTX in reaction to a client's
1053 server_name extension.
1055 New functions (subject to change):
1057 SSL_get_servername()
1058 SSL_get_servername_type()
1061 New CTRL codes and macros (subject to change):
1063 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1064 - SSL_CTX_set_tlsext_servername_callback()
1065 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1066 - SSL_CTX_set_tlsext_servername_arg()
1067 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1069 openssl s_client has a new '-servername ...' option.
1071 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1072 '-key2 ...', '-servername_fatal' (subject to change). This allows
1073 testing the HostName extension for a specific single host name ('-cert'
1074 and '-key' remain fallbacks for handshakes without HostName
1075 negotiation). If the unrecogninzed_name alert has to be sent, this by
1076 default is a warning; it becomes fatal with the '-servername_fatal'
1079 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
1081 *) Add AES and SSE2 assembly language support to VC++ build.
1084 *) Mitigate attack on final subtraction in Montgomery reduction.
1087 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
1088 (which previously caused an internal error).
1091 *) Squeeze another 10% out of IGE mode when in != out.
1094 *) AES IGE mode speedup.
1095 [Dean Gaudet (Google)]
1097 *) Add the Korean symmetric 128-bit cipher SEED (see
1098 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
1099 add SEED ciphersuites from RFC 4162:
1101 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
1102 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
1103 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
1104 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
1106 To minimize changes between patchlevels in the OpenSSL 0.9.8
1107 series, SEED remains excluded from compilation unless OpenSSL
1108 is configured with 'enable-seed'.
1109 [KISA, Bodo Moeller]
1111 *) Mitigate branch prediction attacks, which can be practical if a
1112 single processor is shared, allowing a spy process to extract
1113 information. For detailed background information, see
1114 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
1115 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
1116 and Necessary Software Countermeasures"). The core of the change
1117 are new versions BN_div_no_branch() and
1118 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
1119 respectively, which are slower, but avoid the security-relevant
1120 conditional branches. These are automatically called by BN_div()
1121 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
1122 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
1123 remove a conditional branch.
1125 BN_FLG_CONSTTIME is the new name for the previous
1126 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
1127 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
1128 in the exponent causes BN_mod_exp_mont() to use the alternative
1129 implementation in BN_mod_exp_mont_consttime().) The old name
1130 remains as a deprecated alias.
1132 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
1133 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
1134 constant-time implementations for more than just exponentiation.
1135 Here too the old name is kept as a deprecated alias.
1137 BN_BLINDING_new() will now use BN_dup() for the modulus so that
1138 the BN_BLINDING structure gets an independent copy of the
1139 modulus. This means that the previous "BIGNUM *m" argument to
1140 BN_BLINDING_new() and to BN_BLINDING_create_param() now
1141 essentially becomes "const BIGNUM *m", although we can't actually
1142 change this in the header file before 0.9.9. It allows
1143 RSA_setup_blinding() to use BN_with_flags() on the modulus to
1144 enable BN_FLG_CONSTTIME.
1146 [Matthew D Wood (Intel Corp)]
1148 *) In the SSL/TLS server implementation, be strict about session ID
1149 context matching (which matters if an application uses a single
1150 external cache for different purposes). Previously,
1151 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
1152 set. This did ensure strict client verification, but meant that,
1153 with applications using a single external cache for quite
1154 different requirements, clients could circumvent ciphersuite
1155 restrictions for a given session ID context by starting a session
1156 in a different context.
1159 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
1160 a ciphersuite string such as "DEFAULT:RSA" cannot enable
1161 authentication-only ciphersuites.
1164 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
1165 not complete and could lead to a possible single byte overflow
1166 (CVE-2007-5135) [Ben Laurie]
1168 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
1170 *) Since AES128 and AES256 (and similarly Camellia128 and
1171 Camellia256) share a single mask bit in the logic of
1172 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
1173 kludge to work properly if AES128 is available and AES256 isn't
1174 (or if Camellia128 is available and Camellia256 isn't).
1177 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
1178 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
1179 When a point or a seed is encoded in a BIT STRING, we need to
1180 prevent the removal of trailing zero bits to get the proper DER
1181 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
1182 of a NamedBitList, for which trailing 0 bits need to be removed.)
1185 *) Have SSL/TLS server implementation tolerate "mismatched" record
1186 protocol version while receiving ClientHello even if the
1187 ClientHello is fragmented. (The server can't insist on the
1188 particular protocol version it has chosen before the ServerHello
1189 message has informed the client about his choice.)
1192 *) Add RFC 3779 support.
1193 [Rob Austein for ARIN, Ben Laurie]
1195 *) Load error codes if they are not already present instead of using a
1196 static variable. This allows them to be cleanly unloaded and reloaded.
1197 Improve header file function name parsing.
1200 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
1201 or CAPABILITY handshake as required by RFCs.
1204 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
1206 *) Introduce limits to prevent malicious keys being able to
1207 cause a denial of service. (CVE-2006-2940)
1208 [Steve Henson, Bodo Moeller]
1210 *) Fix ASN.1 parsing of certain invalid structures that can result
1211 in a denial of service. (CVE-2006-2937) [Steve Henson]
1213 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
1214 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
1216 *) Fix SSL client code which could crash if connecting to a
1217 malicious SSLv2 server. (CVE-2006-4343)
1218 [Tavis Ormandy and Will Drewry, Google Security Team]
1220 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
1221 match only those. Before that, "AES256-SHA" would be interpreted
1222 as a pattern and match "AES128-SHA" too (since AES128-SHA got
1223 the same strength classification in 0.9.7h) as we currently only
1224 have a single AES bit in the ciphersuite description bitmap.
1225 That change, however, also applied to ciphersuite strings such as
1226 "RC4-MD5" that intentionally matched multiple ciphersuites --
1227 namely, SSL 2.0 ciphersuites in addition to the more common ones
1228 from SSL 3.0/TLS 1.0.
1230 So we change the selection algorithm again: Naming an explicit
1231 ciphersuite selects this one ciphersuite, and any other similar
1232 ciphersuite (same bitmap) from *other* protocol versions.
1233 Thus, "RC4-MD5" again will properly select both the SSL 2.0
1234 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
1236 Since SSL 2.0 does not have any ciphersuites for which the
1237 128/256 bit distinction would be relevant, this works for now.
1238 The proper fix will be to use different bits for AES128 and
1239 AES256, which would have avoided the problems from the beginning;
1240 however, bits are scarce, so we can only do this in a new release
1241 (not just a patchlevel) when we can change the SSL_CIPHER
1242 definition to split the single 'unsigned long mask' bitmap into
1243 multiple values to extend the available space.
1247 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
1249 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
1250 (CVE-2006-4339) [Ben Laurie and Google Security Team]
1252 *) Add AES IGE and biIGE modes.
1255 *) Change the Unix randomness entropy gathering to use poll() when
1256 possible instead of select(), since the latter has some
1257 undesirable limitations.
1258 [Darryl Miles via Richard Levitte and Bodo Moeller]
1260 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
1261 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
1262 cannot be implicitly activated as part of, e.g., the "AES" alias.
1263 However, please upgrade to OpenSSL 0.9.9[-dev] for
1264 non-experimental use of the ECC ciphersuites to get TLS extension
1265 support, which is required for curve and point format negotiation
1266 to avoid potential handshake problems.
1269 *) Disable rogue ciphersuites:
1271 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
1272 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
1273 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
1275 The latter two were purportedly from
1276 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
1279 Also deactivate the remaining ciphersuites from
1280 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
1281 unofficial, and the ID has long expired.
1284 *) Fix RSA blinding Heisenbug (problems sometimes occured on
1285 dual-core machines) and other potential thread-safety issues.
1288 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
1289 versions), which is now available for royalty-free use
1290 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
1291 Also, add Camellia TLS ciphersuites from RFC 4132.
1293 To minimize changes between patchlevels in the OpenSSL 0.9.8
1294 series, Camellia remains excluded from compilation unless OpenSSL
1295 is configured with 'enable-camellia'.
1298 *) Disable the padding bug check when compression is in use. The padding
1299 bug check assumes the first packet is of even length, this is not
1300 necessarily true if compresssion is enabled and can result in false
1301 positives causing handshake failure. The actual bug test is ancient
1302 code so it is hoped that implementations will either have fixed it by
1303 now or any which still have the bug do not support compression.
1306 Changes between 0.9.8a and 0.9.8b [04 May 2006]
1308 *) When applying a cipher rule check to see if string match is an explicit
1309 cipher suite and only match that one cipher suite if it is.
1312 *) Link in manifests for VC++ if needed.
1313 [Austin Ziegler <halostatue@gmail.com>]
1315 *) Update support for ECC-based TLS ciphersuites according to
1316 draft-ietf-tls-ecc-12.txt with proposed changes (but without
1317 TLS extensions, which are supported starting with the 0.9.9
1318 branch, not in the OpenSSL 0.9.8 branch).
1321 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
1322 opaque EVP_CIPHER_CTX handling.
1325 *) Fixes and enhancements to zlib compression code. We now only use
1326 "zlib1.dll" and use the default __cdecl calling convention on Win32
1327 to conform with the standards mentioned here:
1328 http://www.zlib.net/DLL_FAQ.txt
1329 Static zlib linking now works on Windows and the new --with-zlib-include
1330 --with-zlib-lib options to Configure can be used to supply the location
1331 of the headers and library. Gracefully handle case where zlib library
1335 *) Several fixes and enhancements to the OID generation code. The old code
1336 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
1337 handle numbers larger than ULONG_MAX, truncated printing and had a
1338 non standard OBJ_obj2txt() behaviour.
1341 *) Add support for building of engines under engine/ as shared libraries
1342 under VC++ build system.
1345 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
1346 Hopefully, we will not see any false combination of paths any more.
1349 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
1351 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
1352 (part of SSL_OP_ALL). This option used to disable the
1353 countermeasure against man-in-the-middle protocol-version
1354 rollback in the SSL 2.0 server implementation, which is a bad
1355 idea. (CVE-2005-2969)
1357 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
1358 for Information Security, National Institute of Advanced Industrial
1359 Science and Technology [AIST], Japan)]
1361 *) Add two function to clear and return the verify parameter flags.
1364 *) Keep cipherlists sorted in the source instead of sorting them at
1365 runtime, thus removing the need for a lock.
1368 *) Avoid some small subgroup attacks in Diffie-Hellman.
1369 [Nick Mathewson and Ben Laurie]
1371 *) Add functions for well-known primes.
1374 *) Extended Windows CE support.
1375 [Satoshi Nakamura and Andy Polyakov]
1377 *) Initialize SSL_METHOD structures at compile time instead of during
1378 runtime, thus removing the need for a lock.
1381 *) Make PKCS7_decrypt() work even if no certificate is supplied by
1382 attempting to decrypt each encrypted key in turn. Add support to
1386 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
1388 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
1391 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
1394 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
1395 key into the same file any more.
1398 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
1401 *) Add -utf8 command line and config file option to 'ca'.
1402 [Stefan <stf@udoma.org]
1404 *) Removed the macro des_crypt(), as it seems to conflict with some
1405 libraries. Use DES_crypt().
1408 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
1409 involves renaming the source and generated shared-libs for
1410 both. The engines will accept the corrected or legacy ids
1411 ('ncipher' and '4758_cca' respectively) when binding. NB,
1412 this only applies when building 'shared'.
1413 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
1415 *) Add attribute functions to EVP_PKEY structure. Modify
1416 PKCS12_create() to recognize a CSP name attribute and
1417 use it. Make -CSP option work again in pkcs12 utility.
1420 *) Add new functionality to the bn blinding code:
1421 - automatic re-creation of the BN_BLINDING parameters after
1422 a fixed number of uses (currently 32)
1423 - add new function for parameter creation
1424 - introduce flags to control the update behaviour of the
1425 BN_BLINDING parameters
1426 - hide BN_BLINDING structure
1427 Add a second BN_BLINDING slot to the RSA structure to improve
1428 performance when a single RSA object is shared among several
1432 *) Add support for DTLS.
1433 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
1435 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
1436 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
1439 *) Remove buggy and incompletet DH cert support from
1440 ssl/ssl_rsa.c and ssl/s3_both.c
1443 *) Use SHA-1 instead of MD5 as the default digest algorithm for
1444 the apps/openssl applications.
1447 *) Compile clean with "-Wall -Wmissing-prototypes
1448 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
1449 DEBUG_SAFESTACK must also be set.
1452 *) Change ./Configure so that certain algorithms can be disabled by default.
1453 The new counterpiece to "no-xxx" is "enable-xxx".
1455 The patented RC5 and MDC2 algorithms will now be disabled unless
1456 "enable-rc5" and "enable-mdc2", respectively, are specified.
1458 (IDEA remains enabled despite being patented. This is because IDEA
1459 is frequently required for interoperability, and there is no license
1460 fee for non-commercial use. As before, "no-idea" can be used to
1461 avoid this algorithm.)
1465 *) Add processing of proxy certificates (see RFC 3820). This work was
1466 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
1467 EGEE (Enabling Grids for E-science in Europe).
1470 *) RC4 performance overhaul on modern architectures/implementations, such
1471 as Intel P4, IA-64 and AMD64.
1474 *) New utility extract-section.pl. This can be used specify an alternative
1475 section number in a pod file instead of having to treat each file as
1476 a separate case in Makefile. This can be done by adding two lines to the
1479 =for comment openssl_section:XXX
1481 The blank line is mandatory.
1485 *) New arguments -certform, -keyform and -pass for s_client and s_server
1486 to allow alternative format key and certificate files and passphrase
1490 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
1491 update associated structures and add various utility functions.
1493 Add new policy related verify parameters, include policy checking in
1494 standard verify code. Enhance 'smime' application with extra parameters
1495 to support policy checking and print out.
1498 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
1499 Nehemiah processors. These extensions support AES encryption in hardware
1500 as well as RNG (though RNG support is currently disabled).
1501 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
1503 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
1506 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
1507 [Andy Polyakov and a number of other people]
1509 *) Improved PowerPC platform support. Most notably BIGNUM assembler
1510 implementation contributed by IBM.
1511 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
1513 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
1514 exponent rather than 'unsigned long'. There is a corresponding change to
1515 the new 'rsa_keygen' element of the RSA_METHOD structure.
1516 [Jelte Jansen, Geoff Thorpe]
1518 *) Functionality for creating the initial serial number file is now
1519 moved from CA.pl to the 'ca' utility with a new option -create_serial.
1521 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
1522 number file to 1, which is bound to cause problems. To avoid
1523 the problems while respecting compatibility between different 0.9.7
1524 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
1525 CA.pl for serial number initialization. With the new release 0.9.8,
1526 we can fix the problem directly in the 'ca' utility.)
1529 *) Reduced header interdepencies by declaring more opaque objects in
1530 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
1531 give fewer recursive includes, which could break lazy source code - so
1532 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
1533 developers should define this symbol when building and using openssl to
1534 ensure they track the recommended behaviour, interfaces, [etc], but
1535 backwards-compatible behaviour prevails when this isn't defined.
1538 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
1541 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
1542 This will generate a random key of the appropriate length based on the
1543 cipher context. The EVP_CIPHER can provide its own random key generation
1544 routine to support keys of a specific form. This is used in the des and
1545 3des routines to generate a key of the correct parity. Update S/MIME
1546 code to use new functions and hence generate correct parity DES keys.
1547 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
1548 valid (weak or incorrect parity).
1551 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
1552 as looking them up. This is useful when the verified structure may contain
1553 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
1554 present unless the new PKCS7_NO_CRL flag is asserted.
1557 *) Extend ASN1 oid configuration module. It now additionally accepts the
1560 shortName = some long name, 1.2.3.4
1563 *) Reimplemented the BN_CTX implementation. There is now no more static
1564 limitation on the number of variables it can handle nor the depth of the
1565 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
1566 information can now expand as required, and rather than having a single
1567 static array of bignums, BN_CTX now uses a linked-list of such arrays
1568 allowing it to expand on demand whilst maintaining the usefulness of
1569 BN_CTX's "bundling".
1572 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
1573 to allow all RSA operations to function using a single BN_CTX.
1576 *) Preliminary support for certificate policy evaluation and checking. This
1577 is initially intended to pass the tests outlined in "Conformance Testing
1578 of Relying Party Client Certificate Path Processing Logic" v1.07.
1581 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
1582 remained unused and not that useful. A variety of other little bignum
1583 tweaks and fixes have also been made continuing on from the audit (see
1587 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
1588 associated ASN1, EVP and SSL functions and old ASN1 macros.
1591 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
1592 and this should never fail. So the return value from the use of
1593 BN_set_word() (which can fail due to needless expansion) is now deprecated;
1594 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
1597 *) BN_CTX_get() should return zero-valued bignums, providing the same
1598 initialised value as BN_new().
1599 [Geoff Thorpe, suggested by Ulf Möller]
1601 *) Support for inhibitAnyPolicy certificate extension.
1604 *) An audit of the BIGNUM code is underway, for which debugging code is
1605 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
1606 is considered valid when processing BIGNUMs, and causes execution to
1607 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
1608 further steps are taken to deliberately pollute unused data in BIGNUM
1609 structures to try and expose faulty code further on. For now, openssl will
1610 (in its default mode of operation) continue to tolerate the inconsistent
1611 forms that it has tolerated in the past, but authors and packagers should
1612 consider trying openssl and their own applications when compiled with
1613 these debugging symbols defined. It will help highlight potential bugs in
1614 their own code, and will improve the test coverage for OpenSSL itself. At
1615 some point, these tighter rules will become openssl's default to improve
1616 maintainability, though the assert()s and other overheads will remain only
1617 in debugging configurations. See bn.h for more details.
1618 [Geoff Thorpe, Nils Larsch, Ulf Möller]
1620 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
1621 that can only be obtained through BN_CTX_new() (which implicitly
1622 initialises it). The presence of this function only made it possible
1623 to overwrite an existing structure (and cause memory leaks).
1626 *) Because of the callback-based approach for implementing LHASH as a
1627 template type, lh_insert() adds opaque objects to hash-tables and
1628 lh_doall() or lh_doall_arg() are typically used with a destructor callback
1629 to clean up those corresponding objects before destroying the hash table
1630 (and losing the object pointers). So some over-zealous constifications in
1631 LHASH have been relaxed so that lh_insert() does not take (nor store) the
1632 objects as "const" and the lh_doall[_arg] callback wrappers are not
1633 prototyped to have "const" restrictions on the object pointers they are
1634 given (and so aren't required to cast them away any more).
1637 *) The tmdiff.h API was so ugly and minimal that our own timing utility
1638 (speed) prefers to use its own implementation. The two implementations
1639 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
1640 its object type properly exposed (MS_TM) instead of casting to/from "char
1641 *". This may still change yet if someone realises MS_TM and "ms_time_***"
1642 aren't necessarily the greatest nomenclatures - but this is what was used
1643 internally to the implementation so I've used that for now.
1646 *) Ensure that deprecated functions do not get compiled when
1647 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
1648 the self-tests were still using deprecated key-generation functions so
1649 these have been updated also.
1652 *) Reorganise PKCS#7 code to separate the digest location functionality
1653 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
1654 New function PKCS7_set_digest() to set the digest type for PKCS#7
1655 digestedData type. Add additional code to correctly generate the
1656 digestedData type and add support for this type in PKCS7 initialization
1660 *) New function PKCS7_set0_type_other() this initializes a PKCS7
1661 structure of type "other".
1664 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
1665 sure the loop does correctly stop and breaking ("division by zero")
1666 modulus operations are not performed. The (pre-generated) prime
1667 table crypto/bn/bn_prime.h was already correct, but it could not be
1668 re-generated on some platforms because of the "division by zero"
1669 situation in the script.
1670 [Ralf S. Engelschall]
1672 *) Update support for ECC-based TLS ciphersuites according to
1673 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
1674 SHA-1 now is only used for "small" curves (where the
1675 representation of a field element takes up to 24 bytes); for
1676 larger curves, the field element resulting from ECDH is directly
1677 used as premaster secret.
1678 [Douglas Stebila (Sun Microsystems Laboratories)]
1680 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
1681 curve secp160r1 to the tests.
1682 [Douglas Stebila (Sun Microsystems Laboratories)]
1684 *) Add the possibility to load symbols globally with DSO.
1685 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
1687 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
1688 control of the error stack.
1691 *) Add support for STORE in ENGINE.
1694 *) Add the STORE type. The intention is to provide a common interface
1695 to certificate and key stores, be they simple file-based stores, or
1696 HSM-type store, or LDAP stores, or...
1697 NOTE: The code is currently UNTESTED and isn't really used anywhere.
1700 *) Add a generic structure called OPENSSL_ITEM. This can be used to
1701 pass a list of arguments to any function as well as provide a way
1702 for a function to pass data back to the caller.
1705 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
1706 works like BUF_strdup() but can be used to duplicate a portion of
1707 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
1711 *) Add the function sk_find_ex() which works like sk_find(), but will
1712 return an index to an element even if an exact match couldn't be
1713 found. The index is guaranteed to point at the element where the
1714 searched-for key would be inserted to preserve sorting order.
1717 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
1718 takes an extra flags argument for optional functionality. Currently,
1719 the following flags are defined:
1721 OBJ_BSEARCH_VALUE_ON_NOMATCH
1722 This one gets OBJ_bsearch_ex() to return a pointer to the first
1723 element where the comparing function returns a negative or zero
1726 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
1727 This one gets OBJ_bsearch_ex() to return a pointer to the first
1728 element where the comparing function returns zero. This is useful
1729 if there are more than one element where the comparing function
1733 *) Make it possible to create self-signed certificates with 'openssl ca'
1734 in such a way that the self-signed certificate becomes part of the
1735 CA database and uses the same mechanisms for serial number generation
1736 as all other certificate signing. The new flag '-selfsign' enables
1737 this functionality. Adapt CA.sh and CA.pl.in.
1740 *) Add functionality to check the public key of a certificate request
1741 against a given private. This is useful to check that a certificate
1742 request can be signed by that key (self-signing).
1745 *) Make it possible to have multiple active certificates with the same
1746 subject in the CA index file. This is done only if the keyword
1747 'unique_subject' is set to 'no' in the main CA section (default
1748 if 'CA_default') of the configuration file. The value is saved
1749 with the database itself in a separate index attribute file,
1750 named like the index file with '.attr' appended to the name.
1753 *) Generate muti valued AVAs using '+' notation in config files for
1757 *) Support for nameConstraints certificate extension.
1760 *) Support for policyConstraints certificate extension.
1763 *) Support for policyMappings certificate extension.
1766 *) Make sure the default DSA_METHOD implementation only uses its
1767 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
1768 and change its own handlers to be NULL so as to remove unnecessary
1769 indirection. This lets alternative implementations fallback to the
1770 default implementation more easily.
1773 *) Support for directoryName in GeneralName related extensions
1777 *) Make it possible to link applications using Makefile.shared.
1778 Make that possible even when linking against static libraries!
1781 *) Support for single pass processing for S/MIME signing. This now
1782 means that S/MIME signing can be done from a pipe, in addition
1783 cleartext signing (multipart/signed type) is effectively streaming
1784 and the signed data does not need to be all held in memory.
1786 This is done with a new flag PKCS7_STREAM. When this flag is set
1787 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
1788 is done after the data is output (and digests calculated) in
1789 SMIME_write_PKCS7().
1792 *) Add full support for -rpath/-R, both in shared libraries and
1793 applications, at least on the platforms where it's known how
1797 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
1798 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
1799 will now compute a table of multiples of the generator that
1800 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
1801 faster (notably in the case of a single point multiplication,
1802 scalar * generator).
1803 [Nils Larsch, Bodo Moeller]
1805 *) IPv6 support for certificate extensions. The various extensions
1806 which use the IP:a.b.c.d can now take IPv6 addresses using the
1807 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
1811 *) Added an ENGINE that implements RSA by performing private key
1812 exponentiations with the GMP library. The conversions to and from
1813 GMP's mpz_t format aren't optimised nor are any montgomery forms
1814 cached, and on x86 it appears OpenSSL's own performance has caught up.
1815 However there are likely to be other architectures where GMP could
1816 provide a boost. This ENGINE is not built in by default, but it can be
1817 specified at Configure time and should be accompanied by the necessary
1818 linker additions, eg;
1819 ./config -DOPENSSL_USE_GMP -lgmp
1822 *) "openssl engine" will not display ENGINE/DSO load failure errors when
1823 testing availability of engines with "-t" - the old behaviour is
1824 produced by increasing the feature's verbosity with "-tt".
1827 *) ECDSA routines: under certain error conditions uninitialized BN objects
1828 could be freed. Solution: make sure initialization is performed early
1829 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
1833 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
1834 and DH_METHOD (eg. by ENGINE implementations) to override the normal
1835 software implementations. For DSA and DH, parameter generation can
1836 also be overriden by providing the appropriate method callbacks.
1839 *) Change the "progress" mechanism used in key-generation and
1840 primality testing to functions that take a new BN_GENCB pointer in
1841 place of callback/argument pairs. The new API functions have "_ex"
1842 postfixes and the older functions are reimplemented as wrappers for
1843 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
1844 declarations of the old functions to help (graceful) attempts to
1845 migrate to the new functions. Also, the new key-generation API
1846 functions operate on a caller-supplied key-structure and return
1847 success/failure rather than returning a key or NULL - this is to
1848 help make "keygen" another member function of RSA_METHOD etc.
1850 Example for using the new callback interface:
1852 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
1856 BN_GENCB_set(&my_cb, my_callback, my_arg);
1858 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
1859 /* For the meaning of a, b in calls to my_callback(), see the
1860 * documentation of the function that calls the callback.
1861 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
1862 * my_callback should return 1 if it wants BN_is_prime_ex()
1863 * to continue, or 0 to stop.
1868 *) Change the ZLIB compression method to be stateful, and make it
1869 available to TLS with the number defined in
1870 draft-ietf-tls-compression-04.txt.
1873 *) Add the ASN.1 structures and functions for CertificatePair, which
1874 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
1876 CertificatePair ::= SEQUENCE {
1877 forward [0] Certificate OPTIONAL,
1878 reverse [1] Certificate OPTIONAL,
1879 -- at least one of the pair shall be present -- }
1881 Also implement the PEM functions to read and write certificate
1882 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
1884 This needed to be defined, mostly for the sake of the LDAP
1885 attribute crossCertificatePair, but may prove useful elsewhere as
1889 *) Make it possible to inhibit symlinking of shared libraries in
1890 Makefile.shared, for Cygwin's sake.
1893 *) Extend the BIGNUM API by creating a function
1894 void BN_set_negative(BIGNUM *a, int neg);
1895 and a macro that behave like
1896 int BN_is_negative(const BIGNUM *a);
1898 to avoid the need to access 'a->neg' directly in applications.
1901 *) Implement fast modular reduction for pseudo-Mersenne primes
1902 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
1903 EC_GROUP_new_curve_GFp() will now automatically use this
1905 [Nils Larsch <nla@trustcenter.de>]
1907 *) Add new lock type (CRYPTO_LOCK_BN).
1910 *) Change the ENGINE framework to automatically load engines
1911 dynamically from specific directories unless they could be
1912 found to already be built in or loaded. Move all the
1913 current engines except for the cryptodev one to a new
1915 The engines in engines/ are built as shared libraries if
1916 the "shared" options was given to ./Configure or ./config.
1917 Otherwise, they are inserted in libcrypto.a.
1918 /usr/local/ssl/engines is the default directory for dynamic
1919 engines, but that can be overriden at configure time through
1920 the usual use of --prefix and/or --openssldir, and at run
1921 time with the environment variable OPENSSL_ENGINES.
1922 [Geoff Thorpe and Richard Levitte]
1924 *) Add Makefile.shared, a helper makefile to build shared
1925 libraries. Addapt Makefile.org.
1928 *) Add version info to Win32 DLLs.
1929 [Peter 'Luna' Runestig" <peter@runestig.com>]
1931 *) Add new 'medium level' PKCS#12 API. Certificates and keys
1932 can be added using this API to created arbitrary PKCS#12
1933 files while avoiding the low level API.
1935 New options to PKCS12_create(), key or cert can be NULL and
1936 will then be omitted from the output file. The encryption
1937 algorithm NIDs can be set to -1 for no encryption, the mac
1938 iteration count can be set to 0 to omit the mac.
1940 Enhance pkcs12 utility by making the -nokeys and -nocerts
1941 options work when creating a PKCS#12 file. New option -nomac
1942 to omit the mac, NONE can be set for an encryption algorithm.
1943 New code is modified to use the enhanced PKCS12_create()
1944 instead of the low level API.
1947 *) Extend ASN1 encoder to support indefinite length constructed
1948 encoding. This can output sequences tags and octet strings in
1949 this form. Modify pk7_asn1.c to support indefinite length
1950 encoding. This is experimental and needs additional code to
1951 be useful, such as an ASN1 bio and some enhanced streaming
1954 Extend template encode functionality so that tagging is passed
1955 down to the template encoder.
1958 *) Let 'openssl req' fail if an argument to '-newkey' is not
1959 recognized instead of using RSA as a default.
1962 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
1963 As these are not official, they are not included in "ALL";
1964 the "ECCdraft" ciphersuite group alias can be used to select them.
1965 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
1967 *) Add ECDH engine support.
1968 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
1970 *) Add ECDH in new directory crypto/ecdh/.
1971 [Douglas Stebila (Sun Microsystems Laboratories)]
1973 *) Let BN_rand_range() abort with an error after 100 iterations
1974 without success (which indicates a broken PRNG).
1977 *) Change BN_mod_sqrt() so that it verifies that the input value
1978 is really the square of the return value. (Previously,
1979 BN_mod_sqrt would show GIGO behaviour.)
1982 *) Add named elliptic curves over binary fields from X9.62, SECG,
1983 and WAP/WTLS; add OIDs that were still missing.
1985 [Sheueling Chang Shantz and Douglas Stebila
1986 (Sun Microsystems Laboratories)]
1988 *) Extend the EC library for elliptic curves over binary fields
1989 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
1992 EC_GF2m_simple_method
1996 EC_GROUP_new_curve_GF2m
1997 EC_GROUP_set_curve_GF2m
1998 EC_GROUP_get_curve_GF2m
1999 EC_POINT_set_affine_coordinates_GF2m
2000 EC_POINT_get_affine_coordinates_GF2m
2001 EC_POINT_set_compressed_coordinates_GF2m
2003 Point compression for binary fields is disabled by default for
2004 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
2007 As binary polynomials are represented as BIGNUMs, various members
2008 of the EC_GROUP and EC_POINT data structures can be shared
2009 between the implementations for prime fields and binary fields;
2010 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
2011 are essentially identical to their ..._GFp counterparts.
2012 (For simplicity, the '..._GFp' prefix has been dropped from
2013 various internal method names.)
2015 An internal 'field_div' method (similar to 'field_mul' and
2016 'field_sqr') has been added; this is used only for binary fields.
2018 [Sheueling Chang Shantz and Douglas Stebila
2019 (Sun Microsystems Laboratories)]
2021 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
2022 through methods ('mul', 'precompute_mult').
2024 The generic implementations (now internally called 'ec_wNAF_mul'
2025 and 'ec_wNAF_precomputed_mult') remain the default if these
2026 methods are undefined.
2028 [Sheueling Chang Shantz and Douglas Stebila
2029 (Sun Microsystems Laboratories)]
2031 *) New function EC_GROUP_get_degree, which is defined through
2032 EC_METHOD. For curves over prime fields, this returns the bit
2033 length of the modulus.
2035 [Sheueling Chang Shantz and Douglas Stebila
2036 (Sun Microsystems Laboratories)]
2038 *) New functions EC_GROUP_dup, EC_POINT_dup.
2039 (These simply call ..._new and ..._copy).
2041 [Sheueling Chang Shantz and Douglas Stebila
2042 (Sun Microsystems Laboratories)]
2044 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
2045 Polynomials are represented as BIGNUMs (where the sign bit is not
2046 used) in the following functions [macros]:
2049 BN_GF2m_sub [= BN_GF2m_add]
2050 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
2051 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
2052 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
2054 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
2055 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
2056 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
2057 BN_GF2m_cmp [= BN_ucmp]
2059 (Note that only the 'mod' functions are actually for fields GF(2^m).
2060 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
2062 For some functions, an the irreducible polynomial defining a
2063 field can be given as an 'unsigned int[]' with strictly
2064 decreasing elements giving the indices of those bits that are set;
2065 i.e., p[] represents the polynomial
2066 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
2068 p[0] > p[1] > ... > p[k] = 0.
2069 This applies to the following functions:
2074 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
2075 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
2077 BN_GF2m_mod_sqrt_arr
2078 BN_GF2m_mod_solve_quad_arr
2082 Conversion can be performed by the following functions:
2087 bntest.c has additional tests for binary polynomial arithmetic.
2089 Two implementations for BN_GF2m_mod_div() are available.
2090 The default algorithm simply uses BN_GF2m_mod_inv() and
2091 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
2092 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
2093 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
2095 [Sheueling Chang Shantz and Douglas Stebila
2096 (Sun Microsystems Laboratories)]
2098 *) Add new error code 'ERR_R_DISABLED' that can be used when some
2099 functionality is disabled at compile-time.
2100 [Douglas Stebila <douglas.stebila@sun.com>]
2102 *) Change default behaviour of 'openssl asn1parse' so that more
2103 information is visible when viewing, e.g., a certificate:
2105 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
2106 mode the content of non-printable OCTET STRINGs is output in a
2107 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
2108 avoid the appearance of a printable string.
2109 [Nils Larsch <nla@trustcenter.de>]
2111 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
2113 EC_GROUP_set_asn1_flag()
2114 EC_GROUP_get_asn1_flag()
2115 EC_GROUP_set_point_conversion_form()
2116 EC_GROUP_get_point_conversion_form()
2117 These control ASN1 encoding details:
2118 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
2119 has been set to OPENSSL_EC_NAMED_CURVE.
2120 - Points are encoded in uncompressed form by default; options for
2121 asn1_for are as for point2oct, namely
2122 POINT_CONVERSION_COMPRESSED
2123 POINT_CONVERSION_UNCOMPRESSED
2124 POINT_CONVERSION_HYBRID
2126 Also add 'seed' and 'seed_len' members to EC_GROUP with access
2129 EC_GROUP_get0_seed()
2130 EC_GROUP_get_seed_len()
2131 This is used only for ASN1 purposes (so far).
2132 [Nils Larsch <nla@trustcenter.de>]
2134 *) Add 'field_type' member to EC_METHOD, which holds the NID
2135 of the appropriate field type OID. The new function
2136 EC_METHOD_get_field_type() returns this value.
2137 [Nils Larsch <nla@trustcenter.de>]
2142 EC_POINT_point2hex()
2143 EC_POINT_hex2point()
2144 providing useful interfaces to EC_POINT_point2oct() and
2145 EC_POINT_oct2point().
2146 [Nils Larsch <nla@trustcenter.de>]
2148 *) Change internals of the EC library so that the functions
2149 EC_GROUP_set_generator()
2150 EC_GROUP_get_generator()
2151 EC_GROUP_get_order()
2152 EC_GROUP_get_cofactor()
2153 are implemented directly in crypto/ec/ec_lib.c and not dispatched
2154 to methods, which would lead to unnecessary code duplication when
2155 adding different types of curves.
2156 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
2158 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
2159 arithmetic, and such that modified wNAFs are generated
2160 (which avoid length expansion in many cases).
2163 *) Add a function EC_GROUP_check_discriminant() (defined via
2164 EC_METHOD) that verifies that the curve discriminant is non-zero.
2166 Add a function EC_GROUP_check() that makes some sanity tests
2167 on a EC_GROUP, its generator and order. This includes
2168 EC_GROUP_check_discriminant().
2169 [Nils Larsch <nla@trustcenter.de>]
2171 *) Add ECDSA in new directory crypto/ecdsa/.
2173 Add applications 'openssl ecparam' and 'openssl ecdsa'
2174 (these are based on 'openssl dsaparam' and 'openssl dsa').
2176 ECDSA support is also included in various other files across the
2177 library. Most notably,
2178 - 'openssl req' now has a '-newkey ecdsa:file' option;
2179 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
2180 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
2181 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
2182 them suitable for ECDSA where domain parameters must be
2183 extracted before the specific public key;
2184 - ECDSA engine support has been added.
2185 [Nils Larsch <nla@trustcenter.de>]
2187 *) Include some named elliptic curves, and add OIDs from X9.62,
2188 SECG, and WAP/WTLS. Each curve can be obtained from the new
2190 EC_GROUP_new_by_curve_name(),
2191 and the list of available named curves can be obtained with
2192 EC_get_builtin_curves().
2193 Also add a 'curve_name' member to EC_GROUP objects, which can be
2195 EC_GROUP_set_curve_name()
2196 EC_GROUP_get_curve_name()
2197 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
2199 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
2200 was actually never needed) and in BN_mul(). The removal in BN_mul()
2201 required a small change in bn_mul_part_recursive() and the addition
2202 of the functions bn_cmp_part_words(), bn_sub_part_words() and
2203 bn_add_part_words(), which do the same thing as bn_cmp_words(),
2204 bn_sub_words() and bn_add_words() except they take arrays with
2208 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
2210 *) Cleanse PEM buffers before freeing them since they may contain
2212 [Benjamin Bennett <ben@psc.edu>]
2214 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2215 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2216 authentication-only ciphersuites.
2219 *) Since AES128 and AES256 share a single mask bit in the logic of
2220 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2221 kludge to work properly if AES128 is available and AES256 isn't.
2224 *) Expand security boundary to match 1.1.1 module.
2227 *) Remove redundant features: hash file source, editing of test vectors
2228 modify fipsld to use external fips_premain.c signature.
2231 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
2232 run algorithm test programs.
2235 *) Make algorithm test programs more tolerant of whitespace.
2238 *) Have SSL/TLS server implementation tolerate "mismatched" record
2239 protocol version while receiving ClientHello even if the
2240 ClientHello is fragmented. (The server can't insist on the
2241 particular protocol version it has chosen before the ServerHello
2242 message has informed the client about his choice.)
2245 *) Load error codes if they are not already present instead of using a
2246 static variable. This allows them to be cleanly unloaded and reloaded.
2249 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
2251 *) Introduce limits to prevent malicious keys being able to
2252 cause a denial of service. (CVE-2006-2940)
2253 [Steve Henson, Bodo Moeller]
2255 *) Fix ASN.1 parsing of certain invalid structures that can result
2256 in a denial of service. (CVE-2006-2937) [Steve Henson]
2258 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2259 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2261 *) Fix SSL client code which could crash if connecting to a
2262 malicious SSLv2 server. (CVE-2006-4343)
2263 [Tavis Ormandy and Will Drewry, Google Security Team]
2265 *) Change ciphersuite string processing so that an explicit
2266 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
2267 will no longer include "AES128-SHA"), and any other similar
2268 ciphersuite (same bitmap) from *other* protocol versions (so that
2269 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
2270 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
2271 changes from 0.9.8b and 0.9.8d.
2274 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
2276 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2277 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2279 *) Change the Unix randomness entropy gathering to use poll() when
2280 possible instead of select(), since the latter has some
2281 undesirable limitations.
2282 [Darryl Miles via Richard Levitte and Bodo Moeller]
2284 *) Disable rogue ciphersuites:
2286 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2287 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2288 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2290 The latter two were purportedly from
2291 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2294 Also deactive the remaining ciphersuites from
2295 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2296 unofficial, and the ID has long expired.
2299 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2300 dual-core machines) and other potential thread-safety issues.
2303 Changes between 0.9.7i and 0.9.7j [04 May 2006]
2305 *) Adapt fipsld and the build system to link against the validated FIPS
2306 module in FIPS mode.
2309 *) Fixes for VC++ 2005 build under Windows.
2312 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
2313 from a Windows bash shell such as MSYS. It is autodetected from the
2314 "config" script when run from a VC++ environment. Modify standard VC++
2315 build to use fipscanister.o from the GNU make build.
2318 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
2320 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
2321 The value now differs depending on if you build for FIPS or not.
2322 BEWARE! A program linked with a shared FIPSed libcrypto can't be
2323 safely run with a non-FIPSed libcrypto, as it may crash because of
2324 the difference induced by this change.
2327 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
2329 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2330 (part of SSL_OP_ALL). This option used to disable the
2331 countermeasure against man-in-the-middle protocol-version
2332 rollback in the SSL 2.0 server implementation, which is a bad
2333 idea. (CVE-2005-2969)
2335 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2336 for Information Security, National Institute of Advanced Industrial
2337 Science and Technology [AIST], Japan)]
2339 *) Minimal support for X9.31 signatures and PSS padding modes. This is
2340 mainly for FIPS compliance and not fully integrated at this stage.
2343 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
2344 the exponentiation using a fixed-length exponent. (Otherwise,
2345 the information leaked through timing could expose the secret key
2346 after many signatures; cf. Bleichenbacher's attack on DSA with
2350 *) Make a new fixed-window mod_exp implementation the default for
2351 RSA, DSA, and DH private-key operations so that the sequence of
2352 squares and multiplies and the memory access pattern are
2353 independent of the particular secret key. This will mitigate
2354 cache-timing and potential related attacks.
2356 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
2357 and this is automatically used by BN_mod_exp_mont() if the new flag
2358 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
2359 will use this BN flag for private exponents unless the flag
2360 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
2361 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
2363 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
2365 *) Change the client implementation for SSLv23_method() and
2366 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
2367 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
2368 (Previously, the SSL 2.0 backwards compatible Client Hello
2369 message format would be used even with SSL_OP_NO_SSLv2.)
2372 *) Add support for smime-type MIME parameter in S/MIME messages which some
2376 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
2377 a threadsafe manner. Modify rsa code to use new function and add calls
2378 to dsa and dh code (which had race conditions before).
2381 *) Include the fixed error library code in the C error file definitions
2382 instead of fixing them up at runtime. This keeps the error code
2383 structures constant.
2386 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
2388 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
2391 *) Fixes for newer kerberos headers. NB: the casts are needed because
2392 the 'length' field is signed on one version and unsigned on another
2393 with no (?) obvious way to tell the difference, without these VC++
2394 complains. Also the "definition" of FAR (blank) is no longer included
2395 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
2396 some needed definitions.
2399 *) Undo Cygwin change.
2402 *) Added support for proxy certificates according to RFC 3820.
2403 Because they may be a security thread to unaware applications,
2404 they must be explicitely allowed in run-time. See
2405 docs/HOWTO/proxy_certificates.txt for further information.
2408 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
2410 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
2411 server and client random values. Previously
2412 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
2413 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
2415 This change has negligible security impact because:
2417 1. Server and client random values still have 24 bytes of pseudo random
2420 2. Server and client random values are sent in the clear in the initial
2423 3. The master secret is derived using the premaster secret (48 bytes in
2424 size for static RSA ciphersuites) as well as client server and random
2427 The OpenSSL team would like to thank the UK NISCC for bringing this issue
2430 [Stephen Henson, reported by UK NISCC]
2432 *) Use Windows randomness collection on Cygwin.
2435 *) Fix hang in EGD/PRNGD query when communication socket is closed
2436 prematurely by EGD/PRNGD.
2437 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
2439 *) Prompt for pass phrases when appropriate for PKCS12 input format.
2442 *) Back-port of selected performance improvements from development
2443 branch, as well as improved support for PowerPC platforms.
2446 *) Add lots of checks for memory allocation failure, error codes to indicate
2447 failure and freeing up memory if a failure occurs.
2448 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
2450 *) Add new -passin argument to dgst.
2453 *) Perform some character comparisons of different types in X509_NAME_cmp:
2454 this is needed for some certificates that reencode DNs into UTF8Strings
2455 (in violation of RFC3280) and can't or wont issue name rollover
2459 *) Make an explicit check during certificate validation to see that
2460 the CA setting in each certificate on the chain is correct. As a
2461 side effect always do the following basic checks on extensions,
2462 not just when there's an associated purpose to the check:
2464 - if there is an unhandled critical extension (unless the user
2465 has chosen to ignore this fault)
2466 - if the path length has been exceeded (if one is set at all)
2467 - that certain extensions fit the associated purpose (if one has
2471 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
2473 *) Avoid a race condition when CRLs are checked in a multi threaded
2474 environment. This would happen due to the reordering of the revoked
2475 entries during signature checking and serial number lookup. Now the
2476 encoding is cached and the serial number sort performed under a lock.
2477 Add new STACK function sk_is_sorted().
2480 *) Add Delta CRL to the extension code.
2483 *) Various fixes to s3_pkt.c so alerts are sent properly.
2484 [David Holmes <d.holmes@f5.com>]
2486 *) Reduce the chances of duplicate issuer name and serial numbers (in
2487 violation of RFC3280) using the OpenSSL certificate creation utilities.
2488 This is done by creating a random 64 bit value for the initial serial
2489 number when a serial number file is created or when a self signed
2490 certificate is created using 'openssl req -x509'. The initial serial
2491 number file is created using 'openssl x509 -next_serial' in CA.pl
2492 rather than being initialized to 1.
2495 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
2497 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
2498 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
2499 [Joe Orton, Steve Henson]
2501 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
2503 [Joe Orton, Steve Henson]
2505 *) Make it possible to have multiple active certificates with the same
2506 subject in the CA index file. This is done only if the keyword
2507 'unique_subject' is set to 'no' in the main CA section (default
2508 if 'CA_default') of the configuration file. The value is saved
2509 with the database itself in a separate index attribute file,
2510 named like the index file with '.attr' appended to the name.
2513 *) X509 verify fixes. Disable broken certificate workarounds when
2514 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
2515 keyUsage extension present. Don't accept CRLs with unhandled critical
2516 extensions: since verify currently doesn't process CRL extensions this
2517 rejects a CRL with *any* critical extensions. Add new verify error codes
2521 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
2522 A clarification of RFC2560 will require the use of OCTET STRINGs and
2523 some implementations cannot handle the current raw format. Since OpenSSL
2524 copies and compares OCSP nonces as opaque blobs without any attempt at
2525 parsing them this should not create any compatibility issues.
2528 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
2529 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
2530 this HMAC (and other) operations are several times slower than OpenSSL
2534 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
2535 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2537 *) Use the correct content when signing type "other".
2540 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
2542 *) Fix various bugs revealed by running the NISCC test suite:
2544 Stop out of bounds reads in the ASN1 code when presented with
2545 invalid tags (CVE-2003-0543 and CVE-2003-0544).
2547 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
2549 If verify callback ignores invalid public key errors don't try to check
2550 certificate signature with the NULL public key.
2554 *) New -ignore_err option in ocsp application to stop the server
2555 exiting on the first error in a request.
2558 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
2559 if the server requested one: as stated in TLS 1.0 and SSL 3.0
2563 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
2564 extra data after the compression methods not only for TLS 1.0
2565 but also for SSL 3.0 (as required by the specification).
2566 [Bodo Moeller; problem pointed out by Matthias Loepfe]
2568 *) Change X509_certificate_type() to mark the key as exported/exportable
2569 when it's 512 *bits* long, not 512 bytes.
2572 *) Change AES_cbc_encrypt() so it outputs exact multiple of
2573 blocks during encryption.
2576 *) Various fixes to base64 BIO and non blocking I/O. On write
2577 flushes were not handled properly if the BIO retried. On read
2578 data was not being buffered properly and had various logic bugs.
2579 This also affects blocking I/O when the data being decoded is a
2583 *) Various S/MIME bugfixes and compatibility changes:
2584 output correct application/pkcs7 MIME type if
2585 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
2586 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
2587 of files as .eml work). Correctly handle very long lines in MIME
2591 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
2593 *) Countermeasure against the Klima-Pokorny-Rosa extension of
2594 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
2595 a protocol version number mismatch like a decryption error
2596 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
2599 *) Turn on RSA blinding by default in the default implementation
2600 to avoid a timing attack. Applications that don't want it can call
2601 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
2602 They would be ill-advised to do so in most cases.
2603 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
2605 *) Change RSA blinding code so that it works when the PRNG is not
2606 seeded (in this case, the secret RSA exponent is abused as
2607 an unpredictable seed -- if it is not unpredictable, there
2608 is no point in blinding anyway). Make RSA blinding thread-safe
2609 by remembering the creator's thread ID in rsa->blinding and
2610 having all other threads use local one-time blinding factors
2611 (this requires more computation than sharing rsa->blinding, but
2612 avoids excessive locking; and if an RSA object is not shared
2613 between threads, blinding will still be very fast).
2616 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
2617 ENGINE as defaults for all supported algorithms irrespective of
2618 the 'flags' parameter. 'flags' is now honoured, so applications
2619 should make sure they are passing it correctly.
2622 *) Target "mingw" now allows native Windows code to be generated in
2623 the Cygwin environment as well as with the MinGW compiler.
2626 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
2628 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
2629 via timing by performing a MAC computation even if incorrrect
2630 block cipher padding has been found. This is a countermeasure
2631 against active attacks where the attacker has to distinguish
2632 between bad padding and a MAC verification error. (CVE-2003-0078)
2634 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
2635 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
2636 Martin Vuagnoux (EPFL, Ilion)]
2638 *) Make the no-err option work as intended. The intention with no-err
2639 is not to have the whole error stack handling routines removed from
2640 libcrypto, it's only intended to remove all the function name and
2641 reason texts, thereby removing some of the footprint that may not
2642 be interesting if those errors aren't displayed anyway.
2644 NOTE: it's still possible for any application or module to have it's
2645 own set of error texts inserted. The routines are there, just not
2646 used by default when no-err is given.
2649 *) Add support for FreeBSD on IA64.
2650 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
2652 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
2653 Kerberos function mit_des_cbc_cksum(). Before this change,
2654 the value returned by DES_cbc_cksum() was like the one from
2655 mit_des_cbc_cksum(), except the bytes were swapped.
2656 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
2658 *) Allow an application to disable the automatic SSL chain building.
2659 Before this a rather primitive chain build was always performed in
2660 ssl3_output_cert_chain(): an application had no way to send the
2661 correct chain if the automatic operation produced an incorrect result.
2663 Now the chain builder is disabled if either:
2665 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
2667 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
2669 The reasoning behind this is that an application would not want the
2670 auto chain building to take place if extra chain certificates are
2671 present and it might also want a means of sending no additional
2672 certificates (for example the chain has two certificates and the
2676 *) Add the possibility to build without the ENGINE framework.
2677 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2679 *) Under Win32 gmtime() can return NULL: check return value in
2680 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
2683 *) DSA routines: under certain error conditions uninitialized BN objects
2684 could be freed. Solution: make sure initialization is performed early
2685 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
2686 Nils Larsch <nla@trustcenter.de> via PR#459)
2689 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
2690 checked on reconnect on the client side, therefore session resumption
2691 could still fail with a "ssl session id is different" error. This
2692 behaviour is masked when SSL_OP_ALL is used due to
2693 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
2694 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2695 followup to PR #377.
2698 *) IA-32 assembler support enhancements: unified ELF targets, support
2699 for SCO/Caldera platforms, fix for Cygwin shared build.
2702 *) Add support for FreeBSD on sparc64. As a consequence, support for
2703 FreeBSD on non-x86 processors is separate from x86 processors on
2704 the config script, much like the NetBSD support.
2705 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
2707 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
2709 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
2712 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
2713 code (06) was taken as the first octet of the session ID and the last
2714 octet was ignored consequently. As a result SSLv2 client side session
2715 caching could not have worked due to the session ID mismatch between
2717 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2721 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
2722 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
2726 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
2727 seems that in spite of existing for more than a year, many application
2728 author have done nothing to provide the necessary callbacks, which
2729 means that this particular engine will not work properly anywhere.
2730 This is a very unfortunate situation which forces us, in the name
2731 of usability, to give the hw_ncipher.c a static lock, which is part
2733 NOTE: This is for the 0.9.7 series ONLY. This hack will never
2734 appear in 0.9.8 or later. We EXPECT application authors to have
2735 dealt properly with this when 0.9.8 is released (unless we actually
2736 make such changes in the libcrypto locking code that changes will
2737 have to be made anyway).
2740 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
2741 octets have been read, EOF or an error occurs. Without this change
2742 some truncated ASN1 structures will not produce an error.
2745 *) Disable Heimdal support, since it hasn't been fully implemented.
2746 Still give the possibility to force the use of Heimdal, but with
2747 warnings and a request that patches get sent to openssl-dev.
2750 *) Add the VC-CE target, introduce the WINCE sysname, and add
2751 INSTALL.WCE and appropriate conditionals to make it build.
2752 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2754 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
2755 cygssl-x.y.z.dll, where x, y and z are the major, minor and
2756 edit numbers of the version.
2757 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
2759 *) Introduce safe string copy and catenation functions
2760 (BUF_strlcpy() and BUF_strlcat()).
2761 [Ben Laurie (CHATS) and Richard Levitte]
2763 *) Avoid using fixed-size buffers for one-line DNs.
2764 [Ben Laurie (CHATS)]
2766 *) Add BUF_MEM_grow_clean() to avoid information leakage when
2767 resizing buffers containing secrets, and use where appropriate.
2768 [Ben Laurie (CHATS)]
2770 *) Avoid using fixed size buffers for configuration file location.
2771 [Ben Laurie (CHATS)]
2773 *) Avoid filename truncation for various CA files.
2774 [Ben Laurie (CHATS)]
2776 *) Use sizeof in preference to magic numbers.
2777 [Ben Laurie (CHATS)]
2779 *) Avoid filename truncation in cert requests.
2780 [Ben Laurie (CHATS)]
2782 *) Add assertions to check for (supposedly impossible) buffer
2784 [Ben Laurie (CHATS)]
2786 *) Don't cache truncated DNS entries in the local cache (this could
2787 potentially lead to a spoofing attack).
2788 [Ben Laurie (CHATS)]
2790 *) Fix various buffers to be large enough for hex/decimal
2791 representations in a platform independent manner.
2792 [Ben Laurie (CHATS)]
2794 *) Add CRYPTO_realloc_clean() to avoid information leakage when
2795 resizing buffers containing secrets, and use where appropriate.
2796 [Ben Laurie (CHATS)]
2798 *) Add BIO_indent() to avoid much slightly worrying code to do
2800 [Ben Laurie (CHATS)]
2802 *) Convert sprintf()/BIO_puts() to BIO_printf().
2803 [Ben Laurie (CHATS)]
2805 *) buffer_gets() could terminate with the buffer only half
2807 [Ben Laurie (CHATS)]
2809 *) Add assertions to prevent user-supplied crypto functions from
2810 overflowing internal buffers by having large block sizes, etc.
2811 [Ben Laurie (CHATS)]
2813 *) New OPENSSL_assert() macro (similar to assert(), but enabled
2815 [Ben Laurie (CHATS)]
2817 *) Eliminate unused copy of key in RC4.
2818 [Ben Laurie (CHATS)]
2820 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
2821 [Ben Laurie (CHATS)]
2823 *) Fix off-by-one error in EGD path.
2824 [Ben Laurie (CHATS)]
2826 *) If RANDFILE path is too long, ignore instead of truncating.
2827 [Ben Laurie (CHATS)]
2829 *) Eliminate unused and incorrectly sized X.509 structure
2831 [Ben Laurie (CHATS)]
2833 *) Eliminate unused and dangerous function knumber().
2834 [Ben Laurie (CHATS)]
2836 *) Eliminate unused and dangerous structure, KSSL_ERR.
2837 [Ben Laurie (CHATS)]
2839 *) Protect against overlong session ID context length in an encoded
2840 session object. Since these are local, this does not appear to be
2842 [Ben Laurie (CHATS)]
2844 *) Change from security patch (see 0.9.6e below) that did not affect
2845 the 0.9.6 release series:
2847 Remote buffer overflow in SSL3 protocol - an attacker could
2848 supply an oversized master key in Kerberos-enabled versions.
2850 [Ben Laurie (CHATS)]
2852 *) Change the SSL kerb5 codes to match RFC 2712.
2855 *) Make -nameopt work fully for req and add -reqopt switch.
2856 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
2858 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
2859 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
2861 *) Make sure tests can be performed even if the corresponding algorithms
2862 have been removed entirely. This was also the last step to make
2863 OpenSSL compilable with DJGPP under all reasonable conditions.
2864 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
2866 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
2867 to allow version independent disabling of normally unselected ciphers,
2868 which may be activated as a side-effect of selecting a single cipher.
2870 (E.g., cipher list string "RSA" enables ciphersuites that are left
2871 out of "ALL" because they do not provide symmetric encryption.
2872 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
2873 [Lutz Jaenicke, Bodo Moeller]
2875 *) Add appropriate support for separate platform-dependent build
2876 directories. The recommended way to make a platform-dependent
2877 build directory is the following (tested on Linux), maybe with
2880 # Place yourself outside of the OpenSSL source tree. In
2881 # this example, the environment variable OPENSSL_SOURCE
2882 # is assumed to contain the absolute OpenSSL source directory.
2883 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
2884 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
2885 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
2886 mkdir -p `dirname $F`
2887 ln -s $OPENSSL_SOURCE/$F $F
2890 To be absolutely sure not to disturb the source tree, a "make clean"
2891 is a good thing. If it isn't successfull, don't worry about it,
2892 it probably means the source directory is very clean.
2895 *) Make sure any ENGINE control commands make local copies of string
2896 pointers passed to them whenever necessary. Otherwise it is possible
2897 the caller may have overwritten (or deallocated) the original string
2898 data when a later ENGINE operation tries to use the stored values.
2899 [Götz Babin-Ebell <babinebell@trustcenter.de>]
2901 *) Improve diagnostics in file reading and command-line digests.
2902 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
2904 *) Add AES modes CFB and OFB to the object database. Correct an
2905 error in AES-CFB decryption.
2908 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
2909 allows existing EVP_CIPHER_CTX structures to be reused after
2910 calling EVP_*Final(). This behaviour is used by encryption
2911 BIOs and some applications. This has the side effect that
2912 applications must explicitly clean up cipher contexts with
2913 EVP_CIPHER_CTX_cleanup() or they will leak memory.
2916 *) Check the values of dna and dnb in bn_mul_recursive before calling
2917 bn_mul_comba (a non zero value means the a or b arrays do not contain
2918 n2 elements) and fallback to bn_mul_normal if either is not zero.
2921 *) Fix escaping of non-ASCII characters when using the -subj option
2922 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
2925 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
2926 form for "surname", serialNumber has no short form.
2927 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
2928 therefore remove "mail" short name for "internet 7".
2929 The OID for unique identifiers in X509 certificates is
2930 x500UniqueIdentifier, not uniqueIdentifier.
2931 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
2934 *) Add an "init" command to the ENGINE config module and auto initialize
2935 ENGINEs. Without any "init" command the ENGINE will be initialized
2936 after all ctrl commands have been executed on it. If init=1 the
2937 ENGINE is initailized at that point (ctrls before that point are run
2938 on the uninitialized ENGINE and after on the initialized one). If
2939 init=0 then the ENGINE will not be iniatialized at all.
2942 *) Fix the 'app_verify_callback' interface so that the user-defined
2943 argument is actually passed to the callback: In the
2944 SSL_CTX_set_cert_verify_callback() prototype, the callback
2945 declaration has been changed from
2948 int (*cb)(X509_STORE_CTX *,void *);
2949 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
2950 i=s->ctx->app_verify_callback(&ctx)
2951 has been changed into
2952 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
2954 To update applications using SSL_CTX_set_cert_verify_callback(),
2955 a dummy argument can be added to their callback functions.
2956 [D. K. Smetters <smetters@parc.xerox.com>]
2958 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
2959 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
2961 *) Add and OPENSSL_LOAD_CONF define which will cause
2962 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
2963 This allows older applications to transparently support certain
2964 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
2965 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
2966 load the config file and OPENSSL_add_all_algorithms_conf() which will
2967 always load it have also been added.
2970 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
2971 Adjust NIDs and EVP layer.
2972 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2974 *) Config modules support in openssl utility.
2976 Most commands now load modules from the config file,
2977 though in a few (such as version) this isn't done
2978 because it couldn't be used for anything.
2980 In the case of ca and req the config file used is
2981 the same as the utility itself: that is the -config
2982 command line option can be used to specify an
2986 *) Move default behaviour from OPENSSL_config(). If appname is NULL
2987 use "openssl_conf" if filename is NULL use default openssl config file.
2990 *) Add an argument to OPENSSL_config() to allow the use of an alternative
2991 config section name. Add a new flag to tolerate a missing config file
2992 and move code to CONF_modules_load_file().
2995 *) Support for crypto accelerator cards from Accelerated Encryption
2996 Processing, www.aep.ie. (Use engine 'aep')
2997 The support was copied from 0.9.6c [engine] and adapted/corrected
2998 to work with the new engine framework.
2999 [AEP Inc. and Richard Levitte]
3001 *) Support for SureWare crypto accelerator cards from Baltimore
3002 Technologies. (Use engine 'sureware')
3003 The support was copied from 0.9.6c [engine] and adapted
3004 to work with the new engine framework.
3007 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
3008 make the newer ENGINE framework commands for the CHIL engine work.
3009 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
3011 *) Make it possible to produce shared libraries on ReliantUNIX.
3012 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
3014 *) Add the configuration target debug-linux-ppro.
3015 Make 'openssl rsa' use the general key loading routines
3016 implemented in apps.c, and make those routines able to
3017 handle the key format FORMAT_NETSCAPE and the variant
3019 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3021 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
3022 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3024 *) Add -keyform to rsautl, and document -engine.
3025 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
3027 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
3028 BIO_R_NO_SUCH_FILE error code rather than the generic
3029 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
3032 *) Add new functions
3034 ERR_peek_last_error_line
3035 ERR_peek_last_error_line_data.
3036 These are similar to
3039 ERR_peek_error_line_data,
3040 but report on the latest error recorded rather than the first one
3041 still in the error queue.
3042 [Ben Laurie, Bodo Moeller]
3044 *) default_algorithms option in ENGINE config module. This allows things
3046 default_algorithms = ALL
3047 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
3050 *) Prelminary ENGINE config module.
3053 *) New experimental application configuration code.
3056 *) Change the AES code to follow the same name structure as all other
3057 symmetric ciphers, and behave the same way. Move everything to
3058 the directory crypto/aes, thereby obsoleting crypto/rijndael.
3059 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3061 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
3062 [Ben Laurie and Theo de Raadt]
3064 *) Add option to output public keys in req command.
3065 [Massimiliano Pala madwolf@openca.org]
3067 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
3068 (up to about 10% better than before for P-192 and P-224).
3071 *) New functions/macros
3073 SSL_CTX_set_msg_callback(ctx, cb)
3074 SSL_CTX_set_msg_callback_arg(ctx, arg)
3075 SSL_set_msg_callback(ssl, cb)
3076 SSL_set_msg_callback_arg(ssl, arg)
3078 to request calling a callback function
3080 void cb(int write_p, int version, int content_type,
3081 const void *buf, size_t len, SSL *ssl, void *arg)
3083 whenever a protocol message has been completely received
3084 (write_p == 0) or sent (write_p == 1). Here 'version' is the
3085 protocol version according to which the SSL library interprets
3086 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
3087 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
3088 the content type as defined in the SSL 3.0/TLS 1.0 protocol
3089 specification (change_cipher_spec(20), alert(21), handshake(22)).
3090 'buf' and 'len' point to the actual message, 'ssl' to the
3091 SSL object, and 'arg' is the application-defined value set by
3092 SSL[_CTX]_set_msg_callback_arg().
3094 'openssl s_client' and 'openssl s_server' have new '-msg' options
3095 to enable a callback that displays all protocol messages.
3098 *) Change the shared library support so shared libraries are built as
3099 soon as the corresponding static library is finished, and thereby get
3100 openssl and the test programs linked against the shared library.
3101 This still only happens when the keyword "shard" has been given to
3102 the configuration scripts.
3104 NOTE: shared library support is still an experimental thing, and
3105 backward binary compatibility is still not guaranteed.
3106 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
3108 *) Add support for Subject Information Access extension.
3109 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3111 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
3112 additional bytes when new memory had to be allocated, not just
3113 when reusing an existing buffer.
3116 *) New command line and configuration option 'utf8' for the req command.
3117 This allows field values to be specified as UTF8 strings.
3120 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
3121 runs for the former and machine-readable output for the latter.
3124 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
3125 of the e-mail address in the DN (i.e., it will go into a certificate
3126 extension only). The new configuration file option 'email_in_dn = no'
3127 has the same effect.
3128 [Massimiliano Pala madwolf@openca.org]
3130 *) Change all functions with names starting with des_ to be starting
3131 with DES_ instead. Add wrappers that are compatible with libdes,
3132 but are named _ossl_old_des_*. Finally, add macros that map the
3133 des_* symbols to the corresponding _ossl_old_des_* if libdes
3134 compatibility is desired. If OpenSSL 0.9.6c compatibility is
3135 desired, the des_* symbols will be mapped to DES_*, with one
3138 Since we provide two compatibility mappings, the user needs to
3139 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
3140 compatibility is desired. The default (i.e., when that macro
3141 isn't defined) is OpenSSL 0.9.6c compatibility.
3143 There are also macros that enable and disable the support of old
3144 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
3145 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
3146 are defined, the default will apply: to support the old des routines.
3148 In either case, one must include openssl/des.h to get the correct
3149 definitions. Do not try to just include openssl/des_old.h, that
3152 NOTE: This is a major break of an old API into a new one. Software
3153 authors are encouraged to switch to the DES_ style functions. Some
3154 time in the future, des_old.h and the libdes compatibility functions
3155 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
3156 default), and then completely removed.
3159 *) Test for certificates which contain unsupported critical extensions.
3160 If such a certificate is found during a verify operation it is
3161 rejected by default: this behaviour can be overridden by either
3162 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
3163 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
3164 X509_supported_extension() has also been added which returns 1 if a
3165 particular extension is supported.
3168 *) Modify the behaviour of EVP cipher functions in similar way to digests
3169 to retain compatibility with existing code.
3172 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
3173 compatibility with existing code. In particular the 'ctx' parameter does
3174 not have to be to be initialized before the call to EVP_DigestInit() and
3175 it is tidied up after a call to EVP_DigestFinal(). New function
3176 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
3177 EVP_MD_CTX_copy() changed to not require the destination to be
3178 initialized valid and new function EVP_MD_CTX_copy_ex() added which
3179 requires the destination to be valid.
3181 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
3182 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
3185 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
3186 so that complete 'Handshake' protocol structures are kept in memory
3187 instead of overwriting 'msg_type' and 'length' with 'body' data.
3190 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
3191 [Massimo Santin via Richard Levitte]
3193 *) Major restructuring to the underlying ENGINE code. This includes
3194 reduction of linker bloat, separation of pure "ENGINE" manipulation
3195 (initialisation, etc) from functionality dealing with implementations
3196 of specific crypto iterfaces. This change also introduces integrated
3197 support for symmetric ciphers and digest implementations - so ENGINEs
3198 can now accelerate these by providing EVP_CIPHER and EVP_MD
3199 implementations of their own. This is detailed in crypto/engine/README
3200 as it couldn't be adequately described here. However, there are a few
3201 API changes worth noting - some RSA, DSA, DH, and RAND functions that
3202 were changed in the original introduction of ENGINE code have now
3203 reverted back - the hooking from this code to ENGINE is now a good
3204 deal more passive and at run-time, operations deal directly with
3205 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
3206 dereferencing through an ENGINE pointer any more. Also, the ENGINE
3207 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
3208 they were not being used by the framework as there is no concept of a
3209 BIGNUM_METHOD and they could not be generalised to the new
3210 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
3211 ENGINE_cpy() has been removed as it cannot be consistently defined in
3215 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
3218 *) Change mkdef.pl to sort symbols that get the same entry number,
3219 and make sure the automatically generated functions ERR_load_*
3220 become part of libeay.num as well.
3223 *) New function SSL_renegotiate_pending(). This returns true once
3224 renegotiation has been requested (either SSL_renegotiate() call
3225 or HelloRequest/ClientHello receveived from the peer) and becomes
3226 false once a handshake has been completed.
3227 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
3228 sends a HelloRequest, but does not ensure that a handshake takes
3229 place. SSL_renegotiate_pending() is useful for checking if the
3230 client has followed the request.)
3233 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
3234 By default, clients may request session resumption even during
3235 renegotiation (if session ID contexts permit); with this option,
3236 session resumption is possible only in the first handshake.
3238 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
3239 more bits available for options that should not be part of
3240 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
3243 *) Add some demos for certificate and certificate request creation.
3246 *) Make maximum certificate chain size accepted from the peer application
3247 settable (SSL*_get/set_max_cert_list()), as proposed by
3248 "Douglas E. Engert" <deengert@anl.gov>.
3251 *) Add support for shared libraries for Unixware-7
3252 (Boyd Lynn Gerber <gerberb@zenez.com>).
3255 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
3256 be done prior to destruction. Use this to unload error strings from
3257 ENGINEs that load their own error strings. NB: This adds two new API
3258 functions to "get" and "set" this destroy handler in an ENGINE.
3261 *) Alter all existing ENGINE implementations (except "openssl" and
3262 "openbsd") to dynamically instantiate their own error strings. This
3263 makes them more flexible to be built both as statically-linked ENGINEs
3264 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
3265 Also, add stub code to each that makes building them as self-contained
3266 shared-libraries easier (see README.ENGINE).
3269 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
3270 implementations into applications that are completely implemented in
3271 self-contained shared-libraries. The "dynamic" ENGINE exposes control
3272 commands that can be used to configure what shared-library to load and
3273 to control aspects of the way it is handled. Also, made an update to
3274 the README.ENGINE file that brings its information up-to-date and
3275 provides some information and instructions on the "dynamic" ENGINE
3276 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
3279 *) Make it possible to unload ranges of ERR strings with a new
3280 "ERR_unload_strings" function.
3283 *) Add a copy() function to EVP_MD.
3286 *) Make EVP_MD routines take a context pointer instead of just the
3287 md_data void pointer.
3290 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
3291 that the digest can only process a single chunk of data
3292 (typically because it is provided by a piece of
3293 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
3294 is only going to provide a single chunk of data, and hence the
3295 framework needn't accumulate the data for oneshot drivers.
3298 *) As with "ERR", make it possible to replace the underlying "ex_data"
3299 functions. This change also alters the storage and management of global
3300 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
3301 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
3302 index counters. The API functions that use this state have been changed
3303 to take a "class_index" rather than pointers to the class's local STACK
3304 and counter, and there is now an API function to dynamically create new
3305 classes. This centralisation allows us to (a) plug a lot of the
3306 thread-safety problems that existed, and (b) makes it possible to clean
3307 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
3308 such data would previously have always leaked in application code and
3309 workarounds were in place to make the memory debugging turn a blind eye
3310 to it. Application code that doesn't use this new function will still
3311 leak as before, but their memory debugging output will announce it now
3312 rather than letting it slide.
3314 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
3315 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
3316 has a return value to indicate success or failure.
3319 *) Make it possible to replace the underlying "ERR" functions such that the
3320 global state (2 LHASH tables and 2 locks) is only used by the "default"
3321 implementation. This change also adds two functions to "get" and "set"
3322 the implementation prior to it being automatically set the first time
3323 any other ERR function takes place. Ie. an application can call "get",
3324 pass the return value to a module it has just loaded, and that module
3325 can call its own "set" function using that value. This means the
3326 module's "ERR" operations will use (and modify) the error state in the
3327 application and not in its own statically linked copy of OpenSSL code.
3330 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
3331 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
3332 the operation, and provides a more encapsulated way for external code
3333 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
3334 to use these functions rather than manually incrementing the counts.
3336 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
3339 *) Add EVP test program.
3342 *) Add symmetric cipher support to ENGINE. Expect the API to change!
3345 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
3346 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
3347 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
3348 These allow a CRL to be built without having to access X509_CRL fields
3349 directly. Modify 'ca' application to use new functions.
3352 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
3353 bug workarounds. Rollback attack detection is a security feature.
3354 The problem will only arise on OpenSSL servers when TLSv1 is not
3355 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
3356 Software authors not wanting to support TLSv1 will have special reasons
3357 for their choice and can explicitly enable this option.
3358 [Bodo Moeller, Lutz Jaenicke]
3360 *) Rationalise EVP so it can be extended: don't include a union of
3361 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
3362 (similar to those existing for EVP_CIPHER_CTX).
3367 EVP_MD_CTX_init(&md); /* new function call */
3368 EVP_DigestInit(&md, EVP_sha1());
3369 EVP_DigestUpdate(&md, in, len);
3370 EVP_DigestFinal(&md, out, NULL);
3371 EVP_MD_CTX_cleanup(&md); /* new function call */
3375 *) Make DES key schedule conform to the usual scheme, as well as
3376 correcting its structure. This means that calls to DES functions
3377 now have to pass a pointer to a des_key_schedule instead of a
3378 plain des_key_schedule (which was actually always a pointer
3381 des_key_schedule ks;
3383 des_set_key_checked(..., &ks);
3384 des_ncbc_encrypt(..., &ks, ...);
3386 (Note that a later change renames 'des_...' into 'DES_...'.)
3389 *) Initial reduction of linker bloat: the use of some functions, such as
3390 PEM causes large amounts of unused functions to be linked in due to
3391 poor organisation. For example pem_all.c contains every PEM function
3392 which has a knock on effect of linking in large amounts of (unused)
3393 ASN1 code. Grouping together similar functions and splitting unrelated
3394 functions prevents this.
3397 *) Cleanup of EVP macros.
3400 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
3401 correct _ecb suffix.
3404 *) Add initial OCSP responder support to ocsp application. The
3405 revocation information is handled using the text based index
3406 use by the ca application. The responder can either handle
3407 requests generated internally, supplied in files (for example
3408 via a CGI script) or using an internal minimal server.
3411 *) Add configuration choices to get zlib compression for TLS.
3414 *) Changes to Kerberos SSL for RFC 2712 compliance:
3415 1. Implemented real KerberosWrapper, instead of just using
3416 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
3417 2. Implemented optional authenticator field of KerberosWrapper.
3419 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
3420 and authenticator structs; see crypto/krb5/.
3422 Generalized Kerberos calls to support multiple Kerberos libraries.
3423 [Vern Staats <staatsvr@asc.hpc.mil>,
3424 Jeffrey Altman <jaltman@columbia.edu>
3425 via Richard Levitte]
3427 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
3428 already does with RSA. testdsa.h now has 'priv_key/pub_key'
3429 values for each of the key sizes rather than having just
3430 parameters (and 'speed' generating keys each time).
3433 *) Speed up EVP routines.
3436 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
3437 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
3438 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
3439 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
3441 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
3442 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
3443 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
3446 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
3448 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
3451 *) Added the OS2-EMX target.
3452 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
3454 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
3455 to support NCONF routines in extension code. New function CONF_set_nconf()
3456 to allow functions which take an NCONF to also handle the old LHASH
3457 structure: this means that the old CONF compatible routines can be
3458 retained (in particular wrt extensions) without having to duplicate the
3459 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
3462 *) Enhance the general user interface with mechanisms for inner control
3463 and with possibilities to have yes/no kind of prompts.
3466 *) Change all calls to low level digest routines in the library and
3467 applications to use EVP. Add missing calls to HMAC_cleanup() and
3468 don't assume HMAC_CTX can be copied using memcpy().
3469 [Verdon Walker <VWalker@novell.com>, Steve Henson]
3471 *) Add the possibility to control engines through control names but with
3472 arbitrary arguments instead of just a string.
3473 Change the key loaders to take a UI_METHOD instead of a callback
3474 function pointer. NOTE: this breaks binary compatibility with earlier
3475 versions of OpenSSL [engine].
3476 Adapt the nCipher code for these new conditions and add a card insertion
3480 *) Enhance the general user interface with mechanisms to better support
3481 dialog box interfaces, application-defined prompts, the possibility
3482 to use defaults (for example default passwords from somewhere else)
3483 and interrupts/cancellations.
3486 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
3487 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
3490 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
3491 tidy up some unnecessarily weird code in 'sk_new()').
3492 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
3494 *) Change the key loading routines for ENGINEs to use the same kind
3495 callback (pem_password_cb) as all other routines that need this
3499 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
3500 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
3501 than this minimum value is recommended.
3504 *) New random seeder for OpenVMS, using the system process statistics
3505 that are easily reachable.
3508 *) Windows apparently can't transparently handle global
3509 variables defined in DLLs. Initialisations such as:
3511 const ASN1_ITEM *it = &ASN1_INTEGER_it;
3513 wont compile. This is used by the any applications that need to
3514 declare their own ASN1 modules. This was fixed by adding the option
3515 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
3516 needed for static libraries under Win32.
3519 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
3520 setting of purpose and trust fields. New X509_STORE trust and
3521 purpose functions and tidy up setting in other SSL functions.
3524 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
3525 structure. These are inherited by X509_STORE_CTX when it is
3526 initialised. This allows various defaults to be set in the
3527 X509_STORE structure (such as flags for CRL checking and custom
3528 purpose or trust settings) for functions which only use X509_STORE_CTX
3529 internally such as S/MIME.
3531 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
3532 trust settings if they are not set in X509_STORE. This allows X509_STORE
3533 purposes and trust (in S/MIME for example) to override any set by default.
3535 Add command line options for CRL checking to smime, s_client and s_server
3539 *) Initial CRL based revocation checking. If the CRL checking flag(s)
3540 are set then the CRL is looked up in the X509_STORE structure and
3541 its validity and signature checked, then if the certificate is found
3542 in the CRL the verify fails with a revoked error.
3544 Various new CRL related callbacks added to X509_STORE_CTX structure.
3546 Command line options added to 'verify' application to support this.
3548 This needs some additional work, such as being able to handle multiple
3549 CRLs with different times, extension based lookup (rather than just
3550 by subject name) and ultimately more complete V2 CRL extension
3554 *) Add a general user interface API (crypto/ui/). This is designed
3555 to replace things like des_read_password and friends (backward
3556 compatibility functions using this new API are provided).
3557 The purpose is to remove prompting functions from the DES code
3558 section as well as provide for prompting through dialog boxes in
3559 a window system and the like.
3562 *) Add "ex_data" support to ENGINE so implementations can add state at a
3563 per-structure level rather than having to store it globally.
3566 *) Make it possible for ENGINE structures to be copied when retrieved by
3567 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
3568 This causes the "original" ENGINE structure to act like a template,
3569 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
3570 operational state can be localised to each ENGINE structure, despite the
3571 fact they all share the same "methods". New ENGINE structures returned in
3572 this case have no functional references and the return value is the single
3573 structural reference. This matches the single structural reference returned
3574 by ENGINE_by_id() normally, when it is incremented on the pre-existing
3578 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
3579 needs to match any other type at all we need to manually clear the
3583 *) Changes to the "openssl engine" utility to include;
3584 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
3585 about an ENGINE's available control commands.
3586 - executing control commands from command line arguments using the
3587 '-pre' and '-post' switches. '-post' is only used if '-t' is
3588 specified and the ENGINE is successfully initialised. The syntax for
3589 the individual commands are colon-separated, for example;
3590 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
3593 *) New dynamic control command support for ENGINEs. ENGINEs can now
3594 declare their own commands (numbers), names (strings), descriptions,
3595 and input types for run-time discovery by calling applications. A
3596 subset of these commands are implicitly classed as "executable"
3597 depending on their input type, and only these can be invoked through
3598 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
3599 can be based on user input, config files, etc). The distinction is
3600 that "executable" commands cannot return anything other than a boolean
3601 result and can only support numeric or string input, whereas some
3602 discoverable commands may only be for direct use through
3603 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
3604 pointers, or other custom uses. The "executable" commands are to
3605 support parameterisations of ENGINE behaviour that can be
3606 unambiguously defined by ENGINEs and used consistently across any
3607 OpenSSL-based application. Commands have been added to all the
3608 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
3609 control over shared-library paths without source code alterations.
3612 *) Changed all ENGINE implementations to dynamically allocate their
3613 ENGINEs rather than declaring them statically. Apart from this being
3614 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
3615 this also allows the implementations to compile without using the
3616 internal engine_int.h header.
3619 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
3620 'const' value. Any code that should be able to modify a RAND_METHOD
3621 should already have non-const pointers to it (ie. they should only
3622 modify their own ones).
3625 *) Made a variety of little tweaks to the ENGINE code.
3626 - "atalla" and "ubsec" string definitions were moved from header files
3627 to C code. "nuron" string definitions were placed in variables
3628 rather than hard-coded - allowing parameterisation of these values
3629 later on via ctrl() commands.
3630 - Removed unused "#if 0"'d code.
3631 - Fixed engine list iteration code so it uses ENGINE_free() to release
3632 structural references.
3633 - Constified the RAND_METHOD element of ENGINE structures.
3634 - Constified various get/set functions as appropriate and added
3635 missing functions (including a catch-all ENGINE_cpy that duplicates
3636 all ENGINE values onto a new ENGINE except reference counts/state).
3637 - Removed NULL parameter checks in get/set functions. Setting a method
3638 or function to NULL is a way of cancelling out a previously set
3639 value. Passing a NULL ENGINE parameter is just plain stupid anyway
3640 and doesn't justify the extra error symbols and code.
3641 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
3642 flags from engine_int.h to engine.h.
3643 - Changed prototypes for ENGINE handler functions (init(), finish(),
3644 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
3647 *) Implement binary inversion algorithm for BN_mod_inverse in addition
3648 to the algorithm using long division. The binary algorithm can be
3649 used only if the modulus is odd. On 32-bit systems, it is faster
3650 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
3651 roughly 5-15% for 256-bit moduli), so we use it only for moduli
3652 up to 450 bits. In 64-bit environments, the binary algorithm
3653 appears to be advantageous for much longer moduli; here we use it
3654 for moduli up to 2048 bits.
3657 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
3658 could not support the combine flag in choice fields.
3661 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
3662 extensions from a certificate request to the certificate.
3665 *) Allow multiple 'certopt' and 'nameopt' options to be separated
3666 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
3667 file: this allows the display of the certificate about to be
3668 signed to be customised, to allow certain fields to be included
3669 or excluded and extension details. The old system didn't display
3670 multicharacter strings properly, omitted fields not in the policy
3671 and couldn't display additional details such as extensions.
3674 *) Function EC_POINTs_mul for multiple scalar multiplication
3675 of an arbitrary number of elliptic curve points
3676 \sum scalars[i]*points[i],
3677 optionally including the generator defined for the EC_GROUP:
3678 scalar*generator + \sum scalars[i]*points[i].
3680 EC_POINT_mul is a simple wrapper function for the typical case
3681 that the point list has just one item (besides the optional
3685 *) First EC_METHODs for curves over GF(p):
3687 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
3688 operations and provides various method functions that can also
3689 operate with faster implementations of modular arithmetic.
3691 EC_GFp_mont_method() reuses most functions that are part of
3692 EC_GFp_simple_method, but uses Montgomery arithmetic.
3694 [Bodo Moeller; point addition and point doubling
3695 implementation directly derived from source code provided by
3696 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
3698 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
3699 crypto/ec/ec_lib.c):
3701 Curves are EC_GROUP objects (with an optional group generator)
3702 based on EC_METHODs that are built into the library.
3704 Points are EC_POINT objects based on EC_GROUP objects.
3706 Most of the framework would be able to handle curves over arbitrary
3707 finite fields, but as there are no obvious types for fields other
3708 than GF(p), some functions are limited to that for now.
3711 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
3712 that the file contains a complete HTTP response.
3715 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
3716 change the def and num file printf format specifier from "%-40sXXX"
3717 to "%-39s XXX". The latter will always guarantee a space after the
3718 field while the former will cause them to run together if the field
3719 is 40 of more characters long.
3722 *) Constify the cipher and digest 'method' functions and structures
3723 and modify related functions to take constant EVP_MD and EVP_CIPHER
3727 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
3728 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
3731 *) Modify EVP_Digest*() routines so they now return values. Although the
3732 internal software routines can never fail additional hardware versions
3736 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
3738 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
3739 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
3742 ERR_R_NESTED_ASN1_ERROR
3744 ERR_R_MISSING_ASN1_EOS
3745 were 4 .. 9, conflicting with
3746 ERR_LIB_RSA (= ERR_R_RSA_LIB)
3748 ERR_LIB_PEM (= ERR_R_PEM_LIB).
3749 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
3751 Add new error code 'ERR_R_INTERNAL_ERROR'.
3754 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
3758 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
3759 sets the subject name for a new request or supersedes the
3760 subject name in a given request. Formats that can be parsed are
3761 'CN=Some Name, OU=myOU, C=IT'
3763 'CN=Some Name/OU=myOU/C=IT'.
3765 Add options '-batch' and '-verbose' to 'openssl req'.
3766 [Massimiliano Pala <madwolf@hackmasters.net>]
3768 *) Introduce the possibility to access global variables through
3769 functions on platform were that's the best way to handle exporting
3770 global variables in shared libraries. To enable this functionality,
3771 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
3772 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
3773 is normally done by Configure or something similar).
3775 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
3776 in the source file (foo.c) like this:
3778 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
3779 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
3781 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
3782 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
3784 OPENSSL_DECLARE_GLOBAL(int,foo);
3785 #define foo OPENSSL_GLOBAL_REF(foo)
3786 OPENSSL_DECLARE_GLOBAL(double,bar);
3787 #define bar OPENSSL_GLOBAL_REF(bar)
3789 The #defines are very important, and therefore so is including the
3790 header file everywhere where the defined globals are used.
3792 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
3793 of ASN.1 items, but that structure is a bit different.
3795 The largest change is in util/mkdef.pl which has been enhanced with
3796 better and easier to understand logic to choose which symbols should
3797 go into the Windows .def files as well as a number of fixes and code
3798 cleanup (among others, algorithm keywords are now sorted
3799 lexicographically to avoid constant rewrites).
3802 *) In BN_div() keep a copy of the sign of 'num' before writing the
3803 result to 'rm' because if rm==num the value will be overwritten
3804 and produce the wrong result if 'num' is negative: this caused
3805 problems with BN_mod() and BN_nnmod().
3808 *) Function OCSP_request_verify(). This checks the signature on an
3809 OCSP request and verifies the signer certificate. The signer
3810 certificate is just checked for a generic purpose and OCSP request
3814 *) Add OCSP_check_validity() function to check the validity of OCSP
3815 responses. OCSP responses are prepared in real time and may only
3816 be a few seconds old. Simply checking that the current time lies
3817 between thisUpdate and nextUpdate max reject otherwise valid responses
3818 caused by either OCSP responder or client clock inaccuracy. Instead
3819 we allow thisUpdate and nextUpdate to fall within a certain period of
3820 the current time. The age of the response can also optionally be
3821 checked. Two new options -validity_period and -status_age added to
3825 *) If signature or public key algorithm is unrecognized print out its
3826 OID rather that just UNKNOWN.
3829 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
3830 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
3831 ID to be generated from the issuer certificate alone which can then be
3832 passed to OCSP_id_issuer_cmp().
3835 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
3836 ASN1 modules to export functions returning ASN1_ITEM pointers
3837 instead of the ASN1_ITEM structures themselves. This adds several
3838 new macros which allow the underlying ASN1 function/structure to
3839 be accessed transparently. As a result code should not use ASN1_ITEM
3840 references directly (such as &X509_it) but instead use the relevant
3841 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
3842 use of the new ASN1 code on platforms where exporting structures
3843 is problematical (for example in shared libraries) but exporting
3844 functions returning pointers to structures is not.
3847 *) Add support for overriding the generation of SSL/TLS session IDs.
3848 These callbacks can be registered either in an SSL_CTX or per SSL.
3849 The purpose of this is to allow applications to control, if they wish,
3850 the arbitrary values chosen for use as session IDs, particularly as it
3851 can be useful for session caching in multiple-server environments. A
3852 command-line switch for testing this (and any client code that wishes
3853 to use such a feature) has been added to "s_server".
3854 [Geoff Thorpe, Lutz Jaenicke]
3856 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
3857 of the form '#if defined(...) || defined(...) || ...' and
3858 '#if !defined(...) && !defined(...) && ...'. This also avoids
3859 the growing number of special cases it was previously handling.
3862 *) Make all configuration macros available for application by making
3863 sure they are available in opensslconf.h, by giving them names starting
3864 with "OPENSSL_" to avoid conflicts with other packages and by making
3865 sure e_os2.h will cover all platform-specific cases together with
3867 Additionally, it is now possible to define configuration/platform-
3868 specific names (called "system identities"). In the C code, these
3869 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
3870 macro with the name beginning with "OPENSSL_SYS_", which is determined
3871 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
3875 *) New option -set_serial to 'req' and 'x509' this allows the serial
3876 number to use to be specified on the command line. Previously self
3877 signed certificates were hard coded with serial number 0 and the
3878 CA options of 'x509' had to use a serial number in a file which was
3882 *) New options to 'ca' utility to support V2 CRL entry extensions.
3883 Currently CRL reason, invalidity date and hold instruction are
3884 supported. Add new CRL extensions to V3 code and some new objects.
3887 *) New function EVP_CIPHER_CTX_set_padding() this is used to
3888 disable standard block padding (aka PKCS#5 padding) in the EVP
3889 API, which was previously mandatory. This means that the data is
3890 not padded in any way and so the total length much be a multiple
3891 of the block size, otherwise an error occurs.
3894 *) Initial (incomplete) OCSP SSL support.
3897 *) New function OCSP_parse_url(). This splits up a URL into its host,
3898 port and path components: primarily to parse OCSP URLs. New -url
3899 option to ocsp utility.
3902 *) New nonce behavior. The return value of OCSP_check_nonce() now
3903 reflects the various checks performed. Applications can decide
3904 whether to tolerate certain situations such as an absent nonce
3905 in a response when one was present in a request: the ocsp application
3906 just prints out a warning. New function OCSP_add1_basic_nonce()
3907 this is to allow responders to include a nonce in a response even if
3908 the request is nonce-less.
3911 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
3912 skipped when using openssl x509 multiple times on a single input file,
3913 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
3916 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
3917 set string type: to handle setting ASN1_TIME structures. Fix ca
3918 utility to correctly initialize revocation date of CRLs.
3921 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
3922 the clients preferred ciphersuites and rather use its own preferences.
3923 Should help to work around M$ SGC (Server Gated Cryptography) bug in
3924 Internet Explorer by ensuring unchanged hash method during stepup.
3925 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
3928 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
3929 to aes and add a new 'exist' option to print out symbols that don't
3933 *) Additional options to ocsp utility to allow flags to be set and
3934 additional certificates supplied.
3937 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
3938 OCSP client a number of certificate to only verify the response
3942 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
3943 handle the new API. Currently only ECB, CBC modes supported. Add new
3946 Add TLS AES ciphersuites as described in RFC3268, "Advanced
3947 Encryption Standard (AES) Ciphersuites for Transport Layer
3948 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
3949 not enabled by default and were not part of the "ALL" ciphersuite
3950 alias because they were not yet official; they could be
3951 explicitly requested by specifying the "AESdraft" ciphersuite
3952 group alias. In the final release of OpenSSL 0.9.7, the group
3953 alias is called "AES" and is part of "ALL".)
3954 [Ben Laurie, Steve Henson, Bodo Moeller]
3956 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
3957 request to response.
3960 *) Functions for OCSP responders. OCSP_request_onereq_count(),
3961 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
3962 extract information from a certificate request. OCSP_response_create()
3963 creates a response and optionally adds a basic response structure.
3964 OCSP_basic_add1_status() adds a complete single response to a basic
3965 response and returns the OCSP_SINGLERESP structure just added (to allow
3966 extensions to be included for example). OCSP_basic_add1_cert() adds a
3967 certificate to a basic response and OCSP_basic_sign() signs a basic
3968 response with various flags. New helper functions ASN1_TIME_check()
3969 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
3970 (converts ASN1_TIME to GeneralizedTime).
3973 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
3974 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
3975 structure from a certificate. X509_pubkey_digest() digests the public_key
3976 contents: this is used in various key identifiers.
3979 *) Make sk_sort() tolerate a NULL argument.
3980 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
3982 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
3983 passed by the function are trusted implicitly. If any of them signed the
3984 response then it is assumed to be valid and is not verified.
3987 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
3988 to data. This was previously part of the PKCS7 ASN1 code. This
3989 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
3990 [Steve Henson, reported by Kenneth R. Robinette
3991 <support@securenetterm.com>]
3993 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
3994 routines: without these tracing memory leaks is very painful.
3995 Fix leaks in PKCS12 and PKCS7 routines.
3998 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
3999 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
4000 effectively meant GeneralizedTime would never be used. Now it
4001 is initialised to -1 but X509_time_adj() now has to check the value
4002 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
4003 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
4004 [Steve Henson, reported by Kenneth R. Robinette
4005 <support@securenetterm.com>]
4007 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
4008 result in a zero length in the ASN1_INTEGER structure which was
4009 not consistent with the structure when d2i_ASN1_INTEGER() was used
4010 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
4011 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
4012 where it did not print out a minus for negative ASN1_INTEGER.
4015 *) Add summary printout to ocsp utility. The various functions which
4016 convert status values to strings have been renamed to:
4017 OCSP_response_status_str(), OCSP_cert_status_str() and
4018 OCSP_crl_reason_str() and are no longer static. New options
4019 to verify nonce values and to disable verification. OCSP response
4020 printout format cleaned up.
4023 *) Add additional OCSP certificate checks. These are those specified
4024 in RFC2560. This consists of two separate checks: the CA of the
4025 certificate being checked must either be the OCSP signer certificate
4026 or the issuer of the OCSP signer certificate. In the latter case the
4027 OCSP signer certificate must contain the OCSP signing extended key
4028 usage. This check is performed by attempting to match the OCSP
4029 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
4030 in the OCSP_CERTID structures of the response.
4033 *) Initial OCSP certificate verification added to OCSP_basic_verify()
4034 and related routines. This uses the standard OpenSSL certificate
4035 verify routines to perform initial checks (just CA validity) and
4036 to obtain the certificate chain. Then additional checks will be
4037 performed on the chain. Currently the root CA is checked to see
4038 if it is explicitly trusted for OCSP signing. This is used to set
4039 a root CA as a global signing root: that is any certificate that
4040 chains to that CA is an acceptable OCSP signing certificate.
4043 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
4044 extensions from a separate configuration file.
4045 As when reading extensions from the main configuration file,
4046 the '-extensions ...' option may be used for specifying the
4048 [Massimiliano Pala <madwolf@comune.modena.it>]
4050 *) New OCSP utility. Allows OCSP requests to be generated or
4051 read. The request can be sent to a responder and the output
4052 parsed, outputed or printed in text form. Not complete yet:
4053 still needs to check the OCSP response validity.
4056 *) New subcommands for 'openssl ca':
4057 'openssl ca -status <serial>' prints the status of the cert with
4058 the given serial number (according to the index file).
4059 'openssl ca -updatedb' updates the expiry status of certificates
4061 [Massimiliano Pala <madwolf@comune.modena.it>]
4063 *) New '-newreq-nodes' command option to CA.pl. This is like
4064 '-newreq', but calls 'openssl req' with the '-nodes' option
4065 so that the resulting key is not encrypted.
4066 [Damien Miller <djm@mindrot.org>]
4068 *) New configuration for the GNU Hurd.
4069 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
4071 *) Initial code to implement OCSP basic response verify. This
4072 is currently incomplete. Currently just finds the signer's
4073 certificate and verifies the signature on the response.
4076 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
4077 value of OPENSSLDIR. This is available via the new '-d' option
4078 to 'openssl version', and is also included in 'openssl version -a'.
4081 *) Allowing defining memory allocation callbacks that will be given
4082 file name and line number information in additional arguments
4083 (a const char* and an int). The basic functionality remains, as
4084 well as the original possibility to just replace malloc(),
4085 realloc() and free() by functions that do not know about these
4086 additional arguments. To register and find out the current
4087 settings for extended allocation functions, the following
4088 functions are provided:
4090 CRYPTO_set_mem_ex_functions
4091 CRYPTO_set_locked_mem_ex_functions
4092 CRYPTO_get_mem_ex_functions
4093 CRYPTO_get_locked_mem_ex_functions
4095 These work the same way as CRYPTO_set_mem_functions and friends.
4096 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
4097 extended allocation function is enabled.
4098 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
4099 a conventional allocation function is enabled.
4100 [Richard Levitte, Bodo Moeller]
4102 *) Finish off removing the remaining LHASH function pointer casts.
4103 There should no longer be any prototype-casting required when using
4104 the LHASH abstraction, and any casts that remain are "bugs". See
4105 the callback types and macros at the head of lhash.h for details
4106 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
4109 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
4110 If /dev/[u]random devices are not available or do not return enough
4111 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
4113 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
4114 /etc/entropy will be queried once each in this sequence, quering stops
4115 when enough entropy was collected without querying more sockets.
4118 *) Change the Unix RAND_poll() variant to be able to poll several
4119 random devices, as specified by DEVRANDOM, until a sufficient amount
4120 of data has been collected. We spend at most 10 ms on each file
4121 (select timeout) and read in non-blocking mode. DEVRANDOM now
4122 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
4123 (previously it was just the string "/dev/urandom"), so on typical
4124 platforms the 10 ms delay will never occur.
4125 Also separate out the Unix variant to its own file, rand_unix.c.
4126 For VMS, there's a currently-empty rand_vms.c.
4129 *) Move OCSP client related routines to ocsp_cl.c. These
4130 provide utility functions which an application needing
4131 to issue a request to an OCSP responder and analyse the
4132 response will typically need: as opposed to those which an
4133 OCSP responder itself would need which will be added later.
4135 OCSP_request_sign() signs an OCSP request with an API similar
4136 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
4137 response. OCSP_response_get1_basic() extracts basic response
4138 from response. OCSP_resp_find_status(): finds and extracts status
4139 information from an OCSP_CERTID structure (which will be created
4140 when the request structure is built). These are built from lower
4141 level functions which work on OCSP_SINGLERESP structures but
4142 wont normally be used unless the application wishes to examine
4143 extensions in the OCSP response for example.
4145 Replace nonce routines with a pair of functions.
4146 OCSP_request_add1_nonce() adds a nonce value and optionally
4147 generates a random value. OCSP_check_nonce() checks the
4148 validity of the nonce in an OCSP response.
4151 *) Change function OCSP_request_add() to OCSP_request_add0_id().
4152 This doesn't copy the supplied OCSP_CERTID and avoids the
4153 need to free up the newly created id. Change return type
4154 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
4155 This can then be used to add extensions to the request.
4156 Deleted OCSP_request_new(), since most of its functionality
4157 is now in OCSP_REQUEST_new() (and the case insensitive name
4158 clash) apart from the ability to set the request name which
4159 will be added elsewhere.
4162 *) Update OCSP API. Remove obsolete extensions argument from
4163 various functions. Extensions are now handled using the new
4164 OCSP extension code. New simple OCSP HTTP function which
4165 can be used to send requests and parse the response.
4168 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
4169 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
4170 uses the special reorder version of SET OF to sort the attributes
4171 and reorder them to match the encoded order. This resolves a long
4172 standing problem: a verify on a PKCS7 structure just after signing
4173 it used to fail because the attribute order did not match the
4174 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
4175 it uses the received order. This is necessary to tolerate some broken
4176 software that does not order SET OF. This is handled by encoding
4177 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
4178 to produce the required SET OF.
4181 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
4182 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
4183 files to get correct declarations of the ASN.1 item variables.
4186 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
4187 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
4188 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
4189 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
4190 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
4191 ASN1_ITEM and no wrapper functions.
4194 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
4195 replace the old function pointer based I/O routines. Change most of
4196 the *_d2i_bio() and *_d2i_fp() functions to use these.
4199 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
4200 lines, recognice more "algorithms" that can be deselected, and make
4201 it complain about algorithm deselection that isn't recognised.
4204 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
4205 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
4206 to use new functions. Add NO_ASN1_OLD which can be set to remove
4207 some old style ASN1 functions: this can be used to determine if old
4208 code will still work when these eventually go away.
4211 *) New extension functions for OCSP structures, these follow the
4212 same conventions as certificates and CRLs.
4215 *) New function X509V3_add1_i2d(). This automatically encodes and
4216 adds an extension. Its behaviour can be customised with various
4217 flags to append, replace or delete. Various wrappers added for
4218 certifcates and CRLs.
4221 *) Fix to avoid calling the underlying ASN1 print routine when
4222 an extension cannot be parsed. Correct a typo in the
4223 OCSP_SERVICELOC extension. Tidy up print OCSP format.
4226 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
4227 entries for variables.
4230 *) Add functionality to apps/openssl.c for detecting locking
4231 problems: As the program is single-threaded, all we have
4232 to do is register a locking callback using an array for
4233 storing which locks are currently held by the program.
4236 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
4237 SSL_get_ex_data_X509_STORE_idx(), which is used in
4238 ssl_verify_cert_chain() and thus can be called at any time
4239 during TLS/SSL handshakes so that thread-safety is essential.
4240 Unfortunately, the ex_data design is not at all suited
4241 for multi-threaded use, so it probably should be abolished.
4244 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
4245 [Broadcom, tweaked and integrated by Geoff Thorpe]
4247 *) Move common extension printing code to new function
4248 X509V3_print_extensions(). Reorganise OCSP print routines and
4249 implement some needed OCSP ASN1 functions. Add OCSP extensions.
4252 *) New function X509_signature_print() to remove duplication in some
4256 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
4257 set (this was treated exactly the same as SET OF previously). This
4258 is used to reorder the STACK representing the structure to match the
4259 encoding. This will be used to get round a problem where a PKCS7
4260 structure which was signed could not be verified because the STACK
4261 order did not reflect the encoded order.
4264 *) Reimplement the OCSP ASN1 module using the new code.
4267 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
4268 for its ASN1 operations. The old style function pointers still exist
4269 for now but they will eventually go away.
4272 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
4273 completely replaces the old ASN1 functionality with a table driven
4274 encoder and decoder which interprets an ASN1_ITEM structure describing
4275 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
4276 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
4277 has also been converted to the new form.
4280 *) Change BN_mod_exp_recp so that negative moduli are tolerated
4281 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
4282 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
4283 for negative moduli.
4286 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
4287 of not touching the result's sign bit.
4290 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
4294 *) Changed the LHASH code to use prototypes for callbacks, and created
4295 macros to declare and implement thin (optionally static) functions
4296 that provide type-safety and avoid function pointer casting for the
4297 type-specific callbacks.
4300 *) Added Kerberos Cipher Suites to be used with TLS, as written in
4302 [Veers Staats <staatsvr@asc.hpc.mil>,
4303 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
4305 *) Reformat the FAQ so the different questions and answers can be divided
4306 in sections depending on the subject.
4309 *) Have the zlib compression code load ZLIB.DLL dynamically under
4313 *) New function BN_mod_sqrt for computing square roots modulo a prime
4314 (using the probabilistic Tonelli-Shanks algorithm unless
4315 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
4316 be handled deterministically).
4317 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
4319 *) Make BN_mod_inverse faster by explicitly handling small quotients
4320 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
4321 512 bits], about 30% for larger ones [1024 or 2048 bits].)
4324 *) New function BN_kronecker.
4327 *) Fix BN_gcd so that it works on negative inputs; the result is
4328 positive unless both parameters are zero.
4329 Previously something reasonably close to an infinite loop was
4330 possible because numbers could be growing instead of shrinking
4331 in the implementation of Euclid's algorithm.
4334 *) Fix BN_is_word() and BN_is_one() macros to take into account the
4335 sign of the number in question.
4337 Fix BN_is_word(a,w) to work correctly for w == 0.
4339 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
4340 because its test if the absolute value of 'a' equals 'w'.
4341 Note that BN_abs_is_word does *not* handle w == 0 reliably;
4342 it exists mostly for use in the implementations of BN_is_zero(),
4343 BN_is_one(), and BN_is_word().
4346 *) New function BN_swap.
4349 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
4350 the exponentiation functions are more likely to produce reasonable
4351 results on negative inputs.
4354 *) Change BN_mod_mul so that the result is always non-negative.
4355 Previously, it could be negative if one of the factors was negative;
4356 I don't think anyone really wanted that behaviour.
4359 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
4360 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
4361 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
4362 and add new functions:
4371 BN_mod_lshift1_quick
4375 These functions always generate non-negative results.
4377 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
4378 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
4380 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
4381 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
4382 be reduced modulo m.
4383 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
4386 The following entry accidentily appeared in the CHANGES file
4387 distributed with OpenSSL 0.9.7. The modifications described in
4388 it do *not* apply to OpenSSL 0.9.7.
4390 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4391 was actually never needed) and in BN_mul(). The removal in BN_mul()
4392 required a small change in bn_mul_part_recursive() and the addition
4393 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4394 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4395 bn_sub_words() and bn_add_words() except they take arrays with
4400 *) In 'openssl passwd', verify passwords read from the terminal
4401 unless the '-salt' option is used (which usually means that
4402 verification would just waste user's time since the resulting
4403 hash is going to be compared with some given password hash)
4404 or the new '-noverify' option is used.
4406 This is an incompatible change, but it does not affect
4407 non-interactive use of 'openssl passwd' (passwords on the command
4408 line, '-stdin' option, '-in ...' option) and thus should not
4412 *) Remove all references to RSAref, since there's no more need for it.
4415 *) Make DSO load along a path given through an environment variable
4416 (SHLIB_PATH) with shl_load().
4419 *) Constify the ENGINE code as a result of BIGNUM constification.
4420 Also constify the RSA code and most things related to it. In a
4421 few places, most notable in the depth of the ASN.1 code, ugly
4422 casts back to non-const were required (to be solved at a later
4426 *) Make it so the openssl application has all engines loaded by default.
4429 *) Constify the BIGNUM routines a little more.
4432 *) Add the following functions:
4434 ENGINE_load_cswift()
4436 ENGINE_load_atalla()
4438 ENGINE_load_builtin_engines()
4440 That way, an application can itself choose if external engines that
4441 are built-in in OpenSSL shall ever be used or not. The benefit is
4442 that applications won't have to be linked with libdl or other dso
4443 libraries unless it's really needed.
4445 Changed 'openssl engine' to load all engines on demand.
4446 Changed the engine header files to avoid the duplication of some
4447 declarations (they differed!).
4450 *) 'openssl engine' can now list capabilities.
4453 *) Better error reporting in 'openssl engine'.
4456 *) Never call load_dh_param(NULL) in s_server.
4459 *) Add engine application. It can currently list engines by name and
4460 identity, and test if they are actually available.
4463 *) Improve RPM specification file by forcing symbolic linking and making
4464 sure the installed documentation is also owned by root.root.
4465 [Damien Miller <djm@mindrot.org>]
4467 *) Give the OpenSSL applications more possibilities to make use of
4468 keys (public as well as private) handled by engines.
4471 *) Add OCSP code that comes from CertCo.
4474 *) Add VMS support for the Rijndael code.
4477 *) Added untested support for Nuron crypto accelerator.
4480 *) Add support for external cryptographic devices. This code was
4481 previously distributed separately as the "engine" branch.
4482 [Geoff Thorpe, Richard Levitte]
4484 *) Rework the filename-translation in the DSO code. It is now possible to
4485 have far greater control over how a "name" is turned into a filename
4486 depending on the operating environment and any oddities about the
4487 different shared library filenames on each system.
4490 *) Support threads on FreeBSD-elf in Configure.
4493 *) Fix for SHA1 assembly problem with MASM: it produces
4494 warnings about corrupt line number information when assembling
4495 with debugging information. This is caused by the overlapping
4497 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
4500 NCONF_get_number() has no error checking at all. As a replacement,
4501 NCONF_get_number_e() is defined (_e for "error checking") and is
4502 promoted strongly. The old NCONF_get_number is kept around for
4503 binary backward compatibility.
4504 Make it possible for methods to load from something other than a BIO,
4505 by providing a function pointer that is given a name instead of a BIO.
4506 For example, this could be used to load configuration data from an
4510 *) Fix for non blocking accept BIOs. Added new I/O special reason
4511 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
4512 with non blocking I/O was not possible because no retry code was
4513 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
4517 *) Added the beginnings of Rijndael support.
4520 *) Fix for bug in DirectoryString mask setting. Add support for
4521 X509_NAME_print_ex() in 'req' and X509_print_ex() function
4522 to allow certificate printing to more controllable, additional
4523 'certopt' option to 'x509' to allow new printing options to be
4527 *) Clean old EAY MD5 hack from e_os.h.
4530 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
4532 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4533 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4534 [Joe Orton, Steve Henson]
4536 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
4538 *) Fix additional bug revealed by the NISCC test suite:
4540 Stop bug triggering large recursion when presented with
4541 certain ASN.1 tags (CVE-2003-0851)
4544 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
4546 *) Fix various bugs revealed by running the NISCC test suite:
4548 Stop out of bounds reads in the ASN1 code when presented with
4549 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4551 If verify callback ignores invalid public key errors don't try to check
4552 certificate signature with the NULL public key.
4556 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4557 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4561 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4562 extra data after the compression methods not only for TLS 1.0
4563 but also for SSL 3.0 (as required by the specification).
4564 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4566 *) Change X509_certificate_type() to mark the key as exported/exportable
4567 when it's 512 *bits* long, not 512 bytes.
4570 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
4572 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4573 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4574 a protocol version number mismatch like a decryption error
4575 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4578 *) Turn on RSA blinding by default in the default implementation
4579 to avoid a timing attack. Applications that don't want it can call
4580 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4581 They would be ill-advised to do so in most cases.
4582 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4584 *) Change RSA blinding code so that it works when the PRNG is not
4585 seeded (in this case, the secret RSA exponent is abused as
4586 an unpredictable seed -- if it is not unpredictable, there
4587 is no point in blinding anyway). Make RSA blinding thread-safe
4588 by remembering the creator's thread ID in rsa->blinding and
4589 having all other threads use local one-time blinding factors
4590 (this requires more computation than sharing rsa->blinding, but
4591 avoids excessive locking; and if an RSA object is not shared
4592 between threads, blinding will still be very fast).
4595 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
4597 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4598 via timing by performing a MAC computation even if incorrrect
4599 block cipher padding has been found. This is a countermeasure
4600 against active attacks where the attacker has to distinguish
4601 between bad padding and a MAC verification error. (CVE-2003-0078)
4603 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4604 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4605 Martin Vuagnoux (EPFL, Ilion)]
4607 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
4609 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
4610 memory from it's contents. This is done with a counter that will
4611 place alternating values in each byte. This can be used to solve
4612 two issues: 1) the removal of calls to memset() by highly optimizing
4613 compilers, and 2) cleansing with other values than 0, since those can
4614 be read through on certain media, for example a swap space on disk.
4617 *) Bugfix: client side session caching did not work with external caching,
4618 because the session->cipher setting was not restored when reloading
4619 from the external cache. This problem was masked, when
4620 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
4621 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
4624 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
4625 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
4626 [Zeev Lieber <zeev-l@yahoo.com>]
4628 *) Undo an undocumented change introduced in 0.9.6e which caused
4629 repeated calls to OpenSSL_add_all_ciphers() and
4630 OpenSSL_add_all_digests() to be ignored, even after calling
4634 *) Change the default configuration reader to deal with last line not
4635 being properly terminated.
4638 *) Change X509_NAME_cmp() so it applies the special rules on handling
4639 DN values that are of type PrintableString, as well as RDNs of type
4640 emailAddress where the value has the type ia5String.
4641 [stefank@valicert.com via Richard Levitte]
4643 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
4644 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
4645 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
4646 the bitwise-OR of the two for use by the majority of applications
4647 wanting this behaviour, and update the docs. The documented
4648 behaviour and actual behaviour were inconsistent and had been
4649 changing anyway, so this is more a bug-fix than a behavioural
4651 [Geoff Thorpe, diagnosed by Nadav Har'El]
4653 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
4654 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
4657 *) Fix initialization code race conditions in
4658 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
4659 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
4660 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
4661 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
4662 ssl2_get_cipher_by_char(),
4663 ssl3_get_cipher_by_char().
4664 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
4666 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
4667 the cached sessions are flushed, as the remove_cb() might use ex_data
4668 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
4669 (see [openssl.org #212]).
4670 [Geoff Thorpe, Lutz Jaenicke]
4672 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
4673 length, instead of the encoding length to d2i_ASN1_OBJECT.
4676 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
4678 *) [In 0.9.6g-engine release:]
4679 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
4680 [Lynn Gazis <lgazis@rainbow.com>]
4682 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
4684 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
4685 and get fix the header length calculation.
4686 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
4687 Alon Kantor <alonk@checkpoint.com> (and others),
4690 *) Use proper error handling instead of 'assertions' in buffer
4691 overflow checks added in 0.9.6e. This prevents DoS (the
4692 assertions could call abort()).
4693 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
4695 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
4697 *) Add various sanity checks to asn1_get_length() to reject
4698 the ASN1 length bytes if they exceed sizeof(long), will appear
4699 negative or the content length exceeds the length of the
4701 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4703 *) Fix cipher selection routines: ciphers without encryption had no flags
4704 for the cipher strength set and where therefore not handled correctly
4705 by the selection routines (PR #130).
4708 *) Fix EVP_dsa_sha macro.
4712 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
4713 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
4714 that was added in OpenSSL 0.9.6d.
4716 As the countermeasure turned out to be incompatible with some
4717 broken SSL implementations, the new option is part of SSL_OP_ALL.
4718 SSL_OP_ALL is usually employed when compatibility with weird SSL
4719 implementations is desired (e.g. '-bugs' option to 's_client' and
4720 's_server'), so the new option is automatically set in many
4724 *) Changes in security patch:
4726 Changes marked "(CHATS)" were sponsored by the Defense Advanced
4727 Research Projects Agency (DARPA) and Air Force Research Laboratory,
4728 Air Force Materiel Command, USAF, under agreement number
4731 *) Add various sanity checks to asn1_get_length() to reject
4732 the ASN1 length bytes if they exceed sizeof(long), will appear
4733 negative or the content length exceeds the length of the
4734 supplied buffer. (CVE-2002-0659)
4735 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4737 *) Assertions for various potential buffer overflows, not known to
4739 [Ben Laurie (CHATS)]
4741 *) Various temporary buffers to hold ASCII versions of integers were
4742 too small for 64 bit platforms. (CVE-2002-0655)
4743 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
4745 *) Remote buffer overflow in SSL3 protocol - an attacker could
4746 supply an oversized session ID to a client. (CVE-2002-0656)
4747 [Ben Laurie (CHATS)]
4749 *) Remote buffer overflow in SSL2 protocol - an attacker could
4750 supply an oversized client master key. (CVE-2002-0656)
4751 [Ben Laurie (CHATS)]
4753 Changes between 0.9.6c and 0.9.6d [9 May 2002]
4755 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
4756 encoded as NULL) with id-dsa-with-sha1.
4757 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
4759 *) Check various X509_...() return values in apps/req.c.
4760 [Nils Larsch <nla@trustcenter.de>]
4762 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
4763 an end-of-file condition would erronously be flagged, when the CRLF
4764 was just at the end of a processed block. The bug was discovered when
4765 processing data through a buffering memory BIO handing the data to a
4766 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
4767 <ptsekov@syntrex.com> and Nedelcho Stanev.
4770 *) Implement a countermeasure against a vulnerability recently found
4771 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
4772 before application data chunks to avoid the use of known IVs
4773 with data potentially chosen by the attacker.
4776 *) Fix length checks in ssl3_get_client_hello().
4779 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
4780 to prevent ssl3_read_internal() from incorrectly assuming that
4781 ssl3_read_bytes() found application data while handshake
4782 processing was enabled when in fact s->s3->in_read_app_data was
4783 merely automatically cleared during the initial handshake.
4784 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
4786 *) Fix object definitions for Private and Enterprise: they were not
4787 recognized in their shortname (=lowercase) representation. Extend
4788 obj_dat.pl to issue an error when using undefined keywords instead
4789 of silently ignoring the problem (Svenning Sorensen
4790 <sss@sss.dnsalias.net>).
4793 *) Fix DH_generate_parameters() so that it works for 'non-standard'
4794 generators, i.e. generators other than 2 and 5. (Previously, the
4795 code did not properly initialise the 'add' and 'rem' values to
4796 BN_generate_prime().)
4798 In the new general case, we do not insist that 'generator' is
4799 actually a primitive root: This requirement is rather pointless;
4800 a generator of the order-q subgroup is just as good, if not
4804 *) Map new X509 verification errors to alerts. Discovered and submitted by
4805 Tom Wu <tom@arcot.com>.
4808 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
4809 returning non-zero before the data has been completely received
4810 when using non-blocking I/O.
4811 [Bodo Moeller; problem pointed out by John Hughes]
4813 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
4814 [Ben Laurie, Lutz Jaenicke]
4816 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
4817 Yoram Zahavi <YoramZ@gilian.com>).
4820 *) Add information about CygWin 1.3 and on, and preserve proper
4821 configuration for the versions before that.
4822 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4824 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
4825 check whether we deal with a copy of a session and do not delete from
4826 the cache in this case. Problem reported by "Izhar Shoshani Levi"
4827 <izhar@checkpoint.com>.
4830 *) Do not store session data into the internal session cache, if it
4831 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
4832 flag is set). Proposed by Aslam <aslam@funk.com>.
4835 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
4839 *) [In 0.9.6d-engine release:]
4840 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4841 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4843 *) Add the configuration target linux-s390x.
4844 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
4846 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
4847 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
4848 variable as an indication that a ClientHello message has been
4849 received. As the flag value will be lost between multiple
4850 invocations of ssl3_accept when using non-blocking I/O, the
4851 function may not be aware that a handshake has actually taken
4852 place, thus preventing a new session from being added to the
4855 To avoid this problem, we now set s->new_session to 2 instead of
4856 using a local variable.
4857 [Lutz Jaenicke, Bodo Moeller]
4859 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
4860 if the SSL_R_LENGTH_MISMATCH error is detected.
4861 [Geoff Thorpe, Bodo Moeller]
4863 *) New 'shared_ldflag' column in Configure platform table.
4866 *) Fix EVP_CIPHER_mode macro.
4867 ["Dan S. Camper" <dan@bti.net>]
4869 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
4870 type, we must throw them away by setting rr->length to 0.
4871 [D P Chang <dpc@qualys.com>]
4873 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
4875 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
4876 <Dominikus.Scherkl@biodata.com>. (The previous implementation
4877 worked incorrectly for those cases where range = 10..._2 and
4878 3*range is two bits longer than range.)
4881 *) Only add signing time to PKCS7 structures if it is not already
4885 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
4886 OBJ_ld_ce should be OBJ_id_ce.
4887 Also some ip-pda OIDs in crypto/objects/objects.txt were
4888 incorrect (cf. RFC 3039).
4889 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
4891 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
4892 returns early because it has nothing to do.
4893 [Andy Schneider <andy.schneider@bjss.co.uk>]
4895 *) [In 0.9.6c-engine release:]
4896 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
4897 [Andy Schneider <andy.schneider@bjss.co.uk>]
4899 *) [In 0.9.6c-engine release:]
4900 Add support for Cryptographic Appliance's keyserver technology.
4901 (Use engine 'keyclient')
4902 [Cryptographic Appliances and Geoff Thorpe]
4904 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
4905 is called via tools/c89.sh because arguments have to be
4906 rearranged (all '-L' options must appear before the first object
4908 [Richard Shapiro <rshapiro@abinitio.com>]
4910 *) [In 0.9.6c-engine release:]
4911 Add support for Broadcom crypto accelerator cards, backported
4913 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
4915 *) [In 0.9.6c-engine release:]
4916 Add support for SureWare crypto accelerator cards from
4917 Baltimore Technologies. (Use engine 'sureware')
4918 [Baltimore Technologies and Mark Cox]
4920 *) [In 0.9.6c-engine release:]
4921 Add support for crypto accelerator cards from Accelerated
4922 Encryption Processing, www.aep.ie. (Use engine 'aep')
4923 [AEP Inc. and Mark Cox]
4925 *) Add a configuration entry for gcc on UnixWare.
4926 [Gary Benson <gbenson@redhat.com>]
4928 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
4929 messages are stored in a single piece (fixed-length part and
4930 variable-length part combined) and fix various bugs found on the way.
4933 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
4934 instead. BIO_gethostbyname() does not know what timeouts are
4935 appropriate, so entries would stay in cache even when they have
4937 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
4939 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
4940 faced with a pathologically small ClientHello fragment that does
4941 not contain client_version: Instead of aborting with an error,
4942 simply choose the highest available protocol version (i.e.,
4943 TLS 1.0 unless it is disabled). In practice, ClientHello
4944 messages are never sent like this, but this change gives us
4945 strictly correct behaviour at least for TLS.
4948 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
4949 never resets s->method to s->ctx->method when called from within
4950 one of the SSL handshake functions.
4951 [Bodo Moeller; problem pointed out by Niko Baric]
4953 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
4954 (sent using the client's version number) if client_version is
4955 smaller than the protocol version in use. Also change
4956 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
4957 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
4958 the client will at least see that alert.
4961 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
4965 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
4966 client receives HelloRequest while in a handshake.
4967 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
4969 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
4970 should end in 'break', not 'goto end' which circuments various
4971 cleanups done in state SSL_ST_OK. But session related stuff
4972 must be disabled for SSL_ST_OK in the case that we just sent a
4975 Also avoid some overhead by not calling ssl_init_wbio_buffer()
4976 before just sending a HelloRequest.
4977 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
4979 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
4980 reveal whether illegal block cipher padding was found or a MAC
4981 verification error occured. (Neither SSLerr() codes nor alerts
4982 are directly visible to potential attackers, but the information
4983 may leak via logfiles.)
4985 Similar changes are not required for the SSL 2.0 implementation
4986 because the number of padding bytes is sent in clear for SSL 2.0,
4987 and the extra bytes are just ignored. However ssl/s2_pkt.c
4988 failed to verify that the purported number of padding bytes is in
4992 *) Add OpenUNIX-8 support including shared libraries
4993 (Boyd Lynn Gerber <gerberb@zenez.com>).
4996 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
4997 'wristwatch attack' using huge encoding parameters (cf.
4998 James H. Manger's CRYPTO 2001 paper). Note that the
4999 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
5000 encoding parameters and hence was not vulnerable.
5003 *) BN_sqr() bug fix.
5004 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
5006 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
5007 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
5008 followed by modular reduction.
5009 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
5011 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
5012 equivalent based on BN_pseudo_rand() instead of BN_rand().
5015 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
5016 This function was broken, as the check for a new client hello message
5017 to handle SGC did not allow these large messages.
5018 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
5021 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
5024 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
5025 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
5028 *) Rework the configuration and shared library support for Tru64 Unix.
5029 The configuration part makes use of modern compiler features and
5030 still retains old compiler behavior for those that run older versions
5031 of the OS. The shared library support part includes a variant that
5032 uses the RPATH feature, and is available through the special
5033 configuration target "alpha-cc-rpath", which will never be selected
5035 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
5037 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
5038 with the same message size as in ssl3_get_certificate_request().
5039 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
5040 messages might inadvertently be reject as too long.
5041 [Petr Lampa <lampa@fee.vutbr.cz>]
5043 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
5046 *) Modified SSL library such that the verify_callback that has been set
5047 specificly for an SSL object with SSL_set_verify() is actually being
5048 used. Before the change, a verify_callback set with this function was
5049 ignored and the verify_callback() set in the SSL_CTX at the time of
5050 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
5051 to allow the necessary settings.
5054 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
5055 explicitly to NULL, as at least on Solaris 8 this seems not always to be
5056 done automatically (in contradiction to the requirements of the C
5057 standard). This made problems when used from OpenSSH.
5060 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
5061 dh->length and always used
5063 BN_rand_range(priv_key, dh->p).
5065 BN_rand_range() is not necessary for Diffie-Hellman, and this
5066 specific range makes Diffie-Hellman unnecessarily inefficient if
5067 dh->length (recommended exponent length) is much smaller than the
5068 length of dh->p. We could use BN_rand_range() if the order of
5069 the subgroup was stored in the DH structure, but we only have
5074 BN_rand(priv_key, l, ...)
5076 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
5082 RSA_eay_public_encrypt
5083 RSA_eay_private_decrypt
5084 RSA_eay_private_encrypt (signing)
5085 RSA_eay_public_decrypt (signature verification)
5087 (default implementations for RSA_public_encrypt,
5088 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
5089 always reject numbers >= n.
5092 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
5093 to synchronize access to 'locking_thread'. This is necessary on
5094 systems where access to 'locking_thread' (an 'unsigned long'
5095 variable) is not atomic.
5098 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
5099 *before* setting the 'crypto_lock_rand' flag. The previous code had
5100 a race condition if 0 is a valid thread ID.
5101 [Travis Vitek <vitek@roguewave.com>]
5103 *) Add support for shared libraries under Irix.
5104 [Albert Chin-A-Young <china@thewrittenword.com>]
5106 *) Add configuration option to build on Linux on both big-endian and
5108 [Ralf Baechle <ralf@uni-koblenz.de>]
5110 *) Add the possibility to create shared libraries on HP-UX.
5113 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
5115 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
5116 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
5117 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
5118 PRNG state recovery was possible based on the output of
5119 one PRNG request appropriately sized to gain knowledge on
5120 'md' followed by enough consecutive 1-byte PRNG requests
5121 to traverse all of 'state'.
5123 1. When updating 'md_local' (the current thread's copy of 'md')
5124 during PRNG output generation, hash all of the previous
5125 'md_local' value, not just the half used for PRNG output.
5127 2. Make the number of bytes from 'state' included into the hash
5128 independent from the number of PRNG bytes requested.
5130 The first measure alone would be sufficient to avoid
5131 Markku-Juhani's attack. (Actually it had never occurred
5132 to me that the half of 'md_local' used for chaining was the
5133 half from which PRNG output bytes were taken -- I had always
5134 assumed that the secret half would be used.) The second
5135 measure makes sure that additional data from 'state' is never
5136 mixed into 'md_local' in small portions; this heuristically
5137 further strengthens the PRNG.
5140 *) Fix crypto/bn/asm/mips3.s.
5143 *) When only the key is given to "enc", the IV is undefined. Print out
5144 an error message in this case.
5147 *) Handle special case when X509_NAME is empty in X509 printing routines.
5150 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
5151 positive and less than q.
5154 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
5155 used: it isn't thread safe and the add_lock_callback should handle
5157 [Paul Rose <Paul.Rose@bridge.com>]
5159 *) Verify that incoming data obeys the block size in
5160 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
5164 [Ulf Möller, Bodo Möller]
5166 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
5167 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
5168 when fixing the server behaviour for backwards-compatible 'client
5169 hello' messages. (Note that the attack is impractical against
5170 SSL 3.0 and TLS 1.0 anyway because length and version checking
5171 means that the probability of guessing a valid ciphertext is
5172 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
5175 Before 0.9.5, the countermeasure (hide the error by generating a
5176 random 'decryption result') did not work properly because
5177 ERR_clear_error() was missing, meaning that SSL_get_error() would
5178 detect the supposedly ignored error.
5180 Both problems are now fixed.
5183 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
5184 (previously it was 1024).
5187 *) Fix for compatibility mode trust settings: ignore trust settings
5188 unless some valid trust or reject settings are present.
5191 *) Fix for blowfish EVP: its a variable length cipher.
5194 *) Fix various bugs related to DSA S/MIME verification. Handle missing
5195 parameters in DSA public key structures and return an error in the
5196 DSA routines if parameters are absent.
5199 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
5200 in the current directory if neither $RANDFILE nor $HOME was set.
5201 RAND_file_name() in 0.9.6a returned NULL in this case. This has
5202 caused some confusion to Windows users who haven't defined $HOME.
5203 Thus RAND_file_name() is changed again: e_os.h can define a
5204 DEFAULT_HOME, which will be used if $HOME is not set.
5205 For Windows, we use "C:"; on other platforms, we still require
5206 environment variables.
5208 *) Move 'if (!initialized) RAND_poll()' into regions protected by
5209 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
5210 having multiple threads call RAND_poll() concurrently.
5213 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
5214 combination of a flag and a thread ID variable.
5215 Otherwise while one thread is in ssleay_rand_bytes (which sets the
5216 flag), *other* threads can enter ssleay_add_bytes without obeying
5217 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
5218 that they do not hold after the first thread unsets add_do_not_lock).
5221 *) Change bctest again: '-x' expressions are not available in all
5225 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
5227 *) Fix a couple of memory leaks in PKCS7_dataDecode()
5228 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
5230 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
5231 the default extension for executables, if any. Also, make the perl
5232 scripts that use symlink() to test if it really exists and use "cp"
5233 if it doesn't. All this made OpenSSL compilable and installable in
5237 *) Fix for asn1_GetSequence() for indefinite length constructed data.
5238 If SEQUENCE is length is indefinite just set c->slen to the total
5239 amount of data available.
5240 [Steve Henson, reported by shige@FreeBSD.org]
5241 [This change does not apply to 0.9.7.]
5243 *) Change bctest to avoid here-documents inside command substitution
5244 (workaround for FreeBSD /bin/sh bug).
5245 For compatibility with Ultrix, avoid shell functions (introduced
5246 in the bctest version that searches along $PATH).
5249 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
5250 with des_encrypt() defined on some operating systems, like Solaris
5254 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
5255 On the Importance of Eliminating Errors in Cryptographic
5256 Computations, J. Cryptology 14 (2001) 2, 101-119,
5257 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
5260 *) MIPS assembler BIGNUM division bug fix.
5263 *) Disabled incorrect Alpha assembler code.
5266 *) Fix PKCS#7 decode routines so they correctly update the length
5267 after reading an EOC for the EXPLICIT tag.
5269 [This change does not apply to 0.9.7.]
5271 *) Fix bug in PKCS#12 key generation routines. This was triggered
5272 if a 3DES key was generated with a 0 initial byte. Include
5273 PKCS12_BROKEN_KEYGEN compilation option to retain the old
5274 (but broken) behaviour.
5277 *) Enhance bctest to search for a working bc along $PATH and print
5279 [Tim Rice <tim@multitalents.net> via Richard Levitte]
5281 *) Fix memory leaks in err.c: free err_data string if necessary;
5282 don't write to the wrong index in ERR_set_error_data.
5285 *) Implement ssl23_peek (analogous to ssl23_read), which previously
5289 *) Replace rdtsc with _emit statements for VC++ version 5.
5290 [Jeremy Cooper <jeremy@baymoo.org>]
5292 *) Make it possible to reuse SSLv2 sessions.
5295 *) In copy_email() check for >= 0 as a return value for
5296 X509_NAME_get_index_by_NID() since 0 is a valid index.
5297 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
5299 *) Avoid coredump with unsupported or invalid public keys by checking if
5300 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
5301 PKCS7_verify() fails with non detached data.
5304 *) Don't use getenv in library functions when run as setuid/setgid.
5305 New function OPENSSL_issetugid().
5308 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
5309 due to incorrect handling of multi-threading:
5311 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
5313 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
5315 3. Count how many times MemCheck_off() has been called so that
5316 nested use can be treated correctly. This also avoids
5317 inband-signalling in the previous code (which relied on the
5318 assumption that thread ID 0 is impossible).
5321 *) Add "-rand" option also to s_client and s_server.
5324 *) Fix CPU detection on Irix 6.x.
5325 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
5326 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
5328 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
5331 [This change does not apply to 0.9.7.]
5333 *) Use the cached encoding of an X509_NAME structure rather than
5334 copying it. This is apparently the reason for the libsafe "errors"
5335 but the code is actually correct.
5338 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
5339 Bleichenbacher's DSA attack.
5340 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
5341 to be set and top=0 forces the highest bit to be set; top=-1 is new
5342 and leaves the highest bit random.
5343 [Ulf Moeller, Bodo Moeller]
5345 *) In the NCONF_...-based implementations for CONF_... queries
5346 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
5347 a temporary CONF structure with the data component set to NULL
5348 (which gives segmentation faults in lh_retrieve).
5349 Instead, use NULL for the CONF pointer in CONF_get_string and
5350 CONF_get_number (which may use environment variables) and directly
5351 return NULL from CONF_get_section.
5354 *) Fix potential buffer overrun for EBCDIC.
5357 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
5358 keyUsage if basicConstraints absent for a CA.
5361 *) Make SMIME_write_PKCS7() write mail header values with a format that
5362 is more generally accepted (no spaces before the semicolon), since
5363 some programs can't parse those values properly otherwise. Also make
5364 sure BIO's that break lines after each write do not create invalid
5368 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
5369 macros previously used would not encode an empty SEQUENCE OF
5370 and break the signature.
5372 [This change does not apply to 0.9.7.]
5374 *) Zero the premaster secret after deriving the master secret in
5378 *) Add some EVP_add_digest_alias registrations (as found in
5379 OpenSSL_add_all_digests()) to SSL_library_init()
5380 aka OpenSSL_add_ssl_algorithms(). This provides improved
5381 compatibility with peers using X.509 certificates
5382 with unconventional AlgorithmIdentifier OIDs.
5385 *) Fix for Irix with NO_ASM.
5386 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
5388 *) ./config script fixes.
5389 [Ulf Moeller, Richard Levitte]
5391 *) Fix 'openssl passwd -1'.
5394 *) Change PKCS12_key_gen_asc() so it can cope with non null
5395 terminated strings whose length is passed in the passlen
5396 parameter, for example from PEM callbacks. This was done
5397 by adding an extra length parameter to asc2uni().
5398 [Steve Henson, reported by <oddissey@samsung.co.kr>]
5400 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
5401 call failed, free the DSA structure.
5404 *) Fix to uni2asc() to cope with zero length Unicode strings.
5405 These are present in some PKCS#12 files.
5408 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
5409 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
5410 when writing a 32767 byte record.
5411 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
5413 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
5414 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
5416 (RSA objects have a reference count access to which is protected
5417 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
5418 so they are meant to be shared between threads.)
5419 [Bodo Moeller, Geoff Thorpe; original patch submitted by
5420 "Reddie, Steven" <Steven.Reddie@ca.com>]
5422 *) Fix a deadlock in CRYPTO_mem_leaks().
5425 *) Use better test patterns in bntest.
5428 *) rand_win.c fix for Borland C.
5431 *) BN_rshift bugfix for n == 0.
5434 *) Add a 'bctest' script that checks for some known 'bc' bugs
5435 so that 'make test' does not abort just because 'bc' is broken.
5438 *) Store verify_result within SSL_SESSION also for client side to
5439 avoid potential security hole. (Re-used sessions on the client side
5440 always resulted in verify_result==X509_V_OK, not using the original
5441 result of the server certificate verification.)
5444 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
5445 SSL3_RT_APPLICATION_DATA, return 0.
5446 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
5450 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
5451 releases, have been re-implemented by renaming the previous
5452 implementations of ssl2_read and ssl3_read to ssl2_read_internal
5453 and ssl3_read_internal, respectively, and adding 'peek' parameters
5454 to them. The new ssl[23]_{read,peek} functions are calls to
5455 ssl[23]_read_internal with the 'peek' flag set appropriately.
5456 A 'peek' parameter has also been added to ssl3_read_bytes, which
5457 does the actual work for ssl3_read_internal.
5460 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
5461 the method-specific "init()" handler. Also clean up ex_data after
5462 calling the method-specific "finish()" handler. Previously, this was
5463 happening the other way round.
5466 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
5467 The previous value, 12, was not always sufficient for BN_mod_exp().
5470 *) Make sure that shared libraries get the internal name engine with
5471 the full version number and not just 0. This should mark the
5472 shared libraries as not backward compatible. Of course, this should
5473 be changed again when we can guarantee backward binary compatibility.
5476 *) Fix typo in get_cert_by_subject() in by_dir.c
5477 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
5479 *) Rework the system to generate shared libraries:
5481 - Make note of the expected extension for the shared libraries and
5482 if there is a need for symbolic links from for example libcrypto.so.0
5483 to libcrypto.so.0.9.7. There is extended info in Configure for
5486 - Make as few rebuilds of the shared libraries as possible.
5488 - Still avoid linking the OpenSSL programs with the shared libraries.
5490 - When installing, install the shared libraries separately from the
5494 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
5496 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
5497 and not in SSL_clear because the latter is also used by the
5498 accept/connect functions; previously, the settings made by
5499 SSL_set_read_ahead would be lost during the handshake.
5500 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
5502 *) Correct util/mkdef.pl to be selective about disabled algorithms.
5503 Previously, it would create entries for disableed algorithms no
5507 *) Added several new manual pages for SSL_* function.
5510 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
5512 *) In ssl23_get_client_hello, generate an error message when faced
5513 with an initial SSL 3.0/TLS record that is too small to contain the
5514 first two bytes of the ClientHello message, i.e. client_version.
5515 (Note that this is a pathologic case that probably has never happened
5516 in real life.) The previous approach was to use the version number
5517 from the record header as a substitute; but our protocol choice
5518 should not depend on that one because it is not authenticated
5519 by the Finished messages.
5522 *) More robust randomness gathering functions for Windows.
5523 [Jeffrey Altman <jaltman@columbia.edu>]
5525 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
5526 not set then we don't setup the error code for issuer check errors
5527 to avoid possibly overwriting other errors which the callback does
5528 handle. If an application does set the flag then we assume it knows
5529 what it is doing and can handle the new informational codes
5533 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
5534 a general "ANY" type, as such it should be able to decode anything
5535 including tagged types. However it didn't check the class so it would
5536 wrongly interpret tagged types in the same way as their universal
5537 counterpart and unknown types were just rejected. Changed so that the
5538 tagged and unknown types are handled in the same way as a SEQUENCE:
5539 that is the encoding is stored intact. There is also a new type
5540 "V_ASN1_OTHER" which is used when the class is not universal, in this
5541 case we have no idea what the actual type is so we just lump them all
5545 *) On VMS, stdout may very well lead to a file that is written to
5546 in a record-oriented fashion. That means that every write() will
5547 write a separate record, which will be read separately by the
5548 programs trying to read from it. This can be very confusing.
5550 The solution is to put a BIO filter in the way that will buffer
5551 text until a linefeed is reached, and then write everything a
5552 line at a time, so every record written will be an actual line,
5553 not chunks of lines and not (usually doesn't happen, but I've
5554 seen it once) several lines in one record. BIO_f_linebuffer() is
5557 Currently, it's a VMS-only method, because that's where it has
5558 been tested well enough.
5561 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
5562 it can return incorrect results.
5563 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
5564 but it was in 0.9.6-beta[12].)
5567 *) Disable the check for content being present when verifying detached
5568 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
5569 include zero length content when signing messages.
5572 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
5573 BIO_ctrl (for BIO pairs).
5576 *) Add DSO method for VMS.
5579 *) Bug fix: Montgomery multiplication could produce results with the
5583 *) Add RPM specification openssl.spec and modify it to build three
5584 packages. The default package contains applications, application
5585 documentation and run-time libraries. The devel package contains
5586 include files, static libraries and function documentation. The
5587 doc package contains the contents of the doc directory. The original
5588 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
5591 *) Add a large number of documentation files for many SSL routines.
5592 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
5594 *) Add a configuration entry for Sony News 4.
5595 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
5597 *) Don't set the two most significant bits to one when generating a
5598 random number < q in the DSA library.
5601 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
5602 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
5603 the underlying transport is blocking) if a handshake took place.
5604 (The default behaviour is needed by applications such as s_client
5605 and s_server that use select() to determine when to use SSL_read;
5606 but for applications that know in advance when to expect data, it
5607 just makes things more complicated.)
5610 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
5614 *) Add a few more EBCDIC conditionals that make `req' and `x509'
5615 work better on such systems.
5616 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5618 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
5619 Update PKCS12_parse() so it copies the friendlyName and the
5620 keyid to the certificates aux info.
5623 *) Fix bug in PKCS7_verify() which caused an infinite loop
5624 if there was more than one signature.
5625 [Sven Uszpelkat <su@celocom.de>]
5627 *) Major change in util/mkdef.pl to include extra information
5628 about each symbol, as well as presentig variables as well
5629 as functions. This change means that there's n more need
5630 to rebuild the .num files when some algorithms are excluded.
5633 *) Allow the verify time to be set by an application,
5634 rather than always using the current time.
5637 *) Phase 2 verify code reorganisation. The certificate
5638 verify code now looks up an issuer certificate by a
5639 number of criteria: subject name, authority key id
5640 and key usage. It also verifies self signed certificates
5641 by the same criteria. The main comparison function is
5642 X509_check_issued() which performs these checks.
5644 Lot of changes were necessary in order to support this
5645 without completely rewriting the lookup code.
5647 Authority and subject key identifier are now cached.
5649 The LHASH 'certs' is X509_STORE has now been replaced
5650 by a STACK_OF(X509_OBJECT). This is mainly because an
5651 LHASH can't store or retrieve multiple objects with
5652 the same hash value.
5654 As a result various functions (which were all internal
5655 use only) have changed to handle the new X509_STORE
5656 structure. This will break anything that messed round
5657 with X509_STORE internally.
5659 The functions X509_STORE_add_cert() now checks for an
5660 exact match, rather than just subject name.
5662 The X509_STORE API doesn't directly support the retrieval
5663 of multiple certificates matching a given criteria, however
5664 this can be worked round by performing a lookup first
5665 (which will fill the cache with candidate certificates)
5666 and then examining the cache for matches. This is probably
5667 the best we can do without throwing out X509_LOOKUP
5668 entirely (maybe later...).
5670 The X509_VERIFY_CTX structure has been enhanced considerably.
5672 All certificate lookup operations now go via a get_issuer()
5673 callback. Although this currently uses an X509_STORE it
5674 can be replaced by custom lookups. This is a simple way
5675 to bypass the X509_STORE hackery necessary to make this
5676 work and makes it possible to use more efficient techniques
5677 in future. A very simple version which uses a simple
5678 STACK for its trusted certificate store is also provided
5679 using X509_STORE_CTX_trusted_stack().
5681 The verify_cb() and verify() callbacks now have equivalents
5682 in the X509_STORE_CTX structure.
5684 X509_STORE_CTX also has a 'flags' field which can be used
5685 to customise the verify behaviour.
5688 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
5689 excludes S/MIME capabilities.
5692 *) When a certificate request is read in keep a copy of the
5693 original encoding of the signed data and use it when outputing
5694 again. Signatures then use the original encoding rather than
5695 a decoded, encoded version which may cause problems if the
5696 request is improperly encoded.
5699 *) For consistency with other BIO_puts implementations, call
5700 buffer_write(b, ...) directly in buffer_puts instead of calling
5703 In BIO_puts, increment b->num_write as in BIO_write.
5704 [Peter.Sylvester@EdelWeb.fr]
5706 *) Fix BN_mul_word for the case where the word is 0. (We have to use
5707 BN_zero, we may not return a BIGNUM with an array consisting of
5711 *) Avoid calling abort() from within the library when problems are
5712 detected, except if preprocessor symbols have been defined
5713 (such as REF_CHECK, BN_DEBUG etc.).
5716 *) New openssl application 'rsautl'. This utility can be
5717 used for low level RSA operations. DER public key
5718 BIO/fp routines also added.
5721 *) New Configure entry and patches for compiling on QNX 4.
5722 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
5724 *) A demo state-machine implementation was sponsored by
5725 Nuron (http://www.nuron.com/) and is now available in
5726 demos/state_machine.
5729 *) New options added to the 'dgst' utility for signature
5730 generation and verification.
5733 *) Unrecognized PKCS#7 content types are now handled via a
5734 catch all ASN1_TYPE structure. This allows unsupported
5735 types to be stored as a "blob" and an application can
5736 encode and decode it manually.
5739 *) Fix various signed/unsigned issues to make a_strex.c
5741 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
5743 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
5744 length if passed a buffer. ASN1_INTEGER_to_BN failed
5745 if passed a NULL BN and its argument was negative.
5746 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
5748 *) Modification to PKCS#7 encoding routines to output definite
5749 length encoding. Since currently the whole structures are in
5750 memory there's not real point in using indefinite length
5751 constructed encoding. However if OpenSSL is compiled with
5752 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
5755 *) Added BIO_vprintf() and BIO_vsnprintf().
5758 *) Added more prefixes to parse for in the the strings written
5759 through a logging bio, to cover all the levels that are available
5760 through syslog. The prefixes are now:
5762 PANIC, EMERG, EMR => LOG_EMERG
5763 ALERT, ALR => LOG_ALERT
5764 CRIT, CRI => LOG_CRIT
5765 ERROR, ERR => LOG_ERR
5766 WARNING, WARN, WAR => LOG_WARNING
5767 NOTICE, NOTE, NOT => LOG_NOTICE
5768 INFO, INF => LOG_INFO
5769 DEBUG, DBG => LOG_DEBUG
5771 and as before, if none of those prefixes are present at the
5772 beginning of the string, LOG_ERR is chosen.
5774 On Win32, the LOG_* levels are mapped according to this:
5776 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
5777 LOG_WARNING => EVENTLOG_WARNING_TYPE
5778 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
5782 *) Made it possible to reconfigure with just the configuration
5783 argument "reconf" or "reconfigure". The command line arguments
5784 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
5785 and are retrieved from there when reconfiguring.
5789 [Assar Westerlund <assar@sics.se>, Richard Levitte]
5791 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
5794 *) The obj_dat.pl script was messing up the sorting of object
5795 names. The reason was that it compared the quoted version
5796 of strings as a result "OCSP" > "OCSP Signing" because
5797 " > SPACE. Changed script to store unquoted versions of
5798 names and add quotes on output. It was also omitting some
5799 names from the lookup table if they were given a default
5800 value (that is if SN is missing it is given the same
5801 value as LN and vice versa), these are now added on the
5802 grounds that if an object has a name we should be able to
5803 look it up. Finally added warning output when duplicate
5804 short or long names are found.
5807 *) Changes needed for Tandem NSK.
5808 [Scott Uroff <scott@xypro.com>]
5810 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
5811 RSA_padding_check_SSLv23(), special padding was never detected
5812 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
5813 version rollback attacks was not effective.
5815 In s23_clnt.c, don't use special rollback-attack detection padding
5816 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
5817 client; similarly, in s23_srvr.c, don't do the rollback check if
5818 SSL 2.0 is the only protocol enabled in the server.
5821 *) Make it possible to get hexdumps of unprintable data with 'openssl
5822 asn1parse'. By implication, the functions ASN1_parse_dump() and
5823 BIO_dump_indent() are added.
5826 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
5827 these print out strings and name structures based on various
5828 flags including RFC2253 support and proper handling of
5829 multibyte characters. Added options to the 'x509' utility
5830 to allow the various flags to be set.
5833 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
5834 Also change the functions X509_cmp_current_time() and
5835 X509_gmtime_adj() work with an ASN1_TIME structure,
5836 this will enable certificates using GeneralizedTime in validity
5837 dates to be checked.
5840 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
5841 negative public key encodings) on by default,
5842 NO_NEG_PUBKEY_BUG can be set to disable it.
5845 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
5846 content octets. An i2c_ASN1_OBJECT is unnecessary because
5847 the encoding can be trivially obtained from the structure.
5850 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
5851 not read locks (CRYPTO_r_[un]lock).
5854 *) A first attempt at creating official support for shared
5855 libraries through configuration. I've kept it so the
5856 default is static libraries only, and the OpenSSL programs
5857 are always statically linked for now, but there are
5858 preparations for dynamic linking in place.
5859 This has been tested on Linux and Tru64.
5862 *) Randomness polling function for Win9x, as described in:
5863 Peter Gutmann, Software Generation of Practically Strong
5867 *) Fix so PRNG is seeded in req if using an already existing
5871 *) New options to smime application. -inform and -outform
5872 allow alternative formats for the S/MIME message including
5873 PEM and DER. The -content option allows the content to be
5874 specified separately. This should allow things like Netscape
5875 form signing output easier to verify.
5878 *) Fix the ASN1 encoding of tags using the 'long form'.
5881 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
5882 STRING types. These convert content octets to and from the
5883 underlying type. The actual tag and length octets are
5884 already assumed to have been read in and checked. These
5885 are needed because all other string types have virtually
5886 identical handling apart from the tag. By having versions
5887 of the ASN1 functions that just operate on content octets
5888 IMPLICIT tagging can be handled properly. It also allows
5889 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
5890 and ASN1_INTEGER are identical apart from the tag.
5893 *) Change the handling of OID objects as follows:
5895 - New object identifiers are inserted in objects.txt, following
5896 the syntax given in objects.README.
5897 - objects.pl is used to process obj_mac.num and create a new
5899 - obj_dat.pl is used to create a new obj_dat.h, using the data in
5902 This is currently kind of a hack, and the perl code in objects.pl
5903 isn't very elegant, but it works as I intended. The simplest way
5904 to check that it worked correctly is to look in obj_dat.h and
5905 check the array nid_objs and make sure the objects haven't moved
5906 around (this is important!). Additions are OK, as well as
5907 consistent name changes.
5910 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
5913 *) Addition of the command line parameter '-rand file' to 'openssl req'.
5914 The given file adds to whatever has already been seeded into the
5915 random pool through the RANDFILE configuration file option or
5916 environment variable, or the default random state file.
5919 *) mkstack.pl now sorts each macro group into lexical order.
5920 Previously the output order depended on the order the files
5921 appeared in the directory, resulting in needless rewriting
5925 *) Patches to make OpenSSL compile under Win32 again. Mostly
5926 work arounds for the VC++ problem that it treats func() as
5927 func(void). Also stripped out the parts of mkdef.pl that
5928 added extra typesafe functions: these no longer exist.
5931 *) Reorganisation of the stack code. The macros are now all
5932 collected in safestack.h . Each macro is defined in terms of
5933 a "stack macro" of the form SKM_<name>(type, a, b). The
5934 DEBUG_SAFESTACK is now handled in terms of function casts,
5935 this has the advantage of retaining type safety without the
5936 use of additional functions. If DEBUG_SAFESTACK is not defined
5937 then the non typesafe macros are used instead. Also modified the
5938 mkstack.pl script to handle the new form. Needs testing to see
5939 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
5940 the default if no major problems. Similar behaviour for ASN1_SET_OF
5941 and PKCS12_STACK_OF.
5944 *) When some versions of IIS use the 'NET' form of private key the
5945 key derivation algorithm is different. Normally MD5(password) is
5946 used as a 128 bit RC4 key. In the modified case
5947 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
5948 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
5949 as the old Netscape_RSA functions except they have an additional
5950 'sgckey' parameter which uses the modified algorithm. Also added
5951 an -sgckey command line option to the rsa utility. Thanks to
5952 Adrian Peck <bertie@ncipher.com> for posting details of the modified
5953 algorithm to openssl-dev.
5956 *) The evp_local.h macros were using 'c.##kname' which resulted in
5957 invalid expansion on some systems (SCO 5.0.5 for example).
5958 Corrected to 'c.kname'.
5959 [Phillip Porch <root@theporch.com>]
5961 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
5962 a STACK of email addresses from a certificate or request, these look
5963 in the subject name and the subject alternative name extensions and
5964 omit any duplicate addresses.
5967 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
5968 This makes DSA verification about 2 % faster.
5971 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
5972 (meaning that now 2^5 values will be precomputed, which is only 4 KB
5973 plus overhead for 1024 bit moduli).
5974 This makes exponentiations about 0.5 % faster for 1024 bit
5975 exponents (as measured by "openssl speed rsa2048").
5978 *) Rename memory handling macros to avoid conflicts with other
5980 Malloc => OPENSSL_malloc
5981 Malloc_locked => OPENSSL_malloc_locked
5982 Realloc => OPENSSL_realloc
5983 Free => OPENSSL_free
5986 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
5987 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
5990 *) CygWin32 support.
5991 [John Jarvie <jjarvie@newsguy.com>]
5993 *) The type-safe stack code has been rejigged. It is now only compiled
5994 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
5995 by default all type-specific stack functions are "#define"d back to
5996 standard stack functions. This results in more streamlined output
5997 but retains the type-safety checking possibilities of the original
6001 *) The STACK code has been cleaned up, and certain type declarations
6002 that didn't make a lot of sense have been brought in line. This has
6003 also involved a cleanup of sorts in safestack.h to more correctly
6004 map type-safe stack functions onto their plain stack counterparts.
6005 This work has also resulted in a variety of "const"ifications of
6006 lots of the code, especially "_cmp" operations which should normally
6007 be prototyped with "const" parameters anyway.
6010 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
6011 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
6012 (The PRNG state consists of two parts, the large pool 'state' and 'md',
6013 where all of 'md' is used each time the PRNG is used, but 'state'
6014 is used only indexed by a cyclic counter. As entropy may not be
6015 well distributed from the beginning, 'md' is important as a
6016 chaining variable. However, the output function chains only half
6017 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
6018 all of 'md', and seeding with STATE_SIZE dummy bytes will result
6019 in all of 'state' being rewritten, with the new values depending
6020 on virtually all of 'md'. This overcomes the 80 bit limitation.)
6023 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
6024 the handshake is continued after ssl_verify_cert_chain();
6025 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
6026 can lead to 'unexplainable' connection aborts later.
6027 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
6029 *) Major EVP API cipher revision.
6030 Add hooks for extra EVP features. This allows various cipher
6031 parameters to be set in the EVP interface. Support added for variable
6032 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
6033 setting of RC2 and RC5 parameters.
6035 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
6038 Remove lots of duplicated code from the EVP library. For example *every*
6039 cipher init() function handles the 'iv' in the same way according to the
6040 cipher mode. They also all do nothing if the 'key' parameter is NULL and
6041 for CFB and OFB modes they zero ctx->num.
6043 New functionality allows removal of S/MIME code RC2 hack.
6045 Most of the routines have the same form and so can be declared in terms
6048 By shifting this to the top level EVP_CipherInit() it can be removed from
6049 all individual ciphers. If the cipher wants to handle IVs or keys
6050 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
6053 Change lots of functions like EVP_EncryptUpdate() to now return a
6054 value: although software versions of the algorithms cannot fail
6055 any installed hardware versions can.
6058 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
6059 this option is set, tolerate broken clients that send the negotiated
6060 protocol version number instead of the requested protocol version
6064 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
6065 i.e. non-zero for export ciphersuites, zero otherwise.
6066 Previous versions had this flag inverted, inconsistent with
6067 rsa_tmp_cb (..._TMP_RSA_CB).
6068 [Bodo Moeller; problem reported by Amit Chopra]
6070 *) Add missing DSA library text string. Work around for some IIS
6071 key files with invalid SEQUENCE encoding.
6074 *) Add a document (doc/standards.txt) that list all kinds of standards
6075 and so on that are implemented in OpenSSL.
6078 *) Enhance c_rehash script. Old version would mishandle certificates
6079 with the same subject name hash and wouldn't handle CRLs at all.
6080 Added -fingerprint option to crl utility, to support new c_rehash
6084 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
6087 *) Fix for SSL server purpose checking. Server checking was
6088 rejecting certificates which had extended key usage present
6089 but no ssl client purpose.
6090 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
6092 *) Make PKCS#12 code work with no password. The PKCS#12 spec
6093 is a little unclear about how a blank password is handled.
6094 Since the password in encoded as a BMPString with terminating
6095 double NULL a zero length password would end up as just the
6096 double NULL. However no password at all is different and is
6097 handled differently in the PKCS#12 key generation code. NS
6098 treats a blank password as zero length. MSIE treats it as no
6099 password on export: but it will try both on import. We now do
6100 the same: PKCS12_parse() tries zero length and no password if
6101 the password is set to "" or NULL (NULL is now a valid password:
6102 it wasn't before) as does the pkcs12 application.
6105 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
6106 perror when PEM_read_bio_X509_REQ fails, the error message must
6107 be obtained from the error queue.
6110 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
6111 it in ERR_remove_state if appropriate, and change ERR_get_state
6112 accordingly to avoid race conditions (this is necessary because
6113 thread_hash is no longer constant once set).
6116 *) Bugfix for linux-elf makefile.one.
6119 *) RSA_get_default_method() will now cause a default
6120 RSA_METHOD to be chosen if one doesn't exist already.
6121 Previously this was only set during a call to RSA_new()
6122 or RSA_new_method(NULL) meaning it was possible for
6123 RSA_get_default_method() to return NULL.
6126 *) Added native name translation to the existing DSO code
6127 that will convert (if the flag to do so is set) filenames
6128 that are sufficiently small and have no path information
6129 into a canonical native form. Eg. "blah" converted to
6130 "libblah.so" or "blah.dll" etc.
6133 *) New function ERR_error_string_n(e, buf, len) which is like
6134 ERR_error_string(e, buf), but writes at most 'len' bytes
6135 including the 0 terminator. For ERR_error_string_n, 'buf'
6137 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
6139 *) CONF library reworked to become more general. A new CONF
6140 configuration file reader "class" is implemented as well as a
6141 new functions (NCONF_*, for "New CONF") to handle it. The now
6142 old CONF_* functions are still there, but are reimplemented to
6143 work in terms of the new functions. Also, a set of functions
6144 to handle the internal storage of the configuration data is
6145 provided to make it easier to write new configuration file
6146 reader "classes" (I can definitely see something reading a
6147 configuration file in XML format, for example), called _CONF_*,
6148 or "the configuration storage API"...
6150 The new configuration file reading functions are:
6152 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
6153 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
6155 NCONF_default, NCONF_WIN32
6157 NCONF_dump_fp, NCONF_dump_bio
6159 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
6160 NCONF_new creates a new CONF object. This works in the same way
6161 as other interfaces in OpenSSL, like the BIO interface.
6162 NCONF_dump_* dump the internal storage of the configuration file,
6163 which is useful for debugging. All other functions take the same
6164 arguments as the old CONF_* functions wth the exception of the
6165 first that must be a `CONF *' instead of a `LHASH *'.
6167 To make it easer to use the new classes with the old CONF_* functions,
6168 the function CONF_set_default_method is provided.
6171 *) Add '-tls1' option to 'openssl ciphers', which was already
6172 mentioned in the documentation but had not been implemented.
6173 (This option is not yet really useful because even the additional
6174 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
6177 *) Initial DSO code added into libcrypto for letting OpenSSL (and
6178 OpenSSL-based applications) load shared libraries and bind to
6179 them in a portable way.
6180 [Geoff Thorpe, with contributions from Richard Levitte]
6182 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
6184 *) Make sure _lrotl and _lrotr are only used with MSVC.
6186 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
6187 (the default implementation of RAND_status).
6189 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
6190 to '-clrext' (= clear extensions), as intended and documented.
6191 [Bodo Moeller; inconsistency pointed out by Michael Attili
6194 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
6195 was larger than the MD block size.
6196 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
6198 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
6199 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
6200 using the passed key: if the passed key was a private key the result
6201 of X509_print(), for example, would be to print out all the private key
6205 *) des_quad_cksum() byte order bug fix.
6206 [Ulf Möller, using the problem description in krb4-0.9.7, where
6207 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
6209 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
6211 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
6213 *) For easily testing in shell scripts whether some command
6214 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
6215 returns with exit code 0 iff no command of the given name is available.
6216 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
6217 the output goes to stdout and nothing is printed to stderr.
6218 Additional arguments are always ignored.
6220 Since for each cipher there is a command of the same name,
6221 the 'no-cipher' compilation switches can be tested this way.
6223 ('openssl no-XXX' is not able to detect pseudo-commands such
6224 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
6227 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
6230 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
6231 is set; it will be thrown away anyway because each handshake creates
6233 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
6234 to parameters -- in previous versions (since OpenSSL 0.9.3) the
6235 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
6236 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
6239 *) New s_client option -ign_eof: EOF at stdin is ignored, and
6240 'Q' and 'R' lose their special meanings (quit/renegotiate).
6241 This is part of what -quiet does; unlike -quiet, -ign_eof
6242 does not suppress any output.
6245 *) Add compatibility options to the purpose and trust code. The
6246 purpose X509_PURPOSE_ANY is "any purpose" which automatically
6247 accepts a certificate or CA, this was the previous behaviour,
6248 with all the associated security issues.
6250 X509_TRUST_COMPAT is the old trust behaviour: only and
6251 automatically trust self signed roots in certificate store. A
6252 new trust setting X509_TRUST_DEFAULT is used to specify that
6253 a purpose has no associated trust setting and it should instead
6254 use the value in the default purpose.
6257 *) Fix the PKCS#8 DSA private key code so it decodes keys again
6258 and fix a memory leak.
6261 *) In util/mkerr.pl (which implements 'make errors'), preserve
6262 reason strings from the previous version of the .c file, as
6263 the default to have only downcase letters (and digits) in
6264 automatically generated reasons codes is not always appropriate.
6267 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
6268 using strerror. Previously, ERR_reason_error_string() returned
6269 library names as reason strings for SYSerr; but SYSerr is a special
6270 case where small numbers are errno values, not library numbers.
6273 *) Add '-dsaparam' option to 'openssl dhparam' application. This
6274 converts DSA parameters into DH parameters. (When creating parameters,
6275 DSA_generate_parameters is used.)
6278 *) Include 'length' (recommended exponent length) in C code generated
6279 by 'openssl dhparam -C'.
6282 *) The second argument to set_label in perlasm was already being used
6283 so couldn't be used as a "file scope" flag. Moved to third argument
6287 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
6288 instead of RAND_bytes for encryption IVs and salts.
6291 *) Include RAND_status() into RAND_METHOD instead of implementing
6292 it only for md_rand.c Otherwise replacing the PRNG by calling
6293 RAND_set_rand_method would be impossible.
6296 *) Don't let DSA_generate_key() enter an infinite loop if the random
6297 number generation fails.
6300 *) New 'rand' application for creating pseudo-random output.
6303 *) Added configuration support for Linux/IA64
6304 [Rolf Haberrecker <rolf@suse.de>]
6306 *) Assembler module support for Mingw32.
6309 *) Shared library support for HPUX (in shlib/).
6310 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
6312 *) Shared library support for Solaris gcc.
6313 [Lutz Behnke <behnke@trustcenter.de>]
6315 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
6317 *) PKCS7_encrypt() was adding text MIME headers twice because they
6318 were added manually and by SMIME_crlf_copy().
6321 *) In bntest.c don't call BN_rand with zero bits argument.
6322 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
6324 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
6325 case was implemented. This caused BN_div_recp() to fail occasionally.
6328 *) Add an optional second argument to the set_label() in the perl
6329 assembly language builder. If this argument exists and is set
6330 to 1 it signals that the assembler should use a symbol whose
6331 scope is the entire file, not just the current function. This
6332 is needed with MASM which uses the format label:: for this scope.
6333 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
6335 *) Change the ASN1 types so they are typedefs by default. Before
6336 almost all types were #define'd to ASN1_STRING which was causing
6337 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
6341 *) Change names of new functions to the new get1/get0 naming
6342 convention: After 'get1', the caller owns a reference count
6343 and has to call ..._free; 'get0' returns a pointer to some
6344 data structure without incrementing reference counters.
6345 (Some of the existing 'get' functions increment a reference
6346 counter, some don't.)
6347 Similarly, 'set1' and 'add1' functions increase reference
6348 counters or duplicate objects.
6351 *) Allow for the possibility of temp RSA key generation failure:
6352 the code used to assume it always worked and crashed on failure.
6355 *) Fix potential buffer overrun problem in BIO_printf().
6356 [Ulf Möller, using public domain code by Patrick Powell; problem
6357 pointed out by David Sacerdote <das33@cornell.edu>]
6359 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
6360 RAND_egd() and RAND_status(). In the command line application,
6361 the EGD socket can be specified like a seed file using RANDFILE
6365 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
6366 Some CAs (e.g. Verisign) distribute certificates in this form.
6369 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
6370 list to exclude them. This means that no special compilation option
6371 is needed to use anonymous DH: it just needs to be included in the
6375 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
6376 EVP_MD_type. The old functionality is available in a new macro called
6377 EVP_MD_md(). Change code that uses it and update docs.
6380 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
6381 where the 'void *' argument is replaced by a function pointer argument.
6382 Previously 'void *' was abused to point to functions, which works on
6383 many platforms, but is not correct. As these functions are usually
6384 called by macros defined in OpenSSL header files, most source code
6385 should work without changes.
6388 *) <openssl/opensslconf.h> (which is created by Configure) now contains
6389 sections with information on -D... compiler switches used for
6390 compiling the library so that applications can see them. To enable
6391 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
6392 must be defined. E.g.,
6393 #define OPENSSL_ALGORITHM_DEFINES
6394 #include <openssl/opensslconf.h>
6395 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
6396 [Richard Levitte, Ulf and Bodo Möller]
6398 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
6402 *) Change the 'other' type in certificate aux info to a STACK_OF
6403 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
6404 the required ASN1 format: arbitrary types determined by an OID.
6407 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
6408 argument to 'req'. This is not because the function is newer or
6409 better than others it just uses the work 'NEW' in the certificate
6410 request header lines. Some software needs this.
6413 *) Reorganise password command line arguments: now passwords can be
6414 obtained from various sources. Delete the PEM_cb function and make
6415 it the default behaviour: i.e. if the callback is NULL and the
6416 usrdata argument is not NULL interpret it as a null terminated pass
6417 phrase. If usrdata and the callback are NULL then the pass phrase
6418 is prompted for as usual.
6421 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
6422 the support is automatically enabled. The resulting binaries will
6423 autodetect the card and use it if present.
6424 [Ben Laurie and Compaq Inc.]
6426 *) Work around for Netscape hang bug. This sends certificate request
6427 and server done in one record. Since this is perfectly legal in the
6428 SSL/TLS protocol it isn't a "bug" option and is on by default. See
6429 the bugs/SSLv3 entry for more info.
6432 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
6435 *) Add -rand argument to smime and pkcs12 applications and read/write
6439 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
6442 *) Add command line password options to the remaining applications.
6445 *) Bug fix for BN_div_recp() for numerators with an even number of
6449 *) More tests in bntest.c, and changed test_bn output.
6452 *) ./config recognizes MacOS X now.
6455 *) Bug fix for BN_div() when the first words of num and divsor are
6456 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
6459 *) Add support for various broken PKCS#8 formats, and command line
6460 options to produce them.
6463 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
6464 get temporary BIGNUMs from a BN_CTX.
6467 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
6471 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
6472 include a #define from the old name to the new. The original intent
6473 was that statically linked binaries could for example just call
6474 SSLeay_add_all_ciphers() to just add ciphers to the table and not
6475 link with digests. This never worked becayse SSLeay_add_all_digests()
6476 and SSLeay_add_all_ciphers() were in the same source file so calling
6477 one would link with the other. They are now in separate source files.
6480 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
6483 *) Use a less unusual form of the Miller-Rabin primality test (it used
6484 a binary algorithm for exponentiation integrated into the Miller-Rabin
6485 loop, our standard modexp algorithms are faster).
6488 *) Support for the EBCDIC character set completed.
6489 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
6491 *) Source code cleanups: use const where appropriate, eliminate casts,
6492 use void * instead of char * in lhash.
6495 *) Bugfix: ssl3_send_server_key_exchange was not restartable
6496 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
6497 this the server could overwrite ephemeral keys that the client
6501 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
6502 using 50 iterations of the Rabin-Miller test.
6504 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
6505 iterations of the Rabin-Miller test as required by the appendix
6506 to FIPS PUB 186[-1]) instead of DSA_is_prime.
6507 As BN_is_prime_fasttest includes trial division, DSA parameter
6508 generation becomes much faster.
6510 This implies a change for the callback functions in DSA_is_prime
6511 and DSA_generate_parameters: The callback function is called once
6512 for each positive witness in the Rabin-Miller test, not just
6513 occasionally in the inner loop; and the parameters to the
6514 callback function now provide an iteration count for the outer
6515 loop rather than for the current invocation of the inner loop.
6516 DSA_generate_parameters additionally can call the callback
6517 function with an 'iteration count' of -1, meaning that a
6518 candidate has passed the trial division test (when q is generated
6519 from an application-provided seed, trial division is skipped).
6522 *) New function BN_is_prime_fasttest that optionally does trial
6523 division before starting the Rabin-Miller test and has
6524 an additional BN_CTX * argument (whereas BN_is_prime always
6525 has to allocate at least one BN_CTX).
6526 'callback(1, -1, cb_arg)' is called when a number has passed the
6527 trial division stage.
6530 *) Fix for bug in CRL encoding. The validity dates weren't being handled
6534 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
6537 *) New function BN_pseudo_rand().
6540 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
6541 bignum version of BN_from_montgomery() with the working code from
6542 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
6546 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
6547 made it impossible to use the same SSL_SESSION data structure in
6548 SSL2 clients in multiple threads.
6551 *) The return value of RAND_load_file() no longer counts bytes obtained
6552 by stat(). RAND_load_file(..., -1) is new and uses the complete file
6553 to seed the PRNG (previously an explicit byte count was required).
6554 [Ulf Möller, Bodo Möller]
6556 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
6557 used (char *) instead of (void *) and had casts all over the place.
6560 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
6563 *) Retain source code compatibility for BN_prime_checks macro:
6564 BN_is_prime(..., BN_prime_checks, ...) now uses
6565 BN_prime_checks_for_size to determine the appropriate number of
6566 Rabin-Miller iterations.
6569 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
6570 DH_CHECK_P_NOT_SAFE_PRIME.
6571 (Check if this is true? OpenPGP calls them "strong".)
6574 *) Merge the functionality of "dh" and "gendh" programs into a new program
6575 "dhparam". The old programs are retained for now but will handle DH keys
6576 (instead of parameters) in future.
6579 *) Make the ciphers, s_server and s_client programs check the return values
6580 when a new cipher list is set.
6583 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
6584 ciphers. Before when the 56bit ciphers were enabled the sorting was
6587 The syntax for the cipher sorting has been extended to support sorting by
6588 cipher-strength (using the strength_bits hard coded in the tables).
6589 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
6591 Fix a bug in the cipher-command parser: when supplying a cipher command
6592 string with an "undefined" symbol (neither command nor alphanumeric
6593 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
6594 an error is flagged.
6596 Due to the strength-sorting extension, the code of the
6597 ssl_create_cipher_list() function was completely rearranged. I hope that
6598 the readability was also increased :-)
6599 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6601 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
6602 for the first serial number and places 2 in the serial number file. This
6603 avoids problems when the root CA is created with serial number zero and
6604 the first user certificate has the same issuer name and serial number
6608 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
6609 the new code. Add documentation for this stuff.
6612 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
6613 X509_*() to X509at_*() on the grounds that they don't handle X509
6614 structures and behave in an analagous way to the X509v3 functions:
6615 they shouldn't be called directly but wrapper functions should be used
6618 So we also now have some wrapper functions that call the X509at functions
6619 when passed certificate requests. (TO DO: similar things can be done with
6620 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
6621 things. Some of these need some d2i or i2d and print functionality
6622 because they handle more complex structures.)
6625 *) Add missing #ifndefs that caused missing symbols when building libssl
6626 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
6627 NO_RSA in ssl/s2*.c.
6628 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
6630 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
6631 has a return value which indicates the quality of the random data
6632 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
6633 error queue. New function RAND_pseudo_bytes() generates output that is
6634 guaranteed to be unique but not unpredictable. RAND_add is like
6635 RAND_seed, but takes an extra argument for an entropy estimate
6636 (RAND_seed always assumes full entropy).
6639 *) Do more iterations of Rabin-Miller probable prime test (specifically,
6640 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
6641 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
6642 in crypto/bn/bn_prime.c for the complete table). This guarantees a
6643 false-positive rate of at most 2^-80 for random input.
6646 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
6649 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
6650 in the 0.9.5 release), this returns the chain
6651 from an X509_CTX structure with a dup of the stack and all
6652 the X509 reference counts upped: so the stack will exist
6653 after X509_CTX_cleanup() has been called. Modify pkcs12.c
6656 Also make SSL_SESSION_print() print out the verify return
6660 *) Add manpage for the pkcs12 command. Also change the default
6661 behaviour so MAC iteration counts are used unless the new
6662 -nomaciter option is used. This improves file security and
6663 only older versions of MSIE (4.0 for example) need it.
6666 *) Honor the no-xxx Configure options when creating .DEF files.
6669 *) Add PKCS#10 attributes to field table: challengePassword,
6670 unstructuredName and unstructuredAddress. These are taken from
6671 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
6672 international characters are used.
6674 More changes to X509_ATTRIBUTE code: allow the setting of types
6675 based on strings. Remove the 'loc' parameter when adding
6676 attributes because these will be a SET OF encoding which is sorted
6680 *) Initial changes to the 'req' utility to allow request generation
6681 automation. This will allow an application to just generate a template
6682 file containing all the field values and have req construct the
6685 Initial support for X509_ATTRIBUTE handling. Stacks of these are
6686 used all over the place including certificate requests and PKCS#7
6687 structures. They are currently handled manually where necessary with
6688 some primitive wrappers for PKCS#7. The new functions behave in a
6689 manner analogous to the X509 extension functions: they allow
6690 attributes to be looked up by NID and added.
6692 Later something similar to the X509V3 code would be desirable to
6693 automatically handle the encoding, decoding and printing of the
6694 more complex types. The string types like challengePassword can
6695 be handled by the string table functions.
6697 Also modified the multi byte string table handling. Now there is
6698 a 'global mask' which masks out certain types. The table itself
6699 can use the flag STABLE_NO_MASK to ignore the mask setting: this
6700 is useful when for example there is only one permissible type
6701 (as in countryName) and using the mask might result in no valid
6705 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
6706 SSL_get_peer_finished to allow applications to obtain the latest
6707 Finished messages sent to the peer or expected from the peer,
6708 respectively. (SSL_get_peer_finished is usually the Finished message
6709 actually received from the peer, otherwise the protocol will be aborted.)
6711 As the Finished message are message digests of the complete handshake
6712 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
6713 be used for external authentication procedures when the authentication
6714 provided by SSL/TLS is not desired or is not enough.
6717 *) Enhanced support for Alpha Linux is added. Now ./config checks if
6718 the host supports BWX extension and if Compaq C is present on the
6719 $PATH. Just exploiting of the BWX extension results in 20-30%
6720 performance kick for some algorithms, e.g. DES and RC4 to mention
6721 a couple. Compaq C in turn generates ~20% faster code for MD5 and
6725 *) Add support for MS "fast SGC". This is arguably a violation of the
6726 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
6727 weak crypto and after checking the certificate is SGC a second one
6728 with strong crypto. MS SGC stops the first handshake after receiving
6729 the server certificate message and sends a second client hello. Since
6730 a server will typically do all the time consuming operations before
6731 expecting any further messages from the client (server key exchange
6732 is the most expensive) there is little difference between the two.
6734 To get OpenSSL to support MS SGC we have to permit a second client
6735 hello message after we have sent server done. In addition we have to
6736 reset the MAC if we do get this second client hello.
6739 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
6740 if a DER encoded private key is RSA or DSA traditional format. Changed
6741 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
6742 format DER encoded private key. Newer code should use PKCS#8 format which
6743 has the key type encoded in the ASN1 structure. Added DER private key
6744 support to pkcs8 application.
6747 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
6748 ciphersuites has been selected (as required by the SSL 3/TLS 1
6749 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
6750 is set, we interpret this as a request to violate the specification
6751 (the worst that can happen is a handshake failure, and 'correct'
6752 behaviour would result in a handshake failure anyway).
6755 *) In SSL_CTX_add_session, take into account that there might be multiple
6756 SSL_SESSION structures with the same session ID (e.g. when two threads
6757 concurrently obtain them from an external cache).
6758 The internal cache can handle only one SSL_SESSION with a given ID,
6759 so if there's a conflict, we now throw out the old one to achieve
6763 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
6764 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
6765 some routines that use cipher OIDs: some ciphers do not have OIDs
6766 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
6770 *) Simplify the trust setting structure and code. Now we just have
6771 two sequences of OIDs for trusted and rejected settings. These will
6772 typically have values the same as the extended key usage extension
6773 and any application specific purposes.
6775 The trust checking code now has a default behaviour: it will just
6776 check for an object with the same NID as the passed id. Functions can
6777 be provided to override either the default behaviour or the behaviour
6778 for a given id. SSL client, server and email already have functions
6779 in place for compatibility: they check the NID and also return "trusted"
6780 if the certificate is self signed.
6783 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
6784 traditional format into an EVP_PKEY structure.
6787 *) Add a password callback function PEM_cb() which either prompts for
6788 a password if usr_data is NULL or otherwise assumes it is a null
6789 terminated password. Allow passwords to be passed on command line
6790 environment or config files in a few more utilities.
6793 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
6794 keys. Add some short names for PKCS#8 PBE algorithms and allow them
6795 to be specified on the command line for the pkcs8 and pkcs12 utilities.
6796 Update documentation.
6799 *) Support for ASN1 "NULL" type. This could be handled before by using
6800 ASN1_TYPE but there wasn't any function that would try to read a NULL
6801 and produce an error if it couldn't. For compatibility we also have
6802 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
6803 don't allocate anything because they don't need to.
6806 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
6808 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
6810 *) Rebuild of the memory allocation routines used by OpenSSL code and
6811 possibly others as well. The purpose is to make an interface that
6812 provide hooks so anyone can build a separate set of allocation and
6813 deallocation routines to be used by OpenSSL, for example memory
6814 pool implementations, or something else, which was previously hard
6815 since Malloc(), Realloc() and Free() were defined as macros having
6816 the values malloc, realloc and free, respectively (except for Win32
6817 compilations). The same is provided for memory debugging code.
6818 OpenSSL already comes with functionality to find memory leaks, but
6819 this gives people a chance to debug other memory problems.
6821 With these changes, a new set of functions and macros have appeared:
6823 CRYPTO_set_mem_debug_functions() [F]
6824 CRYPTO_get_mem_debug_functions() [F]
6825 CRYPTO_dbg_set_options() [F]
6826 CRYPTO_dbg_get_options() [F]
6827 CRYPTO_malloc_debug_init() [M]
6829 The memory debug functions are NULL by default, unless the library
6830 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
6831 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
6832 gives the standard debugging functions that come with OpenSSL) or
6833 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
6834 provided by the library user) must be used. When the standard
6835 debugging functions are used, CRYPTO_dbg_set_options can be used to
6836 request additional information:
6837 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
6838 the CRYPTO_MDEBUG_xxx macro when compiling the library.
6840 Also, things like CRYPTO_set_mem_functions will always give the
6841 expected result (the new set of functions is used for allocation
6842 and deallocation) at all times, regardless of platform and compiler
6845 To finish it up, some functions that were never use in any other
6846 way than through macros have a new API and new semantic:
6849 CRYPTO_dbg_realloc()
6852 All macros of value have retained their old syntax.
6853 [Richard Levitte and Bodo Moeller]
6855 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
6856 ordering of SMIMECapabilities wasn't in "strength order" and there
6857 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
6861 *) Some ASN1 types with illegal zero length encoding (INTEGER,
6862 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
6863 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
6865 *) Merge in my S/MIME library for OpenSSL. This provides a simple
6866 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
6867 functionality to handle multipart/signed properly) and a utility
6868 called 'smime' to call all this stuff. This is based on code I
6869 originally wrote for Celo who have kindly allowed it to be
6870 included in OpenSSL.
6873 *) Add variants des_set_key_checked and des_set_key_unchecked of
6874 des_set_key (aka des_key_sched). Global variable des_check_key
6875 decides which of these is called by des_set_key; this way
6876 des_check_key behaves as it always did, but applications and
6877 the library itself, which was buggy for des_check_key == 1,
6878 have a cleaner way to pick the version they need.
6881 *) New function PKCS12_newpass() which changes the password of a
6885 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
6886 dynamic mix. In both cases the ids can be used as an index into the
6887 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
6888 functions so they accept a list of the field values and the
6889 application doesn't need to directly manipulate the X509_TRUST
6893 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
6897 *) Modify the way the V3 extension code looks up extensions. This now
6898 works in a similar way to the object code: we have some "standard"
6899 extensions in a static table which is searched with OBJ_bsearch()
6900 and the application can add dynamic ones if needed. The file
6901 crypto/x509v3/ext_dat.h now has the info: this file needs to be
6902 updated whenever a new extension is added to the core code and kept
6903 in ext_nid order. There is a simple program 'tabtest.c' which checks
6904 this. New extensions are not added too often so this file can readily
6905 be maintained manually.
6907 There are two big advantages in doing things this way. The extensions
6908 can be looked up immediately and no longer need to be "added" using
6909 X509V3_add_standard_extensions(): this function now does nothing.
6910 [Side note: I get *lots* of email saying the extension code doesn't
6911 work because people forget to call this function]
6912 Also no dynamic allocation is done unless new extensions are added:
6913 so if we don't add custom extensions there is no need to call
6914 X509V3_EXT_cleanup().
6917 *) Modify enc utility's salting as follows: make salting the default. Add a
6918 magic header, so unsalted files fail gracefully instead of just decrypting
6919 to garbage. This is because not salting is a big security hole, so people
6920 should be discouraged from doing it.
6923 *) Fixes and enhancements to the 'x509' utility. It allowed a message
6924 digest to be passed on the command line but it only used this
6925 parameter when signing a certificate. Modified so all relevant
6926 operations are affected by the digest parameter including the
6927 -fingerprint and -x509toreq options. Also -x509toreq choked if a
6928 DSA key was used because it didn't fix the digest.
6931 *) Initial certificate chain verify code. Currently tests the untrusted
6932 certificates for consistency with the verify purpose (which is set
6933 when the X509_STORE_CTX structure is set up) and checks the pathlength.
6935 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
6936 this is because it will reject chains with invalid extensions whereas
6937 every previous version of OpenSSL and SSLeay made no checks at all.
6939 Trust code: checks the root CA for the relevant trust settings. Trust
6940 settings have an initial value consistent with the verify purpose: e.g.
6941 if the verify purpose is for SSL client use it expects the CA to be
6942 trusted for SSL client use. However the default value can be changed to
6943 permit custom trust settings: one example of this would be to only trust
6944 certificates from a specific "secure" set of CAs.
6946 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
6947 which should be used for version portability: especially since the
6948 verify structure is likely to change more often now.
6950 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
6951 to set them. If not set then assume SSL clients will verify SSL servers
6954 Two new options to the verify program: -untrusted allows a set of
6955 untrusted certificates to be passed in and -purpose which sets the
6956 intended purpose of the certificate. If a purpose is set then the
6957 new chain verify code is used to check extension consistency.
6960 *) Support for the authority information access extension.
6963 *) Modify RSA and DSA PEM read routines to transparently handle
6964 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
6965 public keys in a format compatible with certificate
6966 SubjectPublicKeyInfo structures. Unfortunately there were already
6967 functions called *_PublicKey_* which used various odd formats so
6968 these are retained for compatibility: however the DSA variants were
6969 never in a public release so they have been deleted. Changed dsa/rsa
6970 utilities to handle the new format: note no releases ever handled public
6971 keys so we should be OK.
6973 The primary motivation for this change is to avoid the same fiasco
6974 that dogs private keys: there are several incompatible private key
6975 formats some of which are standard and some OpenSSL specific and
6976 require various evil hacks to allow partial transparent handling and
6977 even then it doesn't work with DER formats. Given the option anything
6978 other than PKCS#8 should be dumped: but the other formats have to
6979 stay in the name of compatibility.
6981 With public keys and the benefit of hindsight one standard format
6982 is used which works with EVP_PKEY, RSA or DSA structures: though
6983 it clearly returns an error if you try to read the wrong kind of key.
6985 Added a -pubkey option to the 'x509' utility to output the public key.
6986 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
6987 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
6988 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
6989 that do the same as the EVP_PKEY_assign_*() except they up the
6990 reference count of the added key (they don't "swallow" the
6994 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
6995 CRLs would fail if the file contained no certificates or no CRLs:
6996 added a new function to read in both types and return the number
6997 read: this means that if none are read it will be an error. The
6998 DER versions of the certificate and CRL reader would always fail
6999 because it isn't possible to mix certificates and CRLs in DER format
7000 without choking one or the other routine. Changed this to just read
7001 a certificate: this is the best we can do. Also modified the code
7002 in apps/verify.c to take notice of return codes: it was previously
7003 attempting to read in certificates from NULL pointers and ignoring
7004 any errors: this is one reason why the cert and CRL reader seemed
7005 to work. It doesn't check return codes from the default certificate
7006 routines: these may well fail if the certificates aren't installed.
7009 *) Code to support otherName option in GeneralName.
7012 *) First update to verify code. Change the verify utility
7013 so it warns if it is passed a self signed certificate:
7014 for consistency with the normal behaviour. X509_verify
7015 has been modified to it will now verify a self signed
7016 certificate if *exactly* the same certificate appears
7017 in the store: it was previously impossible to trust a
7018 single self signed certificate. This means that:
7019 openssl verify ss.pem
7020 now gives a warning about a self signed certificate but
7021 openssl verify -CAfile ss.pem ss.pem
7025 *) For servers, store verify_result in SSL_SESSION data structure
7026 (and add it to external session representation).
7027 This is needed when client certificate verifications fails,
7028 but an application-provided verification callback (set by
7029 SSL_CTX_set_cert_verify_callback) allows accepting the session
7030 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
7031 but returns 1): When the session is reused, we have to set
7032 ssl->verify_result to the appropriate error code to avoid
7034 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
7036 *) Fix a bug in the new PKCS#7 code: it didn't consider the
7037 case in PKCS7_dataInit() where the signed PKCS7 structure
7038 didn't contain any existing data because it was being created.
7039 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
7041 *) Add a salt to the key derivation routines in enc.c. This
7042 forms the first 8 bytes of the encrypted file. Also add a
7043 -S option to allow a salt to be input on the command line.
7046 *) New function X509_cmp(). Oddly enough there wasn't a function
7047 to compare two certificates. We do this by working out the SHA1
7048 hash and comparing that. X509_cmp() will be needed by the trust
7052 *) SSL_get1_session() is like SSL_get_session(), but increments
7053 the reference count in the SSL_SESSION returned.
7054 [Geoff Thorpe <geoff@eu.c2.net>]
7056 *) Fix for 'req': it was adding a null to request attributes.
7057 Also change the X509_LOOKUP and X509_INFO code to handle
7058 certificate auxiliary information.
7061 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
7065 *) Add the possibility to add extra information to the memory leak
7066 detecting output, to form tracebacks, showing from where each
7067 allocation was originated: CRYPTO_push_info("constant string") adds
7068 the string plus current file name and line number to a per-thread
7069 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
7070 is like calling CYRPTO_pop_info() until the stack is empty.
7071 Also updated memory leak detection code to be multi-thread-safe.
7074 *) Add options -text and -noout to pkcs7 utility and delete the
7075 encryption options which never did anything. Update docs.
7078 *) Add options to some of the utilities to allow the pass phrase
7079 to be included on either the command line (not recommended on
7080 OSes like Unix) or read from the environment. Update the
7081 manpages and fix a few bugs.
7084 *) Add a few manpages for some of the openssl commands.
7087 *) Fix the -revoke option in ca. It was freeing up memory twice,
7088 leaking and not finding already revoked certificates.
7091 *) Extensive changes to support certificate auxiliary information.
7092 This involves the use of X509_CERT_AUX structure and X509_AUX
7093 functions. An X509_AUX function such as PEM_read_X509_AUX()
7094 can still read in a certificate file in the usual way but it
7095 will also read in any additional "auxiliary information". By
7096 doing things this way a fair degree of compatibility can be
7097 retained: existing certificates can have this information added
7098 using the new 'x509' options.
7100 Current auxiliary information includes an "alias" and some trust
7101 settings. The trust settings will ultimately be used in enhanced
7102 certificate chain verification routines: currently a certificate
7103 can only be trusted if it is self signed and then it is trusted
7107 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
7108 The problem was that one of the replacement routines had not been working
7109 since SSLeay releases. For now the offending routine has been replaced
7110 with non-optimised assembler. Even so, this now gives around 95%
7111 performance improvement for 1024 bit RSA signs.
7114 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
7115 handling. Most clients have the effective key size in bits equal to
7116 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
7117 A few however don't do this and instead use the size of the decrypted key
7118 to determine the RC2 key length and the AlgorithmIdentifier to determine
7119 the effective key length. In this case the effective key length can still
7120 be 40 bits but the key length can be 168 bits for example. This is fixed
7121 by manually forcing an RC2 key into the EVP_PKEY structure because the
7122 EVP code can't currently handle unusual RC2 key sizes: it always assumes
7123 the key length and effective key length are equal.
7126 *) Add a bunch of functions that should simplify the creation of
7127 X509_NAME structures. Now you should be able to do:
7128 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
7129 and have it automatically work out the correct field type and fill in
7130 the structures. The more adventurous can try:
7131 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
7132 and it will (hopefully) work out the correct multibyte encoding.
7135 *) Change the 'req' utility to use the new field handling and multibyte
7136 copy routines. Before the DN field creation was handled in an ad hoc
7137 way in req, ca, and x509 which was rather broken and didn't support
7138 BMPStrings or UTF8Strings. Since some software doesn't implement
7139 BMPStrings or UTF8Strings yet, they can be enabled using the config file
7140 using the dirstring_type option. See the new comment in the default
7141 openssl.cnf for more info.
7144 *) Make crypto/rand/md_rand.c more robust:
7145 - Assure unique random numbers after fork().
7146 - Make sure that concurrent threads access the global counter and
7147 md serializably so that we never lose entropy in them
7148 or use exactly the same state in multiple threads.
7149 Access to the large state is not always serializable because
7150 the additional locking could be a performance killer, and
7151 md should be large enough anyway.
7154 *) New file apps/app_rand.c with commonly needed functionality
7155 for handling the random seed file.
7157 Use the random seed file in some applications that previously did not:
7159 dsaparam -genkey (which also ignored its '-rand' option),
7162 x509 (when signing).
7163 Except on systems with /dev/urandom, it is crucial to have a random
7164 seed file at least for key creation, DSA signing, and for DH exchanges;
7165 for RSA signatures we could do without one.
7167 gendh and gendsa (unlike genrsa) used to read only the first byte
7168 of each file listed in the '-rand' option. The function as previously
7169 found in genrsa is now in app_rand.c and is used by all programs
7170 that support '-rand'.
7173 *) In RAND_write_file, use mode 0600 for creating files;
7174 don't just chmod when it may be too late.
7177 *) Report an error from X509_STORE_load_locations
7178 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
7181 *) New function ASN1_mbstring_copy() this copies a string in either
7182 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
7183 into an ASN1_STRING type. A mask of permissible types is passed
7184 and it chooses the "minimal" type to use or an error if not type
7188 *) Add function equivalents to the various macros in asn1.h. The old
7189 macros are retained with an M_ prefix. Code inside the library can
7190 use the M_ macros. External code (including the openssl utility)
7191 should *NOT* in order to be "shared library friendly".
7194 *) Add various functions that can check a certificate's extensions
7195 to see if it usable for various purposes such as SSL client,
7196 server or S/MIME and CAs of these types. This is currently
7197 VERY EXPERIMENTAL but will ultimately be used for certificate chain
7198 verification. Also added a -purpose flag to x509 utility to
7199 print out all the purposes.
7202 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
7206 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
7207 for, obtain and decode and extension and obtain its critical flag.
7208 This allows all the necessary extension code to be handled in a
7209 single function call.
7212 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
7213 platforms. See crypto/rc4/rc4_enc.c for further details.
7216 *) New -noout option to asn1parse. This causes no output to be produced
7217 its main use is when combined with -strparse and -out to extract data
7218 from a file (which may not be in ASN.1 format).
7221 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
7222 when producing the local key id.
7223 [Richard Levitte <levitte@stacken.kth.se>]
7225 *) New option -dhparam in s_server. This allows a DH parameter file to be
7226 stated explicitly. If it is not stated then it tries the first server
7227 certificate file. The previous behaviour hard coded the filename
7231 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
7232 a public key to be input or output. For example:
7233 openssl rsa -in key.pem -pubout -out pubkey.pem
7234 Also added necessary DSA public key functions to handle this.
7237 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
7238 in the message. This was handled by allowing
7239 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
7240 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
7242 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
7243 to the end of the strings whereas this didn't. This would cause problems
7244 if strings read with d2i_ASN1_bytes() were later modified.
7245 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
7247 *) Fix for base64 decode bug. When a base64 bio reads only one line of
7248 data and it contains EOF it will end up returning an error. This is
7249 caused by input 46 bytes long. The cause is due to the way base64
7250 BIOs find the start of base64 encoded data. They do this by trying a
7251 trial decode on each line until they find one that works. When they
7252 do a flag is set and it starts again knowing it can pass all the
7253 data directly through the decoder. Unfortunately it doesn't reset
7254 the context it uses. This means that if EOF is reached an attempt
7255 is made to pass two EOFs through the context and this causes the
7256 resulting error. This can also cause other problems as well. As is
7257 usual with these problems it takes *ages* to find and the fix is
7258 trivial: move one line.
7259 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
7261 *) Ugly workaround to get s_client and s_server working under Windows. The
7262 old code wouldn't work because it needed to select() on sockets and the
7263 tty (for keypresses and to see if data could be written). Win32 only
7264 supports select() on sockets so we select() with a 1s timeout on the
7265 sockets and then see if any characters are waiting to be read, if none
7266 are present then we retry, we also assume we can always write data to
7267 the tty. This isn't nice because the code then blocks until we've
7268 received a complete line of data and it is effectively polling the
7269 keyboard at 1s intervals: however it's quite a bit better than not
7270 working at all :-) A dedicated Windows application might handle this
7271 with an event loop for example.
7274 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
7275 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
7276 will be called when RSA_sign() and RSA_verify() are used. This is useful
7277 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
7278 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
7279 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
7280 This necessitated the support of an extra signature type NID_md5_sha1
7281 for SSL signatures and modifications to the SSL library to use it instead
7282 of calling RSA_public_decrypt() and RSA_private_encrypt().
7285 *) Add new -verify -CAfile and -CApath options to the crl program, these
7286 will lookup a CRL issuers certificate and verify the signature in a
7287 similar way to the verify program. Tidy up the crl program so it
7288 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
7289 less strict. It will now permit CRL extensions even if it is not
7290 a V2 CRL: this will allow it to tolerate some broken CRLs.
7293 *) Initialize all non-automatic variables each time one of the openssl
7294 sub-programs is started (this is necessary as they may be started
7295 multiple times from the "OpenSSL>" prompt).
7296 [Lennart Bang, Bodo Moeller]
7298 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
7299 removing all other RSA functionality (this is what NO_RSA does). This
7300 is so (for example) those in the US can disable those operations covered
7301 by the RSA patent while allowing storage and parsing of RSA keys and RSA
7305 *) Non-copying interface to BIO pairs.
7306 (still largely untested)
7309 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
7310 ASCII string. This was handled independently in various places before.
7313 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
7314 UTF8 strings a character at a time.
7317 *) Use client_version from client hello to select the protocol
7318 (s23_srvr.c) and for RSA client key exchange verification
7319 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
7322 *) Add various utility functions to handle SPKACs, these were previously
7323 handled by poking round in the structure internals. Added new function
7324 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
7325 print, verify and generate SPKACs. Based on an original idea from
7326 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
7329 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
7332 *) Allow the config file extension section to be overwritten on the
7333 command line. Based on an original idea from Massimiliano Pala
7334 <madwolf@comune.modena.it>. The new option is called -extensions
7335 and can be applied to ca, req and x509. Also -reqexts to override
7336 the request extensions in req and -crlexts to override the crl extensions
7340 *) Add new feature to the SPKAC handling in ca. Now you can include
7341 the same field multiple times by preceding it by "XXXX." for example:
7344 this is the same syntax as used in the req config file.
7347 *) Allow certificate extensions to be added to certificate requests. These
7348 are specified in a 'req_extensions' option of the req section of the
7349 config file. They can be printed out with the -text option to req but
7350 are otherwise ignored at present.
7353 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
7354 data read consists of only the final block it would not decrypted because
7355 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
7356 A misplaced 'break' also meant the decrypted final block might not be
7357 copied until the next read.
7360 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
7361 a few extra parameters to the DH structure: these will be useful if
7362 for example we want the value of 'q' or implement X9.42 DH.
7365 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
7366 provides hooks that allow the default DSA functions or functions on a
7367 "per key" basis to be replaced. This allows hardware acceleration and
7368 hardware key storage to be handled without major modification to the
7369 library. Also added low level modexp hooks and CRYPTO_EX structure and
7370 associated functions.
7373 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
7374 as "read only": it can't be written to and the buffer it points to will
7375 not be freed. Reading from a read only BIO is much more efficient than
7376 a normal memory BIO. This was added because there are several times when
7377 an area of memory needs to be read from a BIO. The previous method was
7378 to create a memory BIO and write the data to it, this results in two
7379 copies of the data and an O(n^2) reading algorithm. There is a new
7380 function BIO_new_mem_buf() which creates a read only memory BIO from
7381 an area of memory. Also modified the PKCS#7 routines to use read only
7385 *) Bugfix: ssl23_get_client_hello did not work properly when called in
7386 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
7387 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
7388 but a retry condition occured while trying to read the rest.
7391 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
7392 NID_pkcs7_encrypted by default: this was wrong since this should almost
7393 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
7394 the encrypted data type: this is a more sensible place to put it and it
7395 allows the PKCS#12 code to be tidied up that duplicated this
7399 *) Changed obj_dat.pl script so it takes its input and output files on
7400 the command line. This should avoid shell escape redirection problems
7404 *) Initial support for certificate extension requests, these are included
7405 in things like Xenroll certificate requests. Included functions to allow
7406 extensions to be obtained and added.
7409 *) -crlf option to s_client and s_server for sending newlines as
7410 CRLF (as required by many protocols).
7413 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
7415 *) Install libRSAglue.a when OpenSSL is built with RSAref.
7416 [Ralf S. Engelschall]
7418 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
7419 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
7421 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
7425 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
7426 DH parameters/keys (q is lost during that conversion, but the resulting
7427 DH parameters contain its length).
7429 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
7430 much faster than DH_generate_parameters (which creates parameters
7431 where p = 2*q + 1), and also the smaller q makes DH computations
7432 much more efficient (160-bit exponentiation instead of 1024-bit
7433 exponentiation); so this provides a convenient way to support DHE
7434 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
7435 utter importance to use
7436 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
7438 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
7439 when such DH parameters are used, because otherwise small subgroup
7440 attacks may become possible!
7443 *) Avoid memory leak in i2d_DHparams.
7446 *) Allow the -k option to be used more than once in the enc program:
7447 this allows the same encrypted message to be read by multiple recipients.
7450 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
7451 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
7452 it will always use the numerical form of the OID, even if it has a short
7456 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
7457 method only got called if p,q,dmp1,dmq1,iqmp components were present,
7458 otherwise bn_mod_exp was called. In the case of hardware keys for example
7459 no private key components need be present and it might store extra data
7460 in the RSA structure, which cannot be accessed from bn_mod_exp.
7461 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
7462 private key operations.
7465 *) Added support for SPARC Linux.
7468 *) pem_password_cb function type incompatibly changed from
7469 typedef int pem_password_cb(char *buf, int size, int rwflag);
7471 ....(char *buf, int size, int rwflag, void *userdata);
7472 so that applications can pass data to their callbacks:
7473 The PEM[_ASN1]_{read,write}... functions and macros now take an
7474 additional void * argument, which is just handed through whenever
7475 the password callback is called.
7476 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
7478 New function SSL_CTX_set_default_passwd_cb_userdata.
7480 Compatibility note: As many C implementations push function arguments
7481 onto the stack in reverse order, the new library version is likely to
7482 interoperate with programs that have been compiled with the old
7483 pem_password_cb definition (PEM_whatever takes some data that
7484 happens to be on the stack as its last argument, and the callback
7485 just ignores this garbage); but there is no guarantee whatsoever that
7488 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
7489 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
7490 problems not only on Windows, but also on some Unix platforms.
7491 To avoid problematic command lines, these definitions are now in an
7492 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
7493 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
7496 *) MIPS III/IV assembler module is reimplemented.
7499 *) More DES library cleanups: remove references to srand/rand and
7500 delete an unused file.
7503 *) Add support for the the free Netwide assembler (NASM) under Win32,
7504 since not many people have MASM (ml) and it can be hard to obtain.
7505 This is currently experimental but it seems to work OK and pass all
7506 the tests. Check out INSTALL.W32 for info.
7509 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
7510 without temporary keys kept an extra copy of the server key,
7511 and connections with temporary keys did not free everything in case
7515 *) New function RSA_check_key and new openssl rsa option -check
7516 for verifying the consistency of RSA keys.
7517 [Ulf Moeller, Bodo Moeller]
7519 *) Various changes to make Win32 compile work:
7520 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
7521 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
7522 comparison" warnings.
7523 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
7526 *) Add a debugging option to PKCS#5 v2 key generation function: when
7527 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
7528 derived keys are printed to stderr.
7531 *) Copy the flags in ASN1_STRING_dup().
7532 [Roman E. Pavlov <pre@mo.msk.ru>]
7534 *) The x509 application mishandled signing requests containing DSA
7535 keys when the signing key was also DSA and the parameters didn't match.
7537 It was supposed to omit the parameters when they matched the signing key:
7538 the verifying software was then supposed to automatically use the CA's
7539 parameters if they were absent from the end user certificate.
7541 Omitting parameters is no longer recommended. The test was also
7542 the wrong way round! This was probably due to unusual behaviour in
7543 EVP_cmp_parameters() which returns 1 if the parameters match.
7544 This meant that parameters were omitted when they *didn't* match and
7545 the certificate was useless. Certificates signed with 'ca' didn't have
7547 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
7549 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
7550 The interface is as follows:
7551 Applications can use
7552 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
7553 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
7554 "off" is now the default.
7555 The library internally uses
7556 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
7557 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
7558 to disable memory-checking temporarily.
7560 Some inconsistent states that previously were possible (and were
7561 even the default) are now avoided.
7563 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
7564 with each memory chunk allocated; this is occasionally more helpful
7565 than just having a counter.
7567 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
7569 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
7573 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
7574 which largely parallels "options", but is for changing API behaviour,
7575 whereas "options" are about protocol behaviour.
7576 Initial "mode" flags are:
7578 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
7579 a single record has been written.
7580 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
7581 retries use the same buffer location.
7582 (But all of the contents must be
7586 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
7589 *) Fix problems with no-hmac etc.
7590 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
7592 *) New functions RSA_get_default_method(), RSA_set_method() and
7593 RSA_get_method(). These allows replacement of RSA_METHODs without having
7594 to mess around with the internals of an RSA structure.
7597 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
7598 Also really enable memory leak checks in openssl.c and in some
7600 [Chad C. Mulligan, Bodo Moeller]
7602 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
7603 up the length of negative integers. This has now been simplified to just
7604 store the length when it is first determined and use it later, rather
7605 than trying to keep track of where data is copied and updating it to
7607 [Steve Henson, reported by Brien Wheeler
7608 <bwheeler@authentica-security.com>]
7610 *) Add a new function PKCS7_signatureVerify. This allows the verification
7611 of a PKCS#7 signature but with the signing certificate passed to the
7612 function itself. This contrasts with PKCS7_dataVerify which assumes the
7613 certificate is present in the PKCS#7 structure. This isn't always the
7614 case: certificates can be omitted from a PKCS#7 structure and be
7615 distributed by "out of band" means (such as a certificate database).
7618 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
7619 function prototypes in pem.h, also change util/mkdef.pl to add the
7620 necessary function names.
7623 *) mk1mf.pl (used by Windows builds) did not properly read the
7624 options set by Configure in the top level Makefile, and Configure
7625 was not even able to write more than one option correctly.
7626 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
7629 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
7630 file to be loaded from a BIO or FILE pointer. The BIO version will
7631 for example allow memory BIOs to contain config info.
7634 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
7635 Whoever hopes to achieve shared-library compatibility across versions
7636 must use this, not the compile-time macro.
7637 (Exercise 0.9.4: Which is the minimum library version required by
7639 Note: All this applies only to multi-threaded programs, others don't
7643 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
7644 through a BIO pair triggered the default case, i.e.
7645 SSLerr(...,SSL_R_UNKNOWN_STATE).
7648 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
7649 can use the SSL library even if none of the specific BIOs is
7653 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
7654 for the encoded length.
7655 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
7657 *) Add initial documentation of the X509V3 functions.
7660 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
7661 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
7662 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
7663 secure PKCS#8 private key format with a high iteration count.
7666 *) Fix determination of Perl interpreter: A perl or perl5
7667 _directory_ in $PATH was also accepted as the interpreter.
7668 [Ralf S. Engelschall]
7670 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
7671 wrong with it but it was very old and did things like calling
7672 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
7676 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
7677 to use the new extension code.
7680 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
7681 with macros. This should make it easier to change their form, add extra
7682 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
7686 *) Add to configuration table a new entry that can specify an alternative
7687 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
7688 according to Mark Crispin <MRC@Panda.COM>.
7692 *) DES CBC did not update the IV. Weird.
7695 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
7696 Changing the behaviour of the former might break existing programs --
7697 where IV updating is needed, des_ncbc_encrypt can be used.
7700 *) When bntest is run from "make test" it drives bc to check its
7701 calculations, as well as internally checking them. If an internal check
7702 fails, it needs to cause bc to give a non-zero result or make test carries
7703 on without noticing the failure. Fixed.
7706 *) DES library cleanups.
7709 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
7710 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
7711 ciphers. NOTE: although the key derivation function has been verified
7712 against some published test vectors it has not been extensively tested
7713 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
7717 *) Instead of "mkdir -p", which is not fully portable, use new
7718 Perl script "util/mkdir-p.pl".
7721 *) Rewrite the way password based encryption (PBE) is handled. It used to
7722 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
7723 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
7724 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
7725 the 'parameter' field of the AlgorithmIdentifier is passed to the
7726 underlying key generation function so it must do its own ASN1 parsing.
7727 This has also changed the EVP_PBE_CipherInit() function which now has a
7728 'parameter' argument instead of literal salt and iteration count values
7729 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
7732 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
7733 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
7734 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
7735 KEY" because this clashed with PKCS#8 unencrypted string. Since this
7736 value was just used as a "magic string" and not used directly its
7737 value doesn't matter.
7740 *) Introduce some semblance of const correctness to BN. Shame C doesn't
7744 *) "linux-sparc64" configuration (ultrapenguin).
7745 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
7746 "linux-sparc" configuration.
7747 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
7749 *) config now generates no-xxx options for missing ciphers.
7752 *) Support the EBCDIC character set (work in progress).
7753 File ebcdic.c not yet included because it has a different license.
7754 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7756 *) Support BS2000/OSD-POSIX.
7757 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7759 *) Make callbacks for key generation use void * instead of char *.
7762 *) Make S/MIME samples compile (not yet tested).
7765 *) Additional typesafe stacks.
7768 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
7772 Changes between 0.9.3 and 0.9.3a [29 May 1999]
7774 *) New configuration variant "sco5-gcc".
7776 *) Updated some demos.
7777 [Sean O Riordain, Wade Scholine]
7779 *) Add missing BIO_free at exit of pkcs12 application.
7782 *) Fix memory leak in conf.c.
7785 *) Updates for Win32 to assembler version of MD5.
7788 *) Set #! path to perl in apps/der_chop to where we found it
7789 instead of using a fixed path.
7792 *) SHA library changes for irix64-mips4-cc.
7795 *) Improvements for VMS support.
7799 Changes between 0.9.2b and 0.9.3 [24 May 1999]
7801 *) Bignum library bug fix. IRIX 6 passes "make test" now!
7802 This also avoids the problems with SC4.2 and unpatched SC5.
7803 [Andy Polyakov <appro@fy.chalmers.se>]
7805 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
7806 These are required because of the typesafe stack would otherwise break
7807 existing code. If old code used a structure member which used to be STACK
7808 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
7809 sk_num or sk_value it would produce an error because the num, data members
7810 are not present in STACK_OF. Now it just produces a warning. sk_set
7811 replaces the old method of assigning a value to sk_value
7812 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
7813 that does this will no longer work (and should use sk_set instead) but
7814 this could be regarded as a "questionable" behaviour anyway.
7817 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
7818 correctly handle encrypted S/MIME data.
7821 *) Change type of various DES function arguments from des_cblock
7822 (which means, in function argument declarations, pointer to char)
7823 to des_cblock * (meaning pointer to array with 8 char elements),
7824 which allows the compiler to do more typechecking; it was like
7825 that back in SSLeay, but with lots of ugly casts.
7827 Introduce new type const_des_cblock.
7830 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
7831 problems: find RecipientInfo structure that matches recipient certificate
7832 and initialise the ASN1 structures properly based on passed cipher.
7835 *) Belatedly make the BN tests actually check the results.
7838 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
7839 to and from BNs: it was completely broken. New compilation option
7840 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
7841 key elements as negative integers.
7844 *) Reorganize and speed up MD5.
7845 [Andy Polyakov <appro@fy.chalmers.se>]
7848 [Richard Levitte <richard@levitte.org>]
7850 *) New option -out to asn1parse to allow the parsed structure to be
7851 output to a file. This is most useful when combined with the -strparse
7852 option to examine the output of things like OCTET STRINGS.
7855 *) Make SSL library a little more fool-proof by not requiring any longer
7856 that SSL_set_{accept,connect}_state be called before
7857 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
7858 in many applications because usually everything *appeared* to work as
7859 intended anyway -- now it really works as intended).
7862 *) Move openssl.cnf out of lib/.
7865 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
7866 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
7867 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
7868 [Ralf S. Engelschall]
7870 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
7871 handle PKCS#7 enveloped data properly.
7872 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
7874 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
7875 copying pointers. The cert_st handling is changed by this in
7876 various ways (and thus what used to be known as ctx->default_cert
7877 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
7878 any longer when s->cert does not give us what we need).
7879 ssl_cert_instantiate becomes obsolete by this change.
7880 As soon as we've got the new code right (possibly it already is?),
7881 we have solved a couple of bugs of the earlier code where s->cert
7882 was used as if it could not have been shared with other SSL structures.
7884 Note that using the SSL API in certain dirty ways now will result
7885 in different behaviour than observed with earlier library versions:
7886 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
7887 does not influence s as it used to.
7889 In order to clean up things more thoroughly, inside SSL_SESSION
7890 we don't use CERT any longer, but a new structure SESS_CERT
7891 that holds per-session data (if available); currently, this is
7892 the peer's certificate chain and, for clients, the server's certificate
7893 and temporary key. CERT holds only those values that can have
7894 meaningful defaults in an SSL_CTX.
7897 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
7898 from the internal representation. Various PKCS#7 fixes: remove some
7899 evil casts and set the enc_dig_alg field properly based on the signing
7903 *) Allow PKCS#12 password to be set from the command line or the
7904 environment. Let 'ca' get its config file name from the environment
7905 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
7909 *) Allow certificate policies extension to use an IA5STRING for the
7910 organization field. This is contrary to the PKIX definition but
7911 VeriSign uses it and IE5 only recognises this form. Document 'x509'
7915 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
7916 without disallowing inline assembler and the like for non-pedantic builds.
7919 *) Support Borland C++ builder.
7920 [Janez Jere <jj@void.si>, modified by Ulf Möller]
7925 *) SHA-1 cleanups and performance enhancements.
7926 [Andy Polyakov <appro@fy.chalmers.se>]
7928 *) Sparc v8plus assembler for the bignum library.
7929 [Andy Polyakov <appro@fy.chalmers.se>]
7931 *) Accept any -xxx and +xxx compiler options in Configure.
7934 *) Update HPUX configuration.
7937 *) Add missing sk_<type>_unshift() function to safestack.h
7938 [Ralf S. Engelschall]
7940 *) New function SSL_CTX_use_certificate_chain_file that sets the
7941 "extra_cert"s in addition to the certificate. (This makes sense
7942 only for "PEM" format files, as chains as a whole are not
7946 *) Support verify_depth from the SSL API.
7947 x509_vfy.c had what can be considered an off-by-one-error:
7948 Its depth (which was not part of the external interface)
7949 was actually counting the number of certificates in a chain;
7950 now it really counts the depth.
7953 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
7954 instead of X509err, which often resulted in confusing error
7955 messages since the error codes are not globally unique
7956 (e.g. an alleged error in ssl3_accept when a certificate
7957 didn't match the private key).
7959 *) New function SSL_CTX_set_session_id_context that allows to set a default
7960 value (so that you don't need SSL_set_session_id_context for each
7961 connection using the SSL_CTX).
7964 *) OAEP decoding bug fix.
7967 *) Support INSTALL_PREFIX for package builders, as proposed by
7971 *) New Configure options "threads" and "no-threads". For systems
7972 where the proper compiler options are known (currently Solaris
7973 and Linux), "threads" is the default.
7976 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
7979 *) Install various scripts to $(OPENSSLDIR)/misc, not to
7980 $(INSTALLTOP)/bin -- they shouldn't clutter directories
7981 such as /usr/local/bin.
7984 *) "make linux-shared" to build shared libraries.
7985 [Niels Poppe <niels@netbox.org>]
7987 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
7990 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
7991 extension adding in x509 utility.
7994 *) Remove NOPROTO sections and error code comments.
7997 *) Partial rewrite of the DEF file generator to now parse the ANSI
8001 *) New Configure options --prefix=DIR and --openssldir=DIR.
8004 *) Complete rewrite of the error code script(s). It is all now handled
8005 by one script at the top level which handles error code gathering,
8006 header rewriting and C source file generation. It should be much better
8007 than the old method: it now uses a modified version of Ulf's parser to
8008 read the ANSI prototypes in all header files (thus the old K&R definitions
8009 aren't needed for error creation any more) and do a better job of
8010 translating function codes into names. The old 'ASN1 error code imbedded
8011 in a comment' is no longer necessary and it doesn't use .err files which
8012 have now been deleted. Also the error code call doesn't have to appear all
8013 on one line (which resulted in some large lines...).
8016 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
8019 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
8020 0 (which usually indicates a closed connection), but continue reading.
8023 *) Fix some race conditions.
8026 *) Add support for CRL distribution points extension. Add Certificate
8027 Policies and CRL distribution points documentation.
8030 *) Move the autogenerated header file parts to crypto/opensslconf.h.
8033 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
8034 8 of keying material. Merlin has also confirmed interop with this fix
8035 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
8036 [Merlin Hughes <merlin@baltimore.ie>]
8038 *) Fix lots of warnings.
8039 [Richard Levitte <levitte@stacken.kth.se>]
8041 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
8042 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
8043 [Richard Levitte <levitte@stacken.kth.se>]
8045 *) Fix problems with sizeof(long) == 8.
8046 [Andy Polyakov <appro@fy.chalmers.se>]
8048 *) Change functions to ANSI C.
8051 *) Fix typos in error codes.
8052 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
8054 *) Remove defunct assembler files from Configure.
8057 *) SPARC v8 assembler BIGNUM implementation.
8058 [Andy Polyakov <appro@fy.chalmers.se>]
8060 *) Support for Certificate Policies extension: both print and set.
8061 Various additions to support the r2i method this uses.
8064 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
8065 return a const string when you are expecting an allocated buffer.
8068 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
8069 types DirectoryString and DisplayText.
8072 *) Add code to allow r2i extensions to access the configuration database,
8073 add an LHASH database driver and add several ctx helper functions.
8076 *) Fix an evil bug in bn_expand2() which caused various BN functions to
8077 fail when they extended the size of a BIGNUM.
8080 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
8081 support typesafe stack.
8084 *) Fix typo in SSL_[gs]et_options().
8085 [Nils Frostberg <nils@medcom.se>]
8087 *) Delete various functions and files that belonged to the (now obsolete)
8088 old X509V3 handling code.
8091 *) New Configure option "rsaref".
8094 *) Don't auto-generate pem.h.
8097 *) Introduce type-safe ASN.1 SETs.
8100 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
8101 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
8103 *) Introduce type-safe STACKs. This will almost certainly break lots of code
8104 that links with OpenSSL (well at least cause lots of warnings), but fear
8105 not: the conversion is trivial, and it eliminates loads of evil casts. A
8106 few STACKed things have been converted already. Feel free to convert more.
8107 In the fullness of time, I'll do away with the STACK type altogether.
8110 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
8111 specified in <certfile> by updating the entry in the index.txt file.
8112 This way one no longer has to edit the index.txt file manually for
8113 revoking a certificate. The -revoke option does the gory details now.
8114 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
8116 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
8117 `-text' option at all and this way the `-noout -text' combination was
8118 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
8119 [Ralf S. Engelschall]
8121 *) Make sure a corresponding plain text error message exists for the
8122 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
8123 verify callback function determined that a certificate was revoked.
8124 [Ralf S. Engelschall]
8126 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
8127 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
8128 all available cipers including rc5, which was forgotten until now.
8129 In order to let the testing shell script know which algorithms
8130 are available, a new (up to now undocumented) command
8131 "openssl list-cipher-commands" is used.
8134 *) Bugfix: s_client occasionally would sleep in select() when
8135 it should have checked SSL_pending() first.
8138 *) New functions DSA_do_sign and DSA_do_verify to provide access to
8139 the raw DSA values prior to ASN.1 encoding.
8142 *) Tweaks to Configure
8143 [Niels Poppe <niels@netbox.org>]
8145 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
8149 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
8152 *) New config option to avoid instructions that are illegal on the 80386.
8153 The default code is faster, but requires at least a 486.
8156 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
8157 SSL2_SERVER_VERSION (not used at all) macros, which are now the
8158 same as SSL2_VERSION anyway.
8161 *) New "-showcerts" option for s_client.
8164 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
8165 application. Various cleanups and fixes.
8168 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
8169 modify error routines to work internally. Add error codes and PBE init
8170 to library startup routines.
8173 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
8174 packing functions to asn1 and evp. Changed function names and error
8175 codes along the way.
8178 *) PKCS12 integration: and so it begins... First of several patches to
8179 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
8180 objects to objects.h
8183 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
8184 and display support for Thawte strong extranet extension.
8187 *) Add LinuxPPC support.
8188 [Jeff Dubrule <igor@pobox.org>]
8190 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
8191 bn_div_words in alpha.s.
8192 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
8194 *) Make sure the RSA OAEP test is skipped under -DRSAref because
8195 OAEP isn't supported when OpenSSL is built with RSAref.
8196 [Ulf Moeller <ulf@fitug.de>]
8198 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
8199 so they no longer are missing under -DNOPROTO.
8200 [Soren S. Jorvang <soren@t.dk>]
8203 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
8205 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
8206 doesn't work when the session is reused. Coming soon!
8209 *) Fix a security hole, that allows sessions to be reused in the wrong
8210 context thus bypassing client cert protection! All software that uses
8211 client certs and session caches in multiple contexts NEEDS PATCHING to
8212 allow session reuse! A fuller solution is in the works.
8213 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
8215 *) Some more source tree cleanups (removed obsolete files
8216 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
8217 permission on "config" script to be executable) and a fix for the INSTALL
8219 [Ulf Moeller <ulf@fitug.de>]
8221 *) Remove some legacy and erroneous uses of malloc, free instead of
8223 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
8225 *) Make rsa_oaep_test return non-zero on error.
8226 [Ulf Moeller <ulf@fitug.de>]
8228 *) Add support for native Solaris shared libraries. Configure
8229 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
8230 if someone would make that last step automatic.
8231 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
8233 *) ctx_size was not built with the right compiler during "make links". Fixed.
8236 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
8237 except NULL ciphers". This means the default cipher list will no longer
8238 enable NULL ciphers. They need to be specifically enabled e.g. with
8239 the string "DEFAULT:eNULL".
8242 *) Fix to RSA private encryption routines: if p < q then it would
8243 occasionally produce an invalid result. This will only happen with
8244 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
8247 *) Be less restrictive and allow also `perl util/perlpath.pl
8248 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
8249 because this way one can also use an interpreter named `perl5' (which is
8250 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
8251 installed as `perl').
8252 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8254 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
8255 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8257 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
8258 advapi32.lib to Win32 build and change the pem test comparision
8259 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
8260 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
8261 and crypto/des/ede_cbcm_enc.c.
8264 *) DES quad checksum was broken on big-endian architectures. Fixed.
8267 *) Comment out two functions in bio.h that aren't implemented. Fix up the
8268 Win32 test batch file so it (might) work again. The Win32 test batch file
8269 is horrible: I feel ill....
8272 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
8273 in e_os.h. Audit of header files to check ANSI and non ANSI
8274 sections: 10 functions were absent from non ANSI section and not exported
8275 from Windows DLLs. Fixed up libeay.num for new functions.
8278 *) Make `openssl version' output lines consistent.
8279 [Ralf S. Engelschall]
8281 *) Fix Win32 symbol export lists for BIO functions: Added
8282 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
8283 to ms/libeay{16,32}.def.
8284 [Ralf S. Engelschall]
8286 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
8287 fine under Unix and passes some trivial tests I've now added. But the
8288 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
8289 added to make sure no one expects that this stuff really works in the
8290 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
8291 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
8293 [Ralf S. Engelschall]
8295 *) Fix the generation of two part addresses in perl.
8296 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
8298 *) Add config entry for Linux on MIPS.
8299 [John Tobey <jtobey@channel1.com>]
8301 *) Make links whenever Configure is run, unless we are on Windoze.
8304 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
8305 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
8309 *) Add a useful kludge to allow package maintainers to specify compiler and
8310 other platforms details on the command line without having to patch the
8311 Configure script everytime: One now can use ``perl Configure
8312 <id>:<details>'', i.e. platform ids are allowed to have details appended
8313 to them (seperated by colons). This is treated as there would be a static
8314 pre-configured entry in Configure's %table under key <id> with value
8315 <details> and ``perl Configure <id>'' is called. So, when you want to
8316 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
8317 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
8318 now, which overrides the FreeBSD-elf entry on-the-fly.
8319 [Ralf S. Engelschall]
8321 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
8324 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
8325 on the `perl Configure ...' command line. This way one can compile
8326 OpenSSL libraries with Position Independent Code (PIC) which is needed
8327 for linking it into DSOs.
8328 [Ralf S. Engelschall]
8330 *) Remarkably, export ciphers were totally broken and no-one had noticed!
8334 *) Cleaned up the LICENSE document: The official contact for any license
8335 questions now is the OpenSSL core team under openssl-core@openssl.org.
8336 And add a paragraph about the dual-license situation to make sure people
8337 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
8338 to the OpenSSL toolkit.
8339 [Ralf S. Engelschall]
8341 *) General source tree makefile cleanups: Made `making xxx in yyy...'
8342 display consistent in the source tree and replaced `/bin/rm' by `rm'.
8343 Additonally cleaned up the `make links' target: Remove unnecessary
8344 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
8345 to speed processing and no longer clutter the display with confusing
8346 stuff. Instead only the actually done links are displayed.
8347 [Ralf S. Engelschall]
8349 *) Permit null encryption ciphersuites, used for authentication only. It used
8350 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
8351 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
8355 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
8356 signed attributes when verifying signatures (this would break them),
8357 the detached data encoding was wrong and public keys obtained using
8358 X509_get_pubkey() weren't freed.
8361 *) Add text documentation for the BUFFER functions. Also added a work around
8362 to a Win95 console bug. This was triggered by the password read stuff: the
8363 last character typed gets carried over to the next fread(). If you were
8364 generating a new cert request using 'req' for example then the last
8365 character of the passphrase would be CR which would then enter the first
8369 *) Added the new `Includes OpenSSL Cryptography Software' button as
8370 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
8371 button and can be used by applications based on OpenSSL to show the
8372 relationship to the OpenSSL project.
8373 [Ralf S. Engelschall]
8375 *) Remove confusing variables in function signatures in files
8376 ssl/ssl_lib.c and ssl/ssl.h.
8377 [Lennart Bong <lob@kulthea.stacken.kth.se>]
8379 *) Don't install bss_file.c under PREFIX/include/
8380 [Lennart Bong <lob@kulthea.stacken.kth.se>]
8382 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
8383 functions that return function pointers and has support for NT specific
8384 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
8385 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
8386 unsigned to signed types: this was killing the Win32 compile.
8389 *) Add new certificate file to stack functions,
8390 SSL_add_dir_cert_subjects_to_stack() and
8391 SSL_add_file_cert_subjects_to_stack(). These largely supplant
8392 SSL_load_client_CA_file(), and can be used to add multiple certs easily
8393 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
8394 This means that Apache-SSL and similar packages don't have to mess around
8395 to add as many CAs as they want to the preferred list.
8398 *) Experiment with doxygen documentation. Currently only partially applied to
8400 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
8401 openssl.doxy as the configuration file.
8404 *) Get rid of remaining C++-style comments which strict C compilers hate.
8405 [Ralf S. Engelschall, pointed out by Carlos Amengual]
8407 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
8408 compiled in by default: it has problems with large keys.
8411 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
8412 DH private keys and/or callback functions which directly correspond to
8413 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
8414 is needed for applications which have to configure certificates on a
8415 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
8417 For the RSA certificate situation is makes no difference, but
8418 for the DSA certificate situation this fixes the "no shared cipher"
8419 problem where the OpenSSL cipher selection procedure failed because the
8420 temporary keys were not overtaken from the context and the API provided
8421 no way to reconfigure them.
8422 The new functions now let applications reconfigure the stuff and they
8423 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
8424 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
8425 non-public-API function ssl_cert_instantiate() is used as a helper
8426 function and also to reduce code redundancy inside ssl_rsa.c.
8427 [Ralf S. Engelschall]
8429 *) Move s_server -dcert and -dkey options out of the undocumented feature
8430 area because they are useful for the DSA situation and should be
8431 recognized by the users.
8432 [Ralf S. Engelschall]
8434 *) Fix the cipher decision scheme for export ciphers: the export bits are
8435 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
8436 SSL_EXP_MASK. So, the original variable has to be used instead of the
8437 already masked variable.
8438 [Richard Levitte <levitte@stacken.kth.se>]
8440 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
8441 [Richard Levitte <levitte@stacken.kth.se>]
8443 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
8444 from `int' to `unsigned int' because it's a length and initialized by
8445 EVP_DigestFinal() which expects an `unsigned int *'.
8446 [Richard Levitte <levitte@stacken.kth.se>]
8448 *) Don't hard-code path to Perl interpreter on shebang line of Configure
8449 script. Instead use the usual Shell->Perl transition trick.
8450 [Ralf S. Engelschall]
8452 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
8453 (in addition to RSA certificates) to match the behaviour of `openssl dsa
8454 -noout -modulus' as it's already the case for `openssl rsa -noout
8455 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
8456 currently the public key is printed (a decision which was already done by
8457 `openssl dsa -modulus' in the past) which serves a similar purpose.
8458 Additionally the NO_RSA no longer completely removes the whole -modulus
8459 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
8461 [Ralf S. Engelschall]
8463 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
8464 BIO. See the source (crypto/evp/bio_ok.c) for more info.
8465 [Arne Ansper <arne@ats.cyber.ee>]
8467 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
8468 to be added. Now both 'req' and 'ca' can use new objects defined in the
8472 *) Add cool BIO that does syslog (or event log on NT).
8473 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
8475 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
8476 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
8477 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
8478 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
8481 *) Add preliminary config info for new extension code.
8484 *) Make RSA_NO_PADDING really use no padding.
8485 [Ulf Moeller <ulf@fitug.de>]
8487 *) Generate errors when private/public key check is done.
8490 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
8491 for some CRL extensions and new objects added.
8494 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
8495 key usage extension and fuller support for authority key id.
8498 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
8499 padding method for RSA, which is recommended for new applications in PKCS
8500 #1 v2.0 (RFC 2437, October 1998).
8501 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
8502 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
8503 against Bleichbacher's attack on RSA.
8504 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
8507 *) Updates to the new SSL compression code
8508 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8510 *) Fix so that the version number in the master secret, when passed
8511 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
8512 (because the server will not accept higher), that the version number
8513 is 0x03,0x01, not 0x03,0x00
8514 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8516 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
8517 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
8518 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
8521 *) Support for RAW extensions where an arbitrary extension can be
8522 created by including its DER encoding. See apps/openssl.cnf for
8526 *) Make sure latest Perl versions don't interpret some generated C array
8527 code as Perl array code in the crypto/err/err_genc.pl script.
8528 [Lars Weber <3weber@informatik.uni-hamburg.de>]
8530 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
8531 not many people have the assembler. Various Win32 compilation fixes and
8532 update to the INSTALL.W32 file with (hopefully) more accurate Win32
8536 *) Modify configure script 'Configure' to automatically create crypto/date.h
8537 file under Win32 and also build pem.h from pem.org. New script
8538 util/mkfiles.pl to create the MINFO file on environments that can't do a
8539 'make files': perl util/mkfiles.pl >MINFO should work.
8542 *) Major rework of DES function declarations, in the pursuit of correctness
8543 and purity. As a result, many evil casts evaporated, and some weirdness,
8544 too. You may find this causes warnings in your code. Zapping your evil
8545 casts will probably fix them. Mostly.
8548 *) Fix for a typo in asn1.h. Bug fix to object creation script
8549 obj_dat.pl. It considered a zero in an object definition to mean
8550 "end of object": none of the objects in objects.h have any zeros
8551 so it wasn't spotted.
8552 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
8554 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
8555 Masking (CBCM). In the absence of test vectors, the best I have been able
8556 to do is check that the decrypt undoes the encrypt, so far. Send me test
8557 vectors if you have them.
8560 *) Correct calculation of key length for export ciphers (too much space was
8561 allocated for null ciphers). This has not been tested!
8564 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
8565 message is now correct (it understands "crypto" and "ssl" on its
8566 command line). There is also now an "update" option. This will update
8567 the util/ssleay.num and util/libeay.num files with any new functions.
8569 perl util/mkdef.pl crypto ssl update
8570 it will update them.
8573 *) Overhauled the Perl interface (perl/*):
8574 - ported BN stuff to OpenSSL's different BN library
8575 - made the perl/ source tree CVS-aware
8576 - renamed the package from SSLeay to OpenSSL (the files still contain
8577 their history because I've copied them in the repository)
8578 - removed obsolete files (the test scripts will be replaced
8579 by better Test::Harness variants in the future)
8580 [Ralf S. Engelschall]
8582 *) First cut for a very conservative source tree cleanup:
8583 1. merge various obsolete readme texts into doc/ssleay.txt
8584 where we collect the old documents and readme texts.
8585 2. remove the first part of files where I'm already sure that we no
8586 longer need them because of three reasons: either they are just temporary
8587 files which were left by Eric or they are preserved original files where
8588 I've verified that the diff is also available in the CVS via "cvs diff
8589 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
8590 the crypto/md/ stuff).
8591 [Ralf S. Engelschall]
8593 *) More extension code. Incomplete support for subject and issuer alt
8594 name, issuer and authority key id. Change the i2v function parameters
8595 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
8596 what that's for :-) Fix to ASN1 macro which messed up
8597 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
8600 *) Preliminary support for ENUMERATED type. This is largely copied from the
8604 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
8605 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8607 *) Make sure `make rehash' target really finds the `openssl' program.
8608 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8610 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
8611 like to hear about it if this slows down other processors.
8614 *) Add CygWin32 platform information to Configure script.
8615 [Alan Batie <batie@aahz.jf.intel.com>]
8617 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
8618 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
8620 *) New program nseq to manipulate netscape certificate sequences
8623 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
8627 *) Fixes to BN code. Previously the default was to define BN_RECURSION
8628 but the BN code had some problems that would cause failures when
8629 doing certificate verification and some other functions.
8630 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8632 *) Add ASN1 and PEM code to support netscape certificate sequences.
8635 *) Add ASN1 and PEM code to support netscape certificate sequences.
8638 *) Add several PKIX and private extended key usage OIDs.
8641 *) Modify the 'ca' program to handle the new extension code. Modify
8642 openssl.cnf for new extension format, add comments.
8645 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
8646 and add a sample to openssl.cnf so req -x509 now adds appropriate
8650 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
8651 error code, add initial support to X509_print() and x509 application.
8654 *) Takes a deep breath and start addding X509 V3 extension support code. Add
8655 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
8656 stuff is currently isolated and isn't even compiled yet.
8659 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
8660 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
8661 Removed the versions check from X509 routines when loading extensions:
8662 this allows certain broken certificates that don't set the version
8663 properly to be processed.
8666 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
8667 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
8668 can still be regenerated with "make depend".
8671 *) Spelling mistake in C version of CAST-128.
8672 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
8674 *) Changes to the error generation code. The perl script err-code.pl
8675 now reads in the old error codes and retains the old numbers, only
8676 adding new ones if necessary. It also only changes the .err files if new
8677 codes are added. The makefiles have been modified to only insert errors
8678 when needed (to avoid needlessly modifying header files). This is done
8679 by only inserting errors if the .err file is newer than the auto generated
8680 C file. To rebuild all the error codes from scratch (the old behaviour)
8681 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
8682 or delete all the .err files.
8685 *) CAST-128 was incorrectly implemented for short keys. The C version has
8686 been fixed, but is untested. The assembler versions are also fixed, but
8687 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
8688 to regenerate it if needed.
8689 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
8690 Hagino <itojun@kame.net>]
8692 *) File was opened incorrectly in randfile.c.
8693 [Ulf Möller <ulf@fitug.de>]
8695 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
8696 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
8697 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
8698 al: it's just almost always a UTCTime. Note this patch adds new error
8699 codes so do a "make errors" if there are problems.
8702 *) Correct Linux 1 recognition in config.
8703 [Ulf Möller <ulf@fitug.de>]
8705 *) Remove pointless MD5 hash when using DSA keys in ca.
8706 [Anonymous <nobody@replay.com>]
8708 *) Generate an error if given an empty string as a cert directory. Also
8709 generate an error if handed NULL (previously returned 0 to indicate an
8710 error, but didn't set one).
8711 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
8713 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
8716 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
8717 parameters. This was causing a warning which killed off the Win32 compile.
8720 *) Remove C++ style comments from crypto/bn/bn_local.h.
8721 [Neil Costigan <neil.costigan@celocom.com>]
8723 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
8724 based on a text string, looking up short and long names and finally
8725 "dot" format. The "dot" format stuff didn't work. Added new function
8726 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
8727 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
8728 OID is not part of the table.
8731 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
8732 X509_LOOKUP_by_alias().
8735 *) Sort openssl functions by name.
8738 *) Get the gendsa program working (hopefully) and add it to app list. Remove
8739 encryption from sample DSA keys (in case anyone is interested the password
8743 *) Make _all_ *_free functions accept a NULL pointer.
8744 [Frans Heymans <fheymans@isaserver.be>]
8746 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
8748 [Anonymous <nobody@replay.com>]
8750 *) s_server should send the CAfile as acceptable CAs, not its own cert.
8751 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
8753 *) Don't blow it for numeric -newkey arguments to apps/req.
8754 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
8756 *) Temp key "for export" tests were wrong in s3_srvr.c.
8757 [Anonymous <nobody@replay.com>]
8759 *) Add prototype for temp key callback functions
8760 SSL_CTX_set_tmp_{rsa,dh}_callback().
8763 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
8764 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
8767 *) X509_name_add_entry() freed the wrong thing after an error.
8768 [Arne Ansper <arne@ats.cyber.ee>]
8770 *) rsa_eay.c would attempt to free a NULL context.
8771 [Arne Ansper <arne@ats.cyber.ee>]
8773 *) BIO_s_socket() had a broken should_retry() on Windoze.
8774 [Arne Ansper <arne@ats.cyber.ee>]
8776 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
8777 [Arne Ansper <arne@ats.cyber.ee>]
8779 *) Make sure the already existing X509_STORE->depth variable is initialized
8780 in X509_STORE_new(), but document the fact that this variable is still
8781 unused in the certificate verification process.
8782 [Ralf S. Engelschall]
8784 *) Fix the various library and apps files to free up pkeys obtained from
8785 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
8788 *) Fix reference counting in X509_PUBKEY_get(). This makes
8789 demos/maurice/example2.c work, amongst others, probably.
8790 [Steve Henson and Ben Laurie]
8792 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
8793 `openssl' and second, the shortcut symlinks for the `openssl <command>'
8794 are no longer created. This way we have a single and consistent command
8795 line interface `openssl <command>', similar to `cvs <command>'.
8796 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
8798 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
8799 BIT STRING wrapper always have zero unused bits.
8802 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
8805 *) Make the top-level INSTALL documentation easier to understand.
8808 *) Makefiles updated to exit if an error occurs in a sub-directory
8809 make (including if user presses ^C) [Paul Sutton]
8811 *) Make Montgomery context stuff explicit in RSA data structure.
8814 *) Fix build order of pem and err to allow for generated pem.h.
8817 *) Fix renumbering bug in X509_NAME_delete_entry().
8820 *) Enhanced the err-ins.pl script so it makes the error library number
8821 global and can add a library name. This is needed for external ASN1 and
8822 other error libraries.
8825 *) Fixed sk_insert which never worked properly.
8828 *) Fix ASN1 macros so they can handle indefinite length construted
8829 EXPLICIT tags. Some non standard certificates use these: they can now
8833 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
8834 into a single doc/ssleay.txt bundle. This way the information is still
8835 preserved but no longer messes up this directory. Now it's new room for
8836 the new set of documenation files.
8837 [Ralf S. Engelschall]
8839 *) SETs were incorrectly DER encoded. This was a major pain, because they
8840 shared code with SEQUENCEs, which aren't coded the same. This means that
8841 almost everything to do with SETs or SEQUENCEs has either changed name or
8842 number of arguments.
8843 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
8845 *) Fix test data to work with the above.
8848 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
8849 was already fixed by Eric for 0.9.1 it seems.
8850 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
8852 *) Autodetect FreeBSD3.
8855 *) Fix various bugs in Configure. This affects the following platforms:
8859 unixware-2.0-pentium
8863 *) Eliminate generated files from CVS. Reorder tests to regenerate files
8864 before they are needed.
8867 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
8871 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
8873 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
8874 changed SSLeay to OpenSSL in version strings.
8875 [Ralf S. Engelschall]
8877 *) Some fixups to the top-level documents.
8880 *) Fixed the nasty bug where rsaref.h was not found under compile-time
8881 because the symlink to include/ was missing.
8882 [Ralf S. Engelschall]
8884 *) Incorporated the popular no-RSA/DSA-only patches
8885 which allow to compile a RSA-free SSLeay.
8886 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
8888 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
8889 when "ssleay" is still not found.
8890 [Ralf S. Engelschall]
8892 *) Added more platforms to Configure: Cray T3E, HPUX 11,
8893 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
8895 *) Updated the README file.
8896 [Ralf S. Engelschall]
8898 *) Added various .cvsignore files in the CVS repository subdirs
8899 to make a "cvs update" really silent.
8900 [Ralf S. Engelschall]
8902 *) Recompiled the error-definition header files and added
8903 missing symbols to the Win32 linker tables.
8904 [Ralf S. Engelschall]
8906 *) Cleaned up the top-level documents;
8907 o new files: CHANGES and LICENSE
8908 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
8909 o merged COPYRIGHT into LICENSE
8910 o removed obsolete TODO file
8911 o renamed MICROSOFT to INSTALL.W32
8912 [Ralf S. Engelschall]
8914 *) Removed dummy files from the 0.9.1b source tree:
8915 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
8916 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
8917 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
8918 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
8919 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
8920 [Ralf S. Engelschall]
8922 *) Added various platform portability fixes.
8925 *) The Genesis of the OpenSSL rpject:
8926 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
8927 Young and Tim J. Hudson created while they were working for C2Net until
8929 [The OpenSSL Project]
8932 Changes between 0.9.0b and 0.9.1b [not released]
8934 *) Updated a few CA certificates under certs/
8937 *) Changed some BIGNUM api stuff.
8940 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
8941 DGUX x86, Linux Alpha, etc.
8944 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
8945 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
8949 *) Add -strparse option to asn1pars program which parses nested
8951 [Dr Stephen Henson <shenson@bigfoot.com>]
8953 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
8956 *) DSA fix for "ca" program.
8959 *) Added "-genkey" option to "dsaparam" program.
8962 *) Added RIPE MD160 (rmd160) message digest.
8965 *) Added -a (all) option to "ssleay version" command.
8968 *) Added PLATFORM define which is the id given to Configure.
8971 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
8974 *) Extended the ASN.1 parser routines.
8977 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
8980 *) Added a BN_CTX to the BN library.
8983 *) Fixed the weak key values in DES library
8986 *) Changed API in EVP library for cipher aliases.
8989 *) Added support for RC2/64bit cipher.
8992 *) Converted the lhash library to the crypto/mem.c functions.
8995 *) Added more recognized ASN.1 object ids.
8998 *) Added more RSA padding checks for SSL/TLS.
9001 *) Added BIO proxy/filter functionality.
9004 *) Added extra_certs to SSL_CTX which can be used
9005 send extra CA certificates to the client in the CA cert chain sending
9006 process. It can be configured with SSL_CTX_add_extra_chain_cert().
9009 *) Now Fortezza is denied in the authentication phase because
9010 this is key exchange mechanism is not supported by SSLeay at all.
9013 *) Additional PKCS1 checks.
9016 *) Support the string "TLSv1" for all TLS v1 ciphers.
9019 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
9020 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
9023 *) Fixed a few memory leaks.
9026 *) Fixed various code and comment typos.
9029 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
9030 bytes sent in the client random.
9031 [Edward Bishop <ebishop@spyglass.com>]