5 Changes between 1.0.0s and 1.0.0t [xx XXX xxxx]
9 Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
11 *) Malformed ECParameters causes infinite loop
13 When processing an ECParameters structure OpenSSL enters an infinite loop
14 if the curve specified is over a specially malformed binary polynomial
17 This can be used to perform denial of service against any
18 system which processes public keys, certificate requests or
19 certificates. This includes TLS clients and TLS servers with
20 client authentication enabled.
22 This issue was reported to OpenSSL by Joseph Barr-Pixton.
26 *) Exploitable out-of-bounds read in X509_cmp_time
28 X509_cmp_time does not properly check the length of the ASN1_TIME
29 string and can read a few bytes out of bounds. In addition,
30 X509_cmp_time accepts an arbitrary number of fractional seconds in the
33 An attacker can use this to craft malformed certificates and CRLs of
34 various sizes and potentially cause a segmentation fault, resulting in
35 a DoS on applications that verify certificates or CRLs. TLS clients
36 that verify CRLs are affected. TLS clients and servers with client
37 authentication enabled may be affected if they use custom verification
40 This issue was reported to OpenSSL by Robert Swiecki (Google), and
41 independently by Hanno Böck.
45 *) PKCS7 crash with missing EnvelopedContent
47 The PKCS#7 parsing code does not handle missing inner EncryptedContent
48 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
49 with missing content and trigger a NULL pointer dereference on parsing.
51 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
52 structures from untrusted sources are affected. OpenSSL clients and
53 servers are not affected.
55 This issue was reported to OpenSSL by Michal Zalewski (Google).
59 *) CMS verify infinite loop with unknown hash function
61 When verifying a signedData message the CMS code can enter an infinite loop
62 if presented with an unknown hash function OID. This can be used to perform
63 denial of service against any system which verifies signedData messages using
65 This issue was reported to OpenSSL by Johannes Bauer.
69 *) Race condition handling NewSessionTicket
71 If a NewSessionTicket is received by a multi-threaded client when attempting to
72 reuse a previous ticket then a race condition can occur potentially leading to
73 a double free of the ticket data.
77 Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
79 *) Segmentation fault in ASN1_TYPE_cmp fix
81 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
82 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
83 certificate signature algorithm consistency this can be used to crash any
84 certificate verification operation and exploited in a DoS attack. Any
85 application which performs certificate verification is vulnerable including
86 OpenSSL clients and servers which enable client authentication.
90 *) ASN.1 structure reuse memory corruption fix
92 Reusing a structure in ASN.1 parsing may allow an attacker to cause
93 memory corruption via an invalid write. Such reuse is and has been
94 strongly discouraged and is believed to be rare.
96 Applications that parse structures containing CHOICE or ANY DEFINED BY
97 components may be affected. Certificate parsing (d2i_X509 and related
98 functions) are however not affected. OpenSSL clients and servers are
103 *) PKCS7 NULL pointer dereferences fix
105 The PKCS#7 parsing code does not handle missing outer ContentInfo
106 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
107 missing content and trigger a NULL pointer dereference on parsing.
109 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
110 otherwise parse PKCS#7 structures from untrusted sources are
111 affected. OpenSSL clients and servers are not affected.
113 This issue was reported to OpenSSL by Michal Zalewski (Google).
117 *) DoS via reachable assert in SSLv2 servers fix
119 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
120 servers that both support SSLv2 and enable export cipher suites by sending
121 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
123 This issue was discovered by Sean Burford (Google) and Emilia Käsper
124 (OpenSSL development team).
128 *) Use After Free following d2i_ECPrivatekey error fix
130 A malformed EC private key file consumed via the d2i_ECPrivateKey function
131 could cause a use after free condition. This, in turn, could cause a double
132 free in several private key parsing functions (such as d2i_PrivateKey
133 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
134 for applications that receive EC private keys from untrusted
135 sources. This scenario is considered rare.
137 This issue was discovered by the BoringSSL project and fixed in their
142 *) X509_to_X509_REQ NULL pointer deref fix
144 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
145 the certificate key is invalid. This function is rarely used in practice.
147 This issue was discovered by Brian Carpenter.
151 *) Removed the export ciphers from the DEFAULT ciphers
154 Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
156 *) Build fixes for the Windows and OpenVMS platforms
157 [Matt Caswell and Richard Levitte]
159 Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
161 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
162 message can cause a segmentation fault in OpenSSL due to a NULL pointer
163 dereference. This could lead to a Denial Of Service attack. Thanks to
164 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
168 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
169 dtls1_buffer_record function under certain conditions. In particular this
170 could occur if an attacker sent repeated DTLS records with the same
171 sequence number but for the next epoch. The memory leak could be exploited
172 by an attacker in a Denial of Service attack through memory exhaustion.
173 Thanks to Chris Mueller for reporting this issue.
177 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
178 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
179 method would be set to NULL which could later result in a NULL pointer
180 dereference. Thanks to Frank Schmirler for reporting this issue.
184 *) Abort handshake if server key exchange message is omitted for ephemeral
187 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
188 reporting this issue.
192 *) Remove non-export ephemeral RSA code on client and server. This code
193 violated the TLS standard by allowing the use of temporary RSA keys in
194 non-export ciphersuites and could be used by a server to effectively
195 downgrade the RSA key length used to a value smaller than the server
196 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
197 INRIA or reporting this issue.
201 *) Fixed issue where DH client certificates are accepted without verification.
202 An OpenSSL server will accept a DH certificate for client authentication
203 without the certificate verify message. This effectively allows a client to
204 authenticate without the use of a private key. This only affects servers
205 which trust a client certificate authority which issues certificates
206 containing DH keys: these are extremely rare and hardly ever encountered.
207 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
212 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
213 results on some platforms, including x86_64. This bug occurs at random
214 with a very low probability, and is not known to be exploitable in any
215 way, though its exact impact is difficult to determine. Thanks to Pieter
216 Wuille (Blockstream) who reported this issue and also suggested an initial
217 fix. Further analysis was conducted by the OpenSSL development team and
218 Adam Langley of Google. The final fix was developed by Andy Polyakov of
219 the OpenSSL core team.
223 *) Fix various certificate fingerprint issues.
225 By using non-DER or invalid encodings outside the signed portion of a
226 certificate the fingerprint can be changed without breaking the signature.
227 Although no details of the signed portion of the certificate can be changed
228 this can cause problems with some applications: e.g. those using the
229 certificate fingerprint for blacklists.
231 1. Reject signatures with non zero unused bits.
233 If the BIT STRING containing the signature has non zero unused bits reject
234 the signature. All current signature algorithms require zero unused bits.
236 2. Check certificate algorithm consistency.
238 Check the AlgorithmIdentifier inside TBS matches the one in the
239 certificate signature. NB: this will result in signature failure
240 errors for some broken certificates.
242 Thanks to Konrad Kraszewski from Google for reporting this issue.
244 3. Check DSA/ECDSA signatures use DER.
246 Reencode DSA/ECDSA signatures and compare with the original received
247 signature. Return an error if there is a mismatch.
249 This will reject various cases including garbage after signature
250 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
251 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
252 (negative or with leading zeroes).
254 Further analysis was conducted and fixes were developed by Stephen Henson
255 of the OpenSSL core team.
260 Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
262 *) Session Ticket Memory Leak.
264 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
265 integrity of that ticket is first verified. In the event of a session
266 ticket integrity check failing, OpenSSL will fail to free memory
267 causing a memory leak. By sending a large number of invalid session
268 tickets an attacker could exploit this issue in a Denial Of Service
273 *) Build option no-ssl3 is incomplete.
275 When OpenSSL is configured with "no-ssl3" as a build option, servers
276 could accept and complete a SSL 3.0 handshake, and clients could be
277 configured to send them.
279 [Akamai and the OpenSSL team]
281 *) Add support for TLS_FALLBACK_SCSV.
282 Client applications doing fallback retries should call
283 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
285 [Adam Langley, Bodo Moeller]
287 *) Add additional DigestInfo checks.
289 Reencode DigestInto in DER and check against the original when
290 verifying RSA signature: this will reject any improperly encoded
291 DigestInfo structures.
293 Note: this is a precautionary measure and no attacks are currently known.
297 Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
299 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
300 to a denial of service attack. A malicious server can crash the client
301 with a null pointer dereference (read) by specifying an anonymous (EC)DH
302 ciphersuite and sending carefully crafted handshake messages.
304 Thanks to Felix Gröbert (Google) for discovering and researching this
309 *) By sending carefully crafted DTLS packets an attacker could cause openssl
310 to leak memory. This can be exploited through a Denial of Service attack.
311 Thanks to Adam Langley for discovering and researching this issue.
315 *) An attacker can force openssl to consume large amounts of memory whilst
316 processing DTLS handshake messages. This can be exploited through a
317 Denial of Service attack.
318 Thanks to Adam Langley for discovering and researching this issue.
322 *) An attacker can force an error condition which causes openssl to crash
323 whilst processing DTLS packets due to memory being freed twice. This
324 can be exploited through a Denial of Service attack.
325 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
330 *) If a multithreaded client connects to a malicious server using a resumed
331 session and the server sends an ec point format extension it could write
332 up to 255 bytes to freed memory.
334 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
339 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
340 X509_name_oneline, X509_name_print_ex et al. to leak some information
341 from the stack. Applications may be affected if they echo pretty printing
342 output to the attacker.
344 Thanks to Ivan Fratric (Google) for discovering this issue.
346 [Emilia Käsper, and Steve Henson]
348 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
349 for corner cases. (Certain input points at infinity could lead to
350 bogus results, with non-infinity inputs mapped to infinity too.)
353 Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
355 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
356 handshake can force the use of weak keying material in OpenSSL
357 SSL/TLS clients and servers.
359 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
360 researching this issue. (CVE-2014-0224)
361 [KIKUCHI Masashi, Steve Henson]
363 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
364 OpenSSL DTLS client the code can be made to recurse eventually crashing
367 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
369 [Imre Rad, Steve Henson]
371 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
372 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
373 client or server. This is potentially exploitable to run arbitrary
374 code on a vulnerable client or server.
376 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
377 [Jüri Aedla, Steve Henson]
379 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
380 are subject to a denial of service attack.
382 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
383 this issue. (CVE-2014-3470)
384 [Felix Gröbert, Ivan Fratric, Steve Henson]
386 *) Harmonize version and its documentation. -f flag is used to display
388 [mancha <mancha1@zoho.com>]
390 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
392 [mancha <mancha1@zoho.com>]
394 *) Fix some double frees. These are not thought to be exploitable.
395 [mancha <mancha1@zoho.com>]
397 *) Fix for the attack described in the paper "Recovering OpenSSL
398 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
399 by Yuval Yarom and Naomi Benger. Details can be obtained from:
400 http://eprint.iacr.org/2014/140
402 Thanks to Yuval Yarom and Naomi Benger for discovering this
403 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
404 [Yuval Yarom and Naomi Benger]
406 Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
408 *) Keep original DTLS digest and encryption contexts in retransmission
409 structures so we can use the previous session parameters if they need
410 to be resent. (CVE-2013-6450)
413 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
414 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
415 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
416 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
417 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
418 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
419 [Rob Stradling, Adam Langley]
421 Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
423 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
425 This addresses the flaw in CBC record processing discovered by
426 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
427 at: http://www.isg.rhul.ac.uk/tls/
429 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
430 Security Group at Royal Holloway, University of London
431 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
432 Emilia Käsper for the initial patch.
434 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
436 *) Return an error when checking OCSP signatures when key is NULL.
437 This fixes a DoS attack. (CVE-2013-0166)
440 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
441 the right response is stapled. Also change SSL_get_certificate()
442 so it returns the certificate actually sent.
443 See http://rt.openssl.org/Ticket/Display.html?id=2836.
445 [Rob Stradling <rob.stradling@comodo.com>]
447 *) Fix possible deadlock when decoding public keys.
450 Changes between 1.0.0i and 1.0.0j [10 May 2012]
452 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
455 *) Sanity check record length before skipping explicit IV in DTLS
458 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
459 fuzzing as a service testing platform.
463 *) Initialise tkeylen properly when encrypting CMS messages.
464 Thanks to Solar Designer of Openwall for reporting this issue.
467 Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
469 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
470 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
471 in CRYPTO_realloc_clean.
473 Thanks to Tavis Ormandy, Google Security Team, for discovering this
474 issue and to Adam Langley <agl@chromium.org> for fixing it.
476 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
478 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
480 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
481 in CMS and PKCS7 code. When RSA decryption fails use a random key for
482 content decryption and always return the same error. Note: this attack
483 needs on average 2^20 messages so it only affects automated senders. The
484 old behaviour can be reenabled in the CMS code by setting the
485 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
486 an MMA defence is not necessary.
487 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
488 this issue. (CVE-2012-0884)
491 *) Fix CVE-2011-4619: make sure we really are receiving a
492 client hello before rejecting multiple SGC restarts. Thanks to
493 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
496 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
498 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
499 Thanks to Antonio Martin, Enterprise Secure Access Research and
500 Development, Cisco Systems, Inc. for discovering this bug and
501 preparing a fix. (CVE-2012-0050)
504 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
506 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
507 of the Vaudenay padding oracle attack on CBC mode encryption
508 which enables an efficient plaintext recovery attack against
509 the OpenSSL implementation of DTLS. Their attack exploits timing
510 differences arising during decryption processing. A research
511 paper describing this attack can be found at:
512 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
513 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
514 Security Group at Royal Holloway, University of London
515 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
516 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
517 for preparing the fix. (CVE-2011-4108)
518 [Robin Seggelmann, Michael Tuexen]
520 *) Clear bytes used for block padding of SSL 3.0 records.
522 [Adam Langley (Google)]
524 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
525 Kadianakis <desnacked@gmail.com> for discovering this issue and
526 Adam Langley for preparing the fix. (CVE-2011-4619)
527 [Adam Langley (Google)]
529 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
530 [Andrey Kulikov <amdeich@gmail.com>]
532 *) Prevent malformed RFC3779 data triggering an assertion failure.
533 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
534 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
535 [Rob Austein <sra@hactrn.net>]
537 *) Improved PRNG seeding for VOS.
538 [Paul Green <Paul.Green@stratus.com>]
540 *) Fix ssl_ciph.c set-up race.
541 [Adam Langley (Google)]
543 *) Fix spurious failures in ecdsatest.c.
544 [Emilia Käsper (Google)]
546 *) Fix the BIO_f_buffer() implementation (which was mixing different
547 interpretations of the '..._len' fields).
548 [Adam Langley (Google)]
550 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
551 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
552 threads won't reuse the same blinding coefficients.
554 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
555 lock to call BN_BLINDING_invert_ex, and avoids one use of
556 BN_BLINDING_update for each BN_BLINDING structure (previously,
557 the last update always remained unused).
558 [Emilia Käsper (Google)]
560 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
561 [Bob Buckholz (Google)]
563 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
565 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
566 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
567 [Kaspar Brand <ossl@velox.ch>]
569 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
570 for multi-threaded use of ECDH. (CVE-2011-3210)
571 [Adam Langley (Google)]
573 *) Fix x509_name_ex_d2i memory leak on bad inputs.
576 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
577 signature public key algorithm by using OID xref utilities instead.
578 Before this you could only use some ECC ciphersuites with SHA1 only.
581 *) Add protection against ECDSA timing attacks as mentioned in the paper
582 by Billy Bob Brumley and Nicola Tuveri, see:
584 http://eprint.iacr.org/2011/232.pdf
586 [Billy Bob Brumley and Nicola Tuveri]
588 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
590 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
591 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
593 *) Fix bug in string printing code: if *any* escaping is enabled we must
594 escape the escape character (backslash) or the resulting string is
598 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
600 *) Disable code workaround for ancient and obsolete Netscape browsers
601 and servers: an attacker can use it in a ciphersuite downgrade attack.
602 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
605 *) Fixed J-PAKE implementation error, originally discovered by
606 Sebastien Martini, further info and confirmation from Stefan
607 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
610 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
612 *) Fix extension code to avoid race conditions which can result in a buffer
613 overrun vulnerability: resumed sessions must not be modified as they can
614 be shared by multiple threads. CVE-2010-3864
617 *) Fix WIN32 build system to correctly link an ENGINE directory into
621 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
623 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
625 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
627 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
629 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
630 context. The operation can be customised via the ctrl mechanism in
631 case ENGINEs want to include additional functionality.
634 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
637 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
638 output hashes compatible with older versions of OpenSSL.
639 [Willy Weisz <weisz@vcpc.univie.ac.at>]
641 *) Fix compression algorithm handling: if resuming a session use the
642 compression algorithm of the resumed session instead of determining
643 it from client hello again. Don't allow server to change algorithm.
646 *) Add load_crls() function to apps tidying load_certs() too. Add option
647 to verify utility to allow additional CRLs to be included.
650 *) Update OCSP request code to permit adding custom headers to the request:
651 some responders need this.
654 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
656 [Julia Lawall <julia@diku.dk>]
658 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
659 needlessly dereferenced structures, used obsolete functions and
660 didn't handle all updated verify codes correctly.
663 *) Disable MD2 in the default configuration.
666 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
667 indicate the initial BIO being pushed or popped. This makes it possible
668 to determine whether the BIO is the one explicitly called or as a result
669 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
670 it handles reference counts correctly and doesn't zero out the I/O bio
671 when it is not being explicitly popped. WARNING: applications which
672 included workarounds for the old buggy behaviour will need to be modified
673 or they could free up already freed BIOs.
676 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
677 renaming to all platforms (within the 0.9.8 branch, this was
678 done conditionally on Netware platforms to avoid a name clash).
679 [Guenter <lists@gknw.net>]
681 *) Add ECDHE and PSK support to DTLS.
682 [Michael Tuexen <tuexen@fh-muenster.de>]
684 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
688 *) Add "missing" function EVP_MD_flags() (without this the only way to
689 retrieve a digest flags is by accessing the structure directly. Update
690 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
691 or cipher is registered as in the "from" argument. Print out all
692 registered digests in the dgst usage message instead of manually
693 attempting to work them out.
696 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
697 this allows the use of compression and extensions. Change default cipher
698 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
699 by default unless an application cipher string requests it.
702 *) Alter match criteria in PKCS12_parse(). It used to try to use local
703 key ids to find matching certificates and keys but some PKCS#12 files
704 don't follow the (somewhat unwritten) rules and this strategy fails.
705 Now just gather all certificates together and the first private key
706 then look for the first certificate that matches the key.
709 *) Support use of registered digest and cipher names for dgst and cipher
710 commands instead of having to add each one as a special case. So now
717 openssl dgst -sha256 foo
719 and this works for ENGINE based algorithms too.
723 *) Update Gost ENGINE to support parameter files.
724 [Victor B. Wagner <vitus@cryptocom.ru>]
726 *) Support GeneralizedTime in ca utility.
727 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
729 *) Enhance the hash format used for certificate directory links. The new
730 form uses the canonical encoding (meaning equivalent names will work
731 even if they aren't identical) and uses SHA1 instead of MD5. This form
732 is incompatible with the older format and as a result c_rehash should
733 be used to rebuild symbolic links.
736 *) Make PKCS#8 the default write format for private keys, replacing the
737 traditional format. This form is standardised, more secure and doesn't
738 include an implicit MD5 dependency.
741 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
742 committed to OpenSSL should pass this lot as a minimum.
745 *) Add session ticket override functionality for use by EAP-FAST.
746 [Jouni Malinen <j@w1.fi>]
748 *) Modify HMAC functions to return a value. Since these can be implemented
749 in an ENGINE errors can occur.
752 *) Type-checked OBJ_bsearch_ex.
755 *) Type-checked OBJ_bsearch. Also some constification necessitated
756 by type-checking. Still to come: TXT_DB, bsearch(?),
757 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
761 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
762 seconds to a tm structure directly, instead of going through OS
763 specific date routines. This avoids any issues with OS routines such
764 as the year 2038 bug. New *_adj() functions for ASN1 time structures
765 and X509_time_adj_ex() to cover the extended range. The existing
766 X509_time_adj() is still usable and will no longer have any date issues.
769 *) Delta CRL support. New use deltas option which will attempt to locate
770 and search any appropriate delta CRLs available.
772 This work was sponsored by Google.
775 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
776 code and add additional score elements. Validate alternate CRL paths
777 as part of the CRL checking and indicate a new error "CRL path validation
778 error" in this case. Applications wanting additional details can use
779 the verify callback and check the new "parent" field. If this is not
780 NULL CRL path validation is taking place. Existing applications wont
781 see this because it requires extended CRL support which is off by
784 This work was sponsored by Google.
787 *) Support for freshest CRL extension.
789 This work was sponsored by Google.
792 *) Initial indirect CRL support. Currently only supported in the CRLs
793 passed directly and not via lookup. Process certificate issuer
794 CRL entry extension and lookup CRL entries by bother issuer name
795 and serial number. Check and process CRL issuer entry in IDP extension.
797 This work was sponsored by Google.
800 *) Add support for distinct certificate and CRL paths. The CRL issuer
801 certificate is validated separately in this case. Only enabled if
802 an extended CRL support flag is set: this flag will enable additional
803 CRL functionality in future.
805 This work was sponsored by Google.
808 *) Add support for policy mappings extension.
810 This work was sponsored by Google.
813 *) Fixes to pathlength constraint, self issued certificate handling,
814 policy processing to align with RFC3280 and PKITS tests.
816 This work was sponsored by Google.
819 *) Support for name constraints certificate extension. DN, email, DNS
820 and URI types are currently supported.
822 This work was sponsored by Google.
825 *) To cater for systems that provide a pointer-based thread ID rather
826 than numeric, deprecate the current numeric thread ID mechanism and
827 replace it with a structure and associated callback type. This
828 mechanism allows a numeric "hash" to be extracted from a thread ID in
829 either case, and on platforms where pointers are larger than 'long',
830 mixing is done to help ensure the numeric 'hash' is usable even if it
831 can't be guaranteed unique. The default mechanism is to use "&errno"
832 as a pointer-based thread ID to distinguish between threads.
834 Applications that want to provide their own thread IDs should now use
835 CRYPTO_THREADID_set_callback() to register a callback that will call
836 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
838 Note that ERR_remove_state() is now deprecated, because it is tied
839 to the assumption that thread IDs are numeric. ERR_remove_state(0)
840 to free the current thread's error state should be replaced by
841 ERR_remove_thread_state(NULL).
843 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
844 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
845 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
846 application was previously providing a numeric thread callback that
847 was inappropriate for distinguishing threads, then uniqueness might
848 have been obtained with &errno that happened immediately in the
849 intermediate development versions of OpenSSL; this is no longer the
850 case, the numeric thread callback will now override the automatic use
852 [Geoff Thorpe, with help from Bodo Moeller]
854 *) Initial support for different CRL issuing certificates. This covers a
855 simple case where the self issued certificates in the chain exist and
856 the real CRL issuer is higher in the existing chain.
858 This work was sponsored by Google.
861 *) Removed effectively defunct crypto/store from the build.
864 *) Revamp of STACK to provide stronger type-checking. Still to come:
865 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
866 ASN1_STRING, CONF_VALUE.
869 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
870 RAM on SSL connections. This option can save about 34k per idle SSL.
873 *) Revamp of LHASH to provide stronger type-checking. Still to come:
874 STACK, TXT_DB, bsearch, qsort.
877 *) Initial support for Cryptographic Message Syntax (aka CMS) based
878 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
879 support for data, signedData, compressedData, digestedData and
880 encryptedData, envelopedData types included. Scripts to check against
881 RFC4134 examples draft and interop and consistency checks of many
882 content types and variants.
885 *) Add options to enc utility to support use of zlib compression BIO.
888 *) Extend mk1mf to support importing of options and assembly language
889 files from Configure script, currently only included in VC-WIN32.
890 The assembly language rules can now optionally generate the source
891 files from the associated perl scripts.
894 *) Implement remaining functionality needed to support GOST ciphersuites.
895 Interop testing has been performed using CryptoPro implementations.
896 [Victor B. Wagner <vitus@cryptocom.ru>]
898 *) s390x assembler pack.
901 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
905 *) Implement Opaque PRF Input TLS extension as specified in
906 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
907 official specification yet and no extension type assignment by
908 IANA exists, this extension (for now) will have to be explicitly
909 enabled when building OpenSSL by providing the extension number
910 to use. For example, specify an option
912 -DTLSEXT_TYPE_opaque_prf_input=0x9527
914 to the "config" or "Configure" script to enable the extension,
915 assuming extension number 0x9527 (which is a completely arbitrary
916 and unofficial assignment based on the MD5 hash of the Internet
917 Draft). Note that by doing so, you potentially lose
918 interoperability with other TLS implementations since these might
919 be using the same extension number for other purposes.
921 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
922 opaque PRF input value to use in the handshake. This will create
923 an interal copy of the length-'len' string at 'src', and will
924 return non-zero for success.
926 To get more control and flexibility, provide a callback function
929 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
930 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
934 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
937 Callback function 'cb' will be called in handshakes, and is
938 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
939 Argument 'arg' is for application purposes (the value as given to
940 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
941 be provided to the callback function). The callback function
942 has to return non-zero to report success: usually 1 to use opaque
943 PRF input just if possible, or 2 to enforce use of the opaque PRF
944 input. In the latter case, the library will abort the handshake
945 if opaque PRF input is not successfully negotiated.
947 Arguments 'peerinput' and 'len' given to the callback function
948 will always be NULL and 0 in the case of a client. A server will
949 see the client's opaque PRF input through these variables if
950 available (NULL and 0 otherwise). Note that if the server
951 provides an opaque PRF input, the length must be the same as the
952 length of the client's opaque PRF input.
954 Note that the callback function will only be called when creating
955 a new session (session resumption can resume whatever was
956 previously negotiated), and will not be called in SSL 2.0
957 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
958 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
959 for applications that need to enforce opaque PRF input.
963 *) Update ssl code to support digests other than SHA1+MD5 for handshake
966 [Victor B. Wagner <vitus@cryptocom.ru>]
968 *) Add RFC4507 support to OpenSSL. This includes the corrections in
969 RFC4507bis. The encrypted ticket format is an encrypted encoded
970 SSL_SESSION structure, that way new session features are automatically
973 If a client application caches session in an SSL_SESSION structure
974 support is transparent because tickets are now stored in the encoded
977 The SSL_CTX structure automatically generates keys for ticket
978 protection in servers so again support should be possible
979 with no application modification.
981 If a client or server wishes to disable RFC4507 support then the option
982 SSL_OP_NO_TICKET can be set.
984 Add a TLS extension debugging callback to allow the contents of any client
985 or server extensions to be examined.
987 This work was sponsored by Google.
990 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
991 OpenSSL should now compile cleanly on gcc 4.2
992 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
994 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
995 support including streaming MAC support: this is required for GOST
997 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
999 *) Add option -stream to use PKCS#7 streaming in smime utility. New
1000 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
1001 to output in BER and PEM format.
1004 *) Experimental support for use of HMAC via EVP_PKEY interface. This
1005 allows HMAC to be handled via the EVP_DigestSign*() interface. The
1006 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
1007 ENGINE support for HMAC keys which are unextractable. New -mac and
1008 -macopt options to dgst utility.
1011 *) New option -sigopt to dgst utility. Update dgst to use
1012 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
1013 alternative signing paramaters such as X9.31 or PSS in the dgst
1017 *) Change ssl_cipher_apply_rule(), the internal function that does
1018 the work each time a ciphersuite string requests enabling
1019 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
1020 removing ("!foo+bar") a class of ciphersuites: Now it maintains
1021 the order of disabled ciphersuites such that those ciphersuites
1022 that most recently went from enabled to disabled not only stay
1023 in order with respect to each other, but also have higher priority
1024 than other disabled ciphersuites the next time ciphersuites are
1027 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
1028 the same ciphersuites as with "HIGH" alone, but in a specific
1029 order where the PSK ciphersuites come first (since they are the
1030 most recently disabled ciphersuites when "HIGH" is parsed).
1032 Also, change ssl_create_cipher_list() (using this new
1033 funcionality) such that between otherwise identical
1034 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
1038 *) Change ssl_create_cipher_list() so that it automatically
1039 arranges the ciphersuites in reasonable order before starting
1040 to process the rule string. Thus, the definition for "DEFAULT"
1041 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
1042 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
1043 This makes it much easier to arrive at a reasonable default order
1044 in applications for which anonymous ciphers are OK (meaning
1045 that you can't actually use DEFAULT).
1046 [Bodo Moeller; suggested by Victor Duchovni]
1048 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
1049 processing) into multiple integers instead of setting
1050 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
1051 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
1052 (These masks as well as the individual bit definitions are hidden
1053 away into the non-exported interface ssl/ssl_locl.h, so this
1054 change to the definition of the SSL_CIPHER structure shouldn't
1055 affect applications.) This give us more bits for each of these
1056 categories, so there is no longer a need to coagulate AES128 and
1057 AES256 into a single algorithm bit, and to coagulate Camellia128
1058 and Camellia256 into a single algorithm bit, which has led to all
1061 Thus, among other things, the kludge introduced in 0.9.7m and
1062 0.9.8e for masking out AES256 independently of AES128 or masking
1063 out Camellia256 independently of AES256 is not needed here in 0.9.9.
1065 With the change, we also introduce new ciphersuite aliases that
1066 so far were missing: "AES128", "AES256", "CAMELLIA128", and
1070 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
1071 Use the leftmost N bytes of the signature input if the input is
1072 larger than the prime q (with N being the size in bytes of q).
1075 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
1076 it yet and it is largely untested.
1079 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
1082 *) Initial incomplete changes to avoid need for function casts in OpenSSL
1083 some compilers (gcc 4.2 and later) reject their use. Safestack is
1084 reimplemented. Update ASN1 to avoid use of legacy functions.
1087 *) Win32/64 targets are linked with Winsock2.
1090 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
1091 to external functions. This can be used to increase CRL handling
1092 efficiency especially when CRLs are very large by (for example) storing
1093 the CRL revoked certificates in a database.
1096 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
1097 new CRLs added to a directory can be used. New command line option
1098 -verify_return_error to s_client and s_server. This causes real errors
1099 to be returned by the verify callback instead of carrying on no matter
1100 what. This reflects the way a "real world" verify callback would behave.
1103 *) GOST engine, supporting several GOST algorithms and public key formats.
1104 Kindly donated by Cryptocom.
1107 *) Partial support for Issuing Distribution Point CRL extension. CRLs
1108 partitioned by DP are handled but no indirect CRL or reason partitioning
1109 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
1110 selected via a scoring technique which handles IDP and AKID in CRLs.
1113 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
1114 will ultimately be used for all verify operations: this will remove the
1115 X509_STORE dependency on certificate verification and allow alternative
1116 lookup methods. X509_STORE based implementations of these two callbacks.
1119 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
1120 Modify get_crl() to find a valid (unexpired) CRL if possible.
1123 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
1124 this would be called X509_CRL_cmp() but that name is already used by
1125 a function that just compares CRL issuer names. Cache several CRL
1126 extensions in X509_CRL structure and cache CRLDP in X509.
1129 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
1130 this maps equivalent X509_NAME structures into a consistent structure.
1131 Name comparison can then be performed rapidly using memcmp().
1134 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
1138 *) Allow digests to supply their own micalg string for S/MIME type using
1139 the ctrl EVP_MD_CTRL_MICALG.
1142 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
1143 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
1144 ctrl. It can then customise the structure before and/or after signing
1148 *) New function OBJ_add_sigid() to allow application defined signature OIDs
1149 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
1150 to free up any added signature OIDs.
1153 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
1154 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
1155 digest and cipher tables. New options added to openssl utility:
1156 list-message-digest-algorithms and list-cipher-algorithms.
1159 *) Change the array representation of binary polynomials: the list
1160 of degrees of non-zero coefficients is now terminated with -1.
1161 Previously it was terminated with 0, which was also part of the
1162 value; thus, the array representation was not applicable to
1163 polynomials where t^0 has coefficient zero. This change makes
1164 the array representation useful in a more general context.
1167 *) Various modifications and fixes to SSL/TLS cipher string
1168 handling. For ECC, the code now distinguishes between fixed ECDH
1169 with RSA certificates on the one hand and with ECDSA certificates
1170 on the other hand, since these are separate ciphersuites. The
1171 unused code for Fortezza ciphersuites has been removed.
1173 For consistency with EDH, ephemeral ECDH is now called "EECDH"
1174 (not "ECDHE"). For consistency with the code for DH
1175 certificates, use of ECDH certificates is now considered ECDH
1176 authentication, not RSA or ECDSA authentication (the latter is
1177 merely the CA's signing algorithm and not actively used in the
1180 The temporary ciphersuite alias "ECCdraft" is no longer
1181 available, and ECC ciphersuites are no longer excluded from "ALL"
1182 and "DEFAULT". The following aliases now exist for RFC 4492
1183 ciphersuites, most of these by analogy with the DH case:
1185 kECDHr - ECDH cert, signed with RSA
1186 kECDHe - ECDH cert, signed with ECDSA
1187 kECDH - ECDH cert (signed with either RSA or ECDSA)
1188 kEECDH - ephemeral ECDH
1189 ECDH - ECDH cert or ephemeral ECDH
1195 AECDH - anonymous ECDH
1196 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
1200 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
1201 Use correct micalg parameters depending on digest(s) in signed message.
1204 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
1205 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
1208 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
1209 an engine to register a method. Add ENGINE lookups for methods and
1210 functional reference processing.
1213 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
1214 EVP_{Sign,Verify}* which allow an application to customise the signature
1218 *) New -resign option to smime utility. This adds one or more signers
1219 to an existing PKCS#7 signedData structure. Also -md option to use an
1220 alternative message digest algorithm for signing.
1223 *) Tidy up PKCS#7 routines and add new functions to make it easier to
1224 create PKCS7 structures containing multiple signers. Update smime
1225 application to support multiple signers.
1228 *) New -macalg option to pkcs12 utility to allow setting of an alternative
1232 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
1233 Reorganize PBE internals to lookup from a static table using NIDs,
1234 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
1235 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
1236 PRF which will be automatically used with PBES2.
1239 *) Replace the algorithm specific calls to generate keys in "req" with the
1243 *) Update PKCS#7 enveloped data routines to use new API. This is now
1244 supported by any public key method supporting the encrypt operation. A
1245 ctrl is added to allow the public key algorithm to examine or modify
1246 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
1250 *) Add a ctrl to asn1 method to allow a public key algorithm to express
1251 a default digest type to use. In most cases this will be SHA1 but some
1252 algorithms (such as GOST) need to specify an alternative digest. The
1253 return value indicates how strong the prefernce is 1 means optional and
1254 2 is mandatory (that is it is the only supported type). Modify
1255 ASN1_item_sign() to accept a NULL digest argument to indicate it should
1256 use the default md. Update openssl utilities to use the default digest
1257 type for signing if it is not explicitly indicated.
1260 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
1261 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
1262 signing method from the key type. This effectively removes the link
1263 between digests and public key types.
1266 *) Add an OID cross reference table and utility functions. Its purpose is to
1267 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
1268 rsaEncryption. This will allow some of the algorithm specific hackery
1269 needed to use the correct OID to be removed.
1272 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
1273 structures for PKCS7_sign(). They are now set up by the relevant public
1277 *) Add provisional EC pkey method with support for ECDSA and ECDH.
1280 *) Add support for key derivation (agreement) in the API, DH method and
1284 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
1285 public and private key formats. As a side effect these add additional
1286 command line functionality not previously available: DSA signatures can be
1287 generated and verified using pkeyutl and DH key support and generation in
1292 [Oliver Tappe <zooey@hirschkaefer.de>]
1294 *) New make target "install_html_docs" installs HTML renditions of the
1296 [Oliver Tappe <zooey@hirschkaefer.de>]
1298 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
1299 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
1300 support key and parameter generation and add initial key generation
1301 functionality for RSA.
1304 *) Add functions for main EVP_PKEY_method operations. The undocumented
1305 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
1306 EVP_PKEY_{encrypt,decrypt}_old.
1309 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
1310 key API, doesn't do much yet.
1313 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
1314 public key algorithms. New option to openssl utility:
1315 "list-public-key-algorithms" to print out info.
1318 *) Implement the Supported Elliptic Curves Extension for
1319 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1322 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
1323 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
1326 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
1327 utilities such as rsa, dsa, dsaparam etc except they process any key
1331 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1332 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1333 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
1337 *) Initial support for pluggable public key ASN1.
1338 De-spaghettify the public key ASN1 handling. Move public and private
1339 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
1340 algorithm specific handling to a single module within the relevant
1341 algorithm directory. Add functions to allow (near) opaque processing
1342 of public and private key structures.
1345 *) Implement the Supported Point Formats Extension for
1346 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1349 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
1350 for the psk identity [hint] and the psk callback functions to the
1351 SSL_SESSION, SSL and SSL_CTX structure.
1354 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
1358 SSL_CTX_use_psk_identity_hint
1359 SSL_get_psk_identity_hint
1360 SSL_get_psk_identity
1361 SSL_use_psk_identity_hint
1363 [Mika Kousa and Pasi Eronen of Nokia Corporation]
1365 *) Add RFC 3161 compliant time stamp request creation, response generation
1366 and response verification functionality.
1367 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
1369 *) Add initial support for TLS extensions, specifically for the server_name
1370 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1371 have new members for a host name. The SSL data structure has an
1372 additional member SSL_CTX *initial_ctx so that new sessions can be
1373 stored in that context to allow for session resumption, even after the
1374 SSL has been switched to a new SSL_CTX in reaction to a client's
1375 server_name extension.
1377 New functions (subject to change):
1379 SSL_get_servername()
1380 SSL_get_servername_type()
1383 New CTRL codes and macros (subject to change):
1385 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1386 - SSL_CTX_set_tlsext_servername_callback()
1387 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1388 - SSL_CTX_set_tlsext_servername_arg()
1389 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1391 openssl s_client has a new '-servername ...' option.
1393 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1394 '-key2 ...', '-servername_fatal' (subject to change). This allows
1395 testing the HostName extension for a specific single host name ('-cert'
1396 and '-key' remain fallbacks for handshakes without HostName
1397 negotiation). If the unrecogninzed_name alert has to be sent, this by
1398 default is a warning; it becomes fatal with the '-servername_fatal'
1401 [Peter Sylvester, Remy Allais, Christophe Renou]
1403 *) Whirlpool hash implementation is added.
1406 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1407 bn(64,32). Because of instruction set limitations it doesn't have
1408 any negative impact on performance. This was done mostly in order
1409 to make it possible to share assembler modules, such as bn_mul_mont
1410 implementations, between 32- and 64-bit builds without hassle.
1413 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1414 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1418 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1419 dedicated Montgomery multiplication procedure, is introduced.
1420 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1421 "64-bit" performance on certain 32-bit targets.
1424 *) New option SSL_OP_NO_COMP to disable use of compression selectively
1425 in SSL structures. New SSL ctrl to set maximum send fragment size.
1426 Save memory by seeting the I/O buffer sizes dynamically instead of
1427 using the maximum available value.
1430 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
1431 in addition to the text details.
1434 *) Very, very preliminary EXPERIMENTAL support for printing of general
1435 ASN1 structures. This currently produces rather ugly output and doesn't
1436 handle several customised structures at all.
1439 *) Integrated support for PVK file format and some related formats such
1440 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
1441 these in the 'rsa' and 'dsa' utilities.
1444 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
1447 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
1448 place for the (very old) "NETSCAPE" format certificates which are now
1449 handled using new ASN1 code equivalents.
1452 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
1453 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
1454 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
1457 *) Modify CRL distribution points extension code to print out previously
1458 unsupported fields. Enhance extension setting code to allow setting of
1462 *) Add print and set support for Issuing Distribution Point CRL extension.
1465 *) Change 'Configure' script to enable Camellia by default.
1468 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
1470 *) When rejecting SSL/TLS records due to an incorrect version number, never
1471 update s->server with a new major version number. As of
1472 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
1473 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
1474 the previous behavior could result in a read attempt at NULL when
1475 receiving specific incorrect SSL/TLS records once record payload
1476 protection is active. (CVE-2010-0740)
1477 [Bodo Moeller, Adam Langley <agl@chromium.org>]
1479 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
1480 could be crashed if the relevant tables were not present (e.g. chrooted).
1481 [Tomas Hoger <thoger@redhat.com>]
1483 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
1485 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
1486 [Martin Olsson, Neel Mehta]
1488 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
1489 accommodate for stack sorting, always a write lock!).
1492 *) On some versions of WIN32 Heap32Next is very slow. This can cause
1493 excessive delays in the RAND_poll(): over a minute. As a workaround
1494 include a time check in the inner Heap32Next loop too.
1497 *) The code that handled flushing of data in SSL/TLS originally used the
1498 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
1499 the problem outlined in PR#1949. The fix suggested there however can
1500 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
1501 of Apache). So instead simplify the code to flush unconditionally.
1502 This should be fine since flushing with no data to flush is a no op.
1505 *) Handle TLS versions 2.0 and later properly and correctly use the
1506 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
1507 off ancient servers have a habit of sticking around for a while...
1510 *) Modify compression code so it frees up structures without using the
1511 ex_data callbacks. This works around a problem where some applications
1512 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
1513 restarting) then use compression (e.g. SSL with compression) later.
1514 This results in significant per-connection memory leaks and
1515 has caused some security issues including CVE-2008-1678 and
1519 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
1520 change when encrypting or decrypting.
1523 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
1524 connect and renegotiate with servers which do not support RI.
1525 Until RI is more widely deployed this option is enabled by default.
1528 *) Add "missing" ssl ctrls to clear options and mode.
1531 *) If client attempts to renegotiate and doesn't support RI respond with
1532 a no_renegotiation alert as required by RFC5746. Some renegotiating
1533 TLS clients will continue a connection gracefully when they receive
1534 the alert. Unfortunately OpenSSL mishandled this alert and would hang
1535 waiting for a server hello which it will never receive. Now we treat a
1536 received no_renegotiation alert as a fatal error. This is because
1537 applications requesting a renegotiation might well expect it to succeed
1538 and would have no code in place to handle the server denying it so the
1539 only safe thing to do is to terminate the connection.
1542 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
1543 peer supports secure renegotiation and 0 otherwise. Print out peer
1544 renegotiation support in s_client/s_server.
1547 *) Replace the highly broken and deprecated SPKAC certification method with
1548 the updated NID creation version. This should correctly handle UTF8.
1551 *) Implement RFC5746. Re-enable renegotiation but require the extension
1552 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
1553 turns out to be a bad idea. It has been replaced by
1554 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
1555 SSL_CTX_set_options(). This is really not recommended unless you
1556 know what you are doing.
1557 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
1559 *) Fixes to stateless session resumption handling. Use initial_ctx when
1560 issuing and attempting to decrypt tickets in case it has changed during
1561 servername handling. Use a non-zero length session ID when attempting
1562 stateless session resumption: this makes it possible to determine if
1563 a resumption has occurred immediately after receiving server hello
1564 (several places in OpenSSL subtly assume this) instead of later in
1568 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
1569 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
1570 fixes for a few places where the return code is not checked
1572 [Julia Lawall <julia@diku.dk>]
1574 *) Add --strict-warnings option to Configure script to include devteam
1575 warnings in other configurations.
1578 *) Add support for --libdir option and LIBDIR variable in makefiles. This
1579 makes it possible to install openssl libraries in locations which
1580 have names other than "lib", for example "/usr/lib64" which some
1582 [Steve Henson, based on patch from Jeremy Utley]
1584 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
1585 X690 8.9.12 and can produce some misleading textual output of OIDs.
1586 [Steve Henson, reported by Dan Kaminsky]
1588 *) Delete MD2 from algorithm tables. This follows the recommendation in
1589 several standards that it is not used in new applications due to
1590 several cryptographic weaknesses. For binary compatibility reasons
1591 the MD2 API is still compiled in by default.
1594 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
1598 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
1599 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
1601 [Guenter <lists@gknw.net>]
1603 *) Fix the server certificate chain building code to use X509_verify_cert(),
1604 it used to have an ad-hoc builder which was unable to cope with anything
1605 other than a simple chain.
1606 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
1608 *) Don't check self signed certificate signatures in X509_verify_cert()
1609 by default (a flag can override this): it just wastes time without
1610 adding any security. As a useful side effect self signed root CAs
1611 with non-FIPS digests are now usable in FIPS mode.
1614 *) In dtls1_process_out_of_seq_message() the check if the current message
1615 is already buffered was missing. For every new message was memory
1616 allocated, allowing an attacker to perform an denial of service attack
1617 with sending out of seq handshake messages until there is no memory
1618 left. Additionally every future messege was buffered, even if the
1619 sequence number made no sense and would be part of another handshake.
1620 So only messages with sequence numbers less than 10 in advance will be
1621 buffered. (CVE-2009-1378)
1622 [Robin Seggelmann, discovered by Daniel Mentz]
1624 *) Records are buffered if they arrive with a future epoch to be
1625 processed after finishing the corresponding handshake. There is
1626 currently no limitation to this buffer allowing an attacker to perform
1627 a DOS attack with sending records with future epochs until there is no
1628 memory left. This patch adds the pqueue_size() function to detemine
1629 the size of a buffer and limits the record buffer to 100 entries.
1631 [Robin Seggelmann, discovered by Daniel Mentz]
1633 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
1634 parent structure is freed. (CVE-2009-1379)
1637 *) Handle non-blocking I/O properly in SSL_shutdown() call.
1638 [Darryl Miles <darryl-mailinglists@netbauds.net>]
1641 [Ilya O. <vrghost@gmail.com>]
1643 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
1645 *) Disable renegotiation completely - this fixes a severe security
1646 problem (CVE-2009-3555) at the cost of breaking all
1647 renegotiation. Renegotiation can be re-enabled by setting
1648 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
1649 run-time. This is really not recommended unless you know what
1653 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
1655 *) Don't set val to NULL when freeing up structures, it is freed up by
1656 underlying code. If sizeof(void *) > sizeof(long) this can result in
1657 zeroing past the valid field. (CVE-2009-0789)
1658 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
1660 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
1661 checked correctly. This would allow some invalid signed attributes to
1662 appear to verify correctly. (CVE-2009-0591)
1663 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1665 *) Reject UniversalString and BMPString types with invalid lengths. This
1666 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
1667 a legal length. (CVE-2009-0590)
1670 *) Set S/MIME signing as the default purpose rather than setting it
1671 unconditionally. This allows applications to override it at the store
1675 *) Permit restricted recursion of ASN1 strings. This is needed in practice
1676 to handle some structures.
1679 *) Improve efficiency of mem_gets: don't search whole buffer each time
1681 [Jeremy Shapiro <jnshapir@us.ibm.com>]
1683 *) New -hex option for openssl rand.
1686 *) Print out UTF8String and NumericString when parsing ASN1.
1689 *) Support NumericString type for name components.
1692 *) Allow CC in the environment to override the automatically chosen
1693 compiler. Note that nothing is done to ensure flags work with the
1697 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
1699 *) Properly check EVP_VerifyFinal() and similar return values
1701 [Ben Laurie, Bodo Moeller, Google Security Team]
1703 *) Enable TLS extensions by default.
1706 *) Allow the CHIL engine to be loaded, whether the application is
1707 multithreaded or not. (This does not release the developer from the
1708 obligation to set up the dynamic locking callbacks.)
1709 [Sander Temme <sander@temme.net>]
1711 *) Use correct exit code if there is an error in dgst command.
1712 [Steve Henson; problem pointed out by Roland Dirlewanger]
1714 *) Tweak Configure so that you need to say "experimental-jpake" to enable
1715 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
1718 *) Add experimental JPAKE support, including demo authentication in
1719 s_client and s_server.
1722 *) Set the comparison function in v3_addr_canonize().
1723 [Rob Austein <sra@hactrn.net>]
1725 *) Add support for XMPP STARTTLS in s_client.
1726 [Philip Paeps <philip@freebsd.org>]
1728 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
1729 to ensure that even with this option, only ciphersuites in the
1730 server's preference list will be accepted. (Note that the option
1731 applies only when resuming a session, so the earlier behavior was
1732 just about the algorithm choice for symmetric cryptography.)
1735 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
1737 *) Fix NULL pointer dereference if a DTLS server received
1738 ChangeCipherSpec as first record (CVE-2009-1386).
1741 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
1742 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
1745 *) The fix in 0.9.8c that supposedly got rid of unsafe
1746 double-checked locking was incomplete for RSA blinding,
1747 addressing just one layer of what turns out to have been
1748 doubly unsafe triple-checked locking.
1750 So now fix this for real by retiring the MONT_HELPER macro
1751 in crypto/rsa/rsa_eay.c.
1753 [Bodo Moeller; problem pointed out by Marius Schilder]
1755 *) Various precautionary measures:
1757 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
1759 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
1760 (NB: This would require knowledge of the secret session ticket key
1761 to exploit, in which case you'd be SOL either way.)
1763 - Change bn_nist.c so that it will properly handle input BIGNUMs
1764 outside the expected range.
1766 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
1769 [Neel Mehta, Bodo Moeller]
1771 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
1772 the load fails. Useful for distros.
1773 [Ben Laurie and the FreeBSD team]
1775 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
1778 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
1781 *) Expand ENGINE to support engine supplied SSL client certificate functions.
1783 This work was sponsored by Logica.
1786 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
1787 keystores. Support for SSL/TLS client authentication too.
1788 Not compiled unless enable-capieng specified to Configure.
1790 This work was sponsored by Logica.
1793 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
1794 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
1795 attribute creation routines such as certifcate requests and PKCS#12
1799 Changes between 0.9.8g and 0.9.8h [28 May 2008]
1801 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
1802 handshake which could lead to a cilent crash as found using the
1803 Codenomicon TLS test suite (CVE-2008-1672)
1804 [Steve Henson, Mark Cox]
1806 *) Fix double free in TLS server name extensions which could lead to
1807 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
1810 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
1812 Clear the error queue to ensure that error entries left from
1813 older function calls do not interfere with the correct operation.
1814 [Lutz Jaenicke, Erik de Castro Lopo]
1816 *) Remove root CA certificates of commercial CAs:
1818 The OpenSSL project does not recommend any specific CA and does not
1819 have any policy with respect to including or excluding any CA.
1820 Therefore it does not make any sense to ship an arbitrary selection
1821 of root CA certificates with the OpenSSL software.
1824 *) RSA OAEP patches to fix two separate invalid memory reads.
1825 The first one involves inputs when 'lzero' is greater than
1826 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
1827 before the beginning of from). The second one involves inputs where
1828 the 'db' section contains nothing but zeroes (there is a one-byte
1829 invalid read after the end of 'db').
1830 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1832 *) Partial backport from 0.9.9-dev:
1834 Introduce bn_mul_mont (dedicated Montgomery multiplication
1835 procedure) as a candidate for BIGNUM assembler implementation.
1836 While 0.9.9-dev uses assembler for various architectures, only
1837 x86_64 is available by default here in the 0.9.8 branch, and
1838 32-bit x86 is available through a compile-time setting.
1840 To try the 32-bit x86 assembler implementation, use Configure
1841 option "enable-montasm" (which exists only for this backport).
1843 As "enable-montasm" for 32-bit x86 disclaims code stability
1844 anyway, in this constellation we activate additional code
1845 backported from 0.9.9-dev for further performance improvements,
1846 namely BN_from_montgomery_word. (To enable this otherwise,
1847 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
1849 [Andy Polyakov (backport partially by Bodo Moeller)]
1851 *) Add TLS session ticket callback. This allows an application to set
1852 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
1853 values. This is useful for key rollover for example where several key
1854 sets may exist with different names.
1857 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
1858 This was broken until now in 0.9.8 releases, such that the only way
1859 a registered ENGINE could be used (assuming it initialises
1860 successfully on the host) was to explicitly set it as the default
1861 for the relevant algorithms. This is in contradiction with 0.9.7
1862 behaviour and the documentation. With this fix, when an ENGINE is
1863 registered into a given algorithm's table of implementations, the
1864 'uptodate' flag is reset so that auto-discovery will be used next
1865 time a new context for that algorithm attempts to select an
1867 [Ian Lister (tweaked by Geoff Thorpe)]
1869 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
1870 implemention in the following ways:
1872 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
1875 Lack of BER streaming support means one pass streaming processing is
1876 only supported if data is detached: setting the streaming flag is
1877 ignored for embedded content.
1879 CMS support is disabled by default and must be explicitly enabled
1880 with the enable-cms configuration option.
1883 *) Update the GMP engine glue to do direct copies between BIGNUM and
1884 mpz_t when openssl and GMP use the same limb size. Otherwise the
1885 existing "conversion via a text string export" trick is still used.
1886 [Paul Sheer <paulsheer@gmail.com>]
1888 *) Zlib compression BIO. This is a filter BIO which compressed and
1889 uncompresses any data passed through it.
1892 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
1893 RFC3394 compatible AES key wrapping.
1896 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
1897 sets string data without copying. X509_ALGOR_set0() and
1898 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
1899 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
1900 from an X509_ATTRIBUTE structure optionally checking it occurs only
1901 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
1905 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
1906 to get the expected BN_FLG_CONSTTIME behavior.
1907 [Bodo Moeller (Google)]
1911 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
1912 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
1913 - added some more tests to do_tests.pl
1914 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
1915 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
1916 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
1917 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
1918 - various changes to netware.pl to enable gcc-cross builds on Win32
1920 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
1921 - various changes to fix missing prototype warnings
1922 - fixed x86nasm.pl to create correct asm files for NASM COFF output
1923 - added AES, WHIRLPOOL and CPUID assembler code to build files
1924 - added missing AES assembler make rules to mk1mf.pl
1925 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
1926 [Guenter Knauf <eflash@gmx.net>]
1928 *) Implement certificate status request TLS extension defined in RFC3546.
1929 A client can set the appropriate parameters and receive the encoded
1930 OCSP response via a callback. A server can query the supplied parameters
1931 and set the encoded OCSP response in the callback. Add simplified examples
1932 to s_client and s_server.
1935 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
1937 *) Fix various bugs:
1938 + Binary incompatibility of ssl_ctx_st structure
1939 + DTLS interoperation with non-compliant servers
1940 + Don't call get_session_cb() without proposed session
1941 + Fix ia64 assembler code
1942 [Andy Polyakov, Steve Henson]
1944 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
1946 *) DTLS Handshake overhaul. There were longstanding issues with
1947 OpenSSL DTLS implementation, which were making it impossible for
1948 RFC 4347 compliant client to communicate with OpenSSL server.
1949 Unfortunately just fixing these incompatibilities would "cut off"
1950 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
1951 server keeps tolerating non RFC compliant syntax. The opposite is
1952 not true, 0.9.8f client can not communicate with earlier server.
1953 This update even addresses CVE-2007-4995.
1956 *) Changes to avoid need for function casts in OpenSSL: some compilers
1957 (gcc 4.2 and later) reject their use.
1958 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
1961 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1962 RFC4507bis. The encrypted ticket format is an encrypted encoded
1963 SSL_SESSION structure, that way new session features are automatically
1966 If a client application caches session in an SSL_SESSION structure
1967 support is transparent because tickets are now stored in the encoded
1970 The SSL_CTX structure automatically generates keys for ticket
1971 protection in servers so again support should be possible
1972 with no application modification.
1974 If a client or server wishes to disable RFC4507 support then the option
1975 SSL_OP_NO_TICKET can be set.
1977 Add a TLS extension debugging callback to allow the contents of any client
1978 or server extensions to be examined.
1980 This work was sponsored by Google.
1983 *) Add initial support for TLS extensions, specifically for the server_name
1984 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1985 have new members for a host name. The SSL data structure has an
1986 additional member SSL_CTX *initial_ctx so that new sessions can be
1987 stored in that context to allow for session resumption, even after the
1988 SSL has been switched to a new SSL_CTX in reaction to a client's
1989 server_name extension.
1991 New functions (subject to change):
1993 SSL_get_servername()
1994 SSL_get_servername_type()
1997 New CTRL codes and macros (subject to change):
1999 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2000 - SSL_CTX_set_tlsext_servername_callback()
2001 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2002 - SSL_CTX_set_tlsext_servername_arg()
2003 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2005 openssl s_client has a new '-servername ...' option.
2007 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2008 '-key2 ...', '-servername_fatal' (subject to change). This allows
2009 testing the HostName extension for a specific single host name ('-cert'
2010 and '-key' remain fallbacks for handshakes without HostName
2011 negotiation). If the unrecogninzed_name alert has to be sent, this by
2012 default is a warning; it becomes fatal with the '-servername_fatal'
2015 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
2017 *) Add AES and SSE2 assembly language support to VC++ build.
2020 *) Mitigate attack on final subtraction in Montgomery reduction.
2023 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
2024 (which previously caused an internal error).
2027 *) Squeeze another 10% out of IGE mode when in != out.
2030 *) AES IGE mode speedup.
2031 [Dean Gaudet (Google)]
2033 *) Add the Korean symmetric 128-bit cipher SEED (see
2034 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
2035 add SEED ciphersuites from RFC 4162:
2037 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
2038 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
2039 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
2040 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
2042 To minimize changes between patchlevels in the OpenSSL 0.9.8
2043 series, SEED remains excluded from compilation unless OpenSSL
2044 is configured with 'enable-seed'.
2045 [KISA, Bodo Moeller]
2047 *) Mitigate branch prediction attacks, which can be practical if a
2048 single processor is shared, allowing a spy process to extract
2049 information. For detailed background information, see
2050 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
2051 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
2052 and Necessary Software Countermeasures"). The core of the change
2053 are new versions BN_div_no_branch() and
2054 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
2055 respectively, which are slower, but avoid the security-relevant
2056 conditional branches. These are automatically called by BN_div()
2057 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
2058 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
2059 remove a conditional branch.
2061 BN_FLG_CONSTTIME is the new name for the previous
2062 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
2063 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
2064 in the exponent causes BN_mod_exp_mont() to use the alternative
2065 implementation in BN_mod_exp_mont_consttime().) The old name
2066 remains as a deprecated alias.
2068 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
2069 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
2070 constant-time implementations for more than just exponentiation.
2071 Here too the old name is kept as a deprecated alias.
2073 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2074 the BN_BLINDING structure gets an independent copy of the
2075 modulus. This means that the previous "BIGNUM *m" argument to
2076 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2077 essentially becomes "const BIGNUM *m", although we can't actually
2078 change this in the header file before 0.9.9. It allows
2079 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2080 enable BN_FLG_CONSTTIME.
2082 [Matthew D Wood (Intel Corp)]
2084 *) In the SSL/TLS server implementation, be strict about session ID
2085 context matching (which matters if an application uses a single
2086 external cache for different purposes). Previously,
2087 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2088 set. This did ensure strict client verification, but meant that,
2089 with applications using a single external cache for quite
2090 different requirements, clients could circumvent ciphersuite
2091 restrictions for a given session ID context by starting a session
2092 in a different context.
2095 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2096 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2097 authentication-only ciphersuites.
2100 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
2101 not complete and could lead to a possible single byte overflow
2102 (CVE-2007-5135) [Ben Laurie]
2104 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
2106 *) Since AES128 and AES256 (and similarly Camellia128 and
2107 Camellia256) share a single mask bit in the logic of
2108 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2109 kludge to work properly if AES128 is available and AES256 isn't
2110 (or if Camellia128 is available and Camellia256 isn't).
2113 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
2114 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
2115 When a point or a seed is encoded in a BIT STRING, we need to
2116 prevent the removal of trailing zero bits to get the proper DER
2117 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
2118 of a NamedBitList, for which trailing 0 bits need to be removed.)
2121 *) Have SSL/TLS server implementation tolerate "mismatched" record
2122 protocol version while receiving ClientHello even if the
2123 ClientHello is fragmented. (The server can't insist on the
2124 particular protocol version it has chosen before the ServerHello
2125 message has informed the client about his choice.)
2128 *) Add RFC 3779 support.
2129 [Rob Austein for ARIN, Ben Laurie]
2131 *) Load error codes if they are not already present instead of using a
2132 static variable. This allows them to be cleanly unloaded and reloaded.
2133 Improve header file function name parsing.
2136 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
2137 or CAPABILITY handshake as required by RFCs.
2140 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
2142 *) Introduce limits to prevent malicious keys being able to
2143 cause a denial of service. (CVE-2006-2940)
2144 [Steve Henson, Bodo Moeller]
2146 *) Fix ASN.1 parsing of certain invalid structures that can result
2147 in a denial of service. (CVE-2006-2937) [Steve Henson]
2149 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2150 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2152 *) Fix SSL client code which could crash if connecting to a
2153 malicious SSLv2 server. (CVE-2006-4343)
2154 [Tavis Ormandy and Will Drewry, Google Security Team]
2156 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
2157 match only those. Before that, "AES256-SHA" would be interpreted
2158 as a pattern and match "AES128-SHA" too (since AES128-SHA got
2159 the same strength classification in 0.9.7h) as we currently only
2160 have a single AES bit in the ciphersuite description bitmap.
2161 That change, however, also applied to ciphersuite strings such as
2162 "RC4-MD5" that intentionally matched multiple ciphersuites --
2163 namely, SSL 2.0 ciphersuites in addition to the more common ones
2164 from SSL 3.0/TLS 1.0.
2166 So we change the selection algorithm again: Naming an explicit
2167 ciphersuite selects this one ciphersuite, and any other similar
2168 ciphersuite (same bitmap) from *other* protocol versions.
2169 Thus, "RC4-MD5" again will properly select both the SSL 2.0
2170 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
2172 Since SSL 2.0 does not have any ciphersuites for which the
2173 128/256 bit distinction would be relevant, this works for now.
2174 The proper fix will be to use different bits for AES128 and
2175 AES256, which would have avoided the problems from the beginning;
2176 however, bits are scarce, so we can only do this in a new release
2177 (not just a patchlevel) when we can change the SSL_CIPHER
2178 definition to split the single 'unsigned long mask' bitmap into
2179 multiple values to extend the available space.
2183 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2185 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2186 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2188 *) Add AES IGE and biIGE modes.
2191 *) Change the Unix randomness entropy gathering to use poll() when
2192 possible instead of select(), since the latter has some
2193 undesirable limitations.
2194 [Darryl Miles via Richard Levitte and Bodo Moeller]
2196 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2197 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2198 cannot be implicitly activated as part of, e.g., the "AES" alias.
2199 However, please upgrade to OpenSSL 0.9.9[-dev] for
2200 non-experimental use of the ECC ciphersuites to get TLS extension
2201 support, which is required for curve and point format negotiation
2202 to avoid potential handshake problems.
2205 *) Disable rogue ciphersuites:
2207 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2208 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2209 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2211 The latter two were purportedly from
2212 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2215 Also deactivate the remaining ciphersuites from
2216 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2217 unofficial, and the ID has long expired.
2220 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2221 dual-core machines) and other potential thread-safety issues.
2224 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
2225 versions), which is now available for royalty-free use
2226 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
2227 Also, add Camellia TLS ciphersuites from RFC 4132.
2229 To minimize changes between patchlevels in the OpenSSL 0.9.8
2230 series, Camellia remains excluded from compilation unless OpenSSL
2231 is configured with 'enable-camellia'.
2234 *) Disable the padding bug check when compression is in use. The padding
2235 bug check assumes the first packet is of even length, this is not
2236 necessarily true if compresssion is enabled and can result in false
2237 positives causing handshake failure. The actual bug test is ancient
2238 code so it is hoped that implementations will either have fixed it by
2239 now or any which still have the bug do not support compression.
2242 Changes between 0.9.8a and 0.9.8b [04 May 2006]
2244 *) When applying a cipher rule check to see if string match is an explicit
2245 cipher suite and only match that one cipher suite if it is.
2248 *) Link in manifests for VC++ if needed.
2249 [Austin Ziegler <halostatue@gmail.com>]
2251 *) Update support for ECC-based TLS ciphersuites according to
2252 draft-ietf-tls-ecc-12.txt with proposed changes (but without
2253 TLS extensions, which are supported starting with the 0.9.9
2254 branch, not in the OpenSSL 0.9.8 branch).
2257 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
2258 opaque EVP_CIPHER_CTX handling.
2261 *) Fixes and enhancements to zlib compression code. We now only use
2262 "zlib1.dll" and use the default __cdecl calling convention on Win32
2263 to conform with the standards mentioned here:
2264 http://www.zlib.net/DLL_FAQ.txt
2265 Static zlib linking now works on Windows and the new --with-zlib-include
2266 --with-zlib-lib options to Configure can be used to supply the location
2267 of the headers and library. Gracefully handle case where zlib library
2271 *) Several fixes and enhancements to the OID generation code. The old code
2272 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
2273 handle numbers larger than ULONG_MAX, truncated printing and had a
2274 non standard OBJ_obj2txt() behaviour.
2277 *) Add support for building of engines under engine/ as shared libraries
2278 under VC++ build system.
2281 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
2282 Hopefully, we will not see any false combination of paths any more.
2285 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
2287 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2288 (part of SSL_OP_ALL). This option used to disable the
2289 countermeasure against man-in-the-middle protocol-version
2290 rollback in the SSL 2.0 server implementation, which is a bad
2291 idea. (CVE-2005-2969)
2293 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2294 for Information Security, National Institute of Advanced Industrial
2295 Science and Technology [AIST], Japan)]
2297 *) Add two function to clear and return the verify parameter flags.
2300 *) Keep cipherlists sorted in the source instead of sorting them at
2301 runtime, thus removing the need for a lock.
2304 *) Avoid some small subgroup attacks in Diffie-Hellman.
2305 [Nick Mathewson and Ben Laurie]
2307 *) Add functions for well-known primes.
2310 *) Extended Windows CE support.
2311 [Satoshi Nakamura and Andy Polyakov]
2313 *) Initialize SSL_METHOD structures at compile time instead of during
2314 runtime, thus removing the need for a lock.
2317 *) Make PKCS7_decrypt() work even if no certificate is supplied by
2318 attempting to decrypt each encrypted key in turn. Add support to
2322 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
2324 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
2327 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
2330 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
2331 key into the same file any more.
2334 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
2337 *) Add -utf8 command line and config file option to 'ca'.
2338 [Stefan <stf@udoma.org]
2340 *) Removed the macro des_crypt(), as it seems to conflict with some
2341 libraries. Use DES_crypt().
2344 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
2345 involves renaming the source and generated shared-libs for
2346 both. The engines will accept the corrected or legacy ids
2347 ('ncipher' and '4758_cca' respectively) when binding. NB,
2348 this only applies when building 'shared'.
2349 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
2351 *) Add attribute functions to EVP_PKEY structure. Modify
2352 PKCS12_create() to recognize a CSP name attribute and
2353 use it. Make -CSP option work again in pkcs12 utility.
2356 *) Add new functionality to the bn blinding code:
2357 - automatic re-creation of the BN_BLINDING parameters after
2358 a fixed number of uses (currently 32)
2359 - add new function for parameter creation
2360 - introduce flags to control the update behaviour of the
2361 BN_BLINDING parameters
2362 - hide BN_BLINDING structure
2363 Add a second BN_BLINDING slot to the RSA structure to improve
2364 performance when a single RSA object is shared among several
2368 *) Add support for DTLS.
2369 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
2371 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
2372 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
2375 *) Remove buggy and incompletet DH cert support from
2376 ssl/ssl_rsa.c and ssl/s3_both.c
2379 *) Use SHA-1 instead of MD5 as the default digest algorithm for
2380 the apps/openssl applications.
2383 *) Compile clean with "-Wall -Wmissing-prototypes
2384 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
2385 DEBUG_SAFESTACK must also be set.
2388 *) Change ./Configure so that certain algorithms can be disabled by default.
2389 The new counterpiece to "no-xxx" is "enable-xxx".
2391 The patented RC5 and MDC2 algorithms will now be disabled unless
2392 "enable-rc5" and "enable-mdc2", respectively, are specified.
2394 (IDEA remains enabled despite being patented. This is because IDEA
2395 is frequently required for interoperability, and there is no license
2396 fee for non-commercial use. As before, "no-idea" can be used to
2397 avoid this algorithm.)
2401 *) Add processing of proxy certificates (see RFC 3820). This work was
2402 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
2403 EGEE (Enabling Grids for E-science in Europe).
2406 *) RC4 performance overhaul on modern architectures/implementations, such
2407 as Intel P4, IA-64 and AMD64.
2410 *) New utility extract-section.pl. This can be used specify an alternative
2411 section number in a pod file instead of having to treat each file as
2412 a separate case in Makefile. This can be done by adding two lines to the
2415 =for comment openssl_section:XXX
2417 The blank line is mandatory.
2421 *) New arguments -certform, -keyform and -pass for s_client and s_server
2422 to allow alternative format key and certificate files and passphrase
2426 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
2427 update associated structures and add various utility functions.
2429 Add new policy related verify parameters, include policy checking in
2430 standard verify code. Enhance 'smime' application with extra parameters
2431 to support policy checking and print out.
2434 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
2435 Nehemiah processors. These extensions support AES encryption in hardware
2436 as well as RNG (though RNG support is currently disabled).
2437 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
2439 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
2442 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
2443 [Andy Polyakov and a number of other people]
2445 *) Improved PowerPC platform support. Most notably BIGNUM assembler
2446 implementation contributed by IBM.
2447 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
2449 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
2450 exponent rather than 'unsigned long'. There is a corresponding change to
2451 the new 'rsa_keygen' element of the RSA_METHOD structure.
2452 [Jelte Jansen, Geoff Thorpe]
2454 *) Functionality for creating the initial serial number file is now
2455 moved from CA.pl to the 'ca' utility with a new option -create_serial.
2457 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
2458 number file to 1, which is bound to cause problems. To avoid
2459 the problems while respecting compatibility between different 0.9.7
2460 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
2461 CA.pl for serial number initialization. With the new release 0.9.8,
2462 we can fix the problem directly in the 'ca' utility.)
2465 *) Reduced header interdepencies by declaring more opaque objects in
2466 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
2467 give fewer recursive includes, which could break lazy source code - so
2468 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
2469 developers should define this symbol when building and using openssl to
2470 ensure they track the recommended behaviour, interfaces, [etc], but
2471 backwards-compatible behaviour prevails when this isn't defined.
2474 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
2477 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
2478 This will generate a random key of the appropriate length based on the
2479 cipher context. The EVP_CIPHER can provide its own random key generation
2480 routine to support keys of a specific form. This is used in the des and
2481 3des routines to generate a key of the correct parity. Update S/MIME
2482 code to use new functions and hence generate correct parity DES keys.
2483 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
2484 valid (weak or incorrect parity).
2487 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
2488 as looking them up. This is useful when the verified structure may contain
2489 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
2490 present unless the new PKCS7_NO_CRL flag is asserted.
2493 *) Extend ASN1 oid configuration module. It now additionally accepts the
2496 shortName = some long name, 1.2.3.4
2499 *) Reimplemented the BN_CTX implementation. There is now no more static
2500 limitation on the number of variables it can handle nor the depth of the
2501 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
2502 information can now expand as required, and rather than having a single
2503 static array of bignums, BN_CTX now uses a linked-list of such arrays
2504 allowing it to expand on demand whilst maintaining the usefulness of
2505 BN_CTX's "bundling".
2508 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
2509 to allow all RSA operations to function using a single BN_CTX.
2512 *) Preliminary support for certificate policy evaluation and checking. This
2513 is initially intended to pass the tests outlined in "Conformance Testing
2514 of Relying Party Client Certificate Path Processing Logic" v1.07.
2517 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
2518 remained unused and not that useful. A variety of other little bignum
2519 tweaks and fixes have also been made continuing on from the audit (see
2523 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
2524 associated ASN1, EVP and SSL functions and old ASN1 macros.
2527 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
2528 and this should never fail. So the return value from the use of
2529 BN_set_word() (which can fail due to needless expansion) is now deprecated;
2530 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
2533 *) BN_CTX_get() should return zero-valued bignums, providing the same
2534 initialised value as BN_new().
2535 [Geoff Thorpe, suggested by Ulf Möller]
2537 *) Support for inhibitAnyPolicy certificate extension.
2540 *) An audit of the BIGNUM code is underway, for which debugging code is
2541 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
2542 is considered valid when processing BIGNUMs, and causes execution to
2543 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
2544 further steps are taken to deliberately pollute unused data in BIGNUM
2545 structures to try and expose faulty code further on. For now, openssl will
2546 (in its default mode of operation) continue to tolerate the inconsistent
2547 forms that it has tolerated in the past, but authors and packagers should
2548 consider trying openssl and their own applications when compiled with
2549 these debugging symbols defined. It will help highlight potential bugs in
2550 their own code, and will improve the test coverage for OpenSSL itself. At
2551 some point, these tighter rules will become openssl's default to improve
2552 maintainability, though the assert()s and other overheads will remain only
2553 in debugging configurations. See bn.h for more details.
2554 [Geoff Thorpe, Nils Larsch, Ulf Möller]
2556 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
2557 that can only be obtained through BN_CTX_new() (which implicitly
2558 initialises it). The presence of this function only made it possible
2559 to overwrite an existing structure (and cause memory leaks).
2562 *) Because of the callback-based approach for implementing LHASH as a
2563 template type, lh_insert() adds opaque objects to hash-tables and
2564 lh_doall() or lh_doall_arg() are typically used with a destructor callback
2565 to clean up those corresponding objects before destroying the hash table
2566 (and losing the object pointers). So some over-zealous constifications in
2567 LHASH have been relaxed so that lh_insert() does not take (nor store) the
2568 objects as "const" and the lh_doall[_arg] callback wrappers are not
2569 prototyped to have "const" restrictions on the object pointers they are
2570 given (and so aren't required to cast them away any more).
2573 *) The tmdiff.h API was so ugly and minimal that our own timing utility
2574 (speed) prefers to use its own implementation. The two implementations
2575 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
2576 its object type properly exposed (MS_TM) instead of casting to/from "char
2577 *". This may still change yet if someone realises MS_TM and "ms_time_***"
2578 aren't necessarily the greatest nomenclatures - but this is what was used
2579 internally to the implementation so I've used that for now.
2582 *) Ensure that deprecated functions do not get compiled when
2583 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
2584 the self-tests were still using deprecated key-generation functions so
2585 these have been updated also.
2588 *) Reorganise PKCS#7 code to separate the digest location functionality
2589 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
2590 New function PKCS7_set_digest() to set the digest type for PKCS#7
2591 digestedData type. Add additional code to correctly generate the
2592 digestedData type and add support for this type in PKCS7 initialization
2596 *) New function PKCS7_set0_type_other() this initializes a PKCS7
2597 structure of type "other".
2600 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
2601 sure the loop does correctly stop and breaking ("division by zero")
2602 modulus operations are not performed. The (pre-generated) prime
2603 table crypto/bn/bn_prime.h was already correct, but it could not be
2604 re-generated on some platforms because of the "division by zero"
2605 situation in the script.
2606 [Ralf S. Engelschall]
2608 *) Update support for ECC-based TLS ciphersuites according to
2609 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
2610 SHA-1 now is only used for "small" curves (where the
2611 representation of a field element takes up to 24 bytes); for
2612 larger curves, the field element resulting from ECDH is directly
2613 used as premaster secret.
2614 [Douglas Stebila (Sun Microsystems Laboratories)]
2616 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
2617 curve secp160r1 to the tests.
2618 [Douglas Stebila (Sun Microsystems Laboratories)]
2620 *) Add the possibility to load symbols globally with DSO.
2621 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
2623 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
2624 control of the error stack.
2627 *) Add support for STORE in ENGINE.
2630 *) Add the STORE type. The intention is to provide a common interface
2631 to certificate and key stores, be they simple file-based stores, or
2632 HSM-type store, or LDAP stores, or...
2633 NOTE: The code is currently UNTESTED and isn't really used anywhere.
2636 *) Add a generic structure called OPENSSL_ITEM. This can be used to
2637 pass a list of arguments to any function as well as provide a way
2638 for a function to pass data back to the caller.
2641 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
2642 works like BUF_strdup() but can be used to duplicate a portion of
2643 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
2647 *) Add the function sk_find_ex() which works like sk_find(), but will
2648 return an index to an element even if an exact match couldn't be
2649 found. The index is guaranteed to point at the element where the
2650 searched-for key would be inserted to preserve sorting order.
2653 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
2654 takes an extra flags argument for optional functionality. Currently,
2655 the following flags are defined:
2657 OBJ_BSEARCH_VALUE_ON_NOMATCH
2658 This one gets OBJ_bsearch_ex() to return a pointer to the first
2659 element where the comparing function returns a negative or zero
2662 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
2663 This one gets OBJ_bsearch_ex() to return a pointer to the first
2664 element where the comparing function returns zero. This is useful
2665 if there are more than one element where the comparing function
2669 *) Make it possible to create self-signed certificates with 'openssl ca'
2670 in such a way that the self-signed certificate becomes part of the
2671 CA database and uses the same mechanisms for serial number generation
2672 as all other certificate signing. The new flag '-selfsign' enables
2673 this functionality. Adapt CA.sh and CA.pl.in.
2676 *) Add functionality to check the public key of a certificate request
2677 against a given private. This is useful to check that a certificate
2678 request can be signed by that key (self-signing).
2681 *) Make it possible to have multiple active certificates with the same
2682 subject in the CA index file. This is done only if the keyword
2683 'unique_subject' is set to 'no' in the main CA section (default
2684 if 'CA_default') of the configuration file. The value is saved
2685 with the database itself in a separate index attribute file,
2686 named like the index file with '.attr' appended to the name.
2689 *) Generate muti valued AVAs using '+' notation in config files for
2693 *) Support for nameConstraints certificate extension.
2696 *) Support for policyConstraints certificate extension.
2699 *) Support for policyMappings certificate extension.
2702 *) Make sure the default DSA_METHOD implementation only uses its
2703 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
2704 and change its own handlers to be NULL so as to remove unnecessary
2705 indirection. This lets alternative implementations fallback to the
2706 default implementation more easily.
2709 *) Support for directoryName in GeneralName related extensions
2713 *) Make it possible to link applications using Makefile.shared.
2714 Make that possible even when linking against static libraries!
2717 *) Support for single pass processing for S/MIME signing. This now
2718 means that S/MIME signing can be done from a pipe, in addition
2719 cleartext signing (multipart/signed type) is effectively streaming
2720 and the signed data does not need to be all held in memory.
2722 This is done with a new flag PKCS7_STREAM. When this flag is set
2723 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
2724 is done after the data is output (and digests calculated) in
2725 SMIME_write_PKCS7().
2728 *) Add full support for -rpath/-R, both in shared libraries and
2729 applications, at least on the platforms where it's known how
2733 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
2734 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
2735 will now compute a table of multiples of the generator that
2736 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
2737 faster (notably in the case of a single point multiplication,
2738 scalar * generator).
2739 [Nils Larsch, Bodo Moeller]
2741 *) IPv6 support for certificate extensions. The various extensions
2742 which use the IP:a.b.c.d can now take IPv6 addresses using the
2743 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
2747 *) Added an ENGINE that implements RSA by performing private key
2748 exponentiations with the GMP library. The conversions to and from
2749 GMP's mpz_t format aren't optimised nor are any montgomery forms
2750 cached, and on x86 it appears OpenSSL's own performance has caught up.
2751 However there are likely to be other architectures where GMP could
2752 provide a boost. This ENGINE is not built in by default, but it can be
2753 specified at Configure time and should be accompanied by the necessary
2754 linker additions, eg;
2755 ./config -DOPENSSL_USE_GMP -lgmp
2758 *) "openssl engine" will not display ENGINE/DSO load failure errors when
2759 testing availability of engines with "-t" - the old behaviour is
2760 produced by increasing the feature's verbosity with "-tt".
2763 *) ECDSA routines: under certain error conditions uninitialized BN objects
2764 could be freed. Solution: make sure initialization is performed early
2765 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
2769 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
2770 and DH_METHOD (eg. by ENGINE implementations) to override the normal
2771 software implementations. For DSA and DH, parameter generation can
2772 also be overriden by providing the appropriate method callbacks.
2775 *) Change the "progress" mechanism used in key-generation and
2776 primality testing to functions that take a new BN_GENCB pointer in
2777 place of callback/argument pairs. The new API functions have "_ex"
2778 postfixes and the older functions are reimplemented as wrappers for
2779 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
2780 declarations of the old functions to help (graceful) attempts to
2781 migrate to the new functions. Also, the new key-generation API
2782 functions operate on a caller-supplied key-structure and return
2783 success/failure rather than returning a key or NULL - this is to
2784 help make "keygen" another member function of RSA_METHOD etc.
2786 Example for using the new callback interface:
2788 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
2792 BN_GENCB_set(&my_cb, my_callback, my_arg);
2794 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
2795 /* For the meaning of a, b in calls to my_callback(), see the
2796 * documentation of the function that calls the callback.
2797 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
2798 * my_callback should return 1 if it wants BN_is_prime_ex()
2799 * to continue, or 0 to stop.
2804 *) Change the ZLIB compression method to be stateful, and make it
2805 available to TLS with the number defined in
2806 draft-ietf-tls-compression-04.txt.
2809 *) Add the ASN.1 structures and functions for CertificatePair, which
2810 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
2812 CertificatePair ::= SEQUENCE {
2813 forward [0] Certificate OPTIONAL,
2814 reverse [1] Certificate OPTIONAL,
2815 -- at least one of the pair shall be present -- }
2817 Also implement the PEM functions to read and write certificate
2818 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
2820 This needed to be defined, mostly for the sake of the LDAP
2821 attribute crossCertificatePair, but may prove useful elsewhere as
2825 *) Make it possible to inhibit symlinking of shared libraries in
2826 Makefile.shared, for Cygwin's sake.
2829 *) Extend the BIGNUM API by creating a function
2830 void BN_set_negative(BIGNUM *a, int neg);
2831 and a macro that behave like
2832 int BN_is_negative(const BIGNUM *a);
2834 to avoid the need to access 'a->neg' directly in applications.
2837 *) Implement fast modular reduction for pseudo-Mersenne primes
2838 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
2839 EC_GROUP_new_curve_GFp() will now automatically use this
2841 [Nils Larsch <nla@trustcenter.de>]
2843 *) Add new lock type (CRYPTO_LOCK_BN).
2846 *) Change the ENGINE framework to automatically load engines
2847 dynamically from specific directories unless they could be
2848 found to already be built in or loaded. Move all the
2849 current engines except for the cryptodev one to a new
2851 The engines in engines/ are built as shared libraries if
2852 the "shared" options was given to ./Configure or ./config.
2853 Otherwise, they are inserted in libcrypto.a.
2854 /usr/local/ssl/engines is the default directory for dynamic
2855 engines, but that can be overriden at configure time through
2856 the usual use of --prefix and/or --openssldir, and at run
2857 time with the environment variable OPENSSL_ENGINES.
2858 [Geoff Thorpe and Richard Levitte]
2860 *) Add Makefile.shared, a helper makefile to build shared
2861 libraries. Addapt Makefile.org.
2864 *) Add version info to Win32 DLLs.
2865 [Peter 'Luna' Runestig" <peter@runestig.com>]
2867 *) Add new 'medium level' PKCS#12 API. Certificates and keys
2868 can be added using this API to created arbitrary PKCS#12
2869 files while avoiding the low level API.
2871 New options to PKCS12_create(), key or cert can be NULL and
2872 will then be omitted from the output file. The encryption
2873 algorithm NIDs can be set to -1 for no encryption, the mac
2874 iteration count can be set to 0 to omit the mac.
2876 Enhance pkcs12 utility by making the -nokeys and -nocerts
2877 options work when creating a PKCS#12 file. New option -nomac
2878 to omit the mac, NONE can be set for an encryption algorithm.
2879 New code is modified to use the enhanced PKCS12_create()
2880 instead of the low level API.
2883 *) Extend ASN1 encoder to support indefinite length constructed
2884 encoding. This can output sequences tags and octet strings in
2885 this form. Modify pk7_asn1.c to support indefinite length
2886 encoding. This is experimental and needs additional code to
2887 be useful, such as an ASN1 bio and some enhanced streaming
2890 Extend template encode functionality so that tagging is passed
2891 down to the template encoder.
2894 *) Let 'openssl req' fail if an argument to '-newkey' is not
2895 recognized instead of using RSA as a default.
2898 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
2899 As these are not official, they are not included in "ALL";
2900 the "ECCdraft" ciphersuite group alias can be used to select them.
2901 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
2903 *) Add ECDH engine support.
2904 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
2906 *) Add ECDH in new directory crypto/ecdh/.
2907 [Douglas Stebila (Sun Microsystems Laboratories)]
2909 *) Let BN_rand_range() abort with an error after 100 iterations
2910 without success (which indicates a broken PRNG).
2913 *) Change BN_mod_sqrt() so that it verifies that the input value
2914 is really the square of the return value. (Previously,
2915 BN_mod_sqrt would show GIGO behaviour.)
2918 *) Add named elliptic curves over binary fields from X9.62, SECG,
2919 and WAP/WTLS; add OIDs that were still missing.
2921 [Sheueling Chang Shantz and Douglas Stebila
2922 (Sun Microsystems Laboratories)]
2924 *) Extend the EC library for elliptic curves over binary fields
2925 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
2928 EC_GF2m_simple_method
2932 EC_GROUP_new_curve_GF2m
2933 EC_GROUP_set_curve_GF2m
2934 EC_GROUP_get_curve_GF2m
2935 EC_POINT_set_affine_coordinates_GF2m
2936 EC_POINT_get_affine_coordinates_GF2m
2937 EC_POINT_set_compressed_coordinates_GF2m
2939 Point compression for binary fields is disabled by default for
2940 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
2943 As binary polynomials are represented as BIGNUMs, various members
2944 of the EC_GROUP and EC_POINT data structures can be shared
2945 between the implementations for prime fields and binary fields;
2946 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
2947 are essentially identical to their ..._GFp counterparts.
2948 (For simplicity, the '..._GFp' prefix has been dropped from
2949 various internal method names.)
2951 An internal 'field_div' method (similar to 'field_mul' and
2952 'field_sqr') has been added; this is used only for binary fields.
2954 [Sheueling Chang Shantz and Douglas Stebila
2955 (Sun Microsystems Laboratories)]
2957 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
2958 through methods ('mul', 'precompute_mult').
2960 The generic implementations (now internally called 'ec_wNAF_mul'
2961 and 'ec_wNAF_precomputed_mult') remain the default if these
2962 methods are undefined.
2964 [Sheueling Chang Shantz and Douglas Stebila
2965 (Sun Microsystems Laboratories)]
2967 *) New function EC_GROUP_get_degree, which is defined through
2968 EC_METHOD. For curves over prime fields, this returns the bit
2969 length of the modulus.
2971 [Sheueling Chang Shantz and Douglas Stebila
2972 (Sun Microsystems Laboratories)]
2974 *) New functions EC_GROUP_dup, EC_POINT_dup.
2975 (These simply call ..._new and ..._copy).
2977 [Sheueling Chang Shantz and Douglas Stebila
2978 (Sun Microsystems Laboratories)]
2980 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
2981 Polynomials are represented as BIGNUMs (where the sign bit is not
2982 used) in the following functions [macros]:
2985 BN_GF2m_sub [= BN_GF2m_add]
2986 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
2987 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
2988 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
2990 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
2991 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
2992 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
2993 BN_GF2m_cmp [= BN_ucmp]
2995 (Note that only the 'mod' functions are actually for fields GF(2^m).
2996 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
2998 For some functions, an the irreducible polynomial defining a
2999 field can be given as an 'unsigned int[]' with strictly
3000 decreasing elements giving the indices of those bits that are set;
3001 i.e., p[] represents the polynomial
3002 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
3004 p[0] > p[1] > ... > p[k] = 0.
3005 This applies to the following functions:
3010 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
3011 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
3013 BN_GF2m_mod_sqrt_arr
3014 BN_GF2m_mod_solve_quad_arr
3018 Conversion can be performed by the following functions:
3023 bntest.c has additional tests for binary polynomial arithmetic.
3025 Two implementations for BN_GF2m_mod_div() are available.
3026 The default algorithm simply uses BN_GF2m_mod_inv() and
3027 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
3028 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
3029 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
3031 [Sheueling Chang Shantz and Douglas Stebila
3032 (Sun Microsystems Laboratories)]
3034 *) Add new error code 'ERR_R_DISABLED' that can be used when some
3035 functionality is disabled at compile-time.
3036 [Douglas Stebila <douglas.stebila@sun.com>]
3038 *) Change default behaviour of 'openssl asn1parse' so that more
3039 information is visible when viewing, e.g., a certificate:
3041 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
3042 mode the content of non-printable OCTET STRINGs is output in a
3043 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
3044 avoid the appearance of a printable string.
3045 [Nils Larsch <nla@trustcenter.de>]
3047 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
3049 EC_GROUP_set_asn1_flag()
3050 EC_GROUP_get_asn1_flag()
3051 EC_GROUP_set_point_conversion_form()
3052 EC_GROUP_get_point_conversion_form()
3053 These control ASN1 encoding details:
3054 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
3055 has been set to OPENSSL_EC_NAMED_CURVE.
3056 - Points are encoded in uncompressed form by default; options for
3057 asn1_for are as for point2oct, namely
3058 POINT_CONVERSION_COMPRESSED
3059 POINT_CONVERSION_UNCOMPRESSED
3060 POINT_CONVERSION_HYBRID
3062 Also add 'seed' and 'seed_len' members to EC_GROUP with access
3065 EC_GROUP_get0_seed()
3066 EC_GROUP_get_seed_len()
3067 This is used only for ASN1 purposes (so far).
3068 [Nils Larsch <nla@trustcenter.de>]
3070 *) Add 'field_type' member to EC_METHOD, which holds the NID
3071 of the appropriate field type OID. The new function
3072 EC_METHOD_get_field_type() returns this value.
3073 [Nils Larsch <nla@trustcenter.de>]
3078 EC_POINT_point2hex()
3079 EC_POINT_hex2point()
3080 providing useful interfaces to EC_POINT_point2oct() and
3081 EC_POINT_oct2point().
3082 [Nils Larsch <nla@trustcenter.de>]
3084 *) Change internals of the EC library so that the functions
3085 EC_GROUP_set_generator()
3086 EC_GROUP_get_generator()
3087 EC_GROUP_get_order()
3088 EC_GROUP_get_cofactor()
3089 are implemented directly in crypto/ec/ec_lib.c and not dispatched
3090 to methods, which would lead to unnecessary code duplication when
3091 adding different types of curves.
3092 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
3094 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
3095 arithmetic, and such that modified wNAFs are generated
3096 (which avoid length expansion in many cases).
3099 *) Add a function EC_GROUP_check_discriminant() (defined via
3100 EC_METHOD) that verifies that the curve discriminant is non-zero.
3102 Add a function EC_GROUP_check() that makes some sanity tests
3103 on a EC_GROUP, its generator and order. This includes
3104 EC_GROUP_check_discriminant().
3105 [Nils Larsch <nla@trustcenter.de>]
3107 *) Add ECDSA in new directory crypto/ecdsa/.
3109 Add applications 'openssl ecparam' and 'openssl ecdsa'
3110 (these are based on 'openssl dsaparam' and 'openssl dsa').
3112 ECDSA support is also included in various other files across the
3113 library. Most notably,
3114 - 'openssl req' now has a '-newkey ecdsa:file' option;
3115 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
3116 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
3117 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
3118 them suitable for ECDSA where domain parameters must be
3119 extracted before the specific public key;
3120 - ECDSA engine support has been added.
3121 [Nils Larsch <nla@trustcenter.de>]
3123 *) Include some named elliptic curves, and add OIDs from X9.62,
3124 SECG, and WAP/WTLS. Each curve can be obtained from the new
3126 EC_GROUP_new_by_curve_name(),
3127 and the list of available named curves can be obtained with
3128 EC_get_builtin_curves().
3129 Also add a 'curve_name' member to EC_GROUP objects, which can be
3131 EC_GROUP_set_curve_name()
3132 EC_GROUP_get_curve_name()
3133 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
3135 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3136 was actually never needed) and in BN_mul(). The removal in BN_mul()
3137 required a small change in bn_mul_part_recursive() and the addition
3138 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3139 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3140 bn_sub_words() and bn_add_words() except they take arrays with
3144 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
3146 *) Cleanse PEM buffers before freeing them since they may contain
3148 [Benjamin Bennett <ben@psc.edu>]
3150 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3151 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3152 authentication-only ciphersuites.
3155 *) Since AES128 and AES256 share a single mask bit in the logic of
3156 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3157 kludge to work properly if AES128 is available and AES256 isn't.
3160 *) Expand security boundary to match 1.1.1 module.
3163 *) Remove redundant features: hash file source, editing of test vectors
3164 modify fipsld to use external fips_premain.c signature.
3167 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
3168 run algorithm test programs.
3171 *) Make algorithm test programs more tolerant of whitespace.
3174 *) Have SSL/TLS server implementation tolerate "mismatched" record
3175 protocol version while receiving ClientHello even if the
3176 ClientHello is fragmented. (The server can't insist on the
3177 particular protocol version it has chosen before the ServerHello
3178 message has informed the client about his choice.)
3181 *) Load error codes if they are not already present instead of using a
3182 static variable. This allows them to be cleanly unloaded and reloaded.
3185 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3187 *) Introduce limits to prevent malicious keys being able to
3188 cause a denial of service. (CVE-2006-2940)
3189 [Steve Henson, Bodo Moeller]
3191 *) Fix ASN.1 parsing of certain invalid structures that can result
3192 in a denial of service. (CVE-2006-2937) [Steve Henson]
3194 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3195 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3197 *) Fix SSL client code which could crash if connecting to a
3198 malicious SSLv2 server. (CVE-2006-4343)
3199 [Tavis Ormandy and Will Drewry, Google Security Team]
3201 *) Change ciphersuite string processing so that an explicit
3202 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3203 will no longer include "AES128-SHA"), and any other similar
3204 ciphersuite (same bitmap) from *other* protocol versions (so that
3205 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3206 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3207 changes from 0.9.8b and 0.9.8d.
3210 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
3212 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3213 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3215 *) Change the Unix randomness entropy gathering to use poll() when
3216 possible instead of select(), since the latter has some
3217 undesirable limitations.
3218 [Darryl Miles via Richard Levitte and Bodo Moeller]
3220 *) Disable rogue ciphersuites:
3222 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3223 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3224 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3226 The latter two were purportedly from
3227 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3230 Also deactive the remaining ciphersuites from
3231 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3232 unofficial, and the ID has long expired.
3235 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3236 dual-core machines) and other potential thread-safety issues.
3239 Changes between 0.9.7i and 0.9.7j [04 May 2006]
3241 *) Adapt fipsld and the build system to link against the validated FIPS
3242 module in FIPS mode.
3245 *) Fixes for VC++ 2005 build under Windows.
3248 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
3249 from a Windows bash shell such as MSYS. It is autodetected from the
3250 "config" script when run from a VC++ environment. Modify standard VC++
3251 build to use fipscanister.o from the GNU make build.
3254 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
3256 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
3257 The value now differs depending on if you build for FIPS or not.
3258 BEWARE! A program linked with a shared FIPSed libcrypto can't be
3259 safely run with a non-FIPSed libcrypto, as it may crash because of
3260 the difference induced by this change.
3263 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
3265 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3266 (part of SSL_OP_ALL). This option used to disable the
3267 countermeasure against man-in-the-middle protocol-version
3268 rollback in the SSL 2.0 server implementation, which is a bad
3269 idea. (CVE-2005-2969)
3271 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3272 for Information Security, National Institute of Advanced Industrial
3273 Science and Technology [AIST], Japan)]
3275 *) Minimal support for X9.31 signatures and PSS padding modes. This is
3276 mainly for FIPS compliance and not fully integrated at this stage.
3279 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
3280 the exponentiation using a fixed-length exponent. (Otherwise,
3281 the information leaked through timing could expose the secret key
3282 after many signatures; cf. Bleichenbacher's attack on DSA with
3286 *) Make a new fixed-window mod_exp implementation the default for
3287 RSA, DSA, and DH private-key operations so that the sequence of
3288 squares and multiplies and the memory access pattern are
3289 independent of the particular secret key. This will mitigate
3290 cache-timing and potential related attacks.
3292 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
3293 and this is automatically used by BN_mod_exp_mont() if the new flag
3294 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
3295 will use this BN flag for private exponents unless the flag
3296 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
3297 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
3299 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
3301 *) Change the client implementation for SSLv23_method() and
3302 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
3303 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
3304 (Previously, the SSL 2.0 backwards compatible Client Hello
3305 message format would be used even with SSL_OP_NO_SSLv2.)
3308 *) Add support for smime-type MIME parameter in S/MIME messages which some
3312 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
3313 a threadsafe manner. Modify rsa code to use new function and add calls
3314 to dsa and dh code (which had race conditions before).
3317 *) Include the fixed error library code in the C error file definitions
3318 instead of fixing them up at runtime. This keeps the error code
3319 structures constant.
3322 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
3324 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
3327 *) Fixes for newer kerberos headers. NB: the casts are needed because
3328 the 'length' field is signed on one version and unsigned on another
3329 with no (?) obvious way to tell the difference, without these VC++
3330 complains. Also the "definition" of FAR (blank) is no longer included
3331 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
3332 some needed definitions.
3335 *) Undo Cygwin change.
3338 *) Added support for proxy certificates according to RFC 3820.
3339 Because they may be a security thread to unaware applications,
3340 they must be explicitely allowed in run-time. See
3341 docs/HOWTO/proxy_certificates.txt for further information.
3344 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
3346 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
3347 server and client random values. Previously
3348 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
3349 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
3351 This change has negligible security impact because:
3353 1. Server and client random values still have 24 bytes of pseudo random
3356 2. Server and client random values are sent in the clear in the initial
3359 3. The master secret is derived using the premaster secret (48 bytes in
3360 size for static RSA ciphersuites) as well as client server and random
3363 The OpenSSL team would like to thank the UK NISCC for bringing this issue
3366 [Stephen Henson, reported by UK NISCC]
3368 *) Use Windows randomness collection on Cygwin.
3371 *) Fix hang in EGD/PRNGD query when communication socket is closed
3372 prematurely by EGD/PRNGD.
3373 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
3375 *) Prompt for pass phrases when appropriate for PKCS12 input format.
3378 *) Back-port of selected performance improvements from development
3379 branch, as well as improved support for PowerPC platforms.
3382 *) Add lots of checks for memory allocation failure, error codes to indicate
3383 failure and freeing up memory if a failure occurs.
3384 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
3386 *) Add new -passin argument to dgst.
3389 *) Perform some character comparisons of different types in X509_NAME_cmp:
3390 this is needed for some certificates that reencode DNs into UTF8Strings
3391 (in violation of RFC3280) and can't or wont issue name rollover
3395 *) Make an explicit check during certificate validation to see that
3396 the CA setting in each certificate on the chain is correct. As a
3397 side effect always do the following basic checks on extensions,
3398 not just when there's an associated purpose to the check:
3400 - if there is an unhandled critical extension (unless the user
3401 has chosen to ignore this fault)
3402 - if the path length has been exceeded (if one is set at all)
3403 - that certain extensions fit the associated purpose (if one has
3407 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
3409 *) Avoid a race condition when CRLs are checked in a multi threaded
3410 environment. This would happen due to the reordering of the revoked
3411 entries during signature checking and serial number lookup. Now the
3412 encoding is cached and the serial number sort performed under a lock.
3413 Add new STACK function sk_is_sorted().
3416 *) Add Delta CRL to the extension code.
3419 *) Various fixes to s3_pkt.c so alerts are sent properly.
3420 [David Holmes <d.holmes@f5.com>]
3422 *) Reduce the chances of duplicate issuer name and serial numbers (in
3423 violation of RFC3280) using the OpenSSL certificate creation utilities.
3424 This is done by creating a random 64 bit value for the initial serial
3425 number when a serial number file is created or when a self signed
3426 certificate is created using 'openssl req -x509'. The initial serial
3427 number file is created using 'openssl x509 -next_serial' in CA.pl
3428 rather than being initialized to 1.
3431 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
3433 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3434 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3435 [Joe Orton, Steve Henson]
3437 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
3439 [Joe Orton, Steve Henson]
3441 *) Make it possible to have multiple active certificates with the same
3442 subject in the CA index file. This is done only if the keyword
3443 'unique_subject' is set to 'no' in the main CA section (default
3444 if 'CA_default') of the configuration file. The value is saved
3445 with the database itself in a separate index attribute file,
3446 named like the index file with '.attr' appended to the name.
3449 *) X509 verify fixes. Disable broken certificate workarounds when
3450 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
3451 keyUsage extension present. Don't accept CRLs with unhandled critical
3452 extensions: since verify currently doesn't process CRL extensions this
3453 rejects a CRL with *any* critical extensions. Add new verify error codes
3457 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
3458 A clarification of RFC2560 will require the use of OCTET STRINGs and
3459 some implementations cannot handle the current raw format. Since OpenSSL
3460 copies and compares OCSP nonces as opaque blobs without any attempt at
3461 parsing them this should not create any compatibility issues.
3464 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
3465 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
3466 this HMAC (and other) operations are several times slower than OpenSSL
3470 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
3471 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3473 *) Use the correct content when signing type "other".
3476 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
3478 *) Fix various bugs revealed by running the NISCC test suite:
3480 Stop out of bounds reads in the ASN1 code when presented with
3481 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3483 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
3485 If verify callback ignores invalid public key errors don't try to check
3486 certificate signature with the NULL public key.
3490 *) New -ignore_err option in ocsp application to stop the server
3491 exiting on the first error in a request.
3494 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3495 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3499 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3500 extra data after the compression methods not only for TLS 1.0
3501 but also for SSL 3.0 (as required by the specification).
3502 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3504 *) Change X509_certificate_type() to mark the key as exported/exportable
3505 when it's 512 *bits* long, not 512 bytes.
3508 *) Change AES_cbc_encrypt() so it outputs exact multiple of
3509 blocks during encryption.
3512 *) Various fixes to base64 BIO and non blocking I/O. On write
3513 flushes were not handled properly if the BIO retried. On read
3514 data was not being buffered properly and had various logic bugs.
3515 This also affects blocking I/O when the data being decoded is a
3519 *) Various S/MIME bugfixes and compatibility changes:
3520 output correct application/pkcs7 MIME type if
3521 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
3522 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
3523 of files as .eml work). Correctly handle very long lines in MIME
3527 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
3529 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3530 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3531 a protocol version number mismatch like a decryption error
3532 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3535 *) Turn on RSA blinding by default in the default implementation
3536 to avoid a timing attack. Applications that don't want it can call
3537 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3538 They would be ill-advised to do so in most cases.
3539 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3541 *) Change RSA blinding code so that it works when the PRNG is not
3542 seeded (in this case, the secret RSA exponent is abused as
3543 an unpredictable seed -- if it is not unpredictable, there
3544 is no point in blinding anyway). Make RSA blinding thread-safe
3545 by remembering the creator's thread ID in rsa->blinding and
3546 having all other threads use local one-time blinding factors
3547 (this requires more computation than sharing rsa->blinding, but
3548 avoids excessive locking; and if an RSA object is not shared
3549 between threads, blinding will still be very fast).
3552 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
3553 ENGINE as defaults for all supported algorithms irrespective of
3554 the 'flags' parameter. 'flags' is now honoured, so applications
3555 should make sure they are passing it correctly.
3558 *) Target "mingw" now allows native Windows code to be generated in
3559 the Cygwin environment as well as with the MinGW compiler.
3562 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
3564 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3565 via timing by performing a MAC computation even if incorrrect
3566 block cipher padding has been found. This is a countermeasure
3567 against active attacks where the attacker has to distinguish
3568 between bad padding and a MAC verification error. (CVE-2003-0078)
3570 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3571 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3572 Martin Vuagnoux (EPFL, Ilion)]
3574 *) Make the no-err option work as intended. The intention with no-err
3575 is not to have the whole error stack handling routines removed from
3576 libcrypto, it's only intended to remove all the function name and
3577 reason texts, thereby removing some of the footprint that may not
3578 be interesting if those errors aren't displayed anyway.
3580 NOTE: it's still possible for any application or module to have it's
3581 own set of error texts inserted. The routines are there, just not
3582 used by default when no-err is given.
3585 *) Add support for FreeBSD on IA64.
3586 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
3588 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
3589 Kerberos function mit_des_cbc_cksum(). Before this change,
3590 the value returned by DES_cbc_cksum() was like the one from
3591 mit_des_cbc_cksum(), except the bytes were swapped.
3592 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
3594 *) Allow an application to disable the automatic SSL chain building.
3595 Before this a rather primitive chain build was always performed in
3596 ssl3_output_cert_chain(): an application had no way to send the
3597 correct chain if the automatic operation produced an incorrect result.
3599 Now the chain builder is disabled if either:
3601 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
3603 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
3605 The reasoning behind this is that an application would not want the
3606 auto chain building to take place if extra chain certificates are
3607 present and it might also want a means of sending no additional
3608 certificates (for example the chain has two certificates and the
3612 *) Add the possibility to build without the ENGINE framework.
3613 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3615 *) Under Win32 gmtime() can return NULL: check return value in
3616 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
3619 *) DSA routines: under certain error conditions uninitialized BN objects
3620 could be freed. Solution: make sure initialization is performed early
3621 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
3622 Nils Larsch <nla@trustcenter.de> via PR#459)
3625 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
3626 checked on reconnect on the client side, therefore session resumption
3627 could still fail with a "ssl session id is different" error. This
3628 behaviour is masked when SSL_OP_ALL is used due to
3629 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
3630 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3631 followup to PR #377.
3634 *) IA-32 assembler support enhancements: unified ELF targets, support
3635 for SCO/Caldera platforms, fix for Cygwin shared build.
3638 *) Add support for FreeBSD on sparc64. As a consequence, support for
3639 FreeBSD on non-x86 processors is separate from x86 processors on
3640 the config script, much like the NetBSD support.
3641 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
3643 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
3645 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
3648 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
3649 code (06) was taken as the first octet of the session ID and the last
3650 octet was ignored consequently. As a result SSLv2 client side session
3651 caching could not have worked due to the session ID mismatch between
3653 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3657 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
3658 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
3662 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
3663 seems that in spite of existing for more than a year, many application
3664 author have done nothing to provide the necessary callbacks, which
3665 means that this particular engine will not work properly anywhere.
3666 This is a very unfortunate situation which forces us, in the name
3667 of usability, to give the hw_ncipher.c a static lock, which is part
3669 NOTE: This is for the 0.9.7 series ONLY. This hack will never
3670 appear in 0.9.8 or later. We EXPECT application authors to have
3671 dealt properly with this when 0.9.8 is released (unless we actually
3672 make such changes in the libcrypto locking code that changes will
3673 have to be made anyway).
3676 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
3677 octets have been read, EOF or an error occurs. Without this change
3678 some truncated ASN1 structures will not produce an error.
3681 *) Disable Heimdal support, since it hasn't been fully implemented.
3682 Still give the possibility to force the use of Heimdal, but with
3683 warnings and a request that patches get sent to openssl-dev.
3686 *) Add the VC-CE target, introduce the WINCE sysname, and add
3687 INSTALL.WCE and appropriate conditionals to make it build.
3688 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3690 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
3691 cygssl-x.y.z.dll, where x, y and z are the major, minor and
3692 edit numbers of the version.
3693 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3695 *) Introduce safe string copy and catenation functions
3696 (BUF_strlcpy() and BUF_strlcat()).
3697 [Ben Laurie (CHATS) and Richard Levitte]
3699 *) Avoid using fixed-size buffers for one-line DNs.
3700 [Ben Laurie (CHATS)]
3702 *) Add BUF_MEM_grow_clean() to avoid information leakage when
3703 resizing buffers containing secrets, and use where appropriate.
3704 [Ben Laurie (CHATS)]
3706 *) Avoid using fixed size buffers for configuration file location.
3707 [Ben Laurie (CHATS)]
3709 *) Avoid filename truncation for various CA files.
3710 [Ben Laurie (CHATS)]
3712 *) Use sizeof in preference to magic numbers.
3713 [Ben Laurie (CHATS)]
3715 *) Avoid filename truncation in cert requests.
3716 [Ben Laurie (CHATS)]
3718 *) Add assertions to check for (supposedly impossible) buffer
3720 [Ben Laurie (CHATS)]
3722 *) Don't cache truncated DNS entries in the local cache (this could
3723 potentially lead to a spoofing attack).
3724 [Ben Laurie (CHATS)]
3726 *) Fix various buffers to be large enough for hex/decimal
3727 representations in a platform independent manner.
3728 [Ben Laurie (CHATS)]
3730 *) Add CRYPTO_realloc_clean() to avoid information leakage when
3731 resizing buffers containing secrets, and use where appropriate.
3732 [Ben Laurie (CHATS)]
3734 *) Add BIO_indent() to avoid much slightly worrying code to do
3736 [Ben Laurie (CHATS)]
3738 *) Convert sprintf()/BIO_puts() to BIO_printf().
3739 [Ben Laurie (CHATS)]
3741 *) buffer_gets() could terminate with the buffer only half
3743 [Ben Laurie (CHATS)]
3745 *) Add assertions to prevent user-supplied crypto functions from
3746 overflowing internal buffers by having large block sizes, etc.
3747 [Ben Laurie (CHATS)]
3749 *) New OPENSSL_assert() macro (similar to assert(), but enabled
3751 [Ben Laurie (CHATS)]
3753 *) Eliminate unused copy of key in RC4.
3754 [Ben Laurie (CHATS)]
3756 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
3757 [Ben Laurie (CHATS)]
3759 *) Fix off-by-one error in EGD path.
3760 [Ben Laurie (CHATS)]
3762 *) If RANDFILE path is too long, ignore instead of truncating.
3763 [Ben Laurie (CHATS)]
3765 *) Eliminate unused and incorrectly sized X.509 structure
3767 [Ben Laurie (CHATS)]
3769 *) Eliminate unused and dangerous function knumber().
3770 [Ben Laurie (CHATS)]
3772 *) Eliminate unused and dangerous structure, KSSL_ERR.
3773 [Ben Laurie (CHATS)]
3775 *) Protect against overlong session ID context length in an encoded
3776 session object. Since these are local, this does not appear to be
3778 [Ben Laurie (CHATS)]
3780 *) Change from security patch (see 0.9.6e below) that did not affect
3781 the 0.9.6 release series:
3783 Remote buffer overflow in SSL3 protocol - an attacker could
3784 supply an oversized master key in Kerberos-enabled versions.
3786 [Ben Laurie (CHATS)]
3788 *) Change the SSL kerb5 codes to match RFC 2712.
3791 *) Make -nameopt work fully for req and add -reqopt switch.
3792 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
3794 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
3795 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
3797 *) Make sure tests can be performed even if the corresponding algorithms
3798 have been removed entirely. This was also the last step to make
3799 OpenSSL compilable with DJGPP under all reasonable conditions.
3800 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
3802 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
3803 to allow version independent disabling of normally unselected ciphers,
3804 which may be activated as a side-effect of selecting a single cipher.
3806 (E.g., cipher list string "RSA" enables ciphersuites that are left
3807 out of "ALL" because they do not provide symmetric encryption.
3808 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
3809 [Lutz Jaenicke, Bodo Moeller]
3811 *) Add appropriate support for separate platform-dependent build
3812 directories. The recommended way to make a platform-dependent
3813 build directory is the following (tested on Linux), maybe with
3816 # Place yourself outside of the OpenSSL source tree. In
3817 # this example, the environment variable OPENSSL_SOURCE
3818 # is assumed to contain the absolute OpenSSL source directory.
3819 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
3820 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
3821 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
3822 mkdir -p `dirname $F`
3823 ln -s $OPENSSL_SOURCE/$F $F
3826 To be absolutely sure not to disturb the source tree, a "make clean"
3827 is a good thing. If it isn't successfull, don't worry about it,
3828 it probably means the source directory is very clean.
3831 *) Make sure any ENGINE control commands make local copies of string
3832 pointers passed to them whenever necessary. Otherwise it is possible
3833 the caller may have overwritten (or deallocated) the original string
3834 data when a later ENGINE operation tries to use the stored values.
3835 [Götz Babin-Ebell <babinebell@trustcenter.de>]
3837 *) Improve diagnostics in file reading and command-line digests.
3838 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
3840 *) Add AES modes CFB and OFB to the object database. Correct an
3841 error in AES-CFB decryption.
3844 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
3845 allows existing EVP_CIPHER_CTX structures to be reused after
3846 calling EVP_*Final(). This behaviour is used by encryption
3847 BIOs and some applications. This has the side effect that
3848 applications must explicitly clean up cipher contexts with
3849 EVP_CIPHER_CTX_cleanup() or they will leak memory.
3852 *) Check the values of dna and dnb in bn_mul_recursive before calling
3853 bn_mul_comba (a non zero value means the a or b arrays do not contain
3854 n2 elements) and fallback to bn_mul_normal if either is not zero.
3857 *) Fix escaping of non-ASCII characters when using the -subj option
3858 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
3861 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
3862 form for "surname", serialNumber has no short form.
3863 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
3864 therefore remove "mail" short name for "internet 7".
3865 The OID for unique identifiers in X509 certificates is
3866 x500UniqueIdentifier, not uniqueIdentifier.
3867 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
3870 *) Add an "init" command to the ENGINE config module and auto initialize
3871 ENGINEs. Without any "init" command the ENGINE will be initialized
3872 after all ctrl commands have been executed on it. If init=1 the
3873 ENGINE is initailized at that point (ctrls before that point are run
3874 on the uninitialized ENGINE and after on the initialized one). If
3875 init=0 then the ENGINE will not be iniatialized at all.
3878 *) Fix the 'app_verify_callback' interface so that the user-defined
3879 argument is actually passed to the callback: In the
3880 SSL_CTX_set_cert_verify_callback() prototype, the callback
3881 declaration has been changed from
3884 int (*cb)(X509_STORE_CTX *,void *);
3885 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
3886 i=s->ctx->app_verify_callback(&ctx)
3887 has been changed into
3888 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
3890 To update applications using SSL_CTX_set_cert_verify_callback(),
3891 a dummy argument can be added to their callback functions.
3892 [D. K. Smetters <smetters@parc.xerox.com>]
3894 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
3895 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
3897 *) Add and OPENSSL_LOAD_CONF define which will cause
3898 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
3899 This allows older applications to transparently support certain
3900 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
3901 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
3902 load the config file and OPENSSL_add_all_algorithms_conf() which will
3903 always load it have also been added.
3906 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
3907 Adjust NIDs and EVP layer.
3908 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3910 *) Config modules support in openssl utility.
3912 Most commands now load modules from the config file,
3913 though in a few (such as version) this isn't done
3914 because it couldn't be used for anything.
3916 In the case of ca and req the config file used is
3917 the same as the utility itself: that is the -config
3918 command line option can be used to specify an
3922 *) Move default behaviour from OPENSSL_config(). If appname is NULL
3923 use "openssl_conf" if filename is NULL use default openssl config file.
3926 *) Add an argument to OPENSSL_config() to allow the use of an alternative
3927 config section name. Add a new flag to tolerate a missing config file
3928 and move code to CONF_modules_load_file().
3931 *) Support for crypto accelerator cards from Accelerated Encryption
3932 Processing, www.aep.ie. (Use engine 'aep')
3933 The support was copied from 0.9.6c [engine] and adapted/corrected
3934 to work with the new engine framework.
3935 [AEP Inc. and Richard Levitte]
3937 *) Support for SureWare crypto accelerator cards from Baltimore
3938 Technologies. (Use engine 'sureware')
3939 The support was copied from 0.9.6c [engine] and adapted
3940 to work with the new engine framework.
3943 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
3944 make the newer ENGINE framework commands for the CHIL engine work.
3945 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
3947 *) Make it possible to produce shared libraries on ReliantUNIX.
3948 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
3950 *) Add the configuration target debug-linux-ppro.
3951 Make 'openssl rsa' use the general key loading routines
3952 implemented in apps.c, and make those routines able to
3953 handle the key format FORMAT_NETSCAPE and the variant
3955 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3957 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
3958 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3960 *) Add -keyform to rsautl, and document -engine.
3961 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
3963 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
3964 BIO_R_NO_SUCH_FILE error code rather than the generic
3965 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
3968 *) Add new functions
3970 ERR_peek_last_error_line
3971 ERR_peek_last_error_line_data.
3972 These are similar to
3975 ERR_peek_error_line_data,
3976 but report on the latest error recorded rather than the first one
3977 still in the error queue.
3978 [Ben Laurie, Bodo Moeller]
3980 *) default_algorithms option in ENGINE config module. This allows things
3982 default_algorithms = ALL
3983 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
3986 *) Prelminary ENGINE config module.
3989 *) New experimental application configuration code.
3992 *) Change the AES code to follow the same name structure as all other
3993 symmetric ciphers, and behave the same way. Move everything to
3994 the directory crypto/aes, thereby obsoleting crypto/rijndael.
3995 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3997 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
3998 [Ben Laurie and Theo de Raadt]
4000 *) Add option to output public keys in req command.
4001 [Massimiliano Pala madwolf@openca.org]
4003 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
4004 (up to about 10% better than before for P-192 and P-224).
4007 *) New functions/macros
4009 SSL_CTX_set_msg_callback(ctx, cb)
4010 SSL_CTX_set_msg_callback_arg(ctx, arg)
4011 SSL_set_msg_callback(ssl, cb)
4012 SSL_set_msg_callback_arg(ssl, arg)
4014 to request calling a callback function
4016 void cb(int write_p, int version, int content_type,
4017 const void *buf, size_t len, SSL *ssl, void *arg)
4019 whenever a protocol message has been completely received
4020 (write_p == 0) or sent (write_p == 1). Here 'version' is the
4021 protocol version according to which the SSL library interprets
4022 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
4023 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
4024 the content type as defined in the SSL 3.0/TLS 1.0 protocol
4025 specification (change_cipher_spec(20), alert(21), handshake(22)).
4026 'buf' and 'len' point to the actual message, 'ssl' to the
4027 SSL object, and 'arg' is the application-defined value set by
4028 SSL[_CTX]_set_msg_callback_arg().
4030 'openssl s_client' and 'openssl s_server' have new '-msg' options
4031 to enable a callback that displays all protocol messages.
4034 *) Change the shared library support so shared libraries are built as
4035 soon as the corresponding static library is finished, and thereby get
4036 openssl and the test programs linked against the shared library.
4037 This still only happens when the keyword "shard" has been given to
4038 the configuration scripts.
4040 NOTE: shared library support is still an experimental thing, and
4041 backward binary compatibility is still not guaranteed.
4042 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
4044 *) Add support for Subject Information Access extension.
4045 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4047 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
4048 additional bytes when new memory had to be allocated, not just
4049 when reusing an existing buffer.
4052 *) New command line and configuration option 'utf8' for the req command.
4053 This allows field values to be specified as UTF8 strings.
4056 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
4057 runs for the former and machine-readable output for the latter.
4060 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
4061 of the e-mail address in the DN (i.e., it will go into a certificate
4062 extension only). The new configuration file option 'email_in_dn = no'
4063 has the same effect.
4064 [Massimiliano Pala madwolf@openca.org]
4066 *) Change all functions with names starting with des_ to be starting
4067 with DES_ instead. Add wrappers that are compatible with libdes,
4068 but are named _ossl_old_des_*. Finally, add macros that map the
4069 des_* symbols to the corresponding _ossl_old_des_* if libdes
4070 compatibility is desired. If OpenSSL 0.9.6c compatibility is
4071 desired, the des_* symbols will be mapped to DES_*, with one
4074 Since we provide two compatibility mappings, the user needs to
4075 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
4076 compatibility is desired. The default (i.e., when that macro
4077 isn't defined) is OpenSSL 0.9.6c compatibility.
4079 There are also macros that enable and disable the support of old
4080 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
4081 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
4082 are defined, the default will apply: to support the old des routines.
4084 In either case, one must include openssl/des.h to get the correct
4085 definitions. Do not try to just include openssl/des_old.h, that
4088 NOTE: This is a major break of an old API into a new one. Software
4089 authors are encouraged to switch to the DES_ style functions. Some
4090 time in the future, des_old.h and the libdes compatibility functions
4091 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
4092 default), and then completely removed.
4095 *) Test for certificates which contain unsupported critical extensions.
4096 If such a certificate is found during a verify operation it is
4097 rejected by default: this behaviour can be overridden by either
4098 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
4099 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
4100 X509_supported_extension() has also been added which returns 1 if a
4101 particular extension is supported.
4104 *) Modify the behaviour of EVP cipher functions in similar way to digests
4105 to retain compatibility with existing code.
4108 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
4109 compatibility with existing code. In particular the 'ctx' parameter does
4110 not have to be to be initialized before the call to EVP_DigestInit() and
4111 it is tidied up after a call to EVP_DigestFinal(). New function
4112 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
4113 EVP_MD_CTX_copy() changed to not require the destination to be
4114 initialized valid and new function EVP_MD_CTX_copy_ex() added which
4115 requires the destination to be valid.
4117 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
4118 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
4121 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
4122 so that complete 'Handshake' protocol structures are kept in memory
4123 instead of overwriting 'msg_type' and 'length' with 'body' data.
4126 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
4127 [Massimo Santin via Richard Levitte]
4129 *) Major restructuring to the underlying ENGINE code. This includes
4130 reduction of linker bloat, separation of pure "ENGINE" manipulation
4131 (initialisation, etc) from functionality dealing with implementations
4132 of specific crypto iterfaces. This change also introduces integrated
4133 support for symmetric ciphers and digest implementations - so ENGINEs
4134 can now accelerate these by providing EVP_CIPHER and EVP_MD
4135 implementations of their own. This is detailed in crypto/engine/README
4136 as it couldn't be adequately described here. However, there are a few
4137 API changes worth noting - some RSA, DSA, DH, and RAND functions that
4138 were changed in the original introduction of ENGINE code have now
4139 reverted back - the hooking from this code to ENGINE is now a good
4140 deal more passive and at run-time, operations deal directly with
4141 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
4142 dereferencing through an ENGINE pointer any more. Also, the ENGINE
4143 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
4144 they were not being used by the framework as there is no concept of a
4145 BIGNUM_METHOD and they could not be generalised to the new
4146 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
4147 ENGINE_cpy() has been removed as it cannot be consistently defined in
4151 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
4154 *) Change mkdef.pl to sort symbols that get the same entry number,
4155 and make sure the automatically generated functions ERR_load_*
4156 become part of libeay.num as well.
4159 *) New function SSL_renegotiate_pending(). This returns true once
4160 renegotiation has been requested (either SSL_renegotiate() call
4161 or HelloRequest/ClientHello receveived from the peer) and becomes
4162 false once a handshake has been completed.
4163 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
4164 sends a HelloRequest, but does not ensure that a handshake takes
4165 place. SSL_renegotiate_pending() is useful for checking if the
4166 client has followed the request.)
4169 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
4170 By default, clients may request session resumption even during
4171 renegotiation (if session ID contexts permit); with this option,
4172 session resumption is possible only in the first handshake.
4174 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4175 more bits available for options that should not be part of
4176 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4179 *) Add some demos for certificate and certificate request creation.
4182 *) Make maximum certificate chain size accepted from the peer application
4183 settable (SSL*_get/set_max_cert_list()), as proposed by
4184 "Douglas E. Engert" <deengert@anl.gov>.
4187 *) Add support for shared libraries for Unixware-7
4188 (Boyd Lynn Gerber <gerberb@zenez.com>).
4191 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4192 be done prior to destruction. Use this to unload error strings from
4193 ENGINEs that load their own error strings. NB: This adds two new API
4194 functions to "get" and "set" this destroy handler in an ENGINE.
4197 *) Alter all existing ENGINE implementations (except "openssl" and
4198 "openbsd") to dynamically instantiate their own error strings. This
4199 makes them more flexible to be built both as statically-linked ENGINEs
4200 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4201 Also, add stub code to each that makes building them as self-contained
4202 shared-libraries easier (see README.ENGINE).
4205 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4206 implementations into applications that are completely implemented in
4207 self-contained shared-libraries. The "dynamic" ENGINE exposes control
4208 commands that can be used to configure what shared-library to load and
4209 to control aspects of the way it is handled. Also, made an update to
4210 the README.ENGINE file that brings its information up-to-date and
4211 provides some information and instructions on the "dynamic" ENGINE
4212 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
4215 *) Make it possible to unload ranges of ERR strings with a new
4216 "ERR_unload_strings" function.
4219 *) Add a copy() function to EVP_MD.
4222 *) Make EVP_MD routines take a context pointer instead of just the
4223 md_data void pointer.
4226 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
4227 that the digest can only process a single chunk of data
4228 (typically because it is provided by a piece of
4229 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
4230 is only going to provide a single chunk of data, and hence the
4231 framework needn't accumulate the data for oneshot drivers.
4234 *) As with "ERR", make it possible to replace the underlying "ex_data"
4235 functions. This change also alters the storage and management of global
4236 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
4237 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
4238 index counters. The API functions that use this state have been changed
4239 to take a "class_index" rather than pointers to the class's local STACK
4240 and counter, and there is now an API function to dynamically create new
4241 classes. This centralisation allows us to (a) plug a lot of the
4242 thread-safety problems that existed, and (b) makes it possible to clean
4243 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
4244 such data would previously have always leaked in application code and
4245 workarounds were in place to make the memory debugging turn a blind eye
4246 to it. Application code that doesn't use this new function will still
4247 leak as before, but their memory debugging output will announce it now
4248 rather than letting it slide.
4250 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
4251 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
4252 has a return value to indicate success or failure.
4255 *) Make it possible to replace the underlying "ERR" functions such that the
4256 global state (2 LHASH tables and 2 locks) is only used by the "default"
4257 implementation. This change also adds two functions to "get" and "set"
4258 the implementation prior to it being automatically set the first time
4259 any other ERR function takes place. Ie. an application can call "get",
4260 pass the return value to a module it has just loaded, and that module
4261 can call its own "set" function using that value. This means the
4262 module's "ERR" operations will use (and modify) the error state in the
4263 application and not in its own statically linked copy of OpenSSL code.
4266 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
4267 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
4268 the operation, and provides a more encapsulated way for external code
4269 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
4270 to use these functions rather than manually incrementing the counts.
4272 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
4275 *) Add EVP test program.
4278 *) Add symmetric cipher support to ENGINE. Expect the API to change!
4281 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
4282 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
4283 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
4284 These allow a CRL to be built without having to access X509_CRL fields
4285 directly. Modify 'ca' application to use new functions.
4288 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
4289 bug workarounds. Rollback attack detection is a security feature.
4290 The problem will only arise on OpenSSL servers when TLSv1 is not
4291 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
4292 Software authors not wanting to support TLSv1 will have special reasons
4293 for their choice and can explicitly enable this option.
4294 [Bodo Moeller, Lutz Jaenicke]
4296 *) Rationalise EVP so it can be extended: don't include a union of
4297 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
4298 (similar to those existing for EVP_CIPHER_CTX).
4303 EVP_MD_CTX_init(&md); /* new function call */
4304 EVP_DigestInit(&md, EVP_sha1());
4305 EVP_DigestUpdate(&md, in, len);
4306 EVP_DigestFinal(&md, out, NULL);
4307 EVP_MD_CTX_cleanup(&md); /* new function call */
4311 *) Make DES key schedule conform to the usual scheme, as well as
4312 correcting its structure. This means that calls to DES functions
4313 now have to pass a pointer to a des_key_schedule instead of a
4314 plain des_key_schedule (which was actually always a pointer
4317 des_key_schedule ks;
4319 des_set_key_checked(..., &ks);
4320 des_ncbc_encrypt(..., &ks, ...);
4322 (Note that a later change renames 'des_...' into 'DES_...'.)
4325 *) Initial reduction of linker bloat: the use of some functions, such as
4326 PEM causes large amounts of unused functions to be linked in due to
4327 poor organisation. For example pem_all.c contains every PEM function
4328 which has a knock on effect of linking in large amounts of (unused)
4329 ASN1 code. Grouping together similar functions and splitting unrelated
4330 functions prevents this.
4333 *) Cleanup of EVP macros.
4336 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
4337 correct _ecb suffix.
4340 *) Add initial OCSP responder support to ocsp application. The
4341 revocation information is handled using the text based index
4342 use by the ca application. The responder can either handle
4343 requests generated internally, supplied in files (for example
4344 via a CGI script) or using an internal minimal server.
4347 *) Add configuration choices to get zlib compression for TLS.
4350 *) Changes to Kerberos SSL for RFC 2712 compliance:
4351 1. Implemented real KerberosWrapper, instead of just using
4352 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
4353 2. Implemented optional authenticator field of KerberosWrapper.
4355 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
4356 and authenticator structs; see crypto/krb5/.
4358 Generalized Kerberos calls to support multiple Kerberos libraries.
4359 [Vern Staats <staatsvr@asc.hpc.mil>,
4360 Jeffrey Altman <jaltman@columbia.edu>
4361 via Richard Levitte]
4363 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
4364 already does with RSA. testdsa.h now has 'priv_key/pub_key'
4365 values for each of the key sizes rather than having just
4366 parameters (and 'speed' generating keys each time).
4369 *) Speed up EVP routines.
4372 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
4373 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
4374 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
4375 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
4377 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
4378 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
4379 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
4382 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
4384 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
4387 *) Added the OS2-EMX target.
4388 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
4390 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
4391 to support NCONF routines in extension code. New function CONF_set_nconf()
4392 to allow functions which take an NCONF to also handle the old LHASH
4393 structure: this means that the old CONF compatible routines can be
4394 retained (in particular wrt extensions) without having to duplicate the
4395 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
4398 *) Enhance the general user interface with mechanisms for inner control
4399 and with possibilities to have yes/no kind of prompts.
4402 *) Change all calls to low level digest routines in the library and
4403 applications to use EVP. Add missing calls to HMAC_cleanup() and
4404 don't assume HMAC_CTX can be copied using memcpy().
4405 [Verdon Walker <VWalker@novell.com>, Steve Henson]
4407 *) Add the possibility to control engines through control names but with
4408 arbitrary arguments instead of just a string.
4409 Change the key loaders to take a UI_METHOD instead of a callback
4410 function pointer. NOTE: this breaks binary compatibility with earlier
4411 versions of OpenSSL [engine].
4412 Adapt the nCipher code for these new conditions and add a card insertion
4416 *) Enhance the general user interface with mechanisms to better support
4417 dialog box interfaces, application-defined prompts, the possibility
4418 to use defaults (for example default passwords from somewhere else)
4419 and interrupts/cancellations.
4422 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
4423 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
4426 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
4427 tidy up some unnecessarily weird code in 'sk_new()').
4428 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
4430 *) Change the key loading routines for ENGINEs to use the same kind
4431 callback (pem_password_cb) as all other routines that need this
4435 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
4436 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
4437 than this minimum value is recommended.
4440 *) New random seeder for OpenVMS, using the system process statistics
4441 that are easily reachable.
4444 *) Windows apparently can't transparently handle global
4445 variables defined in DLLs. Initialisations such as:
4447 const ASN1_ITEM *it = &ASN1_INTEGER_it;
4449 wont compile. This is used by the any applications that need to
4450 declare their own ASN1 modules. This was fixed by adding the option
4451 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
4452 needed for static libraries under Win32.
4455 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
4456 setting of purpose and trust fields. New X509_STORE trust and
4457 purpose functions and tidy up setting in other SSL functions.
4460 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
4461 structure. These are inherited by X509_STORE_CTX when it is
4462 initialised. This allows various defaults to be set in the
4463 X509_STORE structure (such as flags for CRL checking and custom
4464 purpose or trust settings) for functions which only use X509_STORE_CTX
4465 internally such as S/MIME.
4467 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
4468 trust settings if they are not set in X509_STORE. This allows X509_STORE
4469 purposes and trust (in S/MIME for example) to override any set by default.
4471 Add command line options for CRL checking to smime, s_client and s_server
4475 *) Initial CRL based revocation checking. If the CRL checking flag(s)
4476 are set then the CRL is looked up in the X509_STORE structure and
4477 its validity and signature checked, then if the certificate is found
4478 in the CRL the verify fails with a revoked error.
4480 Various new CRL related callbacks added to X509_STORE_CTX structure.
4482 Command line options added to 'verify' application to support this.
4484 This needs some additional work, such as being able to handle multiple
4485 CRLs with different times, extension based lookup (rather than just
4486 by subject name) and ultimately more complete V2 CRL extension
4490 *) Add a general user interface API (crypto/ui/). This is designed
4491 to replace things like des_read_password and friends (backward
4492 compatibility functions using this new API are provided).
4493 The purpose is to remove prompting functions from the DES code
4494 section as well as provide for prompting through dialog boxes in
4495 a window system and the like.
4498 *) Add "ex_data" support to ENGINE so implementations can add state at a
4499 per-structure level rather than having to store it globally.
4502 *) Make it possible for ENGINE structures to be copied when retrieved by
4503 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
4504 This causes the "original" ENGINE structure to act like a template,
4505 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
4506 operational state can be localised to each ENGINE structure, despite the
4507 fact they all share the same "methods". New ENGINE structures returned in
4508 this case have no functional references and the return value is the single
4509 structural reference. This matches the single structural reference returned
4510 by ENGINE_by_id() normally, when it is incremented on the pre-existing
4514 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
4515 needs to match any other type at all we need to manually clear the
4519 *) Changes to the "openssl engine" utility to include;
4520 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
4521 about an ENGINE's available control commands.
4522 - executing control commands from command line arguments using the
4523 '-pre' and '-post' switches. '-post' is only used if '-t' is
4524 specified and the ENGINE is successfully initialised. The syntax for
4525 the individual commands are colon-separated, for example;
4526 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
4529 *) New dynamic control command support for ENGINEs. ENGINEs can now
4530 declare their own commands (numbers), names (strings), descriptions,
4531 and input types for run-time discovery by calling applications. A
4532 subset of these commands are implicitly classed as "executable"
4533 depending on their input type, and only these can be invoked through
4534 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
4535 can be based on user input, config files, etc). The distinction is
4536 that "executable" commands cannot return anything other than a boolean
4537 result and can only support numeric or string input, whereas some
4538 discoverable commands may only be for direct use through
4539 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
4540 pointers, or other custom uses. The "executable" commands are to
4541 support parameterisations of ENGINE behaviour that can be
4542 unambiguously defined by ENGINEs and used consistently across any
4543 OpenSSL-based application. Commands have been added to all the
4544 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
4545 control over shared-library paths without source code alterations.
4548 *) Changed all ENGINE implementations to dynamically allocate their
4549 ENGINEs rather than declaring them statically. Apart from this being
4550 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
4551 this also allows the implementations to compile without using the
4552 internal engine_int.h header.
4555 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
4556 'const' value. Any code that should be able to modify a RAND_METHOD
4557 should already have non-const pointers to it (ie. they should only
4558 modify their own ones).
4561 *) Made a variety of little tweaks to the ENGINE code.
4562 - "atalla" and "ubsec" string definitions were moved from header files
4563 to C code. "nuron" string definitions were placed in variables
4564 rather than hard-coded - allowing parameterisation of these values
4565 later on via ctrl() commands.
4566 - Removed unused "#if 0"'d code.
4567 - Fixed engine list iteration code so it uses ENGINE_free() to release
4568 structural references.
4569 - Constified the RAND_METHOD element of ENGINE structures.
4570 - Constified various get/set functions as appropriate and added
4571 missing functions (including a catch-all ENGINE_cpy that duplicates
4572 all ENGINE values onto a new ENGINE except reference counts/state).
4573 - Removed NULL parameter checks in get/set functions. Setting a method
4574 or function to NULL is a way of cancelling out a previously set
4575 value. Passing a NULL ENGINE parameter is just plain stupid anyway
4576 and doesn't justify the extra error symbols and code.
4577 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
4578 flags from engine_int.h to engine.h.
4579 - Changed prototypes for ENGINE handler functions (init(), finish(),
4580 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
4583 *) Implement binary inversion algorithm for BN_mod_inverse in addition
4584 to the algorithm using long division. The binary algorithm can be
4585 used only if the modulus is odd. On 32-bit systems, it is faster
4586 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
4587 roughly 5-15% for 256-bit moduli), so we use it only for moduli
4588 up to 450 bits. In 64-bit environments, the binary algorithm
4589 appears to be advantageous for much longer moduli; here we use it
4590 for moduli up to 2048 bits.
4593 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
4594 could not support the combine flag in choice fields.
4597 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
4598 extensions from a certificate request to the certificate.
4601 *) Allow multiple 'certopt' and 'nameopt' options to be separated
4602 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
4603 file: this allows the display of the certificate about to be
4604 signed to be customised, to allow certain fields to be included
4605 or excluded and extension details. The old system didn't display
4606 multicharacter strings properly, omitted fields not in the policy
4607 and couldn't display additional details such as extensions.
4610 *) Function EC_POINTs_mul for multiple scalar multiplication
4611 of an arbitrary number of elliptic curve points
4612 \sum scalars[i]*points[i],
4613 optionally including the generator defined for the EC_GROUP:
4614 scalar*generator + \sum scalars[i]*points[i].
4616 EC_POINT_mul is a simple wrapper function for the typical case
4617 that the point list has just one item (besides the optional
4621 *) First EC_METHODs for curves over GF(p):
4623 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
4624 operations and provides various method functions that can also
4625 operate with faster implementations of modular arithmetic.
4627 EC_GFp_mont_method() reuses most functions that are part of
4628 EC_GFp_simple_method, but uses Montgomery arithmetic.
4630 [Bodo Moeller; point addition and point doubling
4631 implementation directly derived from source code provided by
4632 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
4634 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
4635 crypto/ec/ec_lib.c):
4637 Curves are EC_GROUP objects (with an optional group generator)
4638 based on EC_METHODs that are built into the library.
4640 Points are EC_POINT objects based on EC_GROUP objects.
4642 Most of the framework would be able to handle curves over arbitrary
4643 finite fields, but as there are no obvious types for fields other
4644 than GF(p), some functions are limited to that for now.
4647 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
4648 that the file contains a complete HTTP response.
4651 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
4652 change the def and num file printf format specifier from "%-40sXXX"
4653 to "%-39s XXX". The latter will always guarantee a space after the
4654 field while the former will cause them to run together if the field
4655 is 40 of more characters long.
4658 *) Constify the cipher and digest 'method' functions and structures
4659 and modify related functions to take constant EVP_MD and EVP_CIPHER
4663 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
4664 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
4667 *) Modify EVP_Digest*() routines so they now return values. Although the
4668 internal software routines can never fail additional hardware versions
4672 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
4674 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
4675 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
4678 ERR_R_NESTED_ASN1_ERROR
4680 ERR_R_MISSING_ASN1_EOS
4681 were 4 .. 9, conflicting with
4682 ERR_LIB_RSA (= ERR_R_RSA_LIB)
4684 ERR_LIB_PEM (= ERR_R_PEM_LIB).
4685 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
4687 Add new error code 'ERR_R_INTERNAL_ERROR'.
4690 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
4694 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
4695 sets the subject name for a new request or supersedes the
4696 subject name in a given request. Formats that can be parsed are
4697 'CN=Some Name, OU=myOU, C=IT'
4699 'CN=Some Name/OU=myOU/C=IT'.
4701 Add options '-batch' and '-verbose' to 'openssl req'.
4702 [Massimiliano Pala <madwolf@hackmasters.net>]
4704 *) Introduce the possibility to access global variables through
4705 functions on platform were that's the best way to handle exporting
4706 global variables in shared libraries. To enable this functionality,
4707 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
4708 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
4709 is normally done by Configure or something similar).
4711 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
4712 in the source file (foo.c) like this:
4714 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
4715 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
4717 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
4718 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
4720 OPENSSL_DECLARE_GLOBAL(int,foo);
4721 #define foo OPENSSL_GLOBAL_REF(foo)
4722 OPENSSL_DECLARE_GLOBAL(double,bar);
4723 #define bar OPENSSL_GLOBAL_REF(bar)
4725 The #defines are very important, and therefore so is including the
4726 header file everywhere where the defined globals are used.
4728 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
4729 of ASN.1 items, but that structure is a bit different.
4731 The largest change is in util/mkdef.pl which has been enhanced with
4732 better and easier to understand logic to choose which symbols should
4733 go into the Windows .def files as well as a number of fixes and code
4734 cleanup (among others, algorithm keywords are now sorted
4735 lexicographically to avoid constant rewrites).
4738 *) In BN_div() keep a copy of the sign of 'num' before writing the
4739 result to 'rm' because if rm==num the value will be overwritten
4740 and produce the wrong result if 'num' is negative: this caused
4741 problems with BN_mod() and BN_nnmod().
4744 *) Function OCSP_request_verify(). This checks the signature on an
4745 OCSP request and verifies the signer certificate. The signer
4746 certificate is just checked for a generic purpose and OCSP request
4750 *) Add OCSP_check_validity() function to check the validity of OCSP
4751 responses. OCSP responses are prepared in real time and may only
4752 be a few seconds old. Simply checking that the current time lies
4753 between thisUpdate and nextUpdate max reject otherwise valid responses
4754 caused by either OCSP responder or client clock inaccuracy. Instead
4755 we allow thisUpdate and nextUpdate to fall within a certain period of
4756 the current time. The age of the response can also optionally be
4757 checked. Two new options -validity_period and -status_age added to
4761 *) If signature or public key algorithm is unrecognized print out its
4762 OID rather that just UNKNOWN.
4765 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
4766 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
4767 ID to be generated from the issuer certificate alone which can then be
4768 passed to OCSP_id_issuer_cmp().
4771 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
4772 ASN1 modules to export functions returning ASN1_ITEM pointers
4773 instead of the ASN1_ITEM structures themselves. This adds several
4774 new macros which allow the underlying ASN1 function/structure to
4775 be accessed transparently. As a result code should not use ASN1_ITEM
4776 references directly (such as &X509_it) but instead use the relevant
4777 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
4778 use of the new ASN1 code on platforms where exporting structures
4779 is problematical (for example in shared libraries) but exporting
4780 functions returning pointers to structures is not.
4783 *) Add support for overriding the generation of SSL/TLS session IDs.
4784 These callbacks can be registered either in an SSL_CTX or per SSL.
4785 The purpose of this is to allow applications to control, if they wish,
4786 the arbitrary values chosen for use as session IDs, particularly as it
4787 can be useful for session caching in multiple-server environments. A
4788 command-line switch for testing this (and any client code that wishes
4789 to use such a feature) has been added to "s_server".
4790 [Geoff Thorpe, Lutz Jaenicke]
4792 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
4793 of the form '#if defined(...) || defined(...) || ...' and
4794 '#if !defined(...) && !defined(...) && ...'. This also avoids
4795 the growing number of special cases it was previously handling.
4798 *) Make all configuration macros available for application by making
4799 sure they are available in opensslconf.h, by giving them names starting
4800 with "OPENSSL_" to avoid conflicts with other packages and by making
4801 sure e_os2.h will cover all platform-specific cases together with
4803 Additionally, it is now possible to define configuration/platform-
4804 specific names (called "system identities"). In the C code, these
4805 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
4806 macro with the name beginning with "OPENSSL_SYS_", which is determined
4807 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
4811 *) New option -set_serial to 'req' and 'x509' this allows the serial
4812 number to use to be specified on the command line. Previously self
4813 signed certificates were hard coded with serial number 0 and the
4814 CA options of 'x509' had to use a serial number in a file which was
4818 *) New options to 'ca' utility to support V2 CRL entry extensions.
4819 Currently CRL reason, invalidity date and hold instruction are
4820 supported. Add new CRL extensions to V3 code and some new objects.
4823 *) New function EVP_CIPHER_CTX_set_padding() this is used to
4824 disable standard block padding (aka PKCS#5 padding) in the EVP
4825 API, which was previously mandatory. This means that the data is
4826 not padded in any way and so the total length much be a multiple
4827 of the block size, otherwise an error occurs.
4830 *) Initial (incomplete) OCSP SSL support.
4833 *) New function OCSP_parse_url(). This splits up a URL into its host,
4834 port and path components: primarily to parse OCSP URLs. New -url
4835 option to ocsp utility.
4838 *) New nonce behavior. The return value of OCSP_check_nonce() now
4839 reflects the various checks performed. Applications can decide
4840 whether to tolerate certain situations such as an absent nonce
4841 in a response when one was present in a request: the ocsp application
4842 just prints out a warning. New function OCSP_add1_basic_nonce()
4843 this is to allow responders to include a nonce in a response even if
4844 the request is nonce-less.
4847 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
4848 skipped when using openssl x509 multiple times on a single input file,
4849 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
4852 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
4853 set string type: to handle setting ASN1_TIME structures. Fix ca
4854 utility to correctly initialize revocation date of CRLs.
4857 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
4858 the clients preferred ciphersuites and rather use its own preferences.
4859 Should help to work around M$ SGC (Server Gated Cryptography) bug in
4860 Internet Explorer by ensuring unchanged hash method during stepup.
4861 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
4864 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
4865 to aes and add a new 'exist' option to print out symbols that don't
4869 *) Additional options to ocsp utility to allow flags to be set and
4870 additional certificates supplied.
4873 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
4874 OCSP client a number of certificate to only verify the response
4878 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
4879 handle the new API. Currently only ECB, CBC modes supported. Add new
4882 Add TLS AES ciphersuites as described in RFC3268, "Advanced
4883 Encryption Standard (AES) Ciphersuites for Transport Layer
4884 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
4885 not enabled by default and were not part of the "ALL" ciphersuite
4886 alias because they were not yet official; they could be
4887 explicitly requested by specifying the "AESdraft" ciphersuite
4888 group alias. In the final release of OpenSSL 0.9.7, the group
4889 alias is called "AES" and is part of "ALL".)
4890 [Ben Laurie, Steve Henson, Bodo Moeller]
4892 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
4893 request to response.
4896 *) Functions for OCSP responders. OCSP_request_onereq_count(),
4897 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
4898 extract information from a certificate request. OCSP_response_create()
4899 creates a response and optionally adds a basic response structure.
4900 OCSP_basic_add1_status() adds a complete single response to a basic
4901 response and returns the OCSP_SINGLERESP structure just added (to allow
4902 extensions to be included for example). OCSP_basic_add1_cert() adds a
4903 certificate to a basic response and OCSP_basic_sign() signs a basic
4904 response with various flags. New helper functions ASN1_TIME_check()
4905 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
4906 (converts ASN1_TIME to GeneralizedTime).
4909 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
4910 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
4911 structure from a certificate. X509_pubkey_digest() digests the public_key
4912 contents: this is used in various key identifiers.
4915 *) Make sk_sort() tolerate a NULL argument.
4916 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
4918 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
4919 passed by the function are trusted implicitly. If any of them signed the
4920 response then it is assumed to be valid and is not verified.
4923 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
4924 to data. This was previously part of the PKCS7 ASN1 code. This
4925 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
4926 [Steve Henson, reported by Kenneth R. Robinette
4927 <support@securenetterm.com>]
4929 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
4930 routines: without these tracing memory leaks is very painful.
4931 Fix leaks in PKCS12 and PKCS7 routines.
4934 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
4935 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
4936 effectively meant GeneralizedTime would never be used. Now it
4937 is initialised to -1 but X509_time_adj() now has to check the value
4938 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
4939 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
4940 [Steve Henson, reported by Kenneth R. Robinette
4941 <support@securenetterm.com>]
4943 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
4944 result in a zero length in the ASN1_INTEGER structure which was
4945 not consistent with the structure when d2i_ASN1_INTEGER() was used
4946 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
4947 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
4948 where it did not print out a minus for negative ASN1_INTEGER.
4951 *) Add summary printout to ocsp utility. The various functions which
4952 convert status values to strings have been renamed to:
4953 OCSP_response_status_str(), OCSP_cert_status_str() and
4954 OCSP_crl_reason_str() and are no longer static. New options
4955 to verify nonce values and to disable verification. OCSP response
4956 printout format cleaned up.
4959 *) Add additional OCSP certificate checks. These are those specified
4960 in RFC2560. This consists of two separate checks: the CA of the
4961 certificate being checked must either be the OCSP signer certificate
4962 or the issuer of the OCSP signer certificate. In the latter case the
4963 OCSP signer certificate must contain the OCSP signing extended key
4964 usage. This check is performed by attempting to match the OCSP
4965 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
4966 in the OCSP_CERTID structures of the response.
4969 *) Initial OCSP certificate verification added to OCSP_basic_verify()
4970 and related routines. This uses the standard OpenSSL certificate
4971 verify routines to perform initial checks (just CA validity) and
4972 to obtain the certificate chain. Then additional checks will be
4973 performed on the chain. Currently the root CA is checked to see
4974 if it is explicitly trusted for OCSP signing. This is used to set
4975 a root CA as a global signing root: that is any certificate that
4976 chains to that CA is an acceptable OCSP signing certificate.
4979 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
4980 extensions from a separate configuration file.
4981 As when reading extensions from the main configuration file,
4982 the '-extensions ...' option may be used for specifying the
4984 [Massimiliano Pala <madwolf@comune.modena.it>]
4986 *) New OCSP utility. Allows OCSP requests to be generated or
4987 read. The request can be sent to a responder and the output
4988 parsed, outputed or printed in text form. Not complete yet:
4989 still needs to check the OCSP response validity.
4992 *) New subcommands for 'openssl ca':
4993 'openssl ca -status <serial>' prints the status of the cert with
4994 the given serial number (according to the index file).
4995 'openssl ca -updatedb' updates the expiry status of certificates
4997 [Massimiliano Pala <madwolf@comune.modena.it>]
4999 *) New '-newreq-nodes' command option to CA.pl. This is like
5000 '-newreq', but calls 'openssl req' with the '-nodes' option
5001 so that the resulting key is not encrypted.
5002 [Damien Miller <djm@mindrot.org>]
5004 *) New configuration for the GNU Hurd.
5005 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
5007 *) Initial code to implement OCSP basic response verify. This
5008 is currently incomplete. Currently just finds the signer's
5009 certificate and verifies the signature on the response.
5012 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
5013 value of OPENSSLDIR. This is available via the new '-d' option
5014 to 'openssl version', and is also included in 'openssl version -a'.
5017 *) Allowing defining memory allocation callbacks that will be given
5018 file name and line number information in additional arguments
5019 (a const char* and an int). The basic functionality remains, as
5020 well as the original possibility to just replace malloc(),
5021 realloc() and free() by functions that do not know about these
5022 additional arguments. To register and find out the current
5023 settings for extended allocation functions, the following
5024 functions are provided:
5026 CRYPTO_set_mem_ex_functions
5027 CRYPTO_set_locked_mem_ex_functions
5028 CRYPTO_get_mem_ex_functions
5029 CRYPTO_get_locked_mem_ex_functions
5031 These work the same way as CRYPTO_set_mem_functions and friends.
5032 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
5033 extended allocation function is enabled.
5034 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
5035 a conventional allocation function is enabled.
5036 [Richard Levitte, Bodo Moeller]
5038 *) Finish off removing the remaining LHASH function pointer casts.
5039 There should no longer be any prototype-casting required when using
5040 the LHASH abstraction, and any casts that remain are "bugs". See
5041 the callback types and macros at the head of lhash.h for details
5042 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
5045 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
5046 If /dev/[u]random devices are not available or do not return enough
5047 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
5049 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
5050 /etc/entropy will be queried once each in this sequence, quering stops
5051 when enough entropy was collected without querying more sockets.
5054 *) Change the Unix RAND_poll() variant to be able to poll several
5055 random devices, as specified by DEVRANDOM, until a sufficient amount
5056 of data has been collected. We spend at most 10 ms on each file
5057 (select timeout) and read in non-blocking mode. DEVRANDOM now
5058 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
5059 (previously it was just the string "/dev/urandom"), so on typical
5060 platforms the 10 ms delay will never occur.
5061 Also separate out the Unix variant to its own file, rand_unix.c.
5062 For VMS, there's a currently-empty rand_vms.c.
5065 *) Move OCSP client related routines to ocsp_cl.c. These
5066 provide utility functions which an application needing
5067 to issue a request to an OCSP responder and analyse the
5068 response will typically need: as opposed to those which an
5069 OCSP responder itself would need which will be added later.
5071 OCSP_request_sign() signs an OCSP request with an API similar
5072 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
5073 response. OCSP_response_get1_basic() extracts basic response
5074 from response. OCSP_resp_find_status(): finds and extracts status
5075 information from an OCSP_CERTID structure (which will be created
5076 when the request structure is built). These are built from lower
5077 level functions which work on OCSP_SINGLERESP structures but
5078 wont normally be used unless the application wishes to examine
5079 extensions in the OCSP response for example.
5081 Replace nonce routines with a pair of functions.
5082 OCSP_request_add1_nonce() adds a nonce value and optionally
5083 generates a random value. OCSP_check_nonce() checks the
5084 validity of the nonce in an OCSP response.
5087 *) Change function OCSP_request_add() to OCSP_request_add0_id().
5088 This doesn't copy the supplied OCSP_CERTID and avoids the
5089 need to free up the newly created id. Change return type
5090 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
5091 This can then be used to add extensions to the request.
5092 Deleted OCSP_request_new(), since most of its functionality
5093 is now in OCSP_REQUEST_new() (and the case insensitive name
5094 clash) apart from the ability to set the request name which
5095 will be added elsewhere.
5098 *) Update OCSP API. Remove obsolete extensions argument from
5099 various functions. Extensions are now handled using the new
5100 OCSP extension code. New simple OCSP HTTP function which
5101 can be used to send requests and parse the response.
5104 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
5105 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
5106 uses the special reorder version of SET OF to sort the attributes
5107 and reorder them to match the encoded order. This resolves a long
5108 standing problem: a verify on a PKCS7 structure just after signing
5109 it used to fail because the attribute order did not match the
5110 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
5111 it uses the received order. This is necessary to tolerate some broken
5112 software that does not order SET OF. This is handled by encoding
5113 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
5114 to produce the required SET OF.
5117 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
5118 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
5119 files to get correct declarations of the ASN.1 item variables.
5122 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
5123 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
5124 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
5125 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
5126 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
5127 ASN1_ITEM and no wrapper functions.
5130 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
5131 replace the old function pointer based I/O routines. Change most of
5132 the *_d2i_bio() and *_d2i_fp() functions to use these.
5135 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
5136 lines, recognice more "algorithms" that can be deselected, and make
5137 it complain about algorithm deselection that isn't recognised.
5140 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
5141 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
5142 to use new functions. Add NO_ASN1_OLD which can be set to remove
5143 some old style ASN1 functions: this can be used to determine if old
5144 code will still work when these eventually go away.
5147 *) New extension functions for OCSP structures, these follow the
5148 same conventions as certificates and CRLs.
5151 *) New function X509V3_add1_i2d(). This automatically encodes and
5152 adds an extension. Its behaviour can be customised with various
5153 flags to append, replace or delete. Various wrappers added for
5154 certifcates and CRLs.
5157 *) Fix to avoid calling the underlying ASN1 print routine when
5158 an extension cannot be parsed. Correct a typo in the
5159 OCSP_SERVICELOC extension. Tidy up print OCSP format.
5162 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
5163 entries for variables.
5166 *) Add functionality to apps/openssl.c for detecting locking
5167 problems: As the program is single-threaded, all we have
5168 to do is register a locking callback using an array for
5169 storing which locks are currently held by the program.
5172 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
5173 SSL_get_ex_data_X509_STORE_idx(), which is used in
5174 ssl_verify_cert_chain() and thus can be called at any time
5175 during TLS/SSL handshakes so that thread-safety is essential.
5176 Unfortunately, the ex_data design is not at all suited
5177 for multi-threaded use, so it probably should be abolished.
5180 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5181 [Broadcom, tweaked and integrated by Geoff Thorpe]
5183 *) Move common extension printing code to new function
5184 X509V3_print_extensions(). Reorganise OCSP print routines and
5185 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5188 *) New function X509_signature_print() to remove duplication in some
5192 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5193 set (this was treated exactly the same as SET OF previously). This
5194 is used to reorder the STACK representing the structure to match the
5195 encoding. This will be used to get round a problem where a PKCS7
5196 structure which was signed could not be verified because the STACK
5197 order did not reflect the encoded order.
5200 *) Reimplement the OCSP ASN1 module using the new code.
5203 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5204 for its ASN1 operations. The old style function pointers still exist
5205 for now but they will eventually go away.
5208 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
5209 completely replaces the old ASN1 functionality with a table driven
5210 encoder and decoder which interprets an ASN1_ITEM structure describing
5211 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
5212 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
5213 has also been converted to the new form.
5216 *) Change BN_mod_exp_recp so that negative moduli are tolerated
5217 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
5218 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
5219 for negative moduli.
5222 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
5223 of not touching the result's sign bit.
5226 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
5230 *) Changed the LHASH code to use prototypes for callbacks, and created
5231 macros to declare and implement thin (optionally static) functions
5232 that provide type-safety and avoid function pointer casting for the
5233 type-specific callbacks.
5236 *) Added Kerberos Cipher Suites to be used with TLS, as written in
5238 [Veers Staats <staatsvr@asc.hpc.mil>,
5239 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
5241 *) Reformat the FAQ so the different questions and answers can be divided
5242 in sections depending on the subject.
5245 *) Have the zlib compression code load ZLIB.DLL dynamically under
5249 *) New function BN_mod_sqrt for computing square roots modulo a prime
5250 (using the probabilistic Tonelli-Shanks algorithm unless
5251 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
5252 be handled deterministically).
5253 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5255 *) Make BN_mod_inverse faster by explicitly handling small quotients
5256 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
5257 512 bits], about 30% for larger ones [1024 or 2048 bits].)
5260 *) New function BN_kronecker.
5263 *) Fix BN_gcd so that it works on negative inputs; the result is
5264 positive unless both parameters are zero.
5265 Previously something reasonably close to an infinite loop was
5266 possible because numbers could be growing instead of shrinking
5267 in the implementation of Euclid's algorithm.
5270 *) Fix BN_is_word() and BN_is_one() macros to take into account the
5271 sign of the number in question.
5273 Fix BN_is_word(a,w) to work correctly for w == 0.
5275 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
5276 because its test if the absolute value of 'a' equals 'w'.
5277 Note that BN_abs_is_word does *not* handle w == 0 reliably;
5278 it exists mostly for use in the implementations of BN_is_zero(),
5279 BN_is_one(), and BN_is_word().
5282 *) New function BN_swap.
5285 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
5286 the exponentiation functions are more likely to produce reasonable
5287 results on negative inputs.
5290 *) Change BN_mod_mul so that the result is always non-negative.
5291 Previously, it could be negative if one of the factors was negative;
5292 I don't think anyone really wanted that behaviour.
5295 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
5296 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
5297 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
5298 and add new functions:
5307 BN_mod_lshift1_quick
5311 These functions always generate non-negative results.
5313 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
5314 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
5316 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
5317 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
5318 be reduced modulo m.
5319 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5322 The following entry accidentily appeared in the CHANGES file
5323 distributed with OpenSSL 0.9.7. The modifications described in
5324 it do *not* apply to OpenSSL 0.9.7.
5326 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5327 was actually never needed) and in BN_mul(). The removal in BN_mul()
5328 required a small change in bn_mul_part_recursive() and the addition
5329 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5330 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5331 bn_sub_words() and bn_add_words() except they take arrays with
5336 *) In 'openssl passwd', verify passwords read from the terminal
5337 unless the '-salt' option is used (which usually means that
5338 verification would just waste user's time since the resulting
5339 hash is going to be compared with some given password hash)
5340 or the new '-noverify' option is used.
5342 This is an incompatible change, but it does not affect
5343 non-interactive use of 'openssl passwd' (passwords on the command
5344 line, '-stdin' option, '-in ...' option) and thus should not
5348 *) Remove all references to RSAref, since there's no more need for it.
5351 *) Make DSO load along a path given through an environment variable
5352 (SHLIB_PATH) with shl_load().
5355 *) Constify the ENGINE code as a result of BIGNUM constification.
5356 Also constify the RSA code and most things related to it. In a
5357 few places, most notable in the depth of the ASN.1 code, ugly
5358 casts back to non-const were required (to be solved at a later
5362 *) Make it so the openssl application has all engines loaded by default.
5365 *) Constify the BIGNUM routines a little more.
5368 *) Add the following functions:
5370 ENGINE_load_cswift()
5372 ENGINE_load_atalla()
5374 ENGINE_load_builtin_engines()
5376 That way, an application can itself choose if external engines that
5377 are built-in in OpenSSL shall ever be used or not. The benefit is
5378 that applications won't have to be linked with libdl or other dso
5379 libraries unless it's really needed.
5381 Changed 'openssl engine' to load all engines on demand.
5382 Changed the engine header files to avoid the duplication of some
5383 declarations (they differed!).
5386 *) 'openssl engine' can now list capabilities.
5389 *) Better error reporting in 'openssl engine'.
5392 *) Never call load_dh_param(NULL) in s_server.
5395 *) Add engine application. It can currently list engines by name and
5396 identity, and test if they are actually available.
5399 *) Improve RPM specification file by forcing symbolic linking and making
5400 sure the installed documentation is also owned by root.root.
5401 [Damien Miller <djm@mindrot.org>]
5403 *) Give the OpenSSL applications more possibilities to make use of
5404 keys (public as well as private) handled by engines.
5407 *) Add OCSP code that comes from CertCo.
5410 *) Add VMS support for the Rijndael code.
5413 *) Added untested support for Nuron crypto accelerator.
5416 *) Add support for external cryptographic devices. This code was
5417 previously distributed separately as the "engine" branch.
5418 [Geoff Thorpe, Richard Levitte]
5420 *) Rework the filename-translation in the DSO code. It is now possible to
5421 have far greater control over how a "name" is turned into a filename
5422 depending on the operating environment and any oddities about the
5423 different shared library filenames on each system.
5426 *) Support threads on FreeBSD-elf in Configure.
5429 *) Fix for SHA1 assembly problem with MASM: it produces
5430 warnings about corrupt line number information when assembling
5431 with debugging information. This is caused by the overlapping
5433 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
5436 NCONF_get_number() has no error checking at all. As a replacement,
5437 NCONF_get_number_e() is defined (_e for "error checking") and is
5438 promoted strongly. The old NCONF_get_number is kept around for
5439 binary backward compatibility.
5440 Make it possible for methods to load from something other than a BIO,
5441 by providing a function pointer that is given a name instead of a BIO.
5442 For example, this could be used to load configuration data from an
5446 *) Fix for non blocking accept BIOs. Added new I/O special reason
5447 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
5448 with non blocking I/O was not possible because no retry code was
5449 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
5453 *) Added the beginnings of Rijndael support.
5456 *) Fix for bug in DirectoryString mask setting. Add support for
5457 X509_NAME_print_ex() in 'req' and X509_print_ex() function
5458 to allow certificate printing to more controllable, additional
5459 'certopt' option to 'x509' to allow new printing options to be
5463 *) Clean old EAY MD5 hack from e_os.h.
5466 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
5468 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5469 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5470 [Joe Orton, Steve Henson]
5472 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
5474 *) Fix additional bug revealed by the NISCC test suite:
5476 Stop bug triggering large recursion when presented with
5477 certain ASN.1 tags (CVE-2003-0851)
5480 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
5482 *) Fix various bugs revealed by running the NISCC test suite:
5484 Stop out of bounds reads in the ASN1 code when presented with
5485 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5487 If verify callback ignores invalid public key errors don't try to check
5488 certificate signature with the NULL public key.
5492 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5493 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5497 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5498 extra data after the compression methods not only for TLS 1.0
5499 but also for SSL 3.0 (as required by the specification).
5500 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5502 *) Change X509_certificate_type() to mark the key as exported/exportable
5503 when it's 512 *bits* long, not 512 bytes.
5506 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
5508 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5509 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5510 a protocol version number mismatch like a decryption error
5511 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5514 *) Turn on RSA blinding by default in the default implementation
5515 to avoid a timing attack. Applications that don't want it can call
5516 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5517 They would be ill-advised to do so in most cases.
5518 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5520 *) Change RSA blinding code so that it works when the PRNG is not
5521 seeded (in this case, the secret RSA exponent is abused as
5522 an unpredictable seed -- if it is not unpredictable, there
5523 is no point in blinding anyway). Make RSA blinding thread-safe
5524 by remembering the creator's thread ID in rsa->blinding and
5525 having all other threads use local one-time blinding factors
5526 (this requires more computation than sharing rsa->blinding, but
5527 avoids excessive locking; and if an RSA object is not shared
5528 between threads, blinding will still be very fast).
5531 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
5533 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5534 via timing by performing a MAC computation even if incorrrect
5535 block cipher padding has been found. This is a countermeasure
5536 against active attacks where the attacker has to distinguish
5537 between bad padding and a MAC verification error. (CVE-2003-0078)
5539 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5540 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5541 Martin Vuagnoux (EPFL, Ilion)]
5543 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
5545 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
5546 memory from it's contents. This is done with a counter that will
5547 place alternating values in each byte. This can be used to solve
5548 two issues: 1) the removal of calls to memset() by highly optimizing
5549 compilers, and 2) cleansing with other values than 0, since those can
5550 be read through on certain media, for example a swap space on disk.
5553 *) Bugfix: client side session caching did not work with external caching,
5554 because the session->cipher setting was not restored when reloading
5555 from the external cache. This problem was masked, when
5556 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
5557 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
5560 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
5561 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
5562 [Zeev Lieber <zeev-l@yahoo.com>]
5564 *) Undo an undocumented change introduced in 0.9.6e which caused
5565 repeated calls to OpenSSL_add_all_ciphers() and
5566 OpenSSL_add_all_digests() to be ignored, even after calling
5570 *) Change the default configuration reader to deal with last line not
5571 being properly terminated.
5574 *) Change X509_NAME_cmp() so it applies the special rules on handling
5575 DN values that are of type PrintableString, as well as RDNs of type
5576 emailAddress where the value has the type ia5String.
5577 [stefank@valicert.com via Richard Levitte]
5579 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
5580 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
5581 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
5582 the bitwise-OR of the two for use by the majority of applications
5583 wanting this behaviour, and update the docs. The documented
5584 behaviour and actual behaviour were inconsistent and had been
5585 changing anyway, so this is more a bug-fix than a behavioural
5587 [Geoff Thorpe, diagnosed by Nadav Har'El]
5589 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
5590 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
5593 *) Fix initialization code race conditions in
5594 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
5595 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
5596 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
5597 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
5598 ssl2_get_cipher_by_char(),
5599 ssl3_get_cipher_by_char().
5600 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
5602 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
5603 the cached sessions are flushed, as the remove_cb() might use ex_data
5604 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
5605 (see [openssl.org #212]).
5606 [Geoff Thorpe, Lutz Jaenicke]
5608 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
5609 length, instead of the encoding length to d2i_ASN1_OBJECT.
5612 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
5614 *) [In 0.9.6g-engine release:]
5615 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
5616 [Lynn Gazis <lgazis@rainbow.com>]
5618 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
5620 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
5621 and get fix the header length calculation.
5622 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
5623 Alon Kantor <alonk@checkpoint.com> (and others),
5626 *) Use proper error handling instead of 'assertions' in buffer
5627 overflow checks added in 0.9.6e. This prevents DoS (the
5628 assertions could call abort()).
5629 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
5631 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
5633 *) Add various sanity checks to asn1_get_length() to reject
5634 the ASN1 length bytes if they exceed sizeof(long), will appear
5635 negative or the content length exceeds the length of the
5637 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5639 *) Fix cipher selection routines: ciphers without encryption had no flags
5640 for the cipher strength set and where therefore not handled correctly
5641 by the selection routines (PR #130).
5644 *) Fix EVP_dsa_sha macro.
5648 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
5649 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
5650 that was added in OpenSSL 0.9.6d.
5652 As the countermeasure turned out to be incompatible with some
5653 broken SSL implementations, the new option is part of SSL_OP_ALL.
5654 SSL_OP_ALL is usually employed when compatibility with weird SSL
5655 implementations is desired (e.g. '-bugs' option to 's_client' and
5656 's_server'), so the new option is automatically set in many
5660 *) Changes in security patch:
5662 Changes marked "(CHATS)" were sponsored by the Defense Advanced
5663 Research Projects Agency (DARPA) and Air Force Research Laboratory,
5664 Air Force Materiel Command, USAF, under agreement number
5667 *) Add various sanity checks to asn1_get_length() to reject
5668 the ASN1 length bytes if they exceed sizeof(long), will appear
5669 negative or the content length exceeds the length of the
5670 supplied buffer. (CVE-2002-0659)
5671 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5673 *) Assertions for various potential buffer overflows, not known to
5675 [Ben Laurie (CHATS)]
5677 *) Various temporary buffers to hold ASCII versions of integers were
5678 too small for 64 bit platforms. (CVE-2002-0655)
5679 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
5681 *) Remote buffer overflow in SSL3 protocol - an attacker could
5682 supply an oversized session ID to a client. (CVE-2002-0656)
5683 [Ben Laurie (CHATS)]
5685 *) Remote buffer overflow in SSL2 protocol - an attacker could
5686 supply an oversized client master key. (CVE-2002-0656)
5687 [Ben Laurie (CHATS)]
5689 Changes between 0.9.6c and 0.9.6d [9 May 2002]
5691 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
5692 encoded as NULL) with id-dsa-with-sha1.
5693 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
5695 *) Check various X509_...() return values in apps/req.c.
5696 [Nils Larsch <nla@trustcenter.de>]
5698 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
5699 an end-of-file condition would erronously be flagged, when the CRLF
5700 was just at the end of a processed block. The bug was discovered when
5701 processing data through a buffering memory BIO handing the data to a
5702 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
5703 <ptsekov@syntrex.com> and Nedelcho Stanev.
5706 *) Implement a countermeasure against a vulnerability recently found
5707 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
5708 before application data chunks to avoid the use of known IVs
5709 with data potentially chosen by the attacker.
5712 *) Fix length checks in ssl3_get_client_hello().
5715 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
5716 to prevent ssl3_read_internal() from incorrectly assuming that
5717 ssl3_read_bytes() found application data while handshake
5718 processing was enabled when in fact s->s3->in_read_app_data was
5719 merely automatically cleared during the initial handshake.
5720 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
5722 *) Fix object definitions for Private and Enterprise: they were not
5723 recognized in their shortname (=lowercase) representation. Extend
5724 obj_dat.pl to issue an error when using undefined keywords instead
5725 of silently ignoring the problem (Svenning Sorensen
5726 <sss@sss.dnsalias.net>).
5729 *) Fix DH_generate_parameters() so that it works for 'non-standard'
5730 generators, i.e. generators other than 2 and 5. (Previously, the
5731 code did not properly initialise the 'add' and 'rem' values to
5732 BN_generate_prime().)
5734 In the new general case, we do not insist that 'generator' is
5735 actually a primitive root: This requirement is rather pointless;
5736 a generator of the order-q subgroup is just as good, if not
5740 *) Map new X509 verification errors to alerts. Discovered and submitted by
5741 Tom Wu <tom@arcot.com>.
5744 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
5745 returning non-zero before the data has been completely received
5746 when using non-blocking I/O.
5747 [Bodo Moeller; problem pointed out by John Hughes]
5749 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
5750 [Ben Laurie, Lutz Jaenicke]
5752 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
5753 Yoram Zahavi <YoramZ@gilian.com>).
5756 *) Add information about CygWin 1.3 and on, and preserve proper
5757 configuration for the versions before that.
5758 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5760 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
5761 check whether we deal with a copy of a session and do not delete from
5762 the cache in this case. Problem reported by "Izhar Shoshani Levi"
5763 <izhar@checkpoint.com>.
5766 *) Do not store session data into the internal session cache, if it
5767 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
5768 flag is set). Proposed by Aslam <aslam@funk.com>.
5771 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
5775 *) [In 0.9.6d-engine release:]
5776 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5777 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5779 *) Add the configuration target linux-s390x.
5780 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
5782 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
5783 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
5784 variable as an indication that a ClientHello message has been
5785 received. As the flag value will be lost between multiple
5786 invocations of ssl3_accept when using non-blocking I/O, the
5787 function may not be aware that a handshake has actually taken
5788 place, thus preventing a new session from being added to the
5791 To avoid this problem, we now set s->new_session to 2 instead of
5792 using a local variable.
5793 [Lutz Jaenicke, Bodo Moeller]
5795 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
5796 if the SSL_R_LENGTH_MISMATCH error is detected.
5797 [Geoff Thorpe, Bodo Moeller]
5799 *) New 'shared_ldflag' column in Configure platform table.
5802 *) Fix EVP_CIPHER_mode macro.
5803 ["Dan S. Camper" <dan@bti.net>]
5805 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
5806 type, we must throw them away by setting rr->length to 0.
5807 [D P Chang <dpc@qualys.com>]
5809 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
5811 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
5812 <Dominikus.Scherkl@biodata.com>. (The previous implementation
5813 worked incorrectly for those cases where range = 10..._2 and
5814 3*range is two bits longer than range.)
5817 *) Only add signing time to PKCS7 structures if it is not already
5821 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
5822 OBJ_ld_ce should be OBJ_id_ce.
5823 Also some ip-pda OIDs in crypto/objects/objects.txt were
5824 incorrect (cf. RFC 3039).
5825 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
5827 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
5828 returns early because it has nothing to do.
5829 [Andy Schneider <andy.schneider@bjss.co.uk>]
5831 *) [In 0.9.6c-engine release:]
5832 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
5833 [Andy Schneider <andy.schneider@bjss.co.uk>]
5835 *) [In 0.9.6c-engine release:]
5836 Add support for Cryptographic Appliance's keyserver technology.
5837 (Use engine 'keyclient')
5838 [Cryptographic Appliances and Geoff Thorpe]
5840 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
5841 is called via tools/c89.sh because arguments have to be
5842 rearranged (all '-L' options must appear before the first object
5844 [Richard Shapiro <rshapiro@abinitio.com>]
5846 *) [In 0.9.6c-engine release:]
5847 Add support for Broadcom crypto accelerator cards, backported
5849 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
5851 *) [In 0.9.6c-engine release:]
5852 Add support for SureWare crypto accelerator cards from
5853 Baltimore Technologies. (Use engine 'sureware')
5854 [Baltimore Technologies and Mark Cox]
5856 *) [In 0.9.6c-engine release:]
5857 Add support for crypto accelerator cards from Accelerated
5858 Encryption Processing, www.aep.ie. (Use engine 'aep')
5859 [AEP Inc. and Mark Cox]
5861 *) Add a configuration entry for gcc on UnixWare.
5862 [Gary Benson <gbenson@redhat.com>]
5864 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
5865 messages are stored in a single piece (fixed-length part and
5866 variable-length part combined) and fix various bugs found on the way.
5869 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
5870 instead. BIO_gethostbyname() does not know what timeouts are
5871 appropriate, so entries would stay in cache even when they have
5873 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
5875 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
5876 faced with a pathologically small ClientHello fragment that does
5877 not contain client_version: Instead of aborting with an error,
5878 simply choose the highest available protocol version (i.e.,
5879 TLS 1.0 unless it is disabled). In practice, ClientHello
5880 messages are never sent like this, but this change gives us
5881 strictly correct behaviour at least for TLS.
5884 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
5885 never resets s->method to s->ctx->method when called from within
5886 one of the SSL handshake functions.
5887 [Bodo Moeller; problem pointed out by Niko Baric]
5889 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
5890 (sent using the client's version number) if client_version is
5891 smaller than the protocol version in use. Also change
5892 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
5893 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
5894 the client will at least see that alert.
5897 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
5901 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
5902 client receives HelloRequest while in a handshake.
5903 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
5905 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
5906 should end in 'break', not 'goto end' which circuments various
5907 cleanups done in state SSL_ST_OK. But session related stuff
5908 must be disabled for SSL_ST_OK in the case that we just sent a
5911 Also avoid some overhead by not calling ssl_init_wbio_buffer()
5912 before just sending a HelloRequest.
5913 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
5915 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
5916 reveal whether illegal block cipher padding was found or a MAC
5917 verification error occured. (Neither SSLerr() codes nor alerts
5918 are directly visible to potential attackers, but the information
5919 may leak via logfiles.)
5921 Similar changes are not required for the SSL 2.0 implementation
5922 because the number of padding bytes is sent in clear for SSL 2.0,
5923 and the extra bytes are just ignored. However ssl/s2_pkt.c
5924 failed to verify that the purported number of padding bytes is in
5928 *) Add OpenUNIX-8 support including shared libraries
5929 (Boyd Lynn Gerber <gerberb@zenez.com>).
5932 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
5933 'wristwatch attack' using huge encoding parameters (cf.
5934 James H. Manger's CRYPTO 2001 paper). Note that the
5935 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
5936 encoding parameters and hence was not vulnerable.
5939 *) BN_sqr() bug fix.
5940 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
5942 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
5943 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
5944 followed by modular reduction.
5945 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
5947 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
5948 equivalent based on BN_pseudo_rand() instead of BN_rand().
5951 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
5952 This function was broken, as the check for a new client hello message
5953 to handle SGC did not allow these large messages.
5954 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
5957 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
5960 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
5961 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
5964 *) Rework the configuration and shared library support for Tru64 Unix.
5965 The configuration part makes use of modern compiler features and
5966 still retains old compiler behavior for those that run older versions
5967 of the OS. The shared library support part includes a variant that
5968 uses the RPATH feature, and is available through the special
5969 configuration target "alpha-cc-rpath", which will never be selected
5971 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
5973 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
5974 with the same message size as in ssl3_get_certificate_request().
5975 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
5976 messages might inadvertently be reject as too long.
5977 [Petr Lampa <lampa@fee.vutbr.cz>]
5979 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
5982 *) Modified SSL library such that the verify_callback that has been set
5983 specificly for an SSL object with SSL_set_verify() is actually being
5984 used. Before the change, a verify_callback set with this function was
5985 ignored and the verify_callback() set in the SSL_CTX at the time of
5986 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
5987 to allow the necessary settings.
5990 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
5991 explicitly to NULL, as at least on Solaris 8 this seems not always to be
5992 done automatically (in contradiction to the requirements of the C
5993 standard). This made problems when used from OpenSSH.
5996 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
5997 dh->length and always used
5999 BN_rand_range(priv_key, dh->p).
6001 BN_rand_range() is not necessary for Diffie-Hellman, and this
6002 specific range makes Diffie-Hellman unnecessarily inefficient if
6003 dh->length (recommended exponent length) is much smaller than the
6004 length of dh->p. We could use BN_rand_range() if the order of
6005 the subgroup was stored in the DH structure, but we only have
6010 BN_rand(priv_key, l, ...)
6012 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
6018 RSA_eay_public_encrypt
6019 RSA_eay_private_decrypt
6020 RSA_eay_private_encrypt (signing)
6021 RSA_eay_public_decrypt (signature verification)
6023 (default implementations for RSA_public_encrypt,
6024 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
6025 always reject numbers >= n.
6028 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
6029 to synchronize access to 'locking_thread'. This is necessary on
6030 systems where access to 'locking_thread' (an 'unsigned long'
6031 variable) is not atomic.
6034 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
6035 *before* setting the 'crypto_lock_rand' flag. The previous code had
6036 a race condition if 0 is a valid thread ID.
6037 [Travis Vitek <vitek@roguewave.com>]
6039 *) Add support for shared libraries under Irix.
6040 [Albert Chin-A-Young <china@thewrittenword.com>]
6042 *) Add configuration option to build on Linux on both big-endian and
6044 [Ralf Baechle <ralf@uni-koblenz.de>]
6046 *) Add the possibility to create shared libraries on HP-UX.
6049 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
6051 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
6052 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
6053 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
6054 PRNG state recovery was possible based on the output of
6055 one PRNG request appropriately sized to gain knowledge on
6056 'md' followed by enough consecutive 1-byte PRNG requests
6057 to traverse all of 'state'.
6059 1. When updating 'md_local' (the current thread's copy of 'md')
6060 during PRNG output generation, hash all of the previous
6061 'md_local' value, not just the half used for PRNG output.
6063 2. Make the number of bytes from 'state' included into the hash
6064 independent from the number of PRNG bytes requested.
6066 The first measure alone would be sufficient to avoid
6067 Markku-Juhani's attack. (Actually it had never occurred
6068 to me that the half of 'md_local' used for chaining was the
6069 half from which PRNG output bytes were taken -- I had always
6070 assumed that the secret half would be used.) The second
6071 measure makes sure that additional data from 'state' is never
6072 mixed into 'md_local' in small portions; this heuristically
6073 further strengthens the PRNG.
6076 *) Fix crypto/bn/asm/mips3.s.
6079 *) When only the key is given to "enc", the IV is undefined. Print out
6080 an error message in this case.
6083 *) Handle special case when X509_NAME is empty in X509 printing routines.
6086 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
6087 positive and less than q.
6090 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
6091 used: it isn't thread safe and the add_lock_callback should handle
6093 [Paul Rose <Paul.Rose@bridge.com>]
6095 *) Verify that incoming data obeys the block size in
6096 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
6100 [Ulf Möller, Bodo Möller]
6102 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
6103 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
6104 when fixing the server behaviour for backwards-compatible 'client
6105 hello' messages. (Note that the attack is impractical against
6106 SSL 3.0 and TLS 1.0 anyway because length and version checking
6107 means that the probability of guessing a valid ciphertext is
6108 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
6111 Before 0.9.5, the countermeasure (hide the error by generating a
6112 random 'decryption result') did not work properly because
6113 ERR_clear_error() was missing, meaning that SSL_get_error() would
6114 detect the supposedly ignored error.
6116 Both problems are now fixed.
6119 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
6120 (previously it was 1024).
6123 *) Fix for compatibility mode trust settings: ignore trust settings
6124 unless some valid trust or reject settings are present.
6127 *) Fix for blowfish EVP: its a variable length cipher.
6130 *) Fix various bugs related to DSA S/MIME verification. Handle missing
6131 parameters in DSA public key structures and return an error in the
6132 DSA routines if parameters are absent.
6135 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
6136 in the current directory if neither $RANDFILE nor $HOME was set.
6137 RAND_file_name() in 0.9.6a returned NULL in this case. This has
6138 caused some confusion to Windows users who haven't defined $HOME.
6139 Thus RAND_file_name() is changed again: e_os.h can define a
6140 DEFAULT_HOME, which will be used if $HOME is not set.
6141 For Windows, we use "C:"; on other platforms, we still require
6142 environment variables.
6144 *) Move 'if (!initialized) RAND_poll()' into regions protected by
6145 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
6146 having multiple threads call RAND_poll() concurrently.
6149 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
6150 combination of a flag and a thread ID variable.
6151 Otherwise while one thread is in ssleay_rand_bytes (which sets the
6152 flag), *other* threads can enter ssleay_add_bytes without obeying
6153 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
6154 that they do not hold after the first thread unsets add_do_not_lock).
6157 *) Change bctest again: '-x' expressions are not available in all
6161 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
6163 *) Fix a couple of memory leaks in PKCS7_dataDecode()
6164 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
6166 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
6167 the default extension for executables, if any. Also, make the perl
6168 scripts that use symlink() to test if it really exists and use "cp"
6169 if it doesn't. All this made OpenSSL compilable and installable in
6173 *) Fix for asn1_GetSequence() for indefinite length constructed data.
6174 If SEQUENCE is length is indefinite just set c->slen to the total
6175 amount of data available.
6176 [Steve Henson, reported by shige@FreeBSD.org]
6177 [This change does not apply to 0.9.7.]
6179 *) Change bctest to avoid here-documents inside command substitution
6180 (workaround for FreeBSD /bin/sh bug).
6181 For compatibility with Ultrix, avoid shell functions (introduced
6182 in the bctest version that searches along $PATH).
6185 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6186 with des_encrypt() defined on some operating systems, like Solaris
6190 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6191 On the Importance of Eliminating Errors in Cryptographic
6192 Computations, J. Cryptology 14 (2001) 2, 101-119,
6193 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6196 *) MIPS assembler BIGNUM division bug fix.
6199 *) Disabled incorrect Alpha assembler code.
6202 *) Fix PKCS#7 decode routines so they correctly update the length
6203 after reading an EOC for the EXPLICIT tag.
6205 [This change does not apply to 0.9.7.]
6207 *) Fix bug in PKCS#12 key generation routines. This was triggered
6208 if a 3DES key was generated with a 0 initial byte. Include
6209 PKCS12_BROKEN_KEYGEN compilation option to retain the old
6210 (but broken) behaviour.
6213 *) Enhance bctest to search for a working bc along $PATH and print
6215 [Tim Rice <tim@multitalents.net> via Richard Levitte]
6217 *) Fix memory leaks in err.c: free err_data string if necessary;
6218 don't write to the wrong index in ERR_set_error_data.
6221 *) Implement ssl23_peek (analogous to ssl23_read), which previously
6225 *) Replace rdtsc with _emit statements for VC++ version 5.
6226 [Jeremy Cooper <jeremy@baymoo.org>]
6228 *) Make it possible to reuse SSLv2 sessions.
6231 *) In copy_email() check for >= 0 as a return value for
6232 X509_NAME_get_index_by_NID() since 0 is a valid index.
6233 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
6235 *) Avoid coredump with unsupported or invalid public keys by checking if
6236 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
6237 PKCS7_verify() fails with non detached data.
6240 *) Don't use getenv in library functions when run as setuid/setgid.
6241 New function OPENSSL_issetugid().
6244 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
6245 due to incorrect handling of multi-threading:
6247 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
6249 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
6251 3. Count how many times MemCheck_off() has been called so that
6252 nested use can be treated correctly. This also avoids
6253 inband-signalling in the previous code (which relied on the
6254 assumption that thread ID 0 is impossible).
6257 *) Add "-rand" option also to s_client and s_server.
6260 *) Fix CPU detection on Irix 6.x.
6261 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
6262 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6264 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
6267 [This change does not apply to 0.9.7.]
6269 *) Use the cached encoding of an X509_NAME structure rather than
6270 copying it. This is apparently the reason for the libsafe "errors"
6271 but the code is actually correct.
6274 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
6275 Bleichenbacher's DSA attack.
6276 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
6277 to be set and top=0 forces the highest bit to be set; top=-1 is new
6278 and leaves the highest bit random.
6279 [Ulf Moeller, Bodo Moeller]
6281 *) In the NCONF_...-based implementations for CONF_... queries
6282 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
6283 a temporary CONF structure with the data component set to NULL
6284 (which gives segmentation faults in lh_retrieve).
6285 Instead, use NULL for the CONF pointer in CONF_get_string and
6286 CONF_get_number (which may use environment variables) and directly
6287 return NULL from CONF_get_section.
6290 *) Fix potential buffer overrun for EBCDIC.
6293 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
6294 keyUsage if basicConstraints absent for a CA.
6297 *) Make SMIME_write_PKCS7() write mail header values with a format that
6298 is more generally accepted (no spaces before the semicolon), since
6299 some programs can't parse those values properly otherwise. Also make
6300 sure BIO's that break lines after each write do not create invalid
6304 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
6305 macros previously used would not encode an empty SEQUENCE OF
6306 and break the signature.
6308 [This change does not apply to 0.9.7.]
6310 *) Zero the premaster secret after deriving the master secret in
6314 *) Add some EVP_add_digest_alias registrations (as found in
6315 OpenSSL_add_all_digests()) to SSL_library_init()
6316 aka OpenSSL_add_ssl_algorithms(). This provides improved
6317 compatibility with peers using X.509 certificates
6318 with unconventional AlgorithmIdentifier OIDs.
6321 *) Fix for Irix with NO_ASM.
6322 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6324 *) ./config script fixes.
6325 [Ulf Moeller, Richard Levitte]
6327 *) Fix 'openssl passwd -1'.
6330 *) Change PKCS12_key_gen_asc() so it can cope with non null
6331 terminated strings whose length is passed in the passlen
6332 parameter, for example from PEM callbacks. This was done
6333 by adding an extra length parameter to asc2uni().
6334 [Steve Henson, reported by <oddissey@samsung.co.kr>]
6336 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
6337 call failed, free the DSA structure.
6340 *) Fix to uni2asc() to cope with zero length Unicode strings.
6341 These are present in some PKCS#12 files.
6344 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
6345 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
6346 when writing a 32767 byte record.
6347 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
6349 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
6350 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
6352 (RSA objects have a reference count access to which is protected
6353 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
6354 so they are meant to be shared between threads.)
6355 [Bodo Moeller, Geoff Thorpe; original patch submitted by
6356 "Reddie, Steven" <Steven.Reddie@ca.com>]
6358 *) Fix a deadlock in CRYPTO_mem_leaks().
6361 *) Use better test patterns in bntest.
6364 *) rand_win.c fix for Borland C.
6367 *) BN_rshift bugfix for n == 0.
6370 *) Add a 'bctest' script that checks for some known 'bc' bugs
6371 so that 'make test' does not abort just because 'bc' is broken.
6374 *) Store verify_result within SSL_SESSION also for client side to
6375 avoid potential security hole. (Re-used sessions on the client side
6376 always resulted in verify_result==X509_V_OK, not using the original
6377 result of the server certificate verification.)
6380 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
6381 SSL3_RT_APPLICATION_DATA, return 0.
6382 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
6386 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
6387 releases, have been re-implemented by renaming the previous
6388 implementations of ssl2_read and ssl3_read to ssl2_read_internal
6389 and ssl3_read_internal, respectively, and adding 'peek' parameters
6390 to them. The new ssl[23]_{read,peek} functions are calls to
6391 ssl[23]_read_internal with the 'peek' flag set appropriately.
6392 A 'peek' parameter has also been added to ssl3_read_bytes, which
6393 does the actual work for ssl3_read_internal.
6396 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
6397 the method-specific "init()" handler. Also clean up ex_data after
6398 calling the method-specific "finish()" handler. Previously, this was
6399 happening the other way round.
6402 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
6403 The previous value, 12, was not always sufficient for BN_mod_exp().
6406 *) Make sure that shared libraries get the internal name engine with
6407 the full version number and not just 0. This should mark the
6408 shared libraries as not backward compatible. Of course, this should
6409 be changed again when we can guarantee backward binary compatibility.
6412 *) Fix typo in get_cert_by_subject() in by_dir.c
6413 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
6415 *) Rework the system to generate shared libraries:
6417 - Make note of the expected extension for the shared libraries and
6418 if there is a need for symbolic links from for example libcrypto.so.0
6419 to libcrypto.so.0.9.7. There is extended info in Configure for
6422 - Make as few rebuilds of the shared libraries as possible.
6424 - Still avoid linking the OpenSSL programs with the shared libraries.
6426 - When installing, install the shared libraries separately from the
6430 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
6432 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
6433 and not in SSL_clear because the latter is also used by the
6434 accept/connect functions; previously, the settings made by
6435 SSL_set_read_ahead would be lost during the handshake.
6436 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
6438 *) Correct util/mkdef.pl to be selective about disabled algorithms.
6439 Previously, it would create entries for disableed algorithms no
6443 *) Added several new manual pages for SSL_* function.
6446 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
6448 *) In ssl23_get_client_hello, generate an error message when faced
6449 with an initial SSL 3.0/TLS record that is too small to contain the
6450 first two bytes of the ClientHello message, i.e. client_version.
6451 (Note that this is a pathologic case that probably has never happened
6452 in real life.) The previous approach was to use the version number
6453 from the record header as a substitute; but our protocol choice
6454 should not depend on that one because it is not authenticated
6455 by the Finished messages.
6458 *) More robust randomness gathering functions for Windows.
6459 [Jeffrey Altman <jaltman@columbia.edu>]
6461 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
6462 not set then we don't setup the error code for issuer check errors
6463 to avoid possibly overwriting other errors which the callback does
6464 handle. If an application does set the flag then we assume it knows
6465 what it is doing and can handle the new informational codes
6469 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
6470 a general "ANY" type, as such it should be able to decode anything
6471 including tagged types. However it didn't check the class so it would
6472 wrongly interpret tagged types in the same way as their universal
6473 counterpart and unknown types were just rejected. Changed so that the
6474 tagged and unknown types are handled in the same way as a SEQUENCE:
6475 that is the encoding is stored intact. There is also a new type
6476 "V_ASN1_OTHER" which is used when the class is not universal, in this
6477 case we have no idea what the actual type is so we just lump them all
6481 *) On VMS, stdout may very well lead to a file that is written to
6482 in a record-oriented fashion. That means that every write() will
6483 write a separate record, which will be read separately by the
6484 programs trying to read from it. This can be very confusing.
6486 The solution is to put a BIO filter in the way that will buffer
6487 text until a linefeed is reached, and then write everything a
6488 line at a time, so every record written will be an actual line,
6489 not chunks of lines and not (usually doesn't happen, but I've
6490 seen it once) several lines in one record. BIO_f_linebuffer() is
6493 Currently, it's a VMS-only method, because that's where it has
6494 been tested well enough.
6497 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
6498 it can return incorrect results.
6499 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
6500 but it was in 0.9.6-beta[12].)
6503 *) Disable the check for content being present when verifying detached
6504 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
6505 include zero length content when signing messages.
6508 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
6509 BIO_ctrl (for BIO pairs).
6512 *) Add DSO method for VMS.
6515 *) Bug fix: Montgomery multiplication could produce results with the
6519 *) Add RPM specification openssl.spec and modify it to build three
6520 packages. The default package contains applications, application
6521 documentation and run-time libraries. The devel package contains
6522 include files, static libraries and function documentation. The
6523 doc package contains the contents of the doc directory. The original
6524 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
6527 *) Add a large number of documentation files for many SSL routines.
6528 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6530 *) Add a configuration entry for Sony News 4.
6531 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
6533 *) Don't set the two most significant bits to one when generating a
6534 random number < q in the DSA library.
6537 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
6538 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
6539 the underlying transport is blocking) if a handshake took place.
6540 (The default behaviour is needed by applications such as s_client
6541 and s_server that use select() to determine when to use SSL_read;
6542 but for applications that know in advance when to expect data, it
6543 just makes things more complicated.)
6546 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
6550 *) Add a few more EBCDIC conditionals that make `req' and `x509'
6551 work better on such systems.
6552 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6554 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
6555 Update PKCS12_parse() so it copies the friendlyName and the
6556 keyid to the certificates aux info.
6559 *) Fix bug in PKCS7_verify() which caused an infinite loop
6560 if there was more than one signature.
6561 [Sven Uszpelkat <su@celocom.de>]
6563 *) Major change in util/mkdef.pl to include extra information
6564 about each symbol, as well as presentig variables as well
6565 as functions. This change means that there's n more need
6566 to rebuild the .num files when some algorithms are excluded.
6569 *) Allow the verify time to be set by an application,
6570 rather than always using the current time.
6573 *) Phase 2 verify code reorganisation. The certificate
6574 verify code now looks up an issuer certificate by a
6575 number of criteria: subject name, authority key id
6576 and key usage. It also verifies self signed certificates
6577 by the same criteria. The main comparison function is
6578 X509_check_issued() which performs these checks.
6580 Lot of changes were necessary in order to support this
6581 without completely rewriting the lookup code.
6583 Authority and subject key identifier are now cached.
6585 The LHASH 'certs' is X509_STORE has now been replaced
6586 by a STACK_OF(X509_OBJECT). This is mainly because an
6587 LHASH can't store or retrieve multiple objects with
6588 the same hash value.
6590 As a result various functions (which were all internal
6591 use only) have changed to handle the new X509_STORE
6592 structure. This will break anything that messed round
6593 with X509_STORE internally.
6595 The functions X509_STORE_add_cert() now checks for an
6596 exact match, rather than just subject name.
6598 The X509_STORE API doesn't directly support the retrieval
6599 of multiple certificates matching a given criteria, however
6600 this can be worked round by performing a lookup first
6601 (which will fill the cache with candidate certificates)
6602 and then examining the cache for matches. This is probably
6603 the best we can do without throwing out X509_LOOKUP
6604 entirely (maybe later...).
6606 The X509_VERIFY_CTX structure has been enhanced considerably.
6608 All certificate lookup operations now go via a get_issuer()
6609 callback. Although this currently uses an X509_STORE it
6610 can be replaced by custom lookups. This is a simple way
6611 to bypass the X509_STORE hackery necessary to make this
6612 work and makes it possible to use more efficient techniques
6613 in future. A very simple version which uses a simple
6614 STACK for its trusted certificate store is also provided
6615 using X509_STORE_CTX_trusted_stack().
6617 The verify_cb() and verify() callbacks now have equivalents
6618 in the X509_STORE_CTX structure.
6620 X509_STORE_CTX also has a 'flags' field which can be used
6621 to customise the verify behaviour.
6624 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
6625 excludes S/MIME capabilities.
6628 *) When a certificate request is read in keep a copy of the
6629 original encoding of the signed data and use it when outputing
6630 again. Signatures then use the original encoding rather than
6631 a decoded, encoded version which may cause problems if the
6632 request is improperly encoded.
6635 *) For consistency with other BIO_puts implementations, call
6636 buffer_write(b, ...) directly in buffer_puts instead of calling
6639 In BIO_puts, increment b->num_write as in BIO_write.
6640 [Peter.Sylvester@EdelWeb.fr]
6642 *) Fix BN_mul_word for the case where the word is 0. (We have to use
6643 BN_zero, we may not return a BIGNUM with an array consisting of
6647 *) Avoid calling abort() from within the library when problems are
6648 detected, except if preprocessor symbols have been defined
6649 (such as REF_CHECK, BN_DEBUG etc.).
6652 *) New openssl application 'rsautl'. This utility can be
6653 used for low level RSA operations. DER public key
6654 BIO/fp routines also added.
6657 *) New Configure entry and patches for compiling on QNX 4.
6658 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
6660 *) A demo state-machine implementation was sponsored by
6661 Nuron (http://www.nuron.com/) and is now available in
6662 demos/state_machine.
6665 *) New options added to the 'dgst' utility for signature
6666 generation and verification.
6669 *) Unrecognized PKCS#7 content types are now handled via a
6670 catch all ASN1_TYPE structure. This allows unsupported
6671 types to be stored as a "blob" and an application can
6672 encode and decode it manually.
6675 *) Fix various signed/unsigned issues to make a_strex.c
6677 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
6679 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
6680 length if passed a buffer. ASN1_INTEGER_to_BN failed
6681 if passed a NULL BN and its argument was negative.
6682 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
6684 *) Modification to PKCS#7 encoding routines to output definite
6685 length encoding. Since currently the whole structures are in
6686 memory there's not real point in using indefinite length
6687 constructed encoding. However if OpenSSL is compiled with
6688 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
6691 *) Added BIO_vprintf() and BIO_vsnprintf().
6694 *) Added more prefixes to parse for in the the strings written
6695 through a logging bio, to cover all the levels that are available
6696 through syslog. The prefixes are now:
6698 PANIC, EMERG, EMR => LOG_EMERG
6699 ALERT, ALR => LOG_ALERT
6700 CRIT, CRI => LOG_CRIT
6701 ERROR, ERR => LOG_ERR
6702 WARNING, WARN, WAR => LOG_WARNING
6703 NOTICE, NOTE, NOT => LOG_NOTICE
6704 INFO, INF => LOG_INFO
6705 DEBUG, DBG => LOG_DEBUG
6707 and as before, if none of those prefixes are present at the
6708 beginning of the string, LOG_ERR is chosen.
6710 On Win32, the LOG_* levels are mapped according to this:
6712 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
6713 LOG_WARNING => EVENTLOG_WARNING_TYPE
6714 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
6718 *) Made it possible to reconfigure with just the configuration
6719 argument "reconf" or "reconfigure". The command line arguments
6720 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
6721 and are retrieved from there when reconfiguring.
6725 [Assar Westerlund <assar@sics.se>, Richard Levitte]
6727 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
6730 *) The obj_dat.pl script was messing up the sorting of object
6731 names. The reason was that it compared the quoted version
6732 of strings as a result "OCSP" > "OCSP Signing" because
6733 " > SPACE. Changed script to store unquoted versions of
6734 names and add quotes on output. It was also omitting some
6735 names from the lookup table if they were given a default
6736 value (that is if SN is missing it is given the same
6737 value as LN and vice versa), these are now added on the
6738 grounds that if an object has a name we should be able to
6739 look it up. Finally added warning output when duplicate
6740 short or long names are found.
6743 *) Changes needed for Tandem NSK.
6744 [Scott Uroff <scott@xypro.com>]
6746 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
6747 RSA_padding_check_SSLv23(), special padding was never detected
6748 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
6749 version rollback attacks was not effective.
6751 In s23_clnt.c, don't use special rollback-attack detection padding
6752 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
6753 client; similarly, in s23_srvr.c, don't do the rollback check if
6754 SSL 2.0 is the only protocol enabled in the server.
6757 *) Make it possible to get hexdumps of unprintable data with 'openssl
6758 asn1parse'. By implication, the functions ASN1_parse_dump() and
6759 BIO_dump_indent() are added.
6762 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
6763 these print out strings and name structures based on various
6764 flags including RFC2253 support and proper handling of
6765 multibyte characters. Added options to the 'x509' utility
6766 to allow the various flags to be set.
6769 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
6770 Also change the functions X509_cmp_current_time() and
6771 X509_gmtime_adj() work with an ASN1_TIME structure,
6772 this will enable certificates using GeneralizedTime in validity
6773 dates to be checked.
6776 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
6777 negative public key encodings) on by default,
6778 NO_NEG_PUBKEY_BUG can be set to disable it.
6781 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
6782 content octets. An i2c_ASN1_OBJECT is unnecessary because
6783 the encoding can be trivially obtained from the structure.
6786 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
6787 not read locks (CRYPTO_r_[un]lock).
6790 *) A first attempt at creating official support for shared
6791 libraries through configuration. I've kept it so the
6792 default is static libraries only, and the OpenSSL programs
6793 are always statically linked for now, but there are
6794 preparations for dynamic linking in place.
6795 This has been tested on Linux and Tru64.
6798 *) Randomness polling function for Win9x, as described in:
6799 Peter Gutmann, Software Generation of Practically Strong
6803 *) Fix so PRNG is seeded in req if using an already existing
6807 *) New options to smime application. -inform and -outform
6808 allow alternative formats for the S/MIME message including
6809 PEM and DER. The -content option allows the content to be
6810 specified separately. This should allow things like Netscape
6811 form signing output easier to verify.
6814 *) Fix the ASN1 encoding of tags using the 'long form'.
6817 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
6818 STRING types. These convert content octets to and from the
6819 underlying type. The actual tag and length octets are
6820 already assumed to have been read in and checked. These
6821 are needed because all other string types have virtually
6822 identical handling apart from the tag. By having versions
6823 of the ASN1 functions that just operate on content octets
6824 IMPLICIT tagging can be handled properly. It also allows
6825 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
6826 and ASN1_INTEGER are identical apart from the tag.
6829 *) Change the handling of OID objects as follows:
6831 - New object identifiers are inserted in objects.txt, following
6832 the syntax given in objects.README.
6833 - objects.pl is used to process obj_mac.num and create a new
6835 - obj_dat.pl is used to create a new obj_dat.h, using the data in
6838 This is currently kind of a hack, and the perl code in objects.pl
6839 isn't very elegant, but it works as I intended. The simplest way
6840 to check that it worked correctly is to look in obj_dat.h and
6841 check the array nid_objs and make sure the objects haven't moved
6842 around (this is important!). Additions are OK, as well as
6843 consistent name changes.
6846 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
6849 *) Addition of the command line parameter '-rand file' to 'openssl req'.
6850 The given file adds to whatever has already been seeded into the
6851 random pool through the RANDFILE configuration file option or
6852 environment variable, or the default random state file.
6855 *) mkstack.pl now sorts each macro group into lexical order.
6856 Previously the output order depended on the order the files
6857 appeared in the directory, resulting in needless rewriting
6861 *) Patches to make OpenSSL compile under Win32 again. Mostly
6862 work arounds for the VC++ problem that it treats func() as
6863 func(void). Also stripped out the parts of mkdef.pl that
6864 added extra typesafe functions: these no longer exist.
6867 *) Reorganisation of the stack code. The macros are now all
6868 collected in safestack.h . Each macro is defined in terms of
6869 a "stack macro" of the form SKM_<name>(type, a, b). The
6870 DEBUG_SAFESTACK is now handled in terms of function casts,
6871 this has the advantage of retaining type safety without the
6872 use of additional functions. If DEBUG_SAFESTACK is not defined
6873 then the non typesafe macros are used instead. Also modified the
6874 mkstack.pl script to handle the new form. Needs testing to see
6875 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
6876 the default if no major problems. Similar behaviour for ASN1_SET_OF
6877 and PKCS12_STACK_OF.
6880 *) When some versions of IIS use the 'NET' form of private key the
6881 key derivation algorithm is different. Normally MD5(password) is
6882 used as a 128 bit RC4 key. In the modified case
6883 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
6884 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
6885 as the old Netscape_RSA functions except they have an additional
6886 'sgckey' parameter which uses the modified algorithm. Also added
6887 an -sgckey command line option to the rsa utility. Thanks to
6888 Adrian Peck <bertie@ncipher.com> for posting details of the modified
6889 algorithm to openssl-dev.
6892 *) The evp_local.h macros were using 'c.##kname' which resulted in
6893 invalid expansion on some systems (SCO 5.0.5 for example).
6894 Corrected to 'c.kname'.
6895 [Phillip Porch <root@theporch.com>]
6897 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
6898 a STACK of email addresses from a certificate or request, these look
6899 in the subject name and the subject alternative name extensions and
6900 omit any duplicate addresses.
6903 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
6904 This makes DSA verification about 2 % faster.
6907 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
6908 (meaning that now 2^5 values will be precomputed, which is only 4 KB
6909 plus overhead for 1024 bit moduli).
6910 This makes exponentiations about 0.5 % faster for 1024 bit
6911 exponents (as measured by "openssl speed rsa2048").
6914 *) Rename memory handling macros to avoid conflicts with other
6916 Malloc => OPENSSL_malloc
6917 Malloc_locked => OPENSSL_malloc_locked
6918 Realloc => OPENSSL_realloc
6919 Free => OPENSSL_free
6922 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
6923 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
6926 *) CygWin32 support.
6927 [John Jarvie <jjarvie@newsguy.com>]
6929 *) The type-safe stack code has been rejigged. It is now only compiled
6930 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
6931 by default all type-specific stack functions are "#define"d back to
6932 standard stack functions. This results in more streamlined output
6933 but retains the type-safety checking possibilities of the original
6937 *) The STACK code has been cleaned up, and certain type declarations
6938 that didn't make a lot of sense have been brought in line. This has
6939 also involved a cleanup of sorts in safestack.h to more correctly
6940 map type-safe stack functions onto their plain stack counterparts.
6941 This work has also resulted in a variety of "const"ifications of
6942 lots of the code, especially "_cmp" operations which should normally
6943 be prototyped with "const" parameters anyway.
6946 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
6947 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
6948 (The PRNG state consists of two parts, the large pool 'state' and 'md',
6949 where all of 'md' is used each time the PRNG is used, but 'state'
6950 is used only indexed by a cyclic counter. As entropy may not be
6951 well distributed from the beginning, 'md' is important as a
6952 chaining variable. However, the output function chains only half
6953 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
6954 all of 'md', and seeding with STATE_SIZE dummy bytes will result
6955 in all of 'state' being rewritten, with the new values depending
6956 on virtually all of 'md'. This overcomes the 80 bit limitation.)
6959 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
6960 the handshake is continued after ssl_verify_cert_chain();
6961 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
6962 can lead to 'unexplainable' connection aborts later.
6963 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
6965 *) Major EVP API cipher revision.
6966 Add hooks for extra EVP features. This allows various cipher
6967 parameters to be set in the EVP interface. Support added for variable
6968 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
6969 setting of RC2 and RC5 parameters.
6971 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
6974 Remove lots of duplicated code from the EVP library. For example *every*
6975 cipher init() function handles the 'iv' in the same way according to the
6976 cipher mode. They also all do nothing if the 'key' parameter is NULL and
6977 for CFB and OFB modes they zero ctx->num.
6979 New functionality allows removal of S/MIME code RC2 hack.
6981 Most of the routines have the same form and so can be declared in terms
6984 By shifting this to the top level EVP_CipherInit() it can be removed from
6985 all individual ciphers. If the cipher wants to handle IVs or keys
6986 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
6989 Change lots of functions like EVP_EncryptUpdate() to now return a
6990 value: although software versions of the algorithms cannot fail
6991 any installed hardware versions can.
6994 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
6995 this option is set, tolerate broken clients that send the negotiated
6996 protocol version number instead of the requested protocol version
7000 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
7001 i.e. non-zero for export ciphersuites, zero otherwise.
7002 Previous versions had this flag inverted, inconsistent with
7003 rsa_tmp_cb (..._TMP_RSA_CB).
7004 [Bodo Moeller; problem reported by Amit Chopra]
7006 *) Add missing DSA library text string. Work around for some IIS
7007 key files with invalid SEQUENCE encoding.
7010 *) Add a document (doc/standards.txt) that list all kinds of standards
7011 and so on that are implemented in OpenSSL.
7014 *) Enhance c_rehash script. Old version would mishandle certificates
7015 with the same subject name hash and wouldn't handle CRLs at all.
7016 Added -fingerprint option to crl utility, to support new c_rehash
7020 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
7023 *) Fix for SSL server purpose checking. Server checking was
7024 rejecting certificates which had extended key usage present
7025 but no ssl client purpose.
7026 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
7028 *) Make PKCS#12 code work with no password. The PKCS#12 spec
7029 is a little unclear about how a blank password is handled.
7030 Since the password in encoded as a BMPString with terminating
7031 double NULL a zero length password would end up as just the
7032 double NULL. However no password at all is different and is
7033 handled differently in the PKCS#12 key generation code. NS
7034 treats a blank password as zero length. MSIE treats it as no
7035 password on export: but it will try both on import. We now do
7036 the same: PKCS12_parse() tries zero length and no password if
7037 the password is set to "" or NULL (NULL is now a valid password:
7038 it wasn't before) as does the pkcs12 application.
7041 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
7042 perror when PEM_read_bio_X509_REQ fails, the error message must
7043 be obtained from the error queue.
7046 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
7047 it in ERR_remove_state if appropriate, and change ERR_get_state
7048 accordingly to avoid race conditions (this is necessary because
7049 thread_hash is no longer constant once set).
7052 *) Bugfix for linux-elf makefile.one.
7055 *) RSA_get_default_method() will now cause a default
7056 RSA_METHOD to be chosen if one doesn't exist already.
7057 Previously this was only set during a call to RSA_new()
7058 or RSA_new_method(NULL) meaning it was possible for
7059 RSA_get_default_method() to return NULL.
7062 *) Added native name translation to the existing DSO code
7063 that will convert (if the flag to do so is set) filenames
7064 that are sufficiently small and have no path information
7065 into a canonical native form. Eg. "blah" converted to
7066 "libblah.so" or "blah.dll" etc.
7069 *) New function ERR_error_string_n(e, buf, len) which is like
7070 ERR_error_string(e, buf), but writes at most 'len' bytes
7071 including the 0 terminator. For ERR_error_string_n, 'buf'
7073 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
7075 *) CONF library reworked to become more general. A new CONF
7076 configuration file reader "class" is implemented as well as a
7077 new functions (NCONF_*, for "New CONF") to handle it. The now
7078 old CONF_* functions are still there, but are reimplemented to
7079 work in terms of the new functions. Also, a set of functions
7080 to handle the internal storage of the configuration data is
7081 provided to make it easier to write new configuration file
7082 reader "classes" (I can definitely see something reading a
7083 configuration file in XML format, for example), called _CONF_*,
7084 or "the configuration storage API"...
7086 The new configuration file reading functions are:
7088 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
7089 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
7091 NCONF_default, NCONF_WIN32
7093 NCONF_dump_fp, NCONF_dump_bio
7095 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
7096 NCONF_new creates a new CONF object. This works in the same way
7097 as other interfaces in OpenSSL, like the BIO interface.
7098 NCONF_dump_* dump the internal storage of the configuration file,
7099 which is useful for debugging. All other functions take the same
7100 arguments as the old CONF_* functions wth the exception of the
7101 first that must be a `CONF *' instead of a `LHASH *'.
7103 To make it easer to use the new classes with the old CONF_* functions,
7104 the function CONF_set_default_method is provided.
7107 *) Add '-tls1' option to 'openssl ciphers', which was already
7108 mentioned in the documentation but had not been implemented.
7109 (This option is not yet really useful because even the additional
7110 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
7113 *) Initial DSO code added into libcrypto for letting OpenSSL (and
7114 OpenSSL-based applications) load shared libraries and bind to
7115 them in a portable way.
7116 [Geoff Thorpe, with contributions from Richard Levitte]
7118 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
7120 *) Make sure _lrotl and _lrotr are only used with MSVC.
7122 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
7123 (the default implementation of RAND_status).
7125 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
7126 to '-clrext' (= clear extensions), as intended and documented.
7127 [Bodo Moeller; inconsistency pointed out by Michael Attili
7130 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
7131 was larger than the MD block size.
7132 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
7134 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
7135 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
7136 using the passed key: if the passed key was a private key the result
7137 of X509_print(), for example, would be to print out all the private key
7141 *) des_quad_cksum() byte order bug fix.
7142 [Ulf Möller, using the problem description in krb4-0.9.7, where
7143 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
7145 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
7147 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
7149 *) For easily testing in shell scripts whether some command
7150 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
7151 returns with exit code 0 iff no command of the given name is available.
7152 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
7153 the output goes to stdout and nothing is printed to stderr.
7154 Additional arguments are always ignored.
7156 Since for each cipher there is a command of the same name,
7157 the 'no-cipher' compilation switches can be tested this way.
7159 ('openssl no-XXX' is not able to detect pseudo-commands such
7160 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
7163 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
7166 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
7167 is set; it will be thrown away anyway because each handshake creates
7169 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
7170 to parameters -- in previous versions (since OpenSSL 0.9.3) the
7171 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
7172 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7175 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7176 'Q' and 'R' lose their special meanings (quit/renegotiate).
7177 This is part of what -quiet does; unlike -quiet, -ign_eof
7178 does not suppress any output.
7181 *) Add compatibility options to the purpose and trust code. The
7182 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7183 accepts a certificate or CA, this was the previous behaviour,
7184 with all the associated security issues.
7186 X509_TRUST_COMPAT is the old trust behaviour: only and
7187 automatically trust self signed roots in certificate store. A
7188 new trust setting X509_TRUST_DEFAULT is used to specify that
7189 a purpose has no associated trust setting and it should instead
7190 use the value in the default purpose.
7193 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7194 and fix a memory leak.
7197 *) In util/mkerr.pl (which implements 'make errors'), preserve
7198 reason strings from the previous version of the .c file, as
7199 the default to have only downcase letters (and digits) in
7200 automatically generated reasons codes is not always appropriate.
7203 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7204 using strerror. Previously, ERR_reason_error_string() returned
7205 library names as reason strings for SYSerr; but SYSerr is a special
7206 case where small numbers are errno values, not library numbers.
7209 *) Add '-dsaparam' option to 'openssl dhparam' application. This
7210 converts DSA parameters into DH parameters. (When creating parameters,
7211 DSA_generate_parameters is used.)
7214 *) Include 'length' (recommended exponent length) in C code generated
7215 by 'openssl dhparam -C'.
7218 *) The second argument to set_label in perlasm was already being used
7219 so couldn't be used as a "file scope" flag. Moved to third argument
7223 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
7224 instead of RAND_bytes for encryption IVs and salts.
7227 *) Include RAND_status() into RAND_METHOD instead of implementing
7228 it only for md_rand.c Otherwise replacing the PRNG by calling
7229 RAND_set_rand_method would be impossible.
7232 *) Don't let DSA_generate_key() enter an infinite loop if the random
7233 number generation fails.
7236 *) New 'rand' application for creating pseudo-random output.
7239 *) Added configuration support for Linux/IA64
7240 [Rolf Haberrecker <rolf@suse.de>]
7242 *) Assembler module support for Mingw32.
7245 *) Shared library support for HPUX (in shlib/).
7246 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
7248 *) Shared library support for Solaris gcc.
7249 [Lutz Behnke <behnke@trustcenter.de>]
7251 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
7253 *) PKCS7_encrypt() was adding text MIME headers twice because they
7254 were added manually and by SMIME_crlf_copy().
7257 *) In bntest.c don't call BN_rand with zero bits argument.
7258 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
7260 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
7261 case was implemented. This caused BN_div_recp() to fail occasionally.
7264 *) Add an optional second argument to the set_label() in the perl
7265 assembly language builder. If this argument exists and is set
7266 to 1 it signals that the assembler should use a symbol whose
7267 scope is the entire file, not just the current function. This
7268 is needed with MASM which uses the format label:: for this scope.
7269 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
7271 *) Change the ASN1 types so they are typedefs by default. Before
7272 almost all types were #define'd to ASN1_STRING which was causing
7273 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
7277 *) Change names of new functions to the new get1/get0 naming
7278 convention: After 'get1', the caller owns a reference count
7279 and has to call ..._free; 'get0' returns a pointer to some
7280 data structure without incrementing reference counters.
7281 (Some of the existing 'get' functions increment a reference
7282 counter, some don't.)
7283 Similarly, 'set1' and 'add1' functions increase reference
7284 counters or duplicate objects.
7287 *) Allow for the possibility of temp RSA key generation failure:
7288 the code used to assume it always worked and crashed on failure.
7291 *) Fix potential buffer overrun problem in BIO_printf().
7292 [Ulf Möller, using public domain code by Patrick Powell; problem
7293 pointed out by David Sacerdote <das33@cornell.edu>]
7295 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
7296 RAND_egd() and RAND_status(). In the command line application,
7297 the EGD socket can be specified like a seed file using RANDFILE
7301 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
7302 Some CAs (e.g. Verisign) distribute certificates in this form.
7305 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
7306 list to exclude them. This means that no special compilation option
7307 is needed to use anonymous DH: it just needs to be included in the
7311 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
7312 EVP_MD_type. The old functionality is available in a new macro called
7313 EVP_MD_md(). Change code that uses it and update docs.
7316 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
7317 where the 'void *' argument is replaced by a function pointer argument.
7318 Previously 'void *' was abused to point to functions, which works on
7319 many platforms, but is not correct. As these functions are usually
7320 called by macros defined in OpenSSL header files, most source code
7321 should work without changes.
7324 *) <openssl/opensslconf.h> (which is created by Configure) now contains
7325 sections with information on -D... compiler switches used for
7326 compiling the library so that applications can see them. To enable
7327 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
7328 must be defined. E.g.,
7329 #define OPENSSL_ALGORITHM_DEFINES
7330 #include <openssl/opensslconf.h>
7331 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
7332 [Richard Levitte, Ulf and Bodo Möller]
7334 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
7338 *) Change the 'other' type in certificate aux info to a STACK_OF
7339 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
7340 the required ASN1 format: arbitrary types determined by an OID.
7343 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
7344 argument to 'req'. This is not because the function is newer or
7345 better than others it just uses the work 'NEW' in the certificate
7346 request header lines. Some software needs this.
7349 *) Reorganise password command line arguments: now passwords can be
7350 obtained from various sources. Delete the PEM_cb function and make
7351 it the default behaviour: i.e. if the callback is NULL and the
7352 usrdata argument is not NULL interpret it as a null terminated pass
7353 phrase. If usrdata and the callback are NULL then the pass phrase
7354 is prompted for as usual.
7357 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
7358 the support is automatically enabled. The resulting binaries will
7359 autodetect the card and use it if present.
7360 [Ben Laurie and Compaq Inc.]
7362 *) Work around for Netscape hang bug. This sends certificate request
7363 and server done in one record. Since this is perfectly legal in the
7364 SSL/TLS protocol it isn't a "bug" option and is on by default. See
7365 the bugs/SSLv3 entry for more info.
7368 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
7371 *) Add -rand argument to smime and pkcs12 applications and read/write
7375 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
7378 *) Add command line password options to the remaining applications.
7381 *) Bug fix for BN_div_recp() for numerators with an even number of
7385 *) More tests in bntest.c, and changed test_bn output.
7388 *) ./config recognizes MacOS X now.
7391 *) Bug fix for BN_div() when the first words of num and divsor are
7392 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
7395 *) Add support for various broken PKCS#8 formats, and command line
7396 options to produce them.
7399 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
7400 get temporary BIGNUMs from a BN_CTX.
7403 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
7407 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
7408 include a #define from the old name to the new. The original intent
7409 was that statically linked binaries could for example just call
7410 SSLeay_add_all_ciphers() to just add ciphers to the table and not
7411 link with digests. This never worked becayse SSLeay_add_all_digests()
7412 and SSLeay_add_all_ciphers() were in the same source file so calling
7413 one would link with the other. They are now in separate source files.
7416 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
7419 *) Use a less unusual form of the Miller-Rabin primality test (it used
7420 a binary algorithm for exponentiation integrated into the Miller-Rabin
7421 loop, our standard modexp algorithms are faster).
7424 *) Support for the EBCDIC character set completed.
7425 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
7427 *) Source code cleanups: use const where appropriate, eliminate casts,
7428 use void * instead of char * in lhash.
7431 *) Bugfix: ssl3_send_server_key_exchange was not restartable
7432 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
7433 this the server could overwrite ephemeral keys that the client
7437 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
7438 using 50 iterations of the Rabin-Miller test.
7440 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
7441 iterations of the Rabin-Miller test as required by the appendix
7442 to FIPS PUB 186[-1]) instead of DSA_is_prime.
7443 As BN_is_prime_fasttest includes trial division, DSA parameter
7444 generation becomes much faster.
7446 This implies a change for the callback functions in DSA_is_prime
7447 and DSA_generate_parameters: The callback function is called once
7448 for each positive witness in the Rabin-Miller test, not just
7449 occasionally in the inner loop; and the parameters to the
7450 callback function now provide an iteration count for the outer
7451 loop rather than for the current invocation of the inner loop.
7452 DSA_generate_parameters additionally can call the callback
7453 function with an 'iteration count' of -1, meaning that a
7454 candidate has passed the trial division test (when q is generated
7455 from an application-provided seed, trial division is skipped).
7458 *) New function BN_is_prime_fasttest that optionally does trial
7459 division before starting the Rabin-Miller test and has
7460 an additional BN_CTX * argument (whereas BN_is_prime always
7461 has to allocate at least one BN_CTX).
7462 'callback(1, -1, cb_arg)' is called when a number has passed the
7463 trial division stage.
7466 *) Fix for bug in CRL encoding. The validity dates weren't being handled
7470 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
7473 *) New function BN_pseudo_rand().
7476 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
7477 bignum version of BN_from_montgomery() with the working code from
7478 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
7482 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
7483 made it impossible to use the same SSL_SESSION data structure in
7484 SSL2 clients in multiple threads.
7487 *) The return value of RAND_load_file() no longer counts bytes obtained
7488 by stat(). RAND_load_file(..., -1) is new and uses the complete file
7489 to seed the PRNG (previously an explicit byte count was required).
7490 [Ulf Möller, Bodo Möller]
7492 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
7493 used (char *) instead of (void *) and had casts all over the place.
7496 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
7499 *) Retain source code compatibility for BN_prime_checks macro:
7500 BN_is_prime(..., BN_prime_checks, ...) now uses
7501 BN_prime_checks_for_size to determine the appropriate number of
7502 Rabin-Miller iterations.
7505 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
7506 DH_CHECK_P_NOT_SAFE_PRIME.
7507 (Check if this is true? OpenPGP calls them "strong".)
7510 *) Merge the functionality of "dh" and "gendh" programs into a new program
7511 "dhparam". The old programs are retained for now but will handle DH keys
7512 (instead of parameters) in future.
7515 *) Make the ciphers, s_server and s_client programs check the return values
7516 when a new cipher list is set.
7519 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
7520 ciphers. Before when the 56bit ciphers were enabled the sorting was
7523 The syntax for the cipher sorting has been extended to support sorting by
7524 cipher-strength (using the strength_bits hard coded in the tables).
7525 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
7527 Fix a bug in the cipher-command parser: when supplying a cipher command
7528 string with an "undefined" symbol (neither command nor alphanumeric
7529 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
7530 an error is flagged.
7532 Due to the strength-sorting extension, the code of the
7533 ssl_create_cipher_list() function was completely rearranged. I hope that
7534 the readability was also increased :-)
7535 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7537 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
7538 for the first serial number and places 2 in the serial number file. This
7539 avoids problems when the root CA is created with serial number zero and
7540 the first user certificate has the same issuer name and serial number
7544 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
7545 the new code. Add documentation for this stuff.
7548 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
7549 X509_*() to X509at_*() on the grounds that they don't handle X509
7550 structures and behave in an analagous way to the X509v3 functions:
7551 they shouldn't be called directly but wrapper functions should be used
7554 So we also now have some wrapper functions that call the X509at functions
7555 when passed certificate requests. (TO DO: similar things can be done with
7556 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
7557 things. Some of these need some d2i or i2d and print functionality
7558 because they handle more complex structures.)
7561 *) Add missing #ifndefs that caused missing symbols when building libssl
7562 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
7563 NO_RSA in ssl/s2*.c.
7564 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
7566 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
7567 has a return value which indicates the quality of the random data
7568 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
7569 error queue. New function RAND_pseudo_bytes() generates output that is
7570 guaranteed to be unique but not unpredictable. RAND_add is like
7571 RAND_seed, but takes an extra argument for an entropy estimate
7572 (RAND_seed always assumes full entropy).
7575 *) Do more iterations of Rabin-Miller probable prime test (specifically,
7576 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
7577 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
7578 in crypto/bn/bn_prime.c for the complete table). This guarantees a
7579 false-positive rate of at most 2^-80 for random input.
7582 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
7585 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
7586 in the 0.9.5 release), this returns the chain
7587 from an X509_CTX structure with a dup of the stack and all
7588 the X509 reference counts upped: so the stack will exist
7589 after X509_CTX_cleanup() has been called. Modify pkcs12.c
7592 Also make SSL_SESSION_print() print out the verify return
7596 *) Add manpage for the pkcs12 command. Also change the default
7597 behaviour so MAC iteration counts are used unless the new
7598 -nomaciter option is used. This improves file security and
7599 only older versions of MSIE (4.0 for example) need it.
7602 *) Honor the no-xxx Configure options when creating .DEF files.
7605 *) Add PKCS#10 attributes to field table: challengePassword,
7606 unstructuredName and unstructuredAddress. These are taken from
7607 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
7608 international characters are used.
7610 More changes to X509_ATTRIBUTE code: allow the setting of types
7611 based on strings. Remove the 'loc' parameter when adding
7612 attributes because these will be a SET OF encoding which is sorted
7616 *) Initial changes to the 'req' utility to allow request generation
7617 automation. This will allow an application to just generate a template
7618 file containing all the field values and have req construct the
7621 Initial support for X509_ATTRIBUTE handling. Stacks of these are
7622 used all over the place including certificate requests and PKCS#7
7623 structures. They are currently handled manually where necessary with
7624 some primitive wrappers for PKCS#7. The new functions behave in a
7625 manner analogous to the X509 extension functions: they allow
7626 attributes to be looked up by NID and added.
7628 Later something similar to the X509V3 code would be desirable to
7629 automatically handle the encoding, decoding and printing of the
7630 more complex types. The string types like challengePassword can
7631 be handled by the string table functions.
7633 Also modified the multi byte string table handling. Now there is
7634 a 'global mask' which masks out certain types. The table itself
7635 can use the flag STABLE_NO_MASK to ignore the mask setting: this
7636 is useful when for example there is only one permissible type
7637 (as in countryName) and using the mask might result in no valid
7641 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
7642 SSL_get_peer_finished to allow applications to obtain the latest
7643 Finished messages sent to the peer or expected from the peer,
7644 respectively. (SSL_get_peer_finished is usually the Finished message
7645 actually received from the peer, otherwise the protocol will be aborted.)
7647 As the Finished message are message digests of the complete handshake
7648 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
7649 be used for external authentication procedures when the authentication
7650 provided by SSL/TLS is not desired or is not enough.
7653 *) Enhanced support for Alpha Linux is added. Now ./config checks if
7654 the host supports BWX extension and if Compaq C is present on the
7655 $PATH. Just exploiting of the BWX extension results in 20-30%
7656 performance kick for some algorithms, e.g. DES and RC4 to mention
7657 a couple. Compaq C in turn generates ~20% faster code for MD5 and
7661 *) Add support for MS "fast SGC". This is arguably a violation of the
7662 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
7663 weak crypto and after checking the certificate is SGC a second one
7664 with strong crypto. MS SGC stops the first handshake after receiving
7665 the server certificate message and sends a second client hello. Since
7666 a server will typically do all the time consuming operations before
7667 expecting any further messages from the client (server key exchange
7668 is the most expensive) there is little difference between the two.
7670 To get OpenSSL to support MS SGC we have to permit a second client
7671 hello message after we have sent server done. In addition we have to
7672 reset the MAC if we do get this second client hello.
7675 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
7676 if a DER encoded private key is RSA or DSA traditional format. Changed
7677 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
7678 format DER encoded private key. Newer code should use PKCS#8 format which
7679 has the key type encoded in the ASN1 structure. Added DER private key
7680 support to pkcs8 application.
7683 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
7684 ciphersuites has been selected (as required by the SSL 3/TLS 1
7685 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
7686 is set, we interpret this as a request to violate the specification
7687 (the worst that can happen is a handshake failure, and 'correct'
7688 behaviour would result in a handshake failure anyway).
7691 *) In SSL_CTX_add_session, take into account that there might be multiple
7692 SSL_SESSION structures with the same session ID (e.g. when two threads
7693 concurrently obtain them from an external cache).
7694 The internal cache can handle only one SSL_SESSION with a given ID,
7695 so if there's a conflict, we now throw out the old one to achieve
7699 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
7700 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
7701 some routines that use cipher OIDs: some ciphers do not have OIDs
7702 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
7706 *) Simplify the trust setting structure and code. Now we just have
7707 two sequences of OIDs for trusted and rejected settings. These will
7708 typically have values the same as the extended key usage extension
7709 and any application specific purposes.
7711 The trust checking code now has a default behaviour: it will just
7712 check for an object with the same NID as the passed id. Functions can
7713 be provided to override either the default behaviour or the behaviour
7714 for a given id. SSL client, server and email already have functions
7715 in place for compatibility: they check the NID and also return "trusted"
7716 if the certificate is self signed.
7719 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
7720 traditional format into an EVP_PKEY structure.
7723 *) Add a password callback function PEM_cb() which either prompts for
7724 a password if usr_data is NULL or otherwise assumes it is a null
7725 terminated password. Allow passwords to be passed on command line
7726 environment or config files in a few more utilities.
7729 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
7730 keys. Add some short names for PKCS#8 PBE algorithms and allow them
7731 to be specified on the command line for the pkcs8 and pkcs12 utilities.
7732 Update documentation.
7735 *) Support for ASN1 "NULL" type. This could be handled before by using
7736 ASN1_TYPE but there wasn't any function that would try to read a NULL
7737 and produce an error if it couldn't. For compatibility we also have
7738 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
7739 don't allocate anything because they don't need to.
7742 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
7744 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
7746 *) Rebuild of the memory allocation routines used by OpenSSL code and
7747 possibly others as well. The purpose is to make an interface that
7748 provide hooks so anyone can build a separate set of allocation and
7749 deallocation routines to be used by OpenSSL, for example memory
7750 pool implementations, or something else, which was previously hard
7751 since Malloc(), Realloc() and Free() were defined as macros having
7752 the values malloc, realloc and free, respectively (except for Win32
7753 compilations). The same is provided for memory debugging code.
7754 OpenSSL already comes with functionality to find memory leaks, but
7755 this gives people a chance to debug other memory problems.
7757 With these changes, a new set of functions and macros have appeared:
7759 CRYPTO_set_mem_debug_functions() [F]
7760 CRYPTO_get_mem_debug_functions() [F]
7761 CRYPTO_dbg_set_options() [F]
7762 CRYPTO_dbg_get_options() [F]
7763 CRYPTO_malloc_debug_init() [M]
7765 The memory debug functions are NULL by default, unless the library
7766 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
7767 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
7768 gives the standard debugging functions that come with OpenSSL) or
7769 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
7770 provided by the library user) must be used. When the standard
7771 debugging functions are used, CRYPTO_dbg_set_options can be used to
7772 request additional information:
7773 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
7774 the CRYPTO_MDEBUG_xxx macro when compiling the library.
7776 Also, things like CRYPTO_set_mem_functions will always give the
7777 expected result (the new set of functions is used for allocation
7778 and deallocation) at all times, regardless of platform and compiler
7781 To finish it up, some functions that were never use in any other
7782 way than through macros have a new API and new semantic:
7785 CRYPTO_dbg_realloc()
7788 All macros of value have retained their old syntax.
7789 [Richard Levitte and Bodo Moeller]
7791 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
7792 ordering of SMIMECapabilities wasn't in "strength order" and there
7793 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
7797 *) Some ASN1 types with illegal zero length encoding (INTEGER,
7798 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
7799 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
7801 *) Merge in my S/MIME library for OpenSSL. This provides a simple
7802 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
7803 functionality to handle multipart/signed properly) and a utility
7804 called 'smime' to call all this stuff. This is based on code I
7805 originally wrote for Celo who have kindly allowed it to be
7806 included in OpenSSL.
7809 *) Add variants des_set_key_checked and des_set_key_unchecked of
7810 des_set_key (aka des_key_sched). Global variable des_check_key
7811 decides which of these is called by des_set_key; this way
7812 des_check_key behaves as it always did, but applications and
7813 the library itself, which was buggy for des_check_key == 1,
7814 have a cleaner way to pick the version they need.
7817 *) New function PKCS12_newpass() which changes the password of a
7821 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
7822 dynamic mix. In both cases the ids can be used as an index into the
7823 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
7824 functions so they accept a list of the field values and the
7825 application doesn't need to directly manipulate the X509_TRUST
7829 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
7833 *) Modify the way the V3 extension code looks up extensions. This now
7834 works in a similar way to the object code: we have some "standard"
7835 extensions in a static table which is searched with OBJ_bsearch()
7836 and the application can add dynamic ones if needed. The file
7837 crypto/x509v3/ext_dat.h now has the info: this file needs to be
7838 updated whenever a new extension is added to the core code and kept
7839 in ext_nid order. There is a simple program 'tabtest.c' which checks
7840 this. New extensions are not added too often so this file can readily
7841 be maintained manually.
7843 There are two big advantages in doing things this way. The extensions
7844 can be looked up immediately and no longer need to be "added" using
7845 X509V3_add_standard_extensions(): this function now does nothing.
7846 [Side note: I get *lots* of email saying the extension code doesn't
7847 work because people forget to call this function]
7848 Also no dynamic allocation is done unless new extensions are added:
7849 so if we don't add custom extensions there is no need to call
7850 X509V3_EXT_cleanup().
7853 *) Modify enc utility's salting as follows: make salting the default. Add a
7854 magic header, so unsalted files fail gracefully instead of just decrypting
7855 to garbage. This is because not salting is a big security hole, so people
7856 should be discouraged from doing it.
7859 *) Fixes and enhancements to the 'x509' utility. It allowed a message
7860 digest to be passed on the command line but it only used this
7861 parameter when signing a certificate. Modified so all relevant
7862 operations are affected by the digest parameter including the
7863 -fingerprint and -x509toreq options. Also -x509toreq choked if a
7864 DSA key was used because it didn't fix the digest.
7867 *) Initial certificate chain verify code. Currently tests the untrusted
7868 certificates for consistency with the verify purpose (which is set
7869 when the X509_STORE_CTX structure is set up) and checks the pathlength.
7871 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
7872 this is because it will reject chains with invalid extensions whereas
7873 every previous version of OpenSSL and SSLeay made no checks at all.
7875 Trust code: checks the root CA for the relevant trust settings. Trust
7876 settings have an initial value consistent with the verify purpose: e.g.
7877 if the verify purpose is for SSL client use it expects the CA to be
7878 trusted for SSL client use. However the default value can be changed to
7879 permit custom trust settings: one example of this would be to only trust
7880 certificates from a specific "secure" set of CAs.
7882 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
7883 which should be used for version portability: especially since the
7884 verify structure is likely to change more often now.
7886 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
7887 to set them. If not set then assume SSL clients will verify SSL servers
7890 Two new options to the verify program: -untrusted allows a set of
7891 untrusted certificates to be passed in and -purpose which sets the
7892 intended purpose of the certificate. If a purpose is set then the
7893 new chain verify code is used to check extension consistency.
7896 *) Support for the authority information access extension.
7899 *) Modify RSA and DSA PEM read routines to transparently handle
7900 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
7901 public keys in a format compatible with certificate
7902 SubjectPublicKeyInfo structures. Unfortunately there were already
7903 functions called *_PublicKey_* which used various odd formats so
7904 these are retained for compatibility: however the DSA variants were
7905 never in a public release so they have been deleted. Changed dsa/rsa
7906 utilities to handle the new format: note no releases ever handled public
7907 keys so we should be OK.
7909 The primary motivation for this change is to avoid the same fiasco
7910 that dogs private keys: there are several incompatible private key
7911 formats some of which are standard and some OpenSSL specific and
7912 require various evil hacks to allow partial transparent handling and
7913 even then it doesn't work with DER formats. Given the option anything
7914 other than PKCS#8 should be dumped: but the other formats have to
7915 stay in the name of compatibility.
7917 With public keys and the benefit of hindsight one standard format
7918 is used which works with EVP_PKEY, RSA or DSA structures: though
7919 it clearly returns an error if you try to read the wrong kind of key.
7921 Added a -pubkey option to the 'x509' utility to output the public key.
7922 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
7923 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
7924 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
7925 that do the same as the EVP_PKEY_assign_*() except they up the
7926 reference count of the added key (they don't "swallow" the
7930 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
7931 CRLs would fail if the file contained no certificates or no CRLs:
7932 added a new function to read in both types and return the number
7933 read: this means that if none are read it will be an error. The
7934 DER versions of the certificate and CRL reader would always fail
7935 because it isn't possible to mix certificates and CRLs in DER format
7936 without choking one or the other routine. Changed this to just read
7937 a certificate: this is the best we can do. Also modified the code
7938 in apps/verify.c to take notice of return codes: it was previously
7939 attempting to read in certificates from NULL pointers and ignoring
7940 any errors: this is one reason why the cert and CRL reader seemed
7941 to work. It doesn't check return codes from the default certificate
7942 routines: these may well fail if the certificates aren't installed.
7945 *) Code to support otherName option in GeneralName.
7948 *) First update to verify code. Change the verify utility
7949 so it warns if it is passed a self signed certificate:
7950 for consistency with the normal behaviour. X509_verify
7951 has been modified to it will now verify a self signed
7952 certificate if *exactly* the same certificate appears
7953 in the store: it was previously impossible to trust a
7954 single self signed certificate. This means that:
7955 openssl verify ss.pem
7956 now gives a warning about a self signed certificate but
7957 openssl verify -CAfile ss.pem ss.pem
7961 *) For servers, store verify_result in SSL_SESSION data structure
7962 (and add it to external session representation).
7963 This is needed when client certificate verifications fails,
7964 but an application-provided verification callback (set by
7965 SSL_CTX_set_cert_verify_callback) allows accepting the session
7966 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
7967 but returns 1): When the session is reused, we have to set
7968 ssl->verify_result to the appropriate error code to avoid
7970 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
7972 *) Fix a bug in the new PKCS#7 code: it didn't consider the
7973 case in PKCS7_dataInit() where the signed PKCS7 structure
7974 didn't contain any existing data because it was being created.
7975 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
7977 *) Add a salt to the key derivation routines in enc.c. This
7978 forms the first 8 bytes of the encrypted file. Also add a
7979 -S option to allow a salt to be input on the command line.
7982 *) New function X509_cmp(). Oddly enough there wasn't a function
7983 to compare two certificates. We do this by working out the SHA1
7984 hash and comparing that. X509_cmp() will be needed by the trust
7988 *) SSL_get1_session() is like SSL_get_session(), but increments
7989 the reference count in the SSL_SESSION returned.
7990 [Geoff Thorpe <geoff@eu.c2.net>]
7992 *) Fix for 'req': it was adding a null to request attributes.
7993 Also change the X509_LOOKUP and X509_INFO code to handle
7994 certificate auxiliary information.
7997 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
8001 *) Add the possibility to add extra information to the memory leak
8002 detecting output, to form tracebacks, showing from where each
8003 allocation was originated: CRYPTO_push_info("constant string") adds
8004 the string plus current file name and line number to a per-thread
8005 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
8006 is like calling CYRPTO_pop_info() until the stack is empty.
8007 Also updated memory leak detection code to be multi-thread-safe.
8010 *) Add options -text and -noout to pkcs7 utility and delete the
8011 encryption options which never did anything. Update docs.
8014 *) Add options to some of the utilities to allow the pass phrase
8015 to be included on either the command line (not recommended on
8016 OSes like Unix) or read from the environment. Update the
8017 manpages and fix a few bugs.
8020 *) Add a few manpages for some of the openssl commands.
8023 *) Fix the -revoke option in ca. It was freeing up memory twice,
8024 leaking and not finding already revoked certificates.
8027 *) Extensive changes to support certificate auxiliary information.
8028 This involves the use of X509_CERT_AUX structure and X509_AUX
8029 functions. An X509_AUX function such as PEM_read_X509_AUX()
8030 can still read in a certificate file in the usual way but it
8031 will also read in any additional "auxiliary information". By
8032 doing things this way a fair degree of compatibility can be
8033 retained: existing certificates can have this information added
8034 using the new 'x509' options.
8036 Current auxiliary information includes an "alias" and some trust
8037 settings. The trust settings will ultimately be used in enhanced
8038 certificate chain verification routines: currently a certificate
8039 can only be trusted if it is self signed and then it is trusted
8043 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
8044 The problem was that one of the replacement routines had not been working
8045 since SSLeay releases. For now the offending routine has been replaced
8046 with non-optimised assembler. Even so, this now gives around 95%
8047 performance improvement for 1024 bit RSA signs.
8050 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
8051 handling. Most clients have the effective key size in bits equal to
8052 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
8053 A few however don't do this and instead use the size of the decrypted key
8054 to determine the RC2 key length and the AlgorithmIdentifier to determine
8055 the effective key length. In this case the effective key length can still
8056 be 40 bits but the key length can be 168 bits for example. This is fixed
8057 by manually forcing an RC2 key into the EVP_PKEY structure because the
8058 EVP code can't currently handle unusual RC2 key sizes: it always assumes
8059 the key length and effective key length are equal.
8062 *) Add a bunch of functions that should simplify the creation of
8063 X509_NAME structures. Now you should be able to do:
8064 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
8065 and have it automatically work out the correct field type and fill in
8066 the structures. The more adventurous can try:
8067 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
8068 and it will (hopefully) work out the correct multibyte encoding.
8071 *) Change the 'req' utility to use the new field handling and multibyte
8072 copy routines. Before the DN field creation was handled in an ad hoc
8073 way in req, ca, and x509 which was rather broken and didn't support
8074 BMPStrings or UTF8Strings. Since some software doesn't implement
8075 BMPStrings or UTF8Strings yet, they can be enabled using the config file
8076 using the dirstring_type option. See the new comment in the default
8077 openssl.cnf for more info.
8080 *) Make crypto/rand/md_rand.c more robust:
8081 - Assure unique random numbers after fork().
8082 - Make sure that concurrent threads access the global counter and
8083 md serializably so that we never lose entropy in them
8084 or use exactly the same state in multiple threads.
8085 Access to the large state is not always serializable because
8086 the additional locking could be a performance killer, and
8087 md should be large enough anyway.
8090 *) New file apps/app_rand.c with commonly needed functionality
8091 for handling the random seed file.
8093 Use the random seed file in some applications that previously did not:
8095 dsaparam -genkey (which also ignored its '-rand' option),
8098 x509 (when signing).
8099 Except on systems with /dev/urandom, it is crucial to have a random
8100 seed file at least for key creation, DSA signing, and for DH exchanges;
8101 for RSA signatures we could do without one.
8103 gendh and gendsa (unlike genrsa) used to read only the first byte
8104 of each file listed in the '-rand' option. The function as previously
8105 found in genrsa is now in app_rand.c and is used by all programs
8106 that support '-rand'.
8109 *) In RAND_write_file, use mode 0600 for creating files;
8110 don't just chmod when it may be too late.
8113 *) Report an error from X509_STORE_load_locations
8114 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
8117 *) New function ASN1_mbstring_copy() this copies a string in either
8118 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
8119 into an ASN1_STRING type. A mask of permissible types is passed
8120 and it chooses the "minimal" type to use or an error if not type
8124 *) Add function equivalents to the various macros in asn1.h. The old
8125 macros are retained with an M_ prefix. Code inside the library can
8126 use the M_ macros. External code (including the openssl utility)
8127 should *NOT* in order to be "shared library friendly".
8130 *) Add various functions that can check a certificate's extensions
8131 to see if it usable for various purposes such as SSL client,
8132 server or S/MIME and CAs of these types. This is currently
8133 VERY EXPERIMENTAL but will ultimately be used for certificate chain
8134 verification. Also added a -purpose flag to x509 utility to
8135 print out all the purposes.
8138 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
8142 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
8143 for, obtain and decode and extension and obtain its critical flag.
8144 This allows all the necessary extension code to be handled in a
8145 single function call.
8148 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
8149 platforms. See crypto/rc4/rc4_enc.c for further details.
8152 *) New -noout option to asn1parse. This causes no output to be produced
8153 its main use is when combined with -strparse and -out to extract data
8154 from a file (which may not be in ASN.1 format).
8157 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
8158 when producing the local key id.
8159 [Richard Levitte <levitte@stacken.kth.se>]
8161 *) New option -dhparam in s_server. This allows a DH parameter file to be
8162 stated explicitly. If it is not stated then it tries the first server
8163 certificate file. The previous behaviour hard coded the filename
8167 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
8168 a public key to be input or output. For example:
8169 openssl rsa -in key.pem -pubout -out pubkey.pem
8170 Also added necessary DSA public key functions to handle this.
8173 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
8174 in the message. This was handled by allowing
8175 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8176 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8178 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8179 to the end of the strings whereas this didn't. This would cause problems
8180 if strings read with d2i_ASN1_bytes() were later modified.
8181 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8183 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8184 data and it contains EOF it will end up returning an error. This is
8185 caused by input 46 bytes long. The cause is due to the way base64
8186 BIOs find the start of base64 encoded data. They do this by trying a
8187 trial decode on each line until they find one that works. When they
8188 do a flag is set and it starts again knowing it can pass all the
8189 data directly through the decoder. Unfortunately it doesn't reset
8190 the context it uses. This means that if EOF is reached an attempt
8191 is made to pass two EOFs through the context and this causes the
8192 resulting error. This can also cause other problems as well. As is
8193 usual with these problems it takes *ages* to find and the fix is
8194 trivial: move one line.
8195 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8197 *) Ugly workaround to get s_client and s_server working under Windows. The
8198 old code wouldn't work because it needed to select() on sockets and the
8199 tty (for keypresses and to see if data could be written). Win32 only
8200 supports select() on sockets so we select() with a 1s timeout on the
8201 sockets and then see if any characters are waiting to be read, if none
8202 are present then we retry, we also assume we can always write data to
8203 the tty. This isn't nice because the code then blocks until we've
8204 received a complete line of data and it is effectively polling the
8205 keyboard at 1s intervals: however it's quite a bit better than not
8206 working at all :-) A dedicated Windows application might handle this
8207 with an event loop for example.
8210 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
8211 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
8212 will be called when RSA_sign() and RSA_verify() are used. This is useful
8213 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
8214 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
8215 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
8216 This necessitated the support of an extra signature type NID_md5_sha1
8217 for SSL signatures and modifications to the SSL library to use it instead
8218 of calling RSA_public_decrypt() and RSA_private_encrypt().
8221 *) Add new -verify -CAfile and -CApath options to the crl program, these
8222 will lookup a CRL issuers certificate and verify the signature in a
8223 similar way to the verify program. Tidy up the crl program so it
8224 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
8225 less strict. It will now permit CRL extensions even if it is not
8226 a V2 CRL: this will allow it to tolerate some broken CRLs.
8229 *) Initialize all non-automatic variables each time one of the openssl
8230 sub-programs is started (this is necessary as they may be started
8231 multiple times from the "OpenSSL>" prompt).
8232 [Lennart Bang, Bodo Moeller]
8234 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
8235 removing all other RSA functionality (this is what NO_RSA does). This
8236 is so (for example) those in the US can disable those operations covered
8237 by the RSA patent while allowing storage and parsing of RSA keys and RSA
8241 *) Non-copying interface to BIO pairs.
8242 (still largely untested)
8245 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
8246 ASCII string. This was handled independently in various places before.
8249 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
8250 UTF8 strings a character at a time.
8253 *) Use client_version from client hello to select the protocol
8254 (s23_srvr.c) and for RSA client key exchange verification
8255 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
8258 *) Add various utility functions to handle SPKACs, these were previously
8259 handled by poking round in the structure internals. Added new function
8260 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
8261 print, verify and generate SPKACs. Based on an original idea from
8262 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
8265 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
8268 *) Allow the config file extension section to be overwritten on the
8269 command line. Based on an original idea from Massimiliano Pala
8270 <madwolf@comune.modena.it>. The new option is called -extensions
8271 and can be applied to ca, req and x509. Also -reqexts to override
8272 the request extensions in req and -crlexts to override the crl extensions
8276 *) Add new feature to the SPKAC handling in ca. Now you can include
8277 the same field multiple times by preceding it by "XXXX." for example:
8280 this is the same syntax as used in the req config file.
8283 *) Allow certificate extensions to be added to certificate requests. These
8284 are specified in a 'req_extensions' option of the req section of the
8285 config file. They can be printed out with the -text option to req but
8286 are otherwise ignored at present.
8289 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
8290 data read consists of only the final block it would not decrypted because
8291 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
8292 A misplaced 'break' also meant the decrypted final block might not be
8293 copied until the next read.
8296 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
8297 a few extra parameters to the DH structure: these will be useful if
8298 for example we want the value of 'q' or implement X9.42 DH.
8301 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
8302 provides hooks that allow the default DSA functions or functions on a
8303 "per key" basis to be replaced. This allows hardware acceleration and
8304 hardware key storage to be handled without major modification to the
8305 library. Also added low level modexp hooks and CRYPTO_EX structure and
8306 associated functions.
8309 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
8310 as "read only": it can't be written to and the buffer it points to will
8311 not be freed. Reading from a read only BIO is much more efficient than
8312 a normal memory BIO. This was added because there are several times when
8313 an area of memory needs to be read from a BIO. The previous method was
8314 to create a memory BIO and write the data to it, this results in two
8315 copies of the data and an O(n^2) reading algorithm. There is a new
8316 function BIO_new_mem_buf() which creates a read only memory BIO from
8317 an area of memory. Also modified the PKCS#7 routines to use read only
8321 *) Bugfix: ssl23_get_client_hello did not work properly when called in
8322 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
8323 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
8324 but a retry condition occured while trying to read the rest.
8327 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
8328 NID_pkcs7_encrypted by default: this was wrong since this should almost
8329 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
8330 the encrypted data type: this is a more sensible place to put it and it
8331 allows the PKCS#12 code to be tidied up that duplicated this
8335 *) Changed obj_dat.pl script so it takes its input and output files on
8336 the command line. This should avoid shell escape redirection problems
8340 *) Initial support for certificate extension requests, these are included
8341 in things like Xenroll certificate requests. Included functions to allow
8342 extensions to be obtained and added.
8345 *) -crlf option to s_client and s_server for sending newlines as
8346 CRLF (as required by many protocols).
8349 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
8351 *) Install libRSAglue.a when OpenSSL is built with RSAref.
8352 [Ralf S. Engelschall]
8354 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
8355 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
8357 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
8361 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
8362 DH parameters/keys (q is lost during that conversion, but the resulting
8363 DH parameters contain its length).
8365 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
8366 much faster than DH_generate_parameters (which creates parameters
8367 where p = 2*q + 1), and also the smaller q makes DH computations
8368 much more efficient (160-bit exponentiation instead of 1024-bit
8369 exponentiation); so this provides a convenient way to support DHE
8370 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
8371 utter importance to use
8372 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8374 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8375 when such DH parameters are used, because otherwise small subgroup
8376 attacks may become possible!
8379 *) Avoid memory leak in i2d_DHparams.
8382 *) Allow the -k option to be used more than once in the enc program:
8383 this allows the same encrypted message to be read by multiple recipients.
8386 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
8387 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
8388 it will always use the numerical form of the OID, even if it has a short
8392 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
8393 method only got called if p,q,dmp1,dmq1,iqmp components were present,
8394 otherwise bn_mod_exp was called. In the case of hardware keys for example
8395 no private key components need be present and it might store extra data
8396 in the RSA structure, which cannot be accessed from bn_mod_exp.
8397 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
8398 private key operations.
8401 *) Added support for SPARC Linux.
8404 *) pem_password_cb function type incompatibly changed from
8405 typedef int pem_password_cb(char *buf, int size, int rwflag);
8407 ....(char *buf, int size, int rwflag, void *userdata);
8408 so that applications can pass data to their callbacks:
8409 The PEM[_ASN1]_{read,write}... functions and macros now take an
8410 additional void * argument, which is just handed through whenever
8411 the password callback is called.
8412 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
8414 New function SSL_CTX_set_default_passwd_cb_userdata.
8416 Compatibility note: As many C implementations push function arguments
8417 onto the stack in reverse order, the new library version is likely to
8418 interoperate with programs that have been compiled with the old
8419 pem_password_cb definition (PEM_whatever takes some data that
8420 happens to be on the stack as its last argument, and the callback
8421 just ignores this garbage); but there is no guarantee whatsoever that
8424 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
8425 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
8426 problems not only on Windows, but also on some Unix platforms.
8427 To avoid problematic command lines, these definitions are now in an
8428 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
8429 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
8432 *) MIPS III/IV assembler module is reimplemented.
8435 *) More DES library cleanups: remove references to srand/rand and
8436 delete an unused file.
8439 *) Add support for the the free Netwide assembler (NASM) under Win32,
8440 since not many people have MASM (ml) and it can be hard to obtain.
8441 This is currently experimental but it seems to work OK and pass all
8442 the tests. Check out INSTALL.W32 for info.
8445 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
8446 without temporary keys kept an extra copy of the server key,
8447 and connections with temporary keys did not free everything in case
8451 *) New function RSA_check_key and new openssl rsa option -check
8452 for verifying the consistency of RSA keys.
8453 [Ulf Moeller, Bodo Moeller]
8455 *) Various changes to make Win32 compile work:
8456 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
8457 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
8458 comparison" warnings.
8459 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
8462 *) Add a debugging option to PKCS#5 v2 key generation function: when
8463 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
8464 derived keys are printed to stderr.
8467 *) Copy the flags in ASN1_STRING_dup().
8468 [Roman E. Pavlov <pre@mo.msk.ru>]
8470 *) The x509 application mishandled signing requests containing DSA
8471 keys when the signing key was also DSA and the parameters didn't match.
8473 It was supposed to omit the parameters when they matched the signing key:
8474 the verifying software was then supposed to automatically use the CA's
8475 parameters if they were absent from the end user certificate.
8477 Omitting parameters is no longer recommended. The test was also
8478 the wrong way round! This was probably due to unusual behaviour in
8479 EVP_cmp_parameters() which returns 1 if the parameters match.
8480 This meant that parameters were omitted when they *didn't* match and
8481 the certificate was useless. Certificates signed with 'ca' didn't have
8483 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
8485 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
8486 The interface is as follows:
8487 Applications can use
8488 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
8489 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
8490 "off" is now the default.
8491 The library internally uses
8492 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
8493 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
8494 to disable memory-checking temporarily.
8496 Some inconsistent states that previously were possible (and were
8497 even the default) are now avoided.
8499 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
8500 with each memory chunk allocated; this is occasionally more helpful
8501 than just having a counter.
8503 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
8505 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
8509 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
8510 which largely parallels "options", but is for changing API behaviour,
8511 whereas "options" are about protocol behaviour.
8512 Initial "mode" flags are:
8514 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
8515 a single record has been written.
8516 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
8517 retries use the same buffer location.
8518 (But all of the contents must be
8522 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
8525 *) Fix problems with no-hmac etc.
8526 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
8528 *) New functions RSA_get_default_method(), RSA_set_method() and
8529 RSA_get_method(). These allows replacement of RSA_METHODs without having
8530 to mess around with the internals of an RSA structure.
8533 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
8534 Also really enable memory leak checks in openssl.c and in some
8536 [Chad C. Mulligan, Bodo Moeller]
8538 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
8539 up the length of negative integers. This has now been simplified to just
8540 store the length when it is first determined and use it later, rather
8541 than trying to keep track of where data is copied and updating it to
8543 [Steve Henson, reported by Brien Wheeler
8544 <bwheeler@authentica-security.com>]
8546 *) Add a new function PKCS7_signatureVerify. This allows the verification
8547 of a PKCS#7 signature but with the signing certificate passed to the
8548 function itself. This contrasts with PKCS7_dataVerify which assumes the
8549 certificate is present in the PKCS#7 structure. This isn't always the
8550 case: certificates can be omitted from a PKCS#7 structure and be
8551 distributed by "out of band" means (such as a certificate database).
8554 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
8555 function prototypes in pem.h, also change util/mkdef.pl to add the
8556 necessary function names.
8559 *) mk1mf.pl (used by Windows builds) did not properly read the
8560 options set by Configure in the top level Makefile, and Configure
8561 was not even able to write more than one option correctly.
8562 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
8565 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
8566 file to be loaded from a BIO or FILE pointer. The BIO version will
8567 for example allow memory BIOs to contain config info.
8570 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
8571 Whoever hopes to achieve shared-library compatibility across versions
8572 must use this, not the compile-time macro.
8573 (Exercise 0.9.4: Which is the minimum library version required by
8575 Note: All this applies only to multi-threaded programs, others don't
8579 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
8580 through a BIO pair triggered the default case, i.e.
8581 SSLerr(...,SSL_R_UNKNOWN_STATE).
8584 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
8585 can use the SSL library even if none of the specific BIOs is
8589 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
8590 for the encoded length.
8591 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
8593 *) Add initial documentation of the X509V3 functions.
8596 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
8597 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
8598 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
8599 secure PKCS#8 private key format with a high iteration count.
8602 *) Fix determination of Perl interpreter: A perl or perl5
8603 _directory_ in $PATH was also accepted as the interpreter.
8604 [Ralf S. Engelschall]
8606 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
8607 wrong with it but it was very old and did things like calling
8608 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
8612 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
8613 to use the new extension code.
8616 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
8617 with macros. This should make it easier to change their form, add extra
8618 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
8622 *) Add to configuration table a new entry that can specify an alternative
8623 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
8624 according to Mark Crispin <MRC@Panda.COM>.
8628 *) DES CBC did not update the IV. Weird.
8631 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
8632 Changing the behaviour of the former might break existing programs --
8633 where IV updating is needed, des_ncbc_encrypt can be used.
8636 *) When bntest is run from "make test" it drives bc to check its
8637 calculations, as well as internally checking them. If an internal check
8638 fails, it needs to cause bc to give a non-zero result or make test carries
8639 on without noticing the failure. Fixed.
8642 *) DES library cleanups.
8645 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
8646 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
8647 ciphers. NOTE: although the key derivation function has been verified
8648 against some published test vectors it has not been extensively tested
8649 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
8653 *) Instead of "mkdir -p", which is not fully portable, use new
8654 Perl script "util/mkdir-p.pl".
8657 *) Rewrite the way password based encryption (PBE) is handled. It used to
8658 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
8659 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
8660 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
8661 the 'parameter' field of the AlgorithmIdentifier is passed to the
8662 underlying key generation function so it must do its own ASN1 parsing.
8663 This has also changed the EVP_PBE_CipherInit() function which now has a
8664 'parameter' argument instead of literal salt and iteration count values
8665 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
8668 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
8669 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
8670 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
8671 KEY" because this clashed with PKCS#8 unencrypted string. Since this
8672 value was just used as a "magic string" and not used directly its
8673 value doesn't matter.
8676 *) Introduce some semblance of const correctness to BN. Shame C doesn't
8680 *) "linux-sparc64" configuration (ultrapenguin).
8681 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
8682 "linux-sparc" configuration.
8683 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
8685 *) config now generates no-xxx options for missing ciphers.
8688 *) Support the EBCDIC character set (work in progress).
8689 File ebcdic.c not yet included because it has a different license.
8690 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8692 *) Support BS2000/OSD-POSIX.
8693 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8695 *) Make callbacks for key generation use void * instead of char *.
8698 *) Make S/MIME samples compile (not yet tested).
8701 *) Additional typesafe stacks.
8704 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
8708 Changes between 0.9.3 and 0.9.3a [29 May 1999]
8710 *) New configuration variant "sco5-gcc".
8712 *) Updated some demos.
8713 [Sean O Riordain, Wade Scholine]
8715 *) Add missing BIO_free at exit of pkcs12 application.
8718 *) Fix memory leak in conf.c.
8721 *) Updates for Win32 to assembler version of MD5.
8724 *) Set #! path to perl in apps/der_chop to where we found it
8725 instead of using a fixed path.
8728 *) SHA library changes for irix64-mips4-cc.
8731 *) Improvements for VMS support.
8735 Changes between 0.9.2b and 0.9.3 [24 May 1999]
8737 *) Bignum library bug fix. IRIX 6 passes "make test" now!
8738 This also avoids the problems with SC4.2 and unpatched SC5.
8739 [Andy Polyakov <appro@fy.chalmers.se>]
8741 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
8742 These are required because of the typesafe stack would otherwise break
8743 existing code. If old code used a structure member which used to be STACK
8744 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
8745 sk_num or sk_value it would produce an error because the num, data members
8746 are not present in STACK_OF. Now it just produces a warning. sk_set
8747 replaces the old method of assigning a value to sk_value
8748 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
8749 that does this will no longer work (and should use sk_set instead) but
8750 this could be regarded as a "questionable" behaviour anyway.
8753 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
8754 correctly handle encrypted S/MIME data.
8757 *) Change type of various DES function arguments from des_cblock
8758 (which means, in function argument declarations, pointer to char)
8759 to des_cblock * (meaning pointer to array with 8 char elements),
8760 which allows the compiler to do more typechecking; it was like
8761 that back in SSLeay, but with lots of ugly casts.
8763 Introduce new type const_des_cblock.
8766 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
8767 problems: find RecipientInfo structure that matches recipient certificate
8768 and initialise the ASN1 structures properly based on passed cipher.
8771 *) Belatedly make the BN tests actually check the results.
8774 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
8775 to and from BNs: it was completely broken. New compilation option
8776 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
8777 key elements as negative integers.
8780 *) Reorganize and speed up MD5.
8781 [Andy Polyakov <appro@fy.chalmers.se>]
8784 [Richard Levitte <richard@levitte.org>]
8786 *) New option -out to asn1parse to allow the parsed structure to be
8787 output to a file. This is most useful when combined with the -strparse
8788 option to examine the output of things like OCTET STRINGS.
8791 *) Make SSL library a little more fool-proof by not requiring any longer
8792 that SSL_set_{accept,connect}_state be called before
8793 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
8794 in many applications because usually everything *appeared* to work as
8795 intended anyway -- now it really works as intended).
8798 *) Move openssl.cnf out of lib/.
8801 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
8802 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
8803 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
8804 [Ralf S. Engelschall]
8806 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
8807 handle PKCS#7 enveloped data properly.
8808 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
8810 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
8811 copying pointers. The cert_st handling is changed by this in
8812 various ways (and thus what used to be known as ctx->default_cert
8813 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
8814 any longer when s->cert does not give us what we need).
8815 ssl_cert_instantiate becomes obsolete by this change.
8816 As soon as we've got the new code right (possibly it already is?),
8817 we have solved a couple of bugs of the earlier code where s->cert
8818 was used as if it could not have been shared with other SSL structures.
8820 Note that using the SSL API in certain dirty ways now will result
8821 in different behaviour than observed with earlier library versions:
8822 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
8823 does not influence s as it used to.
8825 In order to clean up things more thoroughly, inside SSL_SESSION
8826 we don't use CERT any longer, but a new structure SESS_CERT
8827 that holds per-session data (if available); currently, this is
8828 the peer's certificate chain and, for clients, the server's certificate
8829 and temporary key. CERT holds only those values that can have
8830 meaningful defaults in an SSL_CTX.
8833 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
8834 from the internal representation. Various PKCS#7 fixes: remove some
8835 evil casts and set the enc_dig_alg field properly based on the signing
8839 *) Allow PKCS#12 password to be set from the command line or the
8840 environment. Let 'ca' get its config file name from the environment
8841 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
8845 *) Allow certificate policies extension to use an IA5STRING for the
8846 organization field. This is contrary to the PKIX definition but
8847 VeriSign uses it and IE5 only recognises this form. Document 'x509'
8851 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
8852 without disallowing inline assembler and the like for non-pedantic builds.
8855 *) Support Borland C++ builder.
8856 [Janez Jere <jj@void.si>, modified by Ulf Möller]
8861 *) SHA-1 cleanups and performance enhancements.
8862 [Andy Polyakov <appro@fy.chalmers.se>]
8864 *) Sparc v8plus assembler for the bignum library.
8865 [Andy Polyakov <appro@fy.chalmers.se>]
8867 *) Accept any -xxx and +xxx compiler options in Configure.
8870 *) Update HPUX configuration.
8873 *) Add missing sk_<type>_unshift() function to safestack.h
8874 [Ralf S. Engelschall]
8876 *) New function SSL_CTX_use_certificate_chain_file that sets the
8877 "extra_cert"s in addition to the certificate. (This makes sense
8878 only for "PEM" format files, as chains as a whole are not
8882 *) Support verify_depth from the SSL API.
8883 x509_vfy.c had what can be considered an off-by-one-error:
8884 Its depth (which was not part of the external interface)
8885 was actually counting the number of certificates in a chain;
8886 now it really counts the depth.
8889 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
8890 instead of X509err, which often resulted in confusing error
8891 messages since the error codes are not globally unique
8892 (e.g. an alleged error in ssl3_accept when a certificate
8893 didn't match the private key).
8895 *) New function SSL_CTX_set_session_id_context that allows to set a default
8896 value (so that you don't need SSL_set_session_id_context for each
8897 connection using the SSL_CTX).
8900 *) OAEP decoding bug fix.
8903 *) Support INSTALL_PREFIX for package builders, as proposed by
8907 *) New Configure options "threads" and "no-threads". For systems
8908 where the proper compiler options are known (currently Solaris
8909 and Linux), "threads" is the default.
8912 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
8915 *) Install various scripts to $(OPENSSLDIR)/misc, not to
8916 $(INSTALLTOP)/bin -- they shouldn't clutter directories
8917 such as /usr/local/bin.
8920 *) "make linux-shared" to build shared libraries.
8921 [Niels Poppe <niels@netbox.org>]
8923 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
8926 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
8927 extension adding in x509 utility.
8930 *) Remove NOPROTO sections and error code comments.
8933 *) Partial rewrite of the DEF file generator to now parse the ANSI
8937 *) New Configure options --prefix=DIR and --openssldir=DIR.
8940 *) Complete rewrite of the error code script(s). It is all now handled
8941 by one script at the top level which handles error code gathering,
8942 header rewriting and C source file generation. It should be much better
8943 than the old method: it now uses a modified version of Ulf's parser to
8944 read the ANSI prototypes in all header files (thus the old K&R definitions
8945 aren't needed for error creation any more) and do a better job of
8946 translating function codes into names. The old 'ASN1 error code imbedded
8947 in a comment' is no longer necessary and it doesn't use .err files which
8948 have now been deleted. Also the error code call doesn't have to appear all
8949 on one line (which resulted in some large lines...).
8952 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
8955 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
8956 0 (which usually indicates a closed connection), but continue reading.
8959 *) Fix some race conditions.
8962 *) Add support for CRL distribution points extension. Add Certificate
8963 Policies and CRL distribution points documentation.
8966 *) Move the autogenerated header file parts to crypto/opensslconf.h.
8969 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
8970 8 of keying material. Merlin has also confirmed interop with this fix
8971 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
8972 [Merlin Hughes <merlin@baltimore.ie>]
8974 *) Fix lots of warnings.
8975 [Richard Levitte <levitte@stacken.kth.se>]
8977 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
8978 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
8979 [Richard Levitte <levitte@stacken.kth.se>]
8981 *) Fix problems with sizeof(long) == 8.
8982 [Andy Polyakov <appro@fy.chalmers.se>]
8984 *) Change functions to ANSI C.
8987 *) Fix typos in error codes.
8988 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
8990 *) Remove defunct assembler files from Configure.
8993 *) SPARC v8 assembler BIGNUM implementation.
8994 [Andy Polyakov <appro@fy.chalmers.se>]
8996 *) Support for Certificate Policies extension: both print and set.
8997 Various additions to support the r2i method this uses.
9000 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
9001 return a const string when you are expecting an allocated buffer.
9004 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
9005 types DirectoryString and DisplayText.
9008 *) Add code to allow r2i extensions to access the configuration database,
9009 add an LHASH database driver and add several ctx helper functions.
9012 *) Fix an evil bug in bn_expand2() which caused various BN functions to
9013 fail when they extended the size of a BIGNUM.
9016 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
9017 support typesafe stack.
9020 *) Fix typo in SSL_[gs]et_options().
9021 [Nils Frostberg <nils@medcom.se>]
9023 *) Delete various functions and files that belonged to the (now obsolete)
9024 old X509V3 handling code.
9027 *) New Configure option "rsaref".
9030 *) Don't auto-generate pem.h.
9033 *) Introduce type-safe ASN.1 SETs.
9036 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
9037 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
9039 *) Introduce type-safe STACKs. This will almost certainly break lots of code
9040 that links with OpenSSL (well at least cause lots of warnings), but fear
9041 not: the conversion is trivial, and it eliminates loads of evil casts. A
9042 few STACKed things have been converted already. Feel free to convert more.
9043 In the fullness of time, I'll do away with the STACK type altogether.
9046 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
9047 specified in <certfile> by updating the entry in the index.txt file.
9048 This way one no longer has to edit the index.txt file manually for
9049 revoking a certificate. The -revoke option does the gory details now.
9050 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
9052 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
9053 `-text' option at all and this way the `-noout -text' combination was
9054 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
9055 [Ralf S. Engelschall]
9057 *) Make sure a corresponding plain text error message exists for the
9058 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
9059 verify callback function determined that a certificate was revoked.
9060 [Ralf S. Engelschall]
9062 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
9063 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
9064 all available cipers including rc5, which was forgotten until now.
9065 In order to let the testing shell script know which algorithms
9066 are available, a new (up to now undocumented) command
9067 "openssl list-cipher-commands" is used.
9070 *) Bugfix: s_client occasionally would sleep in select() when
9071 it should have checked SSL_pending() first.
9074 *) New functions DSA_do_sign and DSA_do_verify to provide access to
9075 the raw DSA values prior to ASN.1 encoding.
9078 *) Tweaks to Configure
9079 [Niels Poppe <niels@netbox.org>]
9081 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
9085 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
9088 *) New config option to avoid instructions that are illegal on the 80386.
9089 The default code is faster, but requires at least a 486.
9092 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
9093 SSL2_SERVER_VERSION (not used at all) macros, which are now the
9094 same as SSL2_VERSION anyway.
9097 *) New "-showcerts" option for s_client.
9100 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
9101 application. Various cleanups and fixes.
9104 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
9105 modify error routines to work internally. Add error codes and PBE init
9106 to library startup routines.
9109 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
9110 packing functions to asn1 and evp. Changed function names and error
9111 codes along the way.
9114 *) PKCS12 integration: and so it begins... First of several patches to
9115 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
9116 objects to objects.h
9119 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
9120 and display support for Thawte strong extranet extension.
9123 *) Add LinuxPPC support.
9124 [Jeff Dubrule <igor@pobox.org>]
9126 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
9127 bn_div_words in alpha.s.
9128 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
9130 *) Make sure the RSA OAEP test is skipped under -DRSAref because
9131 OAEP isn't supported when OpenSSL is built with RSAref.
9132 [Ulf Moeller <ulf@fitug.de>]
9134 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
9135 so they no longer are missing under -DNOPROTO.
9136 [Soren S. Jorvang <soren@t.dk>]
9139 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
9141 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
9142 doesn't work when the session is reused. Coming soon!
9145 *) Fix a security hole, that allows sessions to be reused in the wrong
9146 context thus bypassing client cert protection! All software that uses
9147 client certs and session caches in multiple contexts NEEDS PATCHING to
9148 allow session reuse! A fuller solution is in the works.
9149 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
9151 *) Some more source tree cleanups (removed obsolete files
9152 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
9153 permission on "config" script to be executable) and a fix for the INSTALL
9155 [Ulf Moeller <ulf@fitug.de>]
9157 *) Remove some legacy and erroneous uses of malloc, free instead of
9159 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
9161 *) Make rsa_oaep_test return non-zero on error.
9162 [Ulf Moeller <ulf@fitug.de>]
9164 *) Add support for native Solaris shared libraries. Configure
9165 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
9166 if someone would make that last step automatic.
9167 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
9169 *) ctx_size was not built with the right compiler during "make links". Fixed.
9172 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
9173 except NULL ciphers". This means the default cipher list will no longer
9174 enable NULL ciphers. They need to be specifically enabled e.g. with
9175 the string "DEFAULT:eNULL".
9178 *) Fix to RSA private encryption routines: if p < q then it would
9179 occasionally produce an invalid result. This will only happen with
9180 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9183 *) Be less restrictive and allow also `perl util/perlpath.pl
9184 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9185 because this way one can also use an interpreter named `perl5' (which is
9186 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9187 installed as `perl').
9188 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9190 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9191 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9193 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9194 advapi32.lib to Win32 build and change the pem test comparision
9195 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9196 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9197 and crypto/des/ede_cbcm_enc.c.
9200 *) DES quad checksum was broken on big-endian architectures. Fixed.
9203 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9204 Win32 test batch file so it (might) work again. The Win32 test batch file
9205 is horrible: I feel ill....
9208 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
9209 in e_os.h. Audit of header files to check ANSI and non ANSI
9210 sections: 10 functions were absent from non ANSI section and not exported
9211 from Windows DLLs. Fixed up libeay.num for new functions.
9214 *) Make `openssl version' output lines consistent.
9215 [Ralf S. Engelschall]
9217 *) Fix Win32 symbol export lists for BIO functions: Added
9218 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
9219 to ms/libeay{16,32}.def.
9220 [Ralf S. Engelschall]
9222 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
9223 fine under Unix and passes some trivial tests I've now added. But the
9224 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
9225 added to make sure no one expects that this stuff really works in the
9226 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
9227 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
9229 [Ralf S. Engelschall]
9231 *) Fix the generation of two part addresses in perl.
9232 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
9234 *) Add config entry for Linux on MIPS.
9235 [John Tobey <jtobey@channel1.com>]
9237 *) Make links whenever Configure is run, unless we are on Windoze.
9240 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
9241 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
9245 *) Add a useful kludge to allow package maintainers to specify compiler and
9246 other platforms details on the command line without having to patch the
9247 Configure script everytime: One now can use ``perl Configure
9248 <id>:<details>'', i.e. platform ids are allowed to have details appended
9249 to them (seperated by colons). This is treated as there would be a static
9250 pre-configured entry in Configure's %table under key <id> with value
9251 <details> and ``perl Configure <id>'' is called. So, when you want to
9252 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
9253 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
9254 now, which overrides the FreeBSD-elf entry on-the-fly.
9255 [Ralf S. Engelschall]
9257 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
9260 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
9261 on the `perl Configure ...' command line. This way one can compile
9262 OpenSSL libraries with Position Independent Code (PIC) which is needed
9263 for linking it into DSOs.
9264 [Ralf S. Engelschall]
9266 *) Remarkably, export ciphers were totally broken and no-one had noticed!
9270 *) Cleaned up the LICENSE document: The official contact for any license
9271 questions now is the OpenSSL core team under openssl-core@openssl.org.
9272 And add a paragraph about the dual-license situation to make sure people
9273 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
9274 to the OpenSSL toolkit.
9275 [Ralf S. Engelschall]
9277 *) General source tree makefile cleanups: Made `making xxx in yyy...'
9278 display consistent in the source tree and replaced `/bin/rm' by `rm'.
9279 Additonally cleaned up the `make links' target: Remove unnecessary
9280 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
9281 to speed processing and no longer clutter the display with confusing
9282 stuff. Instead only the actually done links are displayed.
9283 [Ralf S. Engelschall]
9285 *) Permit null encryption ciphersuites, used for authentication only. It used
9286 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
9287 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
9291 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
9292 signed attributes when verifying signatures (this would break them),
9293 the detached data encoding was wrong and public keys obtained using
9294 X509_get_pubkey() weren't freed.
9297 *) Add text documentation for the BUFFER functions. Also added a work around
9298 to a Win95 console bug. This was triggered by the password read stuff: the
9299 last character typed gets carried over to the next fread(). If you were
9300 generating a new cert request using 'req' for example then the last
9301 character of the passphrase would be CR which would then enter the first
9305 *) Added the new `Includes OpenSSL Cryptography Software' button as
9306 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
9307 button and can be used by applications based on OpenSSL to show the
9308 relationship to the OpenSSL project.
9309 [Ralf S. Engelschall]
9311 *) Remove confusing variables in function signatures in files
9312 ssl/ssl_lib.c and ssl/ssl.h.
9313 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9315 *) Don't install bss_file.c under PREFIX/include/
9316 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9318 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
9319 functions that return function pointers and has support for NT specific
9320 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
9321 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
9322 unsigned to signed types: this was killing the Win32 compile.
9325 *) Add new certificate file to stack functions,
9326 SSL_add_dir_cert_subjects_to_stack() and
9327 SSL_add_file_cert_subjects_to_stack(). These largely supplant
9328 SSL_load_client_CA_file(), and can be used to add multiple certs easily
9329 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
9330 This means that Apache-SSL and similar packages don't have to mess around
9331 to add as many CAs as they want to the preferred list.
9334 *) Experiment with doxygen documentation. Currently only partially applied to
9336 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
9337 openssl.doxy as the configuration file.
9340 *) Get rid of remaining C++-style comments which strict C compilers hate.
9341 [Ralf S. Engelschall, pointed out by Carlos Amengual]
9343 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
9344 compiled in by default: it has problems with large keys.
9347 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
9348 DH private keys and/or callback functions which directly correspond to
9349 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
9350 is needed for applications which have to configure certificates on a
9351 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
9353 For the RSA certificate situation is makes no difference, but
9354 for the DSA certificate situation this fixes the "no shared cipher"
9355 problem where the OpenSSL cipher selection procedure failed because the
9356 temporary keys were not overtaken from the context and the API provided
9357 no way to reconfigure them.
9358 The new functions now let applications reconfigure the stuff and they
9359 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
9360 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
9361 non-public-API function ssl_cert_instantiate() is used as a helper
9362 function and also to reduce code redundancy inside ssl_rsa.c.
9363 [Ralf S. Engelschall]
9365 *) Move s_server -dcert and -dkey options out of the undocumented feature
9366 area because they are useful for the DSA situation and should be
9367 recognized by the users.
9368 [Ralf S. Engelschall]
9370 *) Fix the cipher decision scheme for export ciphers: the export bits are
9371 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
9372 SSL_EXP_MASK. So, the original variable has to be used instead of the
9373 already masked variable.
9374 [Richard Levitte <levitte@stacken.kth.se>]
9376 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
9377 [Richard Levitte <levitte@stacken.kth.se>]
9379 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
9380 from `int' to `unsigned int' because it's a length and initialized by
9381 EVP_DigestFinal() which expects an `unsigned int *'.
9382 [Richard Levitte <levitte@stacken.kth.se>]
9384 *) Don't hard-code path to Perl interpreter on shebang line of Configure
9385 script. Instead use the usual Shell->Perl transition trick.
9386 [Ralf S. Engelschall]
9388 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
9389 (in addition to RSA certificates) to match the behaviour of `openssl dsa
9390 -noout -modulus' as it's already the case for `openssl rsa -noout
9391 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
9392 currently the public key is printed (a decision which was already done by
9393 `openssl dsa -modulus' in the past) which serves a similar purpose.
9394 Additionally the NO_RSA no longer completely removes the whole -modulus
9395 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
9397 [Ralf S. Engelschall]
9399 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
9400 BIO. See the source (crypto/evp/bio_ok.c) for more info.
9401 [Arne Ansper <arne@ats.cyber.ee>]
9403 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
9404 to be added. Now both 'req' and 'ca' can use new objects defined in the
9408 *) Add cool BIO that does syslog (or event log on NT).
9409 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
9411 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
9412 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
9413 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
9414 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
9417 *) Add preliminary config info for new extension code.
9420 *) Make RSA_NO_PADDING really use no padding.
9421 [Ulf Moeller <ulf@fitug.de>]
9423 *) Generate errors when private/public key check is done.
9426 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
9427 for some CRL extensions and new objects added.
9430 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
9431 key usage extension and fuller support for authority key id.
9434 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
9435 padding method for RSA, which is recommended for new applications in PKCS
9436 #1 v2.0 (RFC 2437, October 1998).
9437 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
9438 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
9439 against Bleichbacher's attack on RSA.
9440 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
9443 *) Updates to the new SSL compression code
9444 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9446 *) Fix so that the version number in the master secret, when passed
9447 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
9448 (because the server will not accept higher), that the version number
9449 is 0x03,0x01, not 0x03,0x00
9450 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9452 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
9453 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
9454 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
9457 *) Support for RAW extensions where an arbitrary extension can be
9458 created by including its DER encoding. See apps/openssl.cnf for
9462 *) Make sure latest Perl versions don't interpret some generated C array
9463 code as Perl array code in the crypto/err/err_genc.pl script.
9464 [Lars Weber <3weber@informatik.uni-hamburg.de>]
9466 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
9467 not many people have the assembler. Various Win32 compilation fixes and
9468 update to the INSTALL.W32 file with (hopefully) more accurate Win32
9472 *) Modify configure script 'Configure' to automatically create crypto/date.h
9473 file under Win32 and also build pem.h from pem.org. New script
9474 util/mkfiles.pl to create the MINFO file on environments that can't do a
9475 'make files': perl util/mkfiles.pl >MINFO should work.
9478 *) Major rework of DES function declarations, in the pursuit of correctness
9479 and purity. As a result, many evil casts evaporated, and some weirdness,
9480 too. You may find this causes warnings in your code. Zapping your evil
9481 casts will probably fix them. Mostly.
9484 *) Fix for a typo in asn1.h. Bug fix to object creation script
9485 obj_dat.pl. It considered a zero in an object definition to mean
9486 "end of object": none of the objects in objects.h have any zeros
9487 so it wasn't spotted.
9488 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
9490 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
9491 Masking (CBCM). In the absence of test vectors, the best I have been able
9492 to do is check that the decrypt undoes the encrypt, so far. Send me test
9493 vectors if you have them.
9496 *) Correct calculation of key length for export ciphers (too much space was
9497 allocated for null ciphers). This has not been tested!
9500 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
9501 message is now correct (it understands "crypto" and "ssl" on its
9502 command line). There is also now an "update" option. This will update
9503 the util/ssleay.num and util/libeay.num files with any new functions.
9505 perl util/mkdef.pl crypto ssl update
9506 it will update them.
9509 *) Overhauled the Perl interface (perl/*):
9510 - ported BN stuff to OpenSSL's different BN library
9511 - made the perl/ source tree CVS-aware
9512 - renamed the package from SSLeay to OpenSSL (the files still contain
9513 their history because I've copied them in the repository)
9514 - removed obsolete files (the test scripts will be replaced
9515 by better Test::Harness variants in the future)
9516 [Ralf S. Engelschall]
9518 *) First cut for a very conservative source tree cleanup:
9519 1. merge various obsolete readme texts into doc/ssleay.txt
9520 where we collect the old documents and readme texts.
9521 2. remove the first part of files where I'm already sure that we no
9522 longer need them because of three reasons: either they are just temporary
9523 files which were left by Eric or they are preserved original files where
9524 I've verified that the diff is also available in the CVS via "cvs diff
9525 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
9526 the crypto/md/ stuff).
9527 [Ralf S. Engelschall]
9529 *) More extension code. Incomplete support for subject and issuer alt
9530 name, issuer and authority key id. Change the i2v function parameters
9531 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
9532 what that's for :-) Fix to ASN1 macro which messed up
9533 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
9536 *) Preliminary support for ENUMERATED type. This is largely copied from the
9540 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
9541 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9543 *) Make sure `make rehash' target really finds the `openssl' program.
9544 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9546 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
9547 like to hear about it if this slows down other processors.
9550 *) Add CygWin32 platform information to Configure script.
9551 [Alan Batie <batie@aahz.jf.intel.com>]
9553 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
9554 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
9556 *) New program nseq to manipulate netscape certificate sequences
9559 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
9563 *) Fixes to BN code. Previously the default was to define BN_RECURSION
9564 but the BN code had some problems that would cause failures when
9565 doing certificate verification and some other functions.
9566 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9568 *) Add ASN1 and PEM code to support netscape certificate sequences.
9571 *) Add ASN1 and PEM code to support netscape certificate sequences.
9574 *) Add several PKIX and private extended key usage OIDs.
9577 *) Modify the 'ca' program to handle the new extension code. Modify
9578 openssl.cnf for new extension format, add comments.
9581 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
9582 and add a sample to openssl.cnf so req -x509 now adds appropriate
9586 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
9587 error code, add initial support to X509_print() and x509 application.
9590 *) Takes a deep breath and start addding X509 V3 extension support code. Add
9591 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
9592 stuff is currently isolated and isn't even compiled yet.
9595 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
9596 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
9597 Removed the versions check from X509 routines when loading extensions:
9598 this allows certain broken certificates that don't set the version
9599 properly to be processed.
9602 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
9603 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
9604 can still be regenerated with "make depend".
9607 *) Spelling mistake in C version of CAST-128.
9608 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
9610 *) Changes to the error generation code. The perl script err-code.pl
9611 now reads in the old error codes and retains the old numbers, only
9612 adding new ones if necessary. It also only changes the .err files if new
9613 codes are added. The makefiles have been modified to only insert errors
9614 when needed (to avoid needlessly modifying header files). This is done
9615 by only inserting errors if the .err file is newer than the auto generated
9616 C file. To rebuild all the error codes from scratch (the old behaviour)
9617 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
9618 or delete all the .err files.
9621 *) CAST-128 was incorrectly implemented for short keys. The C version has
9622 been fixed, but is untested. The assembler versions are also fixed, but
9623 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
9624 to regenerate it if needed.
9625 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
9626 Hagino <itojun@kame.net>]
9628 *) File was opened incorrectly in randfile.c.
9629 [Ulf Möller <ulf@fitug.de>]
9631 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
9632 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
9633 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
9634 al: it's just almost always a UTCTime. Note this patch adds new error
9635 codes so do a "make errors" if there are problems.
9638 *) Correct Linux 1 recognition in config.
9639 [Ulf Möller <ulf@fitug.de>]
9641 *) Remove pointless MD5 hash when using DSA keys in ca.
9642 [Anonymous <nobody@replay.com>]
9644 *) Generate an error if given an empty string as a cert directory. Also
9645 generate an error if handed NULL (previously returned 0 to indicate an
9646 error, but didn't set one).
9647 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
9649 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
9652 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
9653 parameters. This was causing a warning which killed off the Win32 compile.
9656 *) Remove C++ style comments from crypto/bn/bn_local.h.
9657 [Neil Costigan <neil.costigan@celocom.com>]
9659 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
9660 based on a text string, looking up short and long names and finally
9661 "dot" format. The "dot" format stuff didn't work. Added new function
9662 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
9663 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
9664 OID is not part of the table.
9667 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
9668 X509_LOOKUP_by_alias().
9671 *) Sort openssl functions by name.
9674 *) Get the gendsa program working (hopefully) and add it to app list. Remove
9675 encryption from sample DSA keys (in case anyone is interested the password
9679 *) Make _all_ *_free functions accept a NULL pointer.
9680 [Frans Heymans <fheymans@isaserver.be>]
9682 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
9684 [Anonymous <nobody@replay.com>]
9686 *) s_server should send the CAfile as acceptable CAs, not its own cert.
9687 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9689 *) Don't blow it for numeric -newkey arguments to apps/req.
9690 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9692 *) Temp key "for export" tests were wrong in s3_srvr.c.
9693 [Anonymous <nobody@replay.com>]
9695 *) Add prototype for temp key callback functions
9696 SSL_CTX_set_tmp_{rsa,dh}_callback().
9699 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
9700 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
9703 *) X509_name_add_entry() freed the wrong thing after an error.
9704 [Arne Ansper <arne@ats.cyber.ee>]
9706 *) rsa_eay.c would attempt to free a NULL context.
9707 [Arne Ansper <arne@ats.cyber.ee>]
9709 *) BIO_s_socket() had a broken should_retry() on Windoze.
9710 [Arne Ansper <arne@ats.cyber.ee>]
9712 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
9713 [Arne Ansper <arne@ats.cyber.ee>]
9715 *) Make sure the already existing X509_STORE->depth variable is initialized
9716 in X509_STORE_new(), but document the fact that this variable is still
9717 unused in the certificate verification process.
9718 [Ralf S. Engelschall]
9720 *) Fix the various library and apps files to free up pkeys obtained from
9721 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
9724 *) Fix reference counting in X509_PUBKEY_get(). This makes
9725 demos/maurice/example2.c work, amongst others, probably.
9726 [Steve Henson and Ben Laurie]
9728 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
9729 `openssl' and second, the shortcut symlinks for the `openssl <command>'
9730 are no longer created. This way we have a single and consistent command
9731 line interface `openssl <command>', similar to `cvs <command>'.
9732 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
9734 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
9735 BIT STRING wrapper always have zero unused bits.
9738 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
9741 *) Make the top-level INSTALL documentation easier to understand.
9744 *) Makefiles updated to exit if an error occurs in a sub-directory
9745 make (including if user presses ^C) [Paul Sutton]
9747 *) Make Montgomery context stuff explicit in RSA data structure.
9750 *) Fix build order of pem and err to allow for generated pem.h.
9753 *) Fix renumbering bug in X509_NAME_delete_entry().
9756 *) Enhanced the err-ins.pl script so it makes the error library number
9757 global and can add a library name. This is needed for external ASN1 and
9758 other error libraries.
9761 *) Fixed sk_insert which never worked properly.
9764 *) Fix ASN1 macros so they can handle indefinite length construted
9765 EXPLICIT tags. Some non standard certificates use these: they can now
9769 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
9770 into a single doc/ssleay.txt bundle. This way the information is still
9771 preserved but no longer messes up this directory. Now it's new room for
9772 the new set of documenation files.
9773 [Ralf S. Engelschall]
9775 *) SETs were incorrectly DER encoded. This was a major pain, because they
9776 shared code with SEQUENCEs, which aren't coded the same. This means that
9777 almost everything to do with SETs or SEQUENCEs has either changed name or
9778 number of arguments.
9779 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
9781 *) Fix test data to work with the above.
9784 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
9785 was already fixed by Eric for 0.9.1 it seems.
9786 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
9788 *) Autodetect FreeBSD3.
9791 *) Fix various bugs in Configure. This affects the following platforms:
9795 unixware-2.0-pentium
9799 *) Eliminate generated files from CVS. Reorder tests to regenerate files
9800 before they are needed.
9803 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
9807 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
9809 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
9810 changed SSLeay to OpenSSL in version strings.
9811 [Ralf S. Engelschall]
9813 *) Some fixups to the top-level documents.
9816 *) Fixed the nasty bug where rsaref.h was not found under compile-time
9817 because the symlink to include/ was missing.
9818 [Ralf S. Engelschall]
9820 *) Incorporated the popular no-RSA/DSA-only patches
9821 which allow to compile a RSA-free SSLeay.
9822 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
9824 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
9825 when "ssleay" is still not found.
9826 [Ralf S. Engelschall]
9828 *) Added more platforms to Configure: Cray T3E, HPUX 11,
9829 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
9831 *) Updated the README file.
9832 [Ralf S. Engelschall]
9834 *) Added various .cvsignore files in the CVS repository subdirs
9835 to make a "cvs update" really silent.
9836 [Ralf S. Engelschall]
9838 *) Recompiled the error-definition header files and added
9839 missing symbols to the Win32 linker tables.
9840 [Ralf S. Engelschall]
9842 *) Cleaned up the top-level documents;
9843 o new files: CHANGES and LICENSE
9844 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
9845 o merged COPYRIGHT into LICENSE
9846 o removed obsolete TODO file
9847 o renamed MICROSOFT to INSTALL.W32
9848 [Ralf S. Engelschall]
9850 *) Removed dummy files from the 0.9.1b source tree:
9851 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
9852 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
9853 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
9854 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
9855 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
9856 [Ralf S. Engelschall]
9858 *) Added various platform portability fixes.
9861 *) The Genesis of the OpenSSL rpject:
9862 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
9863 Young and Tim J. Hudson created while they were working for C2Net until
9865 [The OpenSSL Project]
9868 Changes between 0.9.0b and 0.9.1b [not released]
9870 *) Updated a few CA certificates under certs/
9873 *) Changed some BIGNUM api stuff.
9876 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
9877 DGUX x86, Linux Alpha, etc.
9880 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
9881 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
9885 *) Add -strparse option to asn1pars program which parses nested
9887 [Dr Stephen Henson <shenson@bigfoot.com>]
9889 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
9892 *) DSA fix for "ca" program.
9895 *) Added "-genkey" option to "dsaparam" program.
9898 *) Added RIPE MD160 (rmd160) message digest.
9901 *) Added -a (all) option to "ssleay version" command.
9904 *) Added PLATFORM define which is the id given to Configure.
9907 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
9910 *) Extended the ASN.1 parser routines.
9913 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
9916 *) Added a BN_CTX to the BN library.
9919 *) Fixed the weak key values in DES library
9922 *) Changed API in EVP library for cipher aliases.
9925 *) Added support for RC2/64bit cipher.
9928 *) Converted the lhash library to the crypto/mem.c functions.
9931 *) Added more recognized ASN.1 object ids.
9934 *) Added more RSA padding checks for SSL/TLS.
9937 *) Added BIO proxy/filter functionality.
9940 *) Added extra_certs to SSL_CTX which can be used
9941 send extra CA certificates to the client in the CA cert chain sending
9942 process. It can be configured with SSL_CTX_add_extra_chain_cert().
9945 *) Now Fortezza is denied in the authentication phase because
9946 this is key exchange mechanism is not supported by SSLeay at all.
9949 *) Additional PKCS1 checks.
9952 *) Support the string "TLSv1" for all TLS v1 ciphers.
9955 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
9956 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
9959 *) Fixed a few memory leaks.
9962 *) Fixed various code and comment typos.
9965 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
9966 bytes sent in the client random.
9967 [Edward Bishop <ebishop@spyglass.com>]