5 Changes between 1.0.0r and 1.0.0s [xx XXX xxxx]
9 Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
11 *) Segmentation fault in ASN1_TYPE_cmp fix
13 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
14 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
15 certificate signature algorithm consistency this can be used to crash any
16 certificate verification operation and exploited in a DoS attack. Any
17 application which performs certificate verification is vulnerable including
18 OpenSSL clients and servers which enable client authentication.
22 *) ASN.1 structure reuse memory corruption fix
24 Reusing a structure in ASN.1 parsing may allow an attacker to cause
25 memory corruption via an invalid write. Such reuse is and has been
26 strongly discouraged and is believed to be rare.
28 Applications that parse structures containing CHOICE or ANY DEFINED BY
29 components may be affected. Certificate parsing (d2i_X509 and related
30 functions) are however not affected. OpenSSL clients and servers are
35 *) PKCS7 NULL pointer dereferences fix
37 The PKCS#7 parsing code does not handle missing outer ContentInfo
38 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
39 missing content and trigger a NULL pointer dereference on parsing.
41 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
42 otherwise parse PKCS#7 structures from untrusted sources are
43 affected. OpenSSL clients and servers are not affected.
45 This issue was reported to OpenSSL by Michal Zalewski (Google).
49 *) DoS via reachable assert in SSLv2 servers fix
51 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
52 servers that both support SSLv2 and enable export cipher suites by sending
53 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
55 This issue was discovered by Sean Burford (Google) and Emilia Käsper
56 (OpenSSL development team).
60 *) Use After Free following d2i_ECPrivatekey error fix
62 A malformed EC private key file consumed via the d2i_ECPrivateKey function
63 could cause a use after free condition. This, in turn, could cause a double
64 free in several private key parsing functions (such as d2i_PrivateKey
65 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
66 for applications that receive EC private keys from untrusted
67 sources. This scenario is considered rare.
69 This issue was discovered by the BoringSSL project and fixed in their
74 *) X509_to_X509_REQ NULL pointer deref fix
76 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
77 the certificate key is invalid. This function is rarely used in practice.
79 This issue was discovered by Brian Carpenter.
83 *) Removed the export ciphers from the DEFAULT ciphers
86 Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
88 *) Build fixes for the Windows and OpenVMS platforms
89 [Matt Caswell and Richard Levitte]
91 Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
93 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
94 message can cause a segmentation fault in OpenSSL due to a NULL pointer
95 dereference. This could lead to a Denial Of Service attack. Thanks to
96 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
100 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
101 dtls1_buffer_record function under certain conditions. In particular this
102 could occur if an attacker sent repeated DTLS records with the same
103 sequence number but for the next epoch. The memory leak could be exploited
104 by an attacker in a Denial of Service attack through memory exhaustion.
105 Thanks to Chris Mueller for reporting this issue.
109 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
110 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
111 method would be set to NULL which could later result in a NULL pointer
112 dereference. Thanks to Frank Schmirler for reporting this issue.
116 *) Abort handshake if server key exchange message is omitted for ephemeral
119 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
120 reporting this issue.
124 *) Remove non-export ephemeral RSA code on client and server. This code
125 violated the TLS standard by allowing the use of temporary RSA keys in
126 non-export ciphersuites and could be used by a server to effectively
127 downgrade the RSA key length used to a value smaller than the server
128 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
129 INRIA or reporting this issue.
133 *) Fixed issue where DH client certificates are accepted without verification.
134 An OpenSSL server will accept a DH certificate for client authentication
135 without the certificate verify message. This effectively allows a client to
136 authenticate without the use of a private key. This only affects servers
137 which trust a client certificate authority which issues certificates
138 containing DH keys: these are extremely rare and hardly ever encountered.
139 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
144 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
145 results on some platforms, including x86_64. This bug occurs at random
146 with a very low probability, and is not known to be exploitable in any
147 way, though its exact impact is difficult to determine. Thanks to Pieter
148 Wuille (Blockstream) who reported this issue and also suggested an initial
149 fix. Further analysis was conducted by the OpenSSL development team and
150 Adam Langley of Google. The final fix was developed by Andy Polyakov of
151 the OpenSSL core team.
155 *) Fix various certificate fingerprint issues.
157 By using non-DER or invalid encodings outside the signed portion of a
158 certificate the fingerprint can be changed without breaking the signature.
159 Although no details of the signed portion of the certificate can be changed
160 this can cause problems with some applications: e.g. those using the
161 certificate fingerprint for blacklists.
163 1. Reject signatures with non zero unused bits.
165 If the BIT STRING containing the signature has non zero unused bits reject
166 the signature. All current signature algorithms require zero unused bits.
168 2. Check certificate algorithm consistency.
170 Check the AlgorithmIdentifier inside TBS matches the one in the
171 certificate signature. NB: this will result in signature failure
172 errors for some broken certificates.
174 Thanks to Konrad Kraszewski from Google for reporting this issue.
176 3. Check DSA/ECDSA signatures use DER.
178 Reencode DSA/ECDSA signatures and compare with the original received
179 signature. Return an error if there is a mismatch.
181 This will reject various cases including garbage after signature
182 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
183 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
184 (negative or with leading zeroes).
186 Further analysis was conducted and fixes were developed by Stephen Henson
187 of the OpenSSL core team.
192 Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
194 *) Session Ticket Memory Leak.
196 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
197 integrity of that ticket is first verified. In the event of a session
198 ticket integrity check failing, OpenSSL will fail to free memory
199 causing a memory leak. By sending a large number of invalid session
200 tickets an attacker could exploit this issue in a Denial Of Service
205 *) Build option no-ssl3 is incomplete.
207 When OpenSSL is configured with "no-ssl3" as a build option, servers
208 could accept and complete a SSL 3.0 handshake, and clients could be
209 configured to send them.
211 [Akamai and the OpenSSL team]
213 *) Add support for TLS_FALLBACK_SCSV.
214 Client applications doing fallback retries should call
215 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
217 [Adam Langley, Bodo Moeller]
219 *) Add additional DigestInfo checks.
221 Reencode DigestInto in DER and check against the original when
222 verifying RSA signature: this will reject any improperly encoded
223 DigestInfo structures.
225 Note: this is a precautionary measure and no attacks are currently known.
229 Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
231 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
232 to a denial of service attack. A malicious server can crash the client
233 with a null pointer dereference (read) by specifying an anonymous (EC)DH
234 ciphersuite and sending carefully crafted handshake messages.
236 Thanks to Felix Gröbert (Google) for discovering and researching this
241 *) By sending carefully crafted DTLS packets an attacker could cause openssl
242 to leak memory. This can be exploited through a Denial of Service attack.
243 Thanks to Adam Langley for discovering and researching this issue.
247 *) An attacker can force openssl to consume large amounts of memory whilst
248 processing DTLS handshake messages. This can be exploited through a
249 Denial of Service attack.
250 Thanks to Adam Langley for discovering and researching this issue.
254 *) An attacker can force an error condition which causes openssl to crash
255 whilst processing DTLS packets due to memory being freed twice. This
256 can be exploited through a Denial of Service attack.
257 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
262 *) If a multithreaded client connects to a malicious server using a resumed
263 session and the server sends an ec point format extension it could write
264 up to 255 bytes to freed memory.
266 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
271 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
272 X509_name_oneline, X509_name_print_ex et al. to leak some information
273 from the stack. Applications may be affected if they echo pretty printing
274 output to the attacker.
276 Thanks to Ivan Fratric (Google) for discovering this issue.
278 [Emilia Käsper, and Steve Henson]
280 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
281 for corner cases. (Certain input points at infinity could lead to
282 bogus results, with non-infinity inputs mapped to infinity too.)
285 Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
287 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
288 handshake can force the use of weak keying material in OpenSSL
289 SSL/TLS clients and servers.
291 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
292 researching this issue. (CVE-2014-0224)
293 [KIKUCHI Masashi, Steve Henson]
295 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
296 OpenSSL DTLS client the code can be made to recurse eventually crashing
299 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
301 [Imre Rad, Steve Henson]
303 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
304 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
305 client or server. This is potentially exploitable to run arbitrary
306 code on a vulnerable client or server.
308 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
309 [Jüri Aedla, Steve Henson]
311 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
312 are subject to a denial of service attack.
314 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
315 this issue. (CVE-2014-3470)
316 [Felix Gröbert, Ivan Fratric, Steve Henson]
318 *) Harmonize version and its documentation. -f flag is used to display
320 [mancha <mancha1@zoho.com>]
322 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
324 [mancha <mancha1@zoho.com>]
326 *) Fix some double frees. These are not thought to be exploitable.
327 [mancha <mancha1@zoho.com>]
329 *) Fix for the attack described in the paper "Recovering OpenSSL
330 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
331 by Yuval Yarom and Naomi Benger. Details can be obtained from:
332 http://eprint.iacr.org/2014/140
334 Thanks to Yuval Yarom and Naomi Benger for discovering this
335 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
336 [Yuval Yarom and Naomi Benger]
338 Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
340 *) Keep original DTLS digest and encryption contexts in retransmission
341 structures so we can use the previous session parameters if they need
342 to be resent. (CVE-2013-6450)
345 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
346 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
347 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
348 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
349 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
350 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
351 [Rob Stradling, Adam Langley]
353 Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
355 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
357 This addresses the flaw in CBC record processing discovered by
358 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
359 at: http://www.isg.rhul.ac.uk/tls/
361 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
362 Security Group at Royal Holloway, University of London
363 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
364 Emilia Käsper for the initial patch.
366 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
368 *) Return an error when checking OCSP signatures when key is NULL.
369 This fixes a DoS attack. (CVE-2013-0166)
372 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
373 the right response is stapled. Also change SSL_get_certificate()
374 so it returns the certificate actually sent.
375 See http://rt.openssl.org/Ticket/Display.html?id=2836.
377 [Rob Stradling <rob.stradling@comodo.com>]
379 *) Fix possible deadlock when decoding public keys.
382 Changes between 1.0.0i and 1.0.0j [10 May 2012]
384 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
387 *) Sanity check record length before skipping explicit IV in DTLS
390 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
391 fuzzing as a service testing platform.
395 *) Initialise tkeylen properly when encrypting CMS messages.
396 Thanks to Solar Designer of Openwall for reporting this issue.
399 Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
401 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
402 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
403 in CRYPTO_realloc_clean.
405 Thanks to Tavis Ormandy, Google Security Team, for discovering this
406 issue and to Adam Langley <agl@chromium.org> for fixing it.
408 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
410 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
412 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
413 in CMS and PKCS7 code. When RSA decryption fails use a random key for
414 content decryption and always return the same error. Note: this attack
415 needs on average 2^20 messages so it only affects automated senders. The
416 old behaviour can be reenabled in the CMS code by setting the
417 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
418 an MMA defence is not necessary.
419 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
420 this issue. (CVE-2012-0884)
423 *) Fix CVE-2011-4619: make sure we really are receiving a
424 client hello before rejecting multiple SGC restarts. Thanks to
425 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
428 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
430 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
431 Thanks to Antonio Martin, Enterprise Secure Access Research and
432 Development, Cisco Systems, Inc. for discovering this bug and
433 preparing a fix. (CVE-2012-0050)
436 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
438 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
439 of the Vaudenay padding oracle attack on CBC mode encryption
440 which enables an efficient plaintext recovery attack against
441 the OpenSSL implementation of DTLS. Their attack exploits timing
442 differences arising during decryption processing. A research
443 paper describing this attack can be found at:
444 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
445 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
446 Security Group at Royal Holloway, University of London
447 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
448 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
449 for preparing the fix. (CVE-2011-4108)
450 [Robin Seggelmann, Michael Tuexen]
452 *) Clear bytes used for block padding of SSL 3.0 records.
454 [Adam Langley (Google)]
456 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
457 Kadianakis <desnacked@gmail.com> for discovering this issue and
458 Adam Langley for preparing the fix. (CVE-2011-4619)
459 [Adam Langley (Google)]
461 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
462 [Andrey Kulikov <amdeich@gmail.com>]
464 *) Prevent malformed RFC3779 data triggering an assertion failure.
465 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
466 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
467 [Rob Austein <sra@hactrn.net>]
469 *) Improved PRNG seeding for VOS.
470 [Paul Green <Paul.Green@stratus.com>]
472 *) Fix ssl_ciph.c set-up race.
473 [Adam Langley (Google)]
475 *) Fix spurious failures in ecdsatest.c.
476 [Emilia Käsper (Google)]
478 *) Fix the BIO_f_buffer() implementation (which was mixing different
479 interpretations of the '..._len' fields).
480 [Adam Langley (Google)]
482 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
483 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
484 threads won't reuse the same blinding coefficients.
486 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
487 lock to call BN_BLINDING_invert_ex, and avoids one use of
488 BN_BLINDING_update for each BN_BLINDING structure (previously,
489 the last update always remained unused).
490 [Emilia Käsper (Google)]
492 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
493 [Bob Buckholz (Google)]
495 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
497 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
498 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
499 [Kaspar Brand <ossl@velox.ch>]
501 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
502 for multi-threaded use of ECDH. (CVE-2011-3210)
503 [Adam Langley (Google)]
505 *) Fix x509_name_ex_d2i memory leak on bad inputs.
508 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
509 signature public key algorithm by using OID xref utilities instead.
510 Before this you could only use some ECC ciphersuites with SHA1 only.
513 *) Add protection against ECDSA timing attacks as mentioned in the paper
514 by Billy Bob Brumley and Nicola Tuveri, see:
516 http://eprint.iacr.org/2011/232.pdf
518 [Billy Bob Brumley and Nicola Tuveri]
520 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
522 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
523 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
525 *) Fix bug in string printing code: if *any* escaping is enabled we must
526 escape the escape character (backslash) or the resulting string is
530 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
532 *) Disable code workaround for ancient and obsolete Netscape browsers
533 and servers: an attacker can use it in a ciphersuite downgrade attack.
534 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
537 *) Fixed J-PAKE implementation error, originally discovered by
538 Sebastien Martini, further info and confirmation from Stefan
539 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
542 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
544 *) Fix extension code to avoid race conditions which can result in a buffer
545 overrun vulnerability: resumed sessions must not be modified as they can
546 be shared by multiple threads. CVE-2010-3864
549 *) Fix WIN32 build system to correctly link an ENGINE directory into
553 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
555 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
557 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
559 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
561 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
562 context. The operation can be customised via the ctrl mechanism in
563 case ENGINEs want to include additional functionality.
566 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
569 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
570 output hashes compatible with older versions of OpenSSL.
571 [Willy Weisz <weisz@vcpc.univie.ac.at>]
573 *) Fix compression algorithm handling: if resuming a session use the
574 compression algorithm of the resumed session instead of determining
575 it from client hello again. Don't allow server to change algorithm.
578 *) Add load_crls() function to apps tidying load_certs() too. Add option
579 to verify utility to allow additional CRLs to be included.
582 *) Update OCSP request code to permit adding custom headers to the request:
583 some responders need this.
586 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
588 [Julia Lawall <julia@diku.dk>]
590 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
591 needlessly dereferenced structures, used obsolete functions and
592 didn't handle all updated verify codes correctly.
595 *) Disable MD2 in the default configuration.
598 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
599 indicate the initial BIO being pushed or popped. This makes it possible
600 to determine whether the BIO is the one explicitly called or as a result
601 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
602 it handles reference counts correctly and doesn't zero out the I/O bio
603 when it is not being explicitly popped. WARNING: applications which
604 included workarounds for the old buggy behaviour will need to be modified
605 or they could free up already freed BIOs.
608 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
609 renaming to all platforms (within the 0.9.8 branch, this was
610 done conditionally on Netware platforms to avoid a name clash).
611 [Guenter <lists@gknw.net>]
613 *) Add ECDHE and PSK support to DTLS.
614 [Michael Tuexen <tuexen@fh-muenster.de>]
616 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
620 *) Add "missing" function EVP_MD_flags() (without this the only way to
621 retrieve a digest flags is by accessing the structure directly. Update
622 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
623 or cipher is registered as in the "from" argument. Print out all
624 registered digests in the dgst usage message instead of manually
625 attempting to work them out.
628 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
629 this allows the use of compression and extensions. Change default cipher
630 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
631 by default unless an application cipher string requests it.
634 *) Alter match criteria in PKCS12_parse(). It used to try to use local
635 key ids to find matching certificates and keys but some PKCS#12 files
636 don't follow the (somewhat unwritten) rules and this strategy fails.
637 Now just gather all certificates together and the first private key
638 then look for the first certificate that matches the key.
641 *) Support use of registered digest and cipher names for dgst and cipher
642 commands instead of having to add each one as a special case. So now
649 openssl dgst -sha256 foo
651 and this works for ENGINE based algorithms too.
655 *) Update Gost ENGINE to support parameter files.
656 [Victor B. Wagner <vitus@cryptocom.ru>]
658 *) Support GeneralizedTime in ca utility.
659 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
661 *) Enhance the hash format used for certificate directory links. The new
662 form uses the canonical encoding (meaning equivalent names will work
663 even if they aren't identical) and uses SHA1 instead of MD5. This form
664 is incompatible with the older format and as a result c_rehash should
665 be used to rebuild symbolic links.
668 *) Make PKCS#8 the default write format for private keys, replacing the
669 traditional format. This form is standardised, more secure and doesn't
670 include an implicit MD5 dependency.
673 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
674 committed to OpenSSL should pass this lot as a minimum.
677 *) Add session ticket override functionality for use by EAP-FAST.
678 [Jouni Malinen <j@w1.fi>]
680 *) Modify HMAC functions to return a value. Since these can be implemented
681 in an ENGINE errors can occur.
684 *) Type-checked OBJ_bsearch_ex.
687 *) Type-checked OBJ_bsearch. Also some constification necessitated
688 by type-checking. Still to come: TXT_DB, bsearch(?),
689 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
693 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
694 seconds to a tm structure directly, instead of going through OS
695 specific date routines. This avoids any issues with OS routines such
696 as the year 2038 bug. New *_adj() functions for ASN1 time structures
697 and X509_time_adj_ex() to cover the extended range. The existing
698 X509_time_adj() is still usable and will no longer have any date issues.
701 *) Delta CRL support. New use deltas option which will attempt to locate
702 and search any appropriate delta CRLs available.
704 This work was sponsored by Google.
707 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
708 code and add additional score elements. Validate alternate CRL paths
709 as part of the CRL checking and indicate a new error "CRL path validation
710 error" in this case. Applications wanting additional details can use
711 the verify callback and check the new "parent" field. If this is not
712 NULL CRL path validation is taking place. Existing applications wont
713 see this because it requires extended CRL support which is off by
716 This work was sponsored by Google.
719 *) Support for freshest CRL extension.
721 This work was sponsored by Google.
724 *) Initial indirect CRL support. Currently only supported in the CRLs
725 passed directly and not via lookup. Process certificate issuer
726 CRL entry extension and lookup CRL entries by bother issuer name
727 and serial number. Check and process CRL issuer entry in IDP extension.
729 This work was sponsored by Google.
732 *) Add support for distinct certificate and CRL paths. The CRL issuer
733 certificate is validated separately in this case. Only enabled if
734 an extended CRL support flag is set: this flag will enable additional
735 CRL functionality in future.
737 This work was sponsored by Google.
740 *) Add support for policy mappings extension.
742 This work was sponsored by Google.
745 *) Fixes to pathlength constraint, self issued certificate handling,
746 policy processing to align with RFC3280 and PKITS tests.
748 This work was sponsored by Google.
751 *) Support for name constraints certificate extension. DN, email, DNS
752 and URI types are currently supported.
754 This work was sponsored by Google.
757 *) To cater for systems that provide a pointer-based thread ID rather
758 than numeric, deprecate the current numeric thread ID mechanism and
759 replace it with a structure and associated callback type. This
760 mechanism allows a numeric "hash" to be extracted from a thread ID in
761 either case, and on platforms where pointers are larger than 'long',
762 mixing is done to help ensure the numeric 'hash' is usable even if it
763 can't be guaranteed unique. The default mechanism is to use "&errno"
764 as a pointer-based thread ID to distinguish between threads.
766 Applications that want to provide their own thread IDs should now use
767 CRYPTO_THREADID_set_callback() to register a callback that will call
768 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
770 Note that ERR_remove_state() is now deprecated, because it is tied
771 to the assumption that thread IDs are numeric. ERR_remove_state(0)
772 to free the current thread's error state should be replaced by
773 ERR_remove_thread_state(NULL).
775 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
776 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
777 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
778 application was previously providing a numeric thread callback that
779 was inappropriate for distinguishing threads, then uniqueness might
780 have been obtained with &errno that happened immediately in the
781 intermediate development versions of OpenSSL; this is no longer the
782 case, the numeric thread callback will now override the automatic use
784 [Geoff Thorpe, with help from Bodo Moeller]
786 *) Initial support for different CRL issuing certificates. This covers a
787 simple case where the self issued certificates in the chain exist and
788 the real CRL issuer is higher in the existing chain.
790 This work was sponsored by Google.
793 *) Removed effectively defunct crypto/store from the build.
796 *) Revamp of STACK to provide stronger type-checking. Still to come:
797 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
798 ASN1_STRING, CONF_VALUE.
801 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
802 RAM on SSL connections. This option can save about 34k per idle SSL.
805 *) Revamp of LHASH to provide stronger type-checking. Still to come:
806 STACK, TXT_DB, bsearch, qsort.
809 *) Initial support for Cryptographic Message Syntax (aka CMS) based
810 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
811 support for data, signedData, compressedData, digestedData and
812 encryptedData, envelopedData types included. Scripts to check against
813 RFC4134 examples draft and interop and consistency checks of many
814 content types and variants.
817 *) Add options to enc utility to support use of zlib compression BIO.
820 *) Extend mk1mf to support importing of options and assembly language
821 files from Configure script, currently only included in VC-WIN32.
822 The assembly language rules can now optionally generate the source
823 files from the associated perl scripts.
826 *) Implement remaining functionality needed to support GOST ciphersuites.
827 Interop testing has been performed using CryptoPro implementations.
828 [Victor B. Wagner <vitus@cryptocom.ru>]
830 *) s390x assembler pack.
833 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
837 *) Implement Opaque PRF Input TLS extension as specified in
838 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
839 official specification yet and no extension type assignment by
840 IANA exists, this extension (for now) will have to be explicitly
841 enabled when building OpenSSL by providing the extension number
842 to use. For example, specify an option
844 -DTLSEXT_TYPE_opaque_prf_input=0x9527
846 to the "config" or "Configure" script to enable the extension,
847 assuming extension number 0x9527 (which is a completely arbitrary
848 and unofficial assignment based on the MD5 hash of the Internet
849 Draft). Note that by doing so, you potentially lose
850 interoperability with other TLS implementations since these might
851 be using the same extension number for other purposes.
853 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
854 opaque PRF input value to use in the handshake. This will create
855 an interal copy of the length-'len' string at 'src', and will
856 return non-zero for success.
858 To get more control and flexibility, provide a callback function
861 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
862 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
866 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
869 Callback function 'cb' will be called in handshakes, and is
870 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
871 Argument 'arg' is for application purposes (the value as given to
872 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
873 be provided to the callback function). The callback function
874 has to return non-zero to report success: usually 1 to use opaque
875 PRF input just if possible, or 2 to enforce use of the opaque PRF
876 input. In the latter case, the library will abort the handshake
877 if opaque PRF input is not successfully negotiated.
879 Arguments 'peerinput' and 'len' given to the callback function
880 will always be NULL and 0 in the case of a client. A server will
881 see the client's opaque PRF input through these variables if
882 available (NULL and 0 otherwise). Note that if the server
883 provides an opaque PRF input, the length must be the same as the
884 length of the client's opaque PRF input.
886 Note that the callback function will only be called when creating
887 a new session (session resumption can resume whatever was
888 previously negotiated), and will not be called in SSL 2.0
889 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
890 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
891 for applications that need to enforce opaque PRF input.
895 *) Update ssl code to support digests other than SHA1+MD5 for handshake
898 [Victor B. Wagner <vitus@cryptocom.ru>]
900 *) Add RFC4507 support to OpenSSL. This includes the corrections in
901 RFC4507bis. The encrypted ticket format is an encrypted encoded
902 SSL_SESSION structure, that way new session features are automatically
905 If a client application caches session in an SSL_SESSION structure
906 support is transparent because tickets are now stored in the encoded
909 The SSL_CTX structure automatically generates keys for ticket
910 protection in servers so again support should be possible
911 with no application modification.
913 If a client or server wishes to disable RFC4507 support then the option
914 SSL_OP_NO_TICKET can be set.
916 Add a TLS extension debugging callback to allow the contents of any client
917 or server extensions to be examined.
919 This work was sponsored by Google.
922 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
923 OpenSSL should now compile cleanly on gcc 4.2
924 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
926 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
927 support including streaming MAC support: this is required for GOST
929 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
931 *) Add option -stream to use PKCS#7 streaming in smime utility. New
932 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
933 to output in BER and PEM format.
936 *) Experimental support for use of HMAC via EVP_PKEY interface. This
937 allows HMAC to be handled via the EVP_DigestSign*() interface. The
938 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
939 ENGINE support for HMAC keys which are unextractable. New -mac and
940 -macopt options to dgst utility.
943 *) New option -sigopt to dgst utility. Update dgst to use
944 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
945 alternative signing paramaters such as X9.31 or PSS in the dgst
949 *) Change ssl_cipher_apply_rule(), the internal function that does
950 the work each time a ciphersuite string requests enabling
951 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
952 removing ("!foo+bar") a class of ciphersuites: Now it maintains
953 the order of disabled ciphersuites such that those ciphersuites
954 that most recently went from enabled to disabled not only stay
955 in order with respect to each other, but also have higher priority
956 than other disabled ciphersuites the next time ciphersuites are
959 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
960 the same ciphersuites as with "HIGH" alone, but in a specific
961 order where the PSK ciphersuites come first (since they are the
962 most recently disabled ciphersuites when "HIGH" is parsed).
964 Also, change ssl_create_cipher_list() (using this new
965 funcionality) such that between otherwise identical
966 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
970 *) Change ssl_create_cipher_list() so that it automatically
971 arranges the ciphersuites in reasonable order before starting
972 to process the rule string. Thus, the definition for "DEFAULT"
973 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
974 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
975 This makes it much easier to arrive at a reasonable default order
976 in applications for which anonymous ciphers are OK (meaning
977 that you can't actually use DEFAULT).
978 [Bodo Moeller; suggested by Victor Duchovni]
980 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
981 processing) into multiple integers instead of setting
982 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
983 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
984 (These masks as well as the individual bit definitions are hidden
985 away into the non-exported interface ssl/ssl_locl.h, so this
986 change to the definition of the SSL_CIPHER structure shouldn't
987 affect applications.) This give us more bits for each of these
988 categories, so there is no longer a need to coagulate AES128 and
989 AES256 into a single algorithm bit, and to coagulate Camellia128
990 and Camellia256 into a single algorithm bit, which has led to all
993 Thus, among other things, the kludge introduced in 0.9.7m and
994 0.9.8e for masking out AES256 independently of AES128 or masking
995 out Camellia256 independently of AES256 is not needed here in 0.9.9.
997 With the change, we also introduce new ciphersuite aliases that
998 so far were missing: "AES128", "AES256", "CAMELLIA128", and
1002 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
1003 Use the leftmost N bytes of the signature input if the input is
1004 larger than the prime q (with N being the size in bytes of q).
1007 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
1008 it yet and it is largely untested.
1011 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
1014 *) Initial incomplete changes to avoid need for function casts in OpenSSL
1015 some compilers (gcc 4.2 and later) reject their use. Safestack is
1016 reimplemented. Update ASN1 to avoid use of legacy functions.
1019 *) Win32/64 targets are linked with Winsock2.
1022 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
1023 to external functions. This can be used to increase CRL handling
1024 efficiency especially when CRLs are very large by (for example) storing
1025 the CRL revoked certificates in a database.
1028 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
1029 new CRLs added to a directory can be used. New command line option
1030 -verify_return_error to s_client and s_server. This causes real errors
1031 to be returned by the verify callback instead of carrying on no matter
1032 what. This reflects the way a "real world" verify callback would behave.
1035 *) GOST engine, supporting several GOST algorithms and public key formats.
1036 Kindly donated by Cryptocom.
1039 *) Partial support for Issuing Distribution Point CRL extension. CRLs
1040 partitioned by DP are handled but no indirect CRL or reason partitioning
1041 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
1042 selected via a scoring technique which handles IDP and AKID in CRLs.
1045 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
1046 will ultimately be used for all verify operations: this will remove the
1047 X509_STORE dependency on certificate verification and allow alternative
1048 lookup methods. X509_STORE based implementations of these two callbacks.
1051 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
1052 Modify get_crl() to find a valid (unexpired) CRL if possible.
1055 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
1056 this would be called X509_CRL_cmp() but that name is already used by
1057 a function that just compares CRL issuer names. Cache several CRL
1058 extensions in X509_CRL structure and cache CRLDP in X509.
1061 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
1062 this maps equivalent X509_NAME structures into a consistent structure.
1063 Name comparison can then be performed rapidly using memcmp().
1066 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
1070 *) Allow digests to supply their own micalg string for S/MIME type using
1071 the ctrl EVP_MD_CTRL_MICALG.
1074 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
1075 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
1076 ctrl. It can then customise the structure before and/or after signing
1080 *) New function OBJ_add_sigid() to allow application defined signature OIDs
1081 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
1082 to free up any added signature OIDs.
1085 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
1086 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
1087 digest and cipher tables. New options added to openssl utility:
1088 list-message-digest-algorithms and list-cipher-algorithms.
1091 *) Change the array representation of binary polynomials: the list
1092 of degrees of non-zero coefficients is now terminated with -1.
1093 Previously it was terminated with 0, which was also part of the
1094 value; thus, the array representation was not applicable to
1095 polynomials where t^0 has coefficient zero. This change makes
1096 the array representation useful in a more general context.
1099 *) Various modifications and fixes to SSL/TLS cipher string
1100 handling. For ECC, the code now distinguishes between fixed ECDH
1101 with RSA certificates on the one hand and with ECDSA certificates
1102 on the other hand, since these are separate ciphersuites. The
1103 unused code for Fortezza ciphersuites has been removed.
1105 For consistency with EDH, ephemeral ECDH is now called "EECDH"
1106 (not "ECDHE"). For consistency with the code for DH
1107 certificates, use of ECDH certificates is now considered ECDH
1108 authentication, not RSA or ECDSA authentication (the latter is
1109 merely the CA's signing algorithm and not actively used in the
1112 The temporary ciphersuite alias "ECCdraft" is no longer
1113 available, and ECC ciphersuites are no longer excluded from "ALL"
1114 and "DEFAULT". The following aliases now exist for RFC 4492
1115 ciphersuites, most of these by analogy with the DH case:
1117 kECDHr - ECDH cert, signed with RSA
1118 kECDHe - ECDH cert, signed with ECDSA
1119 kECDH - ECDH cert (signed with either RSA or ECDSA)
1120 kEECDH - ephemeral ECDH
1121 ECDH - ECDH cert or ephemeral ECDH
1127 AECDH - anonymous ECDH
1128 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
1132 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
1133 Use correct micalg parameters depending on digest(s) in signed message.
1136 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
1137 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
1140 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
1141 an engine to register a method. Add ENGINE lookups for methods and
1142 functional reference processing.
1145 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
1146 EVP_{Sign,Verify}* which allow an application to customise the signature
1150 *) New -resign option to smime utility. This adds one or more signers
1151 to an existing PKCS#7 signedData structure. Also -md option to use an
1152 alternative message digest algorithm for signing.
1155 *) Tidy up PKCS#7 routines and add new functions to make it easier to
1156 create PKCS7 structures containing multiple signers. Update smime
1157 application to support multiple signers.
1160 *) New -macalg option to pkcs12 utility to allow setting of an alternative
1164 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
1165 Reorganize PBE internals to lookup from a static table using NIDs,
1166 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
1167 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
1168 PRF which will be automatically used with PBES2.
1171 *) Replace the algorithm specific calls to generate keys in "req" with the
1175 *) Update PKCS#7 enveloped data routines to use new API. This is now
1176 supported by any public key method supporting the encrypt operation. A
1177 ctrl is added to allow the public key algorithm to examine or modify
1178 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
1182 *) Add a ctrl to asn1 method to allow a public key algorithm to express
1183 a default digest type to use. In most cases this will be SHA1 but some
1184 algorithms (such as GOST) need to specify an alternative digest. The
1185 return value indicates how strong the prefernce is 1 means optional and
1186 2 is mandatory (that is it is the only supported type). Modify
1187 ASN1_item_sign() to accept a NULL digest argument to indicate it should
1188 use the default md. Update openssl utilities to use the default digest
1189 type for signing if it is not explicitly indicated.
1192 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
1193 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
1194 signing method from the key type. This effectively removes the link
1195 between digests and public key types.
1198 *) Add an OID cross reference table and utility functions. Its purpose is to
1199 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
1200 rsaEncryption. This will allow some of the algorithm specific hackery
1201 needed to use the correct OID to be removed.
1204 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
1205 structures for PKCS7_sign(). They are now set up by the relevant public
1209 *) Add provisional EC pkey method with support for ECDSA and ECDH.
1212 *) Add support for key derivation (agreement) in the API, DH method and
1216 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
1217 public and private key formats. As a side effect these add additional
1218 command line functionality not previously available: DSA signatures can be
1219 generated and verified using pkeyutl and DH key support and generation in
1224 [Oliver Tappe <zooey@hirschkaefer.de>]
1226 *) New make target "install_html_docs" installs HTML renditions of the
1228 [Oliver Tappe <zooey@hirschkaefer.de>]
1230 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
1231 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
1232 support key and parameter generation and add initial key generation
1233 functionality for RSA.
1236 *) Add functions for main EVP_PKEY_method operations. The undocumented
1237 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
1238 EVP_PKEY_{encrypt,decrypt}_old.
1241 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
1242 key API, doesn't do much yet.
1245 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
1246 public key algorithms. New option to openssl utility:
1247 "list-public-key-algorithms" to print out info.
1250 *) Implement the Supported Elliptic Curves Extension for
1251 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1254 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
1255 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
1258 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
1259 utilities such as rsa, dsa, dsaparam etc except they process any key
1263 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1264 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1265 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
1269 *) Initial support for pluggable public key ASN1.
1270 De-spaghettify the public key ASN1 handling. Move public and private
1271 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
1272 algorithm specific handling to a single module within the relevant
1273 algorithm directory. Add functions to allow (near) opaque processing
1274 of public and private key structures.
1277 *) Implement the Supported Point Formats Extension for
1278 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1281 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
1282 for the psk identity [hint] and the psk callback functions to the
1283 SSL_SESSION, SSL and SSL_CTX structure.
1286 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
1290 SSL_CTX_use_psk_identity_hint
1291 SSL_get_psk_identity_hint
1292 SSL_get_psk_identity
1293 SSL_use_psk_identity_hint
1295 [Mika Kousa and Pasi Eronen of Nokia Corporation]
1297 *) Add RFC 3161 compliant time stamp request creation, response generation
1298 and response verification functionality.
1299 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
1301 *) Add initial support for TLS extensions, specifically for the server_name
1302 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1303 have new members for a host name. The SSL data structure has an
1304 additional member SSL_CTX *initial_ctx so that new sessions can be
1305 stored in that context to allow for session resumption, even after the
1306 SSL has been switched to a new SSL_CTX in reaction to a client's
1307 server_name extension.
1309 New functions (subject to change):
1311 SSL_get_servername()
1312 SSL_get_servername_type()
1315 New CTRL codes and macros (subject to change):
1317 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1318 - SSL_CTX_set_tlsext_servername_callback()
1319 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1320 - SSL_CTX_set_tlsext_servername_arg()
1321 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1323 openssl s_client has a new '-servername ...' option.
1325 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1326 '-key2 ...', '-servername_fatal' (subject to change). This allows
1327 testing the HostName extension for a specific single host name ('-cert'
1328 and '-key' remain fallbacks for handshakes without HostName
1329 negotiation). If the unrecogninzed_name alert has to be sent, this by
1330 default is a warning; it becomes fatal with the '-servername_fatal'
1333 [Peter Sylvester, Remy Allais, Christophe Renou]
1335 *) Whirlpool hash implementation is added.
1338 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1339 bn(64,32). Because of instruction set limitations it doesn't have
1340 any negative impact on performance. This was done mostly in order
1341 to make it possible to share assembler modules, such as bn_mul_mont
1342 implementations, between 32- and 64-bit builds without hassle.
1345 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1346 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1350 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1351 dedicated Montgomery multiplication procedure, is introduced.
1352 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1353 "64-bit" performance on certain 32-bit targets.
1356 *) New option SSL_OP_NO_COMP to disable use of compression selectively
1357 in SSL structures. New SSL ctrl to set maximum send fragment size.
1358 Save memory by seeting the I/O buffer sizes dynamically instead of
1359 using the maximum available value.
1362 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
1363 in addition to the text details.
1366 *) Very, very preliminary EXPERIMENTAL support for printing of general
1367 ASN1 structures. This currently produces rather ugly output and doesn't
1368 handle several customised structures at all.
1371 *) Integrated support for PVK file format and some related formats such
1372 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
1373 these in the 'rsa' and 'dsa' utilities.
1376 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
1379 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
1380 place for the (very old) "NETSCAPE" format certificates which are now
1381 handled using new ASN1 code equivalents.
1384 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
1385 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
1386 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
1389 *) Modify CRL distribution points extension code to print out previously
1390 unsupported fields. Enhance extension setting code to allow setting of
1394 *) Add print and set support for Issuing Distribution Point CRL extension.
1397 *) Change 'Configure' script to enable Camellia by default.
1400 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
1402 *) When rejecting SSL/TLS records due to an incorrect version number, never
1403 update s->server with a new major version number. As of
1404 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
1405 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
1406 the previous behavior could result in a read attempt at NULL when
1407 receiving specific incorrect SSL/TLS records once record payload
1408 protection is active. (CVE-2010-0740)
1409 [Bodo Moeller, Adam Langley <agl@chromium.org>]
1411 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
1412 could be crashed if the relevant tables were not present (e.g. chrooted).
1413 [Tomas Hoger <thoger@redhat.com>]
1415 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
1417 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
1418 [Martin Olsson, Neel Mehta]
1420 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
1421 accommodate for stack sorting, always a write lock!).
1424 *) On some versions of WIN32 Heap32Next is very slow. This can cause
1425 excessive delays in the RAND_poll(): over a minute. As a workaround
1426 include a time check in the inner Heap32Next loop too.
1429 *) The code that handled flushing of data in SSL/TLS originally used the
1430 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
1431 the problem outlined in PR#1949. The fix suggested there however can
1432 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
1433 of Apache). So instead simplify the code to flush unconditionally.
1434 This should be fine since flushing with no data to flush is a no op.
1437 *) Handle TLS versions 2.0 and later properly and correctly use the
1438 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
1439 off ancient servers have a habit of sticking around for a while...
1442 *) Modify compression code so it frees up structures without using the
1443 ex_data callbacks. This works around a problem where some applications
1444 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
1445 restarting) then use compression (e.g. SSL with compression) later.
1446 This results in significant per-connection memory leaks and
1447 has caused some security issues including CVE-2008-1678 and
1451 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
1452 change when encrypting or decrypting.
1455 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
1456 connect and renegotiate with servers which do not support RI.
1457 Until RI is more widely deployed this option is enabled by default.
1460 *) Add "missing" ssl ctrls to clear options and mode.
1463 *) If client attempts to renegotiate and doesn't support RI respond with
1464 a no_renegotiation alert as required by RFC5746. Some renegotiating
1465 TLS clients will continue a connection gracefully when they receive
1466 the alert. Unfortunately OpenSSL mishandled this alert and would hang
1467 waiting for a server hello which it will never receive. Now we treat a
1468 received no_renegotiation alert as a fatal error. This is because
1469 applications requesting a renegotiation might well expect it to succeed
1470 and would have no code in place to handle the server denying it so the
1471 only safe thing to do is to terminate the connection.
1474 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
1475 peer supports secure renegotiation and 0 otherwise. Print out peer
1476 renegotiation support in s_client/s_server.
1479 *) Replace the highly broken and deprecated SPKAC certification method with
1480 the updated NID creation version. This should correctly handle UTF8.
1483 *) Implement RFC5746. Re-enable renegotiation but require the extension
1484 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
1485 turns out to be a bad idea. It has been replaced by
1486 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
1487 SSL_CTX_set_options(). This is really not recommended unless you
1488 know what you are doing.
1489 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
1491 *) Fixes to stateless session resumption handling. Use initial_ctx when
1492 issuing and attempting to decrypt tickets in case it has changed during
1493 servername handling. Use a non-zero length session ID when attempting
1494 stateless session resumption: this makes it possible to determine if
1495 a resumption has occurred immediately after receiving server hello
1496 (several places in OpenSSL subtly assume this) instead of later in
1500 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
1501 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
1502 fixes for a few places where the return code is not checked
1504 [Julia Lawall <julia@diku.dk>]
1506 *) Add --strict-warnings option to Configure script to include devteam
1507 warnings in other configurations.
1510 *) Add support for --libdir option and LIBDIR variable in makefiles. This
1511 makes it possible to install openssl libraries in locations which
1512 have names other than "lib", for example "/usr/lib64" which some
1514 [Steve Henson, based on patch from Jeremy Utley]
1516 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
1517 X690 8.9.12 and can produce some misleading textual output of OIDs.
1518 [Steve Henson, reported by Dan Kaminsky]
1520 *) Delete MD2 from algorithm tables. This follows the recommendation in
1521 several standards that it is not used in new applications due to
1522 several cryptographic weaknesses. For binary compatibility reasons
1523 the MD2 API is still compiled in by default.
1526 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
1530 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
1531 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
1533 [Guenter <lists@gknw.net>]
1535 *) Fix the server certificate chain building code to use X509_verify_cert(),
1536 it used to have an ad-hoc builder which was unable to cope with anything
1537 other than a simple chain.
1538 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
1540 *) Don't check self signed certificate signatures in X509_verify_cert()
1541 by default (a flag can override this): it just wastes time without
1542 adding any security. As a useful side effect self signed root CAs
1543 with non-FIPS digests are now usable in FIPS mode.
1546 *) In dtls1_process_out_of_seq_message() the check if the current message
1547 is already buffered was missing. For every new message was memory
1548 allocated, allowing an attacker to perform an denial of service attack
1549 with sending out of seq handshake messages until there is no memory
1550 left. Additionally every future messege was buffered, even if the
1551 sequence number made no sense and would be part of another handshake.
1552 So only messages with sequence numbers less than 10 in advance will be
1553 buffered. (CVE-2009-1378)
1554 [Robin Seggelmann, discovered by Daniel Mentz]
1556 *) Records are buffered if they arrive with a future epoch to be
1557 processed after finishing the corresponding handshake. There is
1558 currently no limitation to this buffer allowing an attacker to perform
1559 a DOS attack with sending records with future epochs until there is no
1560 memory left. This patch adds the pqueue_size() function to detemine
1561 the size of a buffer and limits the record buffer to 100 entries.
1563 [Robin Seggelmann, discovered by Daniel Mentz]
1565 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
1566 parent structure is freed. (CVE-2009-1379)
1569 *) Handle non-blocking I/O properly in SSL_shutdown() call.
1570 [Darryl Miles <darryl-mailinglists@netbauds.net>]
1573 [Ilya O. <vrghost@gmail.com>]
1575 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
1577 *) Disable renegotiation completely - this fixes a severe security
1578 problem (CVE-2009-3555) at the cost of breaking all
1579 renegotiation. Renegotiation can be re-enabled by setting
1580 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
1581 run-time. This is really not recommended unless you know what
1585 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
1587 *) Don't set val to NULL when freeing up structures, it is freed up by
1588 underlying code. If sizeof(void *) > sizeof(long) this can result in
1589 zeroing past the valid field. (CVE-2009-0789)
1590 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
1592 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
1593 checked correctly. This would allow some invalid signed attributes to
1594 appear to verify correctly. (CVE-2009-0591)
1595 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1597 *) Reject UniversalString and BMPString types with invalid lengths. This
1598 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
1599 a legal length. (CVE-2009-0590)
1602 *) Set S/MIME signing as the default purpose rather than setting it
1603 unconditionally. This allows applications to override it at the store
1607 *) Permit restricted recursion of ASN1 strings. This is needed in practice
1608 to handle some structures.
1611 *) Improve efficiency of mem_gets: don't search whole buffer each time
1613 [Jeremy Shapiro <jnshapir@us.ibm.com>]
1615 *) New -hex option for openssl rand.
1618 *) Print out UTF8String and NumericString when parsing ASN1.
1621 *) Support NumericString type for name components.
1624 *) Allow CC in the environment to override the automatically chosen
1625 compiler. Note that nothing is done to ensure flags work with the
1629 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
1631 *) Properly check EVP_VerifyFinal() and similar return values
1633 [Ben Laurie, Bodo Moeller, Google Security Team]
1635 *) Enable TLS extensions by default.
1638 *) Allow the CHIL engine to be loaded, whether the application is
1639 multithreaded or not. (This does not release the developer from the
1640 obligation to set up the dynamic locking callbacks.)
1641 [Sander Temme <sander@temme.net>]
1643 *) Use correct exit code if there is an error in dgst command.
1644 [Steve Henson; problem pointed out by Roland Dirlewanger]
1646 *) Tweak Configure so that you need to say "experimental-jpake" to enable
1647 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
1650 *) Add experimental JPAKE support, including demo authentication in
1651 s_client and s_server.
1654 *) Set the comparison function in v3_addr_canonize().
1655 [Rob Austein <sra@hactrn.net>]
1657 *) Add support for XMPP STARTTLS in s_client.
1658 [Philip Paeps <philip@freebsd.org>]
1660 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
1661 to ensure that even with this option, only ciphersuites in the
1662 server's preference list will be accepted. (Note that the option
1663 applies only when resuming a session, so the earlier behavior was
1664 just about the algorithm choice for symmetric cryptography.)
1667 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
1669 *) Fix NULL pointer dereference if a DTLS server received
1670 ChangeCipherSpec as first record (CVE-2009-1386).
1673 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
1674 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
1677 *) The fix in 0.9.8c that supposedly got rid of unsafe
1678 double-checked locking was incomplete for RSA blinding,
1679 addressing just one layer of what turns out to have been
1680 doubly unsafe triple-checked locking.
1682 So now fix this for real by retiring the MONT_HELPER macro
1683 in crypto/rsa/rsa_eay.c.
1685 [Bodo Moeller; problem pointed out by Marius Schilder]
1687 *) Various precautionary measures:
1689 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
1691 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
1692 (NB: This would require knowledge of the secret session ticket key
1693 to exploit, in which case you'd be SOL either way.)
1695 - Change bn_nist.c so that it will properly handle input BIGNUMs
1696 outside the expected range.
1698 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
1701 [Neel Mehta, Bodo Moeller]
1703 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
1704 the load fails. Useful for distros.
1705 [Ben Laurie and the FreeBSD team]
1707 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
1710 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
1713 *) Expand ENGINE to support engine supplied SSL client certificate functions.
1715 This work was sponsored by Logica.
1718 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
1719 keystores. Support for SSL/TLS client authentication too.
1720 Not compiled unless enable-capieng specified to Configure.
1722 This work was sponsored by Logica.
1725 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
1726 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
1727 attribute creation routines such as certifcate requests and PKCS#12
1731 Changes between 0.9.8g and 0.9.8h [28 May 2008]
1733 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
1734 handshake which could lead to a cilent crash as found using the
1735 Codenomicon TLS test suite (CVE-2008-1672)
1736 [Steve Henson, Mark Cox]
1738 *) Fix double free in TLS server name extensions which could lead to
1739 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
1742 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
1744 Clear the error queue to ensure that error entries left from
1745 older function calls do not interfere with the correct operation.
1746 [Lutz Jaenicke, Erik de Castro Lopo]
1748 *) Remove root CA certificates of commercial CAs:
1750 The OpenSSL project does not recommend any specific CA and does not
1751 have any policy with respect to including or excluding any CA.
1752 Therefore it does not make any sense to ship an arbitrary selection
1753 of root CA certificates with the OpenSSL software.
1756 *) RSA OAEP patches to fix two separate invalid memory reads.
1757 The first one involves inputs when 'lzero' is greater than
1758 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
1759 before the beginning of from). The second one involves inputs where
1760 the 'db' section contains nothing but zeroes (there is a one-byte
1761 invalid read after the end of 'db').
1762 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1764 *) Partial backport from 0.9.9-dev:
1766 Introduce bn_mul_mont (dedicated Montgomery multiplication
1767 procedure) as a candidate for BIGNUM assembler implementation.
1768 While 0.9.9-dev uses assembler for various architectures, only
1769 x86_64 is available by default here in the 0.9.8 branch, and
1770 32-bit x86 is available through a compile-time setting.
1772 To try the 32-bit x86 assembler implementation, use Configure
1773 option "enable-montasm" (which exists only for this backport).
1775 As "enable-montasm" for 32-bit x86 disclaims code stability
1776 anyway, in this constellation we activate additional code
1777 backported from 0.9.9-dev for further performance improvements,
1778 namely BN_from_montgomery_word. (To enable this otherwise,
1779 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
1781 [Andy Polyakov (backport partially by Bodo Moeller)]
1783 *) Add TLS session ticket callback. This allows an application to set
1784 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
1785 values. This is useful for key rollover for example where several key
1786 sets may exist with different names.
1789 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
1790 This was broken until now in 0.9.8 releases, such that the only way
1791 a registered ENGINE could be used (assuming it initialises
1792 successfully on the host) was to explicitly set it as the default
1793 for the relevant algorithms. This is in contradiction with 0.9.7
1794 behaviour and the documentation. With this fix, when an ENGINE is
1795 registered into a given algorithm's table of implementations, the
1796 'uptodate' flag is reset so that auto-discovery will be used next
1797 time a new context for that algorithm attempts to select an
1799 [Ian Lister (tweaked by Geoff Thorpe)]
1801 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
1802 implemention in the following ways:
1804 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
1807 Lack of BER streaming support means one pass streaming processing is
1808 only supported if data is detached: setting the streaming flag is
1809 ignored for embedded content.
1811 CMS support is disabled by default and must be explicitly enabled
1812 with the enable-cms configuration option.
1815 *) Update the GMP engine glue to do direct copies between BIGNUM and
1816 mpz_t when openssl and GMP use the same limb size. Otherwise the
1817 existing "conversion via a text string export" trick is still used.
1818 [Paul Sheer <paulsheer@gmail.com>]
1820 *) Zlib compression BIO. This is a filter BIO which compressed and
1821 uncompresses any data passed through it.
1824 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
1825 RFC3394 compatible AES key wrapping.
1828 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
1829 sets string data without copying. X509_ALGOR_set0() and
1830 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
1831 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
1832 from an X509_ATTRIBUTE structure optionally checking it occurs only
1833 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
1837 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
1838 to get the expected BN_FLG_CONSTTIME behavior.
1839 [Bodo Moeller (Google)]
1843 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
1844 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
1845 - added some more tests to do_tests.pl
1846 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
1847 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
1848 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
1849 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
1850 - various changes to netware.pl to enable gcc-cross builds on Win32
1852 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
1853 - various changes to fix missing prototype warnings
1854 - fixed x86nasm.pl to create correct asm files for NASM COFF output
1855 - added AES, WHIRLPOOL and CPUID assembler code to build files
1856 - added missing AES assembler make rules to mk1mf.pl
1857 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
1858 [Guenter Knauf <eflash@gmx.net>]
1860 *) Implement certificate status request TLS extension defined in RFC3546.
1861 A client can set the appropriate parameters and receive the encoded
1862 OCSP response via a callback. A server can query the supplied parameters
1863 and set the encoded OCSP response in the callback. Add simplified examples
1864 to s_client and s_server.
1867 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
1869 *) Fix various bugs:
1870 + Binary incompatibility of ssl_ctx_st structure
1871 + DTLS interoperation with non-compliant servers
1872 + Don't call get_session_cb() without proposed session
1873 + Fix ia64 assembler code
1874 [Andy Polyakov, Steve Henson]
1876 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
1878 *) DTLS Handshake overhaul. There were longstanding issues with
1879 OpenSSL DTLS implementation, which were making it impossible for
1880 RFC 4347 compliant client to communicate with OpenSSL server.
1881 Unfortunately just fixing these incompatibilities would "cut off"
1882 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
1883 server keeps tolerating non RFC compliant syntax. The opposite is
1884 not true, 0.9.8f client can not communicate with earlier server.
1885 This update even addresses CVE-2007-4995.
1888 *) Changes to avoid need for function casts in OpenSSL: some compilers
1889 (gcc 4.2 and later) reject their use.
1890 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
1893 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1894 RFC4507bis. The encrypted ticket format is an encrypted encoded
1895 SSL_SESSION structure, that way new session features are automatically
1898 If a client application caches session in an SSL_SESSION structure
1899 support is transparent because tickets are now stored in the encoded
1902 The SSL_CTX structure automatically generates keys for ticket
1903 protection in servers so again support should be possible
1904 with no application modification.
1906 If a client or server wishes to disable RFC4507 support then the option
1907 SSL_OP_NO_TICKET can be set.
1909 Add a TLS extension debugging callback to allow the contents of any client
1910 or server extensions to be examined.
1912 This work was sponsored by Google.
1915 *) Add initial support for TLS extensions, specifically for the server_name
1916 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1917 have new members for a host name. The SSL data structure has an
1918 additional member SSL_CTX *initial_ctx so that new sessions can be
1919 stored in that context to allow for session resumption, even after the
1920 SSL has been switched to a new SSL_CTX in reaction to a client's
1921 server_name extension.
1923 New functions (subject to change):
1925 SSL_get_servername()
1926 SSL_get_servername_type()
1929 New CTRL codes and macros (subject to change):
1931 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1932 - SSL_CTX_set_tlsext_servername_callback()
1933 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1934 - SSL_CTX_set_tlsext_servername_arg()
1935 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1937 openssl s_client has a new '-servername ...' option.
1939 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1940 '-key2 ...', '-servername_fatal' (subject to change). This allows
1941 testing the HostName extension for a specific single host name ('-cert'
1942 and '-key' remain fallbacks for handshakes without HostName
1943 negotiation). If the unrecogninzed_name alert has to be sent, this by
1944 default is a warning; it becomes fatal with the '-servername_fatal'
1947 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
1949 *) Add AES and SSE2 assembly language support to VC++ build.
1952 *) Mitigate attack on final subtraction in Montgomery reduction.
1955 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
1956 (which previously caused an internal error).
1959 *) Squeeze another 10% out of IGE mode when in != out.
1962 *) AES IGE mode speedup.
1963 [Dean Gaudet (Google)]
1965 *) Add the Korean symmetric 128-bit cipher SEED (see
1966 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
1967 add SEED ciphersuites from RFC 4162:
1969 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
1970 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
1971 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
1972 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
1974 To minimize changes between patchlevels in the OpenSSL 0.9.8
1975 series, SEED remains excluded from compilation unless OpenSSL
1976 is configured with 'enable-seed'.
1977 [KISA, Bodo Moeller]
1979 *) Mitigate branch prediction attacks, which can be practical if a
1980 single processor is shared, allowing a spy process to extract
1981 information. For detailed background information, see
1982 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
1983 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
1984 and Necessary Software Countermeasures"). The core of the change
1985 are new versions BN_div_no_branch() and
1986 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
1987 respectively, which are slower, but avoid the security-relevant
1988 conditional branches. These are automatically called by BN_div()
1989 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
1990 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
1991 remove a conditional branch.
1993 BN_FLG_CONSTTIME is the new name for the previous
1994 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
1995 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
1996 in the exponent causes BN_mod_exp_mont() to use the alternative
1997 implementation in BN_mod_exp_mont_consttime().) The old name
1998 remains as a deprecated alias.
2000 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
2001 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
2002 constant-time implementations for more than just exponentiation.
2003 Here too the old name is kept as a deprecated alias.
2005 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2006 the BN_BLINDING structure gets an independent copy of the
2007 modulus. This means that the previous "BIGNUM *m" argument to
2008 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2009 essentially becomes "const BIGNUM *m", although we can't actually
2010 change this in the header file before 0.9.9. It allows
2011 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2012 enable BN_FLG_CONSTTIME.
2014 [Matthew D Wood (Intel Corp)]
2016 *) In the SSL/TLS server implementation, be strict about session ID
2017 context matching (which matters if an application uses a single
2018 external cache for different purposes). Previously,
2019 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2020 set. This did ensure strict client verification, but meant that,
2021 with applications using a single external cache for quite
2022 different requirements, clients could circumvent ciphersuite
2023 restrictions for a given session ID context by starting a session
2024 in a different context.
2027 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2028 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2029 authentication-only ciphersuites.
2032 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
2033 not complete and could lead to a possible single byte overflow
2034 (CVE-2007-5135) [Ben Laurie]
2036 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
2038 *) Since AES128 and AES256 (and similarly Camellia128 and
2039 Camellia256) share a single mask bit in the logic of
2040 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2041 kludge to work properly if AES128 is available and AES256 isn't
2042 (or if Camellia128 is available and Camellia256 isn't).
2045 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
2046 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
2047 When a point or a seed is encoded in a BIT STRING, we need to
2048 prevent the removal of trailing zero bits to get the proper DER
2049 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
2050 of a NamedBitList, for which trailing 0 bits need to be removed.)
2053 *) Have SSL/TLS server implementation tolerate "mismatched" record
2054 protocol version while receiving ClientHello even if the
2055 ClientHello is fragmented. (The server can't insist on the
2056 particular protocol version it has chosen before the ServerHello
2057 message has informed the client about his choice.)
2060 *) Add RFC 3779 support.
2061 [Rob Austein for ARIN, Ben Laurie]
2063 *) Load error codes if they are not already present instead of using a
2064 static variable. This allows them to be cleanly unloaded and reloaded.
2065 Improve header file function name parsing.
2068 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
2069 or CAPABILITY handshake as required by RFCs.
2072 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
2074 *) Introduce limits to prevent malicious keys being able to
2075 cause a denial of service. (CVE-2006-2940)
2076 [Steve Henson, Bodo Moeller]
2078 *) Fix ASN.1 parsing of certain invalid structures that can result
2079 in a denial of service. (CVE-2006-2937) [Steve Henson]
2081 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2082 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2084 *) Fix SSL client code which could crash if connecting to a
2085 malicious SSLv2 server. (CVE-2006-4343)
2086 [Tavis Ormandy and Will Drewry, Google Security Team]
2088 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
2089 match only those. Before that, "AES256-SHA" would be interpreted
2090 as a pattern and match "AES128-SHA" too (since AES128-SHA got
2091 the same strength classification in 0.9.7h) as we currently only
2092 have a single AES bit in the ciphersuite description bitmap.
2093 That change, however, also applied to ciphersuite strings such as
2094 "RC4-MD5" that intentionally matched multiple ciphersuites --
2095 namely, SSL 2.0 ciphersuites in addition to the more common ones
2096 from SSL 3.0/TLS 1.0.
2098 So we change the selection algorithm again: Naming an explicit
2099 ciphersuite selects this one ciphersuite, and any other similar
2100 ciphersuite (same bitmap) from *other* protocol versions.
2101 Thus, "RC4-MD5" again will properly select both the SSL 2.0
2102 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
2104 Since SSL 2.0 does not have any ciphersuites for which the
2105 128/256 bit distinction would be relevant, this works for now.
2106 The proper fix will be to use different bits for AES128 and
2107 AES256, which would have avoided the problems from the beginning;
2108 however, bits are scarce, so we can only do this in a new release
2109 (not just a patchlevel) when we can change the SSL_CIPHER
2110 definition to split the single 'unsigned long mask' bitmap into
2111 multiple values to extend the available space.
2115 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2117 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2118 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2120 *) Add AES IGE and biIGE modes.
2123 *) Change the Unix randomness entropy gathering to use poll() when
2124 possible instead of select(), since the latter has some
2125 undesirable limitations.
2126 [Darryl Miles via Richard Levitte and Bodo Moeller]
2128 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2129 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2130 cannot be implicitly activated as part of, e.g., the "AES" alias.
2131 However, please upgrade to OpenSSL 0.9.9[-dev] for
2132 non-experimental use of the ECC ciphersuites to get TLS extension
2133 support, which is required for curve and point format negotiation
2134 to avoid potential handshake problems.
2137 *) Disable rogue ciphersuites:
2139 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2140 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2141 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2143 The latter two were purportedly from
2144 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2147 Also deactivate the remaining ciphersuites from
2148 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2149 unofficial, and the ID has long expired.
2152 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2153 dual-core machines) and other potential thread-safety issues.
2156 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
2157 versions), which is now available for royalty-free use
2158 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
2159 Also, add Camellia TLS ciphersuites from RFC 4132.
2161 To minimize changes between patchlevels in the OpenSSL 0.9.8
2162 series, Camellia remains excluded from compilation unless OpenSSL
2163 is configured with 'enable-camellia'.
2166 *) Disable the padding bug check when compression is in use. The padding
2167 bug check assumes the first packet is of even length, this is not
2168 necessarily true if compresssion is enabled and can result in false
2169 positives causing handshake failure. The actual bug test is ancient
2170 code so it is hoped that implementations will either have fixed it by
2171 now or any which still have the bug do not support compression.
2174 Changes between 0.9.8a and 0.9.8b [04 May 2006]
2176 *) When applying a cipher rule check to see if string match is an explicit
2177 cipher suite and only match that one cipher suite if it is.
2180 *) Link in manifests for VC++ if needed.
2181 [Austin Ziegler <halostatue@gmail.com>]
2183 *) Update support for ECC-based TLS ciphersuites according to
2184 draft-ietf-tls-ecc-12.txt with proposed changes (but without
2185 TLS extensions, which are supported starting with the 0.9.9
2186 branch, not in the OpenSSL 0.9.8 branch).
2189 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
2190 opaque EVP_CIPHER_CTX handling.
2193 *) Fixes and enhancements to zlib compression code. We now only use
2194 "zlib1.dll" and use the default __cdecl calling convention on Win32
2195 to conform with the standards mentioned here:
2196 http://www.zlib.net/DLL_FAQ.txt
2197 Static zlib linking now works on Windows and the new --with-zlib-include
2198 --with-zlib-lib options to Configure can be used to supply the location
2199 of the headers and library. Gracefully handle case where zlib library
2203 *) Several fixes and enhancements to the OID generation code. The old code
2204 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
2205 handle numbers larger than ULONG_MAX, truncated printing and had a
2206 non standard OBJ_obj2txt() behaviour.
2209 *) Add support for building of engines under engine/ as shared libraries
2210 under VC++ build system.
2213 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
2214 Hopefully, we will not see any false combination of paths any more.
2217 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
2219 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2220 (part of SSL_OP_ALL). This option used to disable the
2221 countermeasure against man-in-the-middle protocol-version
2222 rollback in the SSL 2.0 server implementation, which is a bad
2223 idea. (CVE-2005-2969)
2225 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2226 for Information Security, National Institute of Advanced Industrial
2227 Science and Technology [AIST], Japan)]
2229 *) Add two function to clear and return the verify parameter flags.
2232 *) Keep cipherlists sorted in the source instead of sorting them at
2233 runtime, thus removing the need for a lock.
2236 *) Avoid some small subgroup attacks in Diffie-Hellman.
2237 [Nick Mathewson and Ben Laurie]
2239 *) Add functions for well-known primes.
2242 *) Extended Windows CE support.
2243 [Satoshi Nakamura and Andy Polyakov]
2245 *) Initialize SSL_METHOD structures at compile time instead of during
2246 runtime, thus removing the need for a lock.
2249 *) Make PKCS7_decrypt() work even if no certificate is supplied by
2250 attempting to decrypt each encrypted key in turn. Add support to
2254 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
2256 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
2259 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
2262 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
2263 key into the same file any more.
2266 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
2269 *) Add -utf8 command line and config file option to 'ca'.
2270 [Stefan <stf@udoma.org]
2272 *) Removed the macro des_crypt(), as it seems to conflict with some
2273 libraries. Use DES_crypt().
2276 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
2277 involves renaming the source and generated shared-libs for
2278 both. The engines will accept the corrected or legacy ids
2279 ('ncipher' and '4758_cca' respectively) when binding. NB,
2280 this only applies when building 'shared'.
2281 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
2283 *) Add attribute functions to EVP_PKEY structure. Modify
2284 PKCS12_create() to recognize a CSP name attribute and
2285 use it. Make -CSP option work again in pkcs12 utility.
2288 *) Add new functionality to the bn blinding code:
2289 - automatic re-creation of the BN_BLINDING parameters after
2290 a fixed number of uses (currently 32)
2291 - add new function for parameter creation
2292 - introduce flags to control the update behaviour of the
2293 BN_BLINDING parameters
2294 - hide BN_BLINDING structure
2295 Add a second BN_BLINDING slot to the RSA structure to improve
2296 performance when a single RSA object is shared among several
2300 *) Add support for DTLS.
2301 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
2303 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
2304 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
2307 *) Remove buggy and incompletet DH cert support from
2308 ssl/ssl_rsa.c and ssl/s3_both.c
2311 *) Use SHA-1 instead of MD5 as the default digest algorithm for
2312 the apps/openssl applications.
2315 *) Compile clean with "-Wall -Wmissing-prototypes
2316 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
2317 DEBUG_SAFESTACK must also be set.
2320 *) Change ./Configure so that certain algorithms can be disabled by default.
2321 The new counterpiece to "no-xxx" is "enable-xxx".
2323 The patented RC5 and MDC2 algorithms will now be disabled unless
2324 "enable-rc5" and "enable-mdc2", respectively, are specified.
2326 (IDEA remains enabled despite being patented. This is because IDEA
2327 is frequently required for interoperability, and there is no license
2328 fee for non-commercial use. As before, "no-idea" can be used to
2329 avoid this algorithm.)
2333 *) Add processing of proxy certificates (see RFC 3820). This work was
2334 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
2335 EGEE (Enabling Grids for E-science in Europe).
2338 *) RC4 performance overhaul on modern architectures/implementations, such
2339 as Intel P4, IA-64 and AMD64.
2342 *) New utility extract-section.pl. This can be used specify an alternative
2343 section number in a pod file instead of having to treat each file as
2344 a separate case in Makefile. This can be done by adding two lines to the
2347 =for comment openssl_section:XXX
2349 The blank line is mandatory.
2353 *) New arguments -certform, -keyform and -pass for s_client and s_server
2354 to allow alternative format key and certificate files and passphrase
2358 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
2359 update associated structures and add various utility functions.
2361 Add new policy related verify parameters, include policy checking in
2362 standard verify code. Enhance 'smime' application with extra parameters
2363 to support policy checking and print out.
2366 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
2367 Nehemiah processors. These extensions support AES encryption in hardware
2368 as well as RNG (though RNG support is currently disabled).
2369 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
2371 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
2374 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
2375 [Andy Polyakov and a number of other people]
2377 *) Improved PowerPC platform support. Most notably BIGNUM assembler
2378 implementation contributed by IBM.
2379 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
2381 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
2382 exponent rather than 'unsigned long'. There is a corresponding change to
2383 the new 'rsa_keygen' element of the RSA_METHOD structure.
2384 [Jelte Jansen, Geoff Thorpe]
2386 *) Functionality for creating the initial serial number file is now
2387 moved from CA.pl to the 'ca' utility with a new option -create_serial.
2389 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
2390 number file to 1, which is bound to cause problems. To avoid
2391 the problems while respecting compatibility between different 0.9.7
2392 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
2393 CA.pl for serial number initialization. With the new release 0.9.8,
2394 we can fix the problem directly in the 'ca' utility.)
2397 *) Reduced header interdepencies by declaring more opaque objects in
2398 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
2399 give fewer recursive includes, which could break lazy source code - so
2400 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
2401 developers should define this symbol when building and using openssl to
2402 ensure they track the recommended behaviour, interfaces, [etc], but
2403 backwards-compatible behaviour prevails when this isn't defined.
2406 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
2409 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
2410 This will generate a random key of the appropriate length based on the
2411 cipher context. The EVP_CIPHER can provide its own random key generation
2412 routine to support keys of a specific form. This is used in the des and
2413 3des routines to generate a key of the correct parity. Update S/MIME
2414 code to use new functions and hence generate correct parity DES keys.
2415 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
2416 valid (weak or incorrect parity).
2419 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
2420 as looking them up. This is useful when the verified structure may contain
2421 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
2422 present unless the new PKCS7_NO_CRL flag is asserted.
2425 *) Extend ASN1 oid configuration module. It now additionally accepts the
2428 shortName = some long name, 1.2.3.4
2431 *) Reimplemented the BN_CTX implementation. There is now no more static
2432 limitation on the number of variables it can handle nor the depth of the
2433 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
2434 information can now expand as required, and rather than having a single
2435 static array of bignums, BN_CTX now uses a linked-list of such arrays
2436 allowing it to expand on demand whilst maintaining the usefulness of
2437 BN_CTX's "bundling".
2440 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
2441 to allow all RSA operations to function using a single BN_CTX.
2444 *) Preliminary support for certificate policy evaluation and checking. This
2445 is initially intended to pass the tests outlined in "Conformance Testing
2446 of Relying Party Client Certificate Path Processing Logic" v1.07.
2449 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
2450 remained unused and not that useful. A variety of other little bignum
2451 tweaks and fixes have also been made continuing on from the audit (see
2455 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
2456 associated ASN1, EVP and SSL functions and old ASN1 macros.
2459 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
2460 and this should never fail. So the return value from the use of
2461 BN_set_word() (which can fail due to needless expansion) is now deprecated;
2462 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
2465 *) BN_CTX_get() should return zero-valued bignums, providing the same
2466 initialised value as BN_new().
2467 [Geoff Thorpe, suggested by Ulf Möller]
2469 *) Support for inhibitAnyPolicy certificate extension.
2472 *) An audit of the BIGNUM code is underway, for which debugging code is
2473 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
2474 is considered valid when processing BIGNUMs, and causes execution to
2475 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
2476 further steps are taken to deliberately pollute unused data in BIGNUM
2477 structures to try and expose faulty code further on. For now, openssl will
2478 (in its default mode of operation) continue to tolerate the inconsistent
2479 forms that it has tolerated in the past, but authors and packagers should
2480 consider trying openssl and their own applications when compiled with
2481 these debugging symbols defined. It will help highlight potential bugs in
2482 their own code, and will improve the test coverage for OpenSSL itself. At
2483 some point, these tighter rules will become openssl's default to improve
2484 maintainability, though the assert()s and other overheads will remain only
2485 in debugging configurations. See bn.h for more details.
2486 [Geoff Thorpe, Nils Larsch, Ulf Möller]
2488 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
2489 that can only be obtained through BN_CTX_new() (which implicitly
2490 initialises it). The presence of this function only made it possible
2491 to overwrite an existing structure (and cause memory leaks).
2494 *) Because of the callback-based approach for implementing LHASH as a
2495 template type, lh_insert() adds opaque objects to hash-tables and
2496 lh_doall() or lh_doall_arg() are typically used with a destructor callback
2497 to clean up those corresponding objects before destroying the hash table
2498 (and losing the object pointers). So some over-zealous constifications in
2499 LHASH have been relaxed so that lh_insert() does not take (nor store) the
2500 objects as "const" and the lh_doall[_arg] callback wrappers are not
2501 prototyped to have "const" restrictions on the object pointers they are
2502 given (and so aren't required to cast them away any more).
2505 *) The tmdiff.h API was so ugly and minimal that our own timing utility
2506 (speed) prefers to use its own implementation. The two implementations
2507 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
2508 its object type properly exposed (MS_TM) instead of casting to/from "char
2509 *". This may still change yet if someone realises MS_TM and "ms_time_***"
2510 aren't necessarily the greatest nomenclatures - but this is what was used
2511 internally to the implementation so I've used that for now.
2514 *) Ensure that deprecated functions do not get compiled when
2515 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
2516 the self-tests were still using deprecated key-generation functions so
2517 these have been updated also.
2520 *) Reorganise PKCS#7 code to separate the digest location functionality
2521 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
2522 New function PKCS7_set_digest() to set the digest type for PKCS#7
2523 digestedData type. Add additional code to correctly generate the
2524 digestedData type and add support for this type in PKCS7 initialization
2528 *) New function PKCS7_set0_type_other() this initializes a PKCS7
2529 structure of type "other".
2532 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
2533 sure the loop does correctly stop and breaking ("division by zero")
2534 modulus operations are not performed. The (pre-generated) prime
2535 table crypto/bn/bn_prime.h was already correct, but it could not be
2536 re-generated on some platforms because of the "division by zero"
2537 situation in the script.
2538 [Ralf S. Engelschall]
2540 *) Update support for ECC-based TLS ciphersuites according to
2541 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
2542 SHA-1 now is only used for "small" curves (where the
2543 representation of a field element takes up to 24 bytes); for
2544 larger curves, the field element resulting from ECDH is directly
2545 used as premaster secret.
2546 [Douglas Stebila (Sun Microsystems Laboratories)]
2548 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
2549 curve secp160r1 to the tests.
2550 [Douglas Stebila (Sun Microsystems Laboratories)]
2552 *) Add the possibility to load symbols globally with DSO.
2553 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
2555 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
2556 control of the error stack.
2559 *) Add support for STORE in ENGINE.
2562 *) Add the STORE type. The intention is to provide a common interface
2563 to certificate and key stores, be they simple file-based stores, or
2564 HSM-type store, or LDAP stores, or...
2565 NOTE: The code is currently UNTESTED and isn't really used anywhere.
2568 *) Add a generic structure called OPENSSL_ITEM. This can be used to
2569 pass a list of arguments to any function as well as provide a way
2570 for a function to pass data back to the caller.
2573 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
2574 works like BUF_strdup() but can be used to duplicate a portion of
2575 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
2579 *) Add the function sk_find_ex() which works like sk_find(), but will
2580 return an index to an element even if an exact match couldn't be
2581 found. The index is guaranteed to point at the element where the
2582 searched-for key would be inserted to preserve sorting order.
2585 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
2586 takes an extra flags argument for optional functionality. Currently,
2587 the following flags are defined:
2589 OBJ_BSEARCH_VALUE_ON_NOMATCH
2590 This one gets OBJ_bsearch_ex() to return a pointer to the first
2591 element where the comparing function returns a negative or zero
2594 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
2595 This one gets OBJ_bsearch_ex() to return a pointer to the first
2596 element where the comparing function returns zero. This is useful
2597 if there are more than one element where the comparing function
2601 *) Make it possible to create self-signed certificates with 'openssl ca'
2602 in such a way that the self-signed certificate becomes part of the
2603 CA database and uses the same mechanisms for serial number generation
2604 as all other certificate signing. The new flag '-selfsign' enables
2605 this functionality. Adapt CA.sh and CA.pl.in.
2608 *) Add functionality to check the public key of a certificate request
2609 against a given private. This is useful to check that a certificate
2610 request can be signed by that key (self-signing).
2613 *) Make it possible to have multiple active certificates with the same
2614 subject in the CA index file. This is done only if the keyword
2615 'unique_subject' is set to 'no' in the main CA section (default
2616 if 'CA_default') of the configuration file. The value is saved
2617 with the database itself in a separate index attribute file,
2618 named like the index file with '.attr' appended to the name.
2621 *) Generate muti valued AVAs using '+' notation in config files for
2625 *) Support for nameConstraints certificate extension.
2628 *) Support for policyConstraints certificate extension.
2631 *) Support for policyMappings certificate extension.
2634 *) Make sure the default DSA_METHOD implementation only uses its
2635 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
2636 and change its own handlers to be NULL so as to remove unnecessary
2637 indirection. This lets alternative implementations fallback to the
2638 default implementation more easily.
2641 *) Support for directoryName in GeneralName related extensions
2645 *) Make it possible to link applications using Makefile.shared.
2646 Make that possible even when linking against static libraries!
2649 *) Support for single pass processing for S/MIME signing. This now
2650 means that S/MIME signing can be done from a pipe, in addition
2651 cleartext signing (multipart/signed type) is effectively streaming
2652 and the signed data does not need to be all held in memory.
2654 This is done with a new flag PKCS7_STREAM. When this flag is set
2655 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
2656 is done after the data is output (and digests calculated) in
2657 SMIME_write_PKCS7().
2660 *) Add full support for -rpath/-R, both in shared libraries and
2661 applications, at least on the platforms where it's known how
2665 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
2666 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
2667 will now compute a table of multiples of the generator that
2668 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
2669 faster (notably in the case of a single point multiplication,
2670 scalar * generator).
2671 [Nils Larsch, Bodo Moeller]
2673 *) IPv6 support for certificate extensions. The various extensions
2674 which use the IP:a.b.c.d can now take IPv6 addresses using the
2675 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
2679 *) Added an ENGINE that implements RSA by performing private key
2680 exponentiations with the GMP library. The conversions to and from
2681 GMP's mpz_t format aren't optimised nor are any montgomery forms
2682 cached, and on x86 it appears OpenSSL's own performance has caught up.
2683 However there are likely to be other architectures where GMP could
2684 provide a boost. This ENGINE is not built in by default, but it can be
2685 specified at Configure time and should be accompanied by the necessary
2686 linker additions, eg;
2687 ./config -DOPENSSL_USE_GMP -lgmp
2690 *) "openssl engine" will not display ENGINE/DSO load failure errors when
2691 testing availability of engines with "-t" - the old behaviour is
2692 produced by increasing the feature's verbosity with "-tt".
2695 *) ECDSA routines: under certain error conditions uninitialized BN objects
2696 could be freed. Solution: make sure initialization is performed early
2697 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
2701 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
2702 and DH_METHOD (eg. by ENGINE implementations) to override the normal
2703 software implementations. For DSA and DH, parameter generation can
2704 also be overriden by providing the appropriate method callbacks.
2707 *) Change the "progress" mechanism used in key-generation and
2708 primality testing to functions that take a new BN_GENCB pointer in
2709 place of callback/argument pairs. The new API functions have "_ex"
2710 postfixes and the older functions are reimplemented as wrappers for
2711 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
2712 declarations of the old functions to help (graceful) attempts to
2713 migrate to the new functions. Also, the new key-generation API
2714 functions operate on a caller-supplied key-structure and return
2715 success/failure rather than returning a key or NULL - this is to
2716 help make "keygen" another member function of RSA_METHOD etc.
2718 Example for using the new callback interface:
2720 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
2724 BN_GENCB_set(&my_cb, my_callback, my_arg);
2726 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
2727 /* For the meaning of a, b in calls to my_callback(), see the
2728 * documentation of the function that calls the callback.
2729 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
2730 * my_callback should return 1 if it wants BN_is_prime_ex()
2731 * to continue, or 0 to stop.
2736 *) Change the ZLIB compression method to be stateful, and make it
2737 available to TLS with the number defined in
2738 draft-ietf-tls-compression-04.txt.
2741 *) Add the ASN.1 structures and functions for CertificatePair, which
2742 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
2744 CertificatePair ::= SEQUENCE {
2745 forward [0] Certificate OPTIONAL,
2746 reverse [1] Certificate OPTIONAL,
2747 -- at least one of the pair shall be present -- }
2749 Also implement the PEM functions to read and write certificate
2750 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
2752 This needed to be defined, mostly for the sake of the LDAP
2753 attribute crossCertificatePair, but may prove useful elsewhere as
2757 *) Make it possible to inhibit symlinking of shared libraries in
2758 Makefile.shared, for Cygwin's sake.
2761 *) Extend the BIGNUM API by creating a function
2762 void BN_set_negative(BIGNUM *a, int neg);
2763 and a macro that behave like
2764 int BN_is_negative(const BIGNUM *a);
2766 to avoid the need to access 'a->neg' directly in applications.
2769 *) Implement fast modular reduction for pseudo-Mersenne primes
2770 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
2771 EC_GROUP_new_curve_GFp() will now automatically use this
2773 [Nils Larsch <nla@trustcenter.de>]
2775 *) Add new lock type (CRYPTO_LOCK_BN).
2778 *) Change the ENGINE framework to automatically load engines
2779 dynamically from specific directories unless they could be
2780 found to already be built in or loaded. Move all the
2781 current engines except for the cryptodev one to a new
2783 The engines in engines/ are built as shared libraries if
2784 the "shared" options was given to ./Configure or ./config.
2785 Otherwise, they are inserted in libcrypto.a.
2786 /usr/local/ssl/engines is the default directory for dynamic
2787 engines, but that can be overriden at configure time through
2788 the usual use of --prefix and/or --openssldir, and at run
2789 time with the environment variable OPENSSL_ENGINES.
2790 [Geoff Thorpe and Richard Levitte]
2792 *) Add Makefile.shared, a helper makefile to build shared
2793 libraries. Addapt Makefile.org.
2796 *) Add version info to Win32 DLLs.
2797 [Peter 'Luna' Runestig" <peter@runestig.com>]
2799 *) Add new 'medium level' PKCS#12 API. Certificates and keys
2800 can be added using this API to created arbitrary PKCS#12
2801 files while avoiding the low level API.
2803 New options to PKCS12_create(), key or cert can be NULL and
2804 will then be omitted from the output file. The encryption
2805 algorithm NIDs can be set to -1 for no encryption, the mac
2806 iteration count can be set to 0 to omit the mac.
2808 Enhance pkcs12 utility by making the -nokeys and -nocerts
2809 options work when creating a PKCS#12 file. New option -nomac
2810 to omit the mac, NONE can be set for an encryption algorithm.
2811 New code is modified to use the enhanced PKCS12_create()
2812 instead of the low level API.
2815 *) Extend ASN1 encoder to support indefinite length constructed
2816 encoding. This can output sequences tags and octet strings in
2817 this form. Modify pk7_asn1.c to support indefinite length
2818 encoding. This is experimental and needs additional code to
2819 be useful, such as an ASN1 bio and some enhanced streaming
2822 Extend template encode functionality so that tagging is passed
2823 down to the template encoder.
2826 *) Let 'openssl req' fail if an argument to '-newkey' is not
2827 recognized instead of using RSA as a default.
2830 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
2831 As these are not official, they are not included in "ALL";
2832 the "ECCdraft" ciphersuite group alias can be used to select them.
2833 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
2835 *) Add ECDH engine support.
2836 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
2838 *) Add ECDH in new directory crypto/ecdh/.
2839 [Douglas Stebila (Sun Microsystems Laboratories)]
2841 *) Let BN_rand_range() abort with an error after 100 iterations
2842 without success (which indicates a broken PRNG).
2845 *) Change BN_mod_sqrt() so that it verifies that the input value
2846 is really the square of the return value. (Previously,
2847 BN_mod_sqrt would show GIGO behaviour.)
2850 *) Add named elliptic curves over binary fields from X9.62, SECG,
2851 and WAP/WTLS; add OIDs that were still missing.
2853 [Sheueling Chang Shantz and Douglas Stebila
2854 (Sun Microsystems Laboratories)]
2856 *) Extend the EC library for elliptic curves over binary fields
2857 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
2860 EC_GF2m_simple_method
2864 EC_GROUP_new_curve_GF2m
2865 EC_GROUP_set_curve_GF2m
2866 EC_GROUP_get_curve_GF2m
2867 EC_POINT_set_affine_coordinates_GF2m
2868 EC_POINT_get_affine_coordinates_GF2m
2869 EC_POINT_set_compressed_coordinates_GF2m
2871 Point compression for binary fields is disabled by default for
2872 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
2875 As binary polynomials are represented as BIGNUMs, various members
2876 of the EC_GROUP and EC_POINT data structures can be shared
2877 between the implementations for prime fields and binary fields;
2878 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
2879 are essentially identical to their ..._GFp counterparts.
2880 (For simplicity, the '..._GFp' prefix has been dropped from
2881 various internal method names.)
2883 An internal 'field_div' method (similar to 'field_mul' and
2884 'field_sqr') has been added; this is used only for binary fields.
2886 [Sheueling Chang Shantz and Douglas Stebila
2887 (Sun Microsystems Laboratories)]
2889 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
2890 through methods ('mul', 'precompute_mult').
2892 The generic implementations (now internally called 'ec_wNAF_mul'
2893 and 'ec_wNAF_precomputed_mult') remain the default if these
2894 methods are undefined.
2896 [Sheueling Chang Shantz and Douglas Stebila
2897 (Sun Microsystems Laboratories)]
2899 *) New function EC_GROUP_get_degree, which is defined through
2900 EC_METHOD. For curves over prime fields, this returns the bit
2901 length of the modulus.
2903 [Sheueling Chang Shantz and Douglas Stebila
2904 (Sun Microsystems Laboratories)]
2906 *) New functions EC_GROUP_dup, EC_POINT_dup.
2907 (These simply call ..._new and ..._copy).
2909 [Sheueling Chang Shantz and Douglas Stebila
2910 (Sun Microsystems Laboratories)]
2912 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
2913 Polynomials are represented as BIGNUMs (where the sign bit is not
2914 used) in the following functions [macros]:
2917 BN_GF2m_sub [= BN_GF2m_add]
2918 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
2919 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
2920 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
2922 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
2923 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
2924 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
2925 BN_GF2m_cmp [= BN_ucmp]
2927 (Note that only the 'mod' functions are actually for fields GF(2^m).
2928 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
2930 For some functions, an the irreducible polynomial defining a
2931 field can be given as an 'unsigned int[]' with strictly
2932 decreasing elements giving the indices of those bits that are set;
2933 i.e., p[] represents the polynomial
2934 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
2936 p[0] > p[1] > ... > p[k] = 0.
2937 This applies to the following functions:
2942 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
2943 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
2945 BN_GF2m_mod_sqrt_arr
2946 BN_GF2m_mod_solve_quad_arr
2950 Conversion can be performed by the following functions:
2955 bntest.c has additional tests for binary polynomial arithmetic.
2957 Two implementations for BN_GF2m_mod_div() are available.
2958 The default algorithm simply uses BN_GF2m_mod_inv() and
2959 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
2960 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
2961 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
2963 [Sheueling Chang Shantz and Douglas Stebila
2964 (Sun Microsystems Laboratories)]
2966 *) Add new error code 'ERR_R_DISABLED' that can be used when some
2967 functionality is disabled at compile-time.
2968 [Douglas Stebila <douglas.stebila@sun.com>]
2970 *) Change default behaviour of 'openssl asn1parse' so that more
2971 information is visible when viewing, e.g., a certificate:
2973 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
2974 mode the content of non-printable OCTET STRINGs is output in a
2975 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
2976 avoid the appearance of a printable string.
2977 [Nils Larsch <nla@trustcenter.de>]
2979 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
2981 EC_GROUP_set_asn1_flag()
2982 EC_GROUP_get_asn1_flag()
2983 EC_GROUP_set_point_conversion_form()
2984 EC_GROUP_get_point_conversion_form()
2985 These control ASN1 encoding details:
2986 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
2987 has been set to OPENSSL_EC_NAMED_CURVE.
2988 - Points are encoded in uncompressed form by default; options for
2989 asn1_for are as for point2oct, namely
2990 POINT_CONVERSION_COMPRESSED
2991 POINT_CONVERSION_UNCOMPRESSED
2992 POINT_CONVERSION_HYBRID
2994 Also add 'seed' and 'seed_len' members to EC_GROUP with access
2997 EC_GROUP_get0_seed()
2998 EC_GROUP_get_seed_len()
2999 This is used only for ASN1 purposes (so far).
3000 [Nils Larsch <nla@trustcenter.de>]
3002 *) Add 'field_type' member to EC_METHOD, which holds the NID
3003 of the appropriate field type OID. The new function
3004 EC_METHOD_get_field_type() returns this value.
3005 [Nils Larsch <nla@trustcenter.de>]
3010 EC_POINT_point2hex()
3011 EC_POINT_hex2point()
3012 providing useful interfaces to EC_POINT_point2oct() and
3013 EC_POINT_oct2point().
3014 [Nils Larsch <nla@trustcenter.de>]
3016 *) Change internals of the EC library so that the functions
3017 EC_GROUP_set_generator()
3018 EC_GROUP_get_generator()
3019 EC_GROUP_get_order()
3020 EC_GROUP_get_cofactor()
3021 are implemented directly in crypto/ec/ec_lib.c and not dispatched
3022 to methods, which would lead to unnecessary code duplication when
3023 adding different types of curves.
3024 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
3026 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
3027 arithmetic, and such that modified wNAFs are generated
3028 (which avoid length expansion in many cases).
3031 *) Add a function EC_GROUP_check_discriminant() (defined via
3032 EC_METHOD) that verifies that the curve discriminant is non-zero.
3034 Add a function EC_GROUP_check() that makes some sanity tests
3035 on a EC_GROUP, its generator and order. This includes
3036 EC_GROUP_check_discriminant().
3037 [Nils Larsch <nla@trustcenter.de>]
3039 *) Add ECDSA in new directory crypto/ecdsa/.
3041 Add applications 'openssl ecparam' and 'openssl ecdsa'
3042 (these are based on 'openssl dsaparam' and 'openssl dsa').
3044 ECDSA support is also included in various other files across the
3045 library. Most notably,
3046 - 'openssl req' now has a '-newkey ecdsa:file' option;
3047 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
3048 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
3049 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
3050 them suitable for ECDSA where domain parameters must be
3051 extracted before the specific public key;
3052 - ECDSA engine support has been added.
3053 [Nils Larsch <nla@trustcenter.de>]
3055 *) Include some named elliptic curves, and add OIDs from X9.62,
3056 SECG, and WAP/WTLS. Each curve can be obtained from the new
3058 EC_GROUP_new_by_curve_name(),
3059 and the list of available named curves can be obtained with
3060 EC_get_builtin_curves().
3061 Also add a 'curve_name' member to EC_GROUP objects, which can be
3063 EC_GROUP_set_curve_name()
3064 EC_GROUP_get_curve_name()
3065 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
3067 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3068 was actually never needed) and in BN_mul(). The removal in BN_mul()
3069 required a small change in bn_mul_part_recursive() and the addition
3070 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3071 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3072 bn_sub_words() and bn_add_words() except they take arrays with
3076 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
3078 *) Cleanse PEM buffers before freeing them since they may contain
3080 [Benjamin Bennett <ben@psc.edu>]
3082 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3083 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3084 authentication-only ciphersuites.
3087 *) Since AES128 and AES256 share a single mask bit in the logic of
3088 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3089 kludge to work properly if AES128 is available and AES256 isn't.
3092 *) Expand security boundary to match 1.1.1 module.
3095 *) Remove redundant features: hash file source, editing of test vectors
3096 modify fipsld to use external fips_premain.c signature.
3099 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
3100 run algorithm test programs.
3103 *) Make algorithm test programs more tolerant of whitespace.
3106 *) Have SSL/TLS server implementation tolerate "mismatched" record
3107 protocol version while receiving ClientHello even if the
3108 ClientHello is fragmented. (The server can't insist on the
3109 particular protocol version it has chosen before the ServerHello
3110 message has informed the client about his choice.)
3113 *) Load error codes if they are not already present instead of using a
3114 static variable. This allows them to be cleanly unloaded and reloaded.
3117 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3119 *) Introduce limits to prevent malicious keys being able to
3120 cause a denial of service. (CVE-2006-2940)
3121 [Steve Henson, Bodo Moeller]
3123 *) Fix ASN.1 parsing of certain invalid structures that can result
3124 in a denial of service. (CVE-2006-2937) [Steve Henson]
3126 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3127 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3129 *) Fix SSL client code which could crash if connecting to a
3130 malicious SSLv2 server. (CVE-2006-4343)
3131 [Tavis Ormandy and Will Drewry, Google Security Team]
3133 *) Change ciphersuite string processing so that an explicit
3134 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3135 will no longer include "AES128-SHA"), and any other similar
3136 ciphersuite (same bitmap) from *other* protocol versions (so that
3137 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3138 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3139 changes from 0.9.8b and 0.9.8d.
3142 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
3144 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3145 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3147 *) Change the Unix randomness entropy gathering to use poll() when
3148 possible instead of select(), since the latter has some
3149 undesirable limitations.
3150 [Darryl Miles via Richard Levitte and Bodo Moeller]
3152 *) Disable rogue ciphersuites:
3154 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3155 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3156 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3158 The latter two were purportedly from
3159 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3162 Also deactive the remaining ciphersuites from
3163 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3164 unofficial, and the ID has long expired.
3167 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3168 dual-core machines) and other potential thread-safety issues.
3171 Changes between 0.9.7i and 0.9.7j [04 May 2006]
3173 *) Adapt fipsld and the build system to link against the validated FIPS
3174 module in FIPS mode.
3177 *) Fixes for VC++ 2005 build under Windows.
3180 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
3181 from a Windows bash shell such as MSYS. It is autodetected from the
3182 "config" script when run from a VC++ environment. Modify standard VC++
3183 build to use fipscanister.o from the GNU make build.
3186 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
3188 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
3189 The value now differs depending on if you build for FIPS or not.
3190 BEWARE! A program linked with a shared FIPSed libcrypto can't be
3191 safely run with a non-FIPSed libcrypto, as it may crash because of
3192 the difference induced by this change.
3195 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
3197 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3198 (part of SSL_OP_ALL). This option used to disable the
3199 countermeasure against man-in-the-middle protocol-version
3200 rollback in the SSL 2.0 server implementation, which is a bad
3201 idea. (CVE-2005-2969)
3203 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3204 for Information Security, National Institute of Advanced Industrial
3205 Science and Technology [AIST], Japan)]
3207 *) Minimal support for X9.31 signatures and PSS padding modes. This is
3208 mainly for FIPS compliance and not fully integrated at this stage.
3211 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
3212 the exponentiation using a fixed-length exponent. (Otherwise,
3213 the information leaked through timing could expose the secret key
3214 after many signatures; cf. Bleichenbacher's attack on DSA with
3218 *) Make a new fixed-window mod_exp implementation the default for
3219 RSA, DSA, and DH private-key operations so that the sequence of
3220 squares and multiplies and the memory access pattern are
3221 independent of the particular secret key. This will mitigate
3222 cache-timing and potential related attacks.
3224 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
3225 and this is automatically used by BN_mod_exp_mont() if the new flag
3226 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
3227 will use this BN flag for private exponents unless the flag
3228 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
3229 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
3231 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
3233 *) Change the client implementation for SSLv23_method() and
3234 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
3235 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
3236 (Previously, the SSL 2.0 backwards compatible Client Hello
3237 message format would be used even with SSL_OP_NO_SSLv2.)
3240 *) Add support for smime-type MIME parameter in S/MIME messages which some
3244 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
3245 a threadsafe manner. Modify rsa code to use new function and add calls
3246 to dsa and dh code (which had race conditions before).
3249 *) Include the fixed error library code in the C error file definitions
3250 instead of fixing them up at runtime. This keeps the error code
3251 structures constant.
3254 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
3256 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
3259 *) Fixes for newer kerberos headers. NB: the casts are needed because
3260 the 'length' field is signed on one version and unsigned on another
3261 with no (?) obvious way to tell the difference, without these VC++
3262 complains. Also the "definition" of FAR (blank) is no longer included
3263 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
3264 some needed definitions.
3267 *) Undo Cygwin change.
3270 *) Added support for proxy certificates according to RFC 3820.
3271 Because they may be a security thread to unaware applications,
3272 they must be explicitely allowed in run-time. See
3273 docs/HOWTO/proxy_certificates.txt for further information.
3276 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
3278 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
3279 server and client random values. Previously
3280 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
3281 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
3283 This change has negligible security impact because:
3285 1. Server and client random values still have 24 bytes of pseudo random
3288 2. Server and client random values are sent in the clear in the initial
3291 3. The master secret is derived using the premaster secret (48 bytes in
3292 size for static RSA ciphersuites) as well as client server and random
3295 The OpenSSL team would like to thank the UK NISCC for bringing this issue
3298 [Stephen Henson, reported by UK NISCC]
3300 *) Use Windows randomness collection on Cygwin.
3303 *) Fix hang in EGD/PRNGD query when communication socket is closed
3304 prematurely by EGD/PRNGD.
3305 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
3307 *) Prompt for pass phrases when appropriate for PKCS12 input format.
3310 *) Back-port of selected performance improvements from development
3311 branch, as well as improved support for PowerPC platforms.
3314 *) Add lots of checks for memory allocation failure, error codes to indicate
3315 failure and freeing up memory if a failure occurs.
3316 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
3318 *) Add new -passin argument to dgst.
3321 *) Perform some character comparisons of different types in X509_NAME_cmp:
3322 this is needed for some certificates that reencode DNs into UTF8Strings
3323 (in violation of RFC3280) and can't or wont issue name rollover
3327 *) Make an explicit check during certificate validation to see that
3328 the CA setting in each certificate on the chain is correct. As a
3329 side effect always do the following basic checks on extensions,
3330 not just when there's an associated purpose to the check:
3332 - if there is an unhandled critical extension (unless the user
3333 has chosen to ignore this fault)
3334 - if the path length has been exceeded (if one is set at all)
3335 - that certain extensions fit the associated purpose (if one has
3339 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
3341 *) Avoid a race condition when CRLs are checked in a multi threaded
3342 environment. This would happen due to the reordering of the revoked
3343 entries during signature checking and serial number lookup. Now the
3344 encoding is cached and the serial number sort performed under a lock.
3345 Add new STACK function sk_is_sorted().
3348 *) Add Delta CRL to the extension code.
3351 *) Various fixes to s3_pkt.c so alerts are sent properly.
3352 [David Holmes <d.holmes@f5.com>]
3354 *) Reduce the chances of duplicate issuer name and serial numbers (in
3355 violation of RFC3280) using the OpenSSL certificate creation utilities.
3356 This is done by creating a random 64 bit value for the initial serial
3357 number when a serial number file is created or when a self signed
3358 certificate is created using 'openssl req -x509'. The initial serial
3359 number file is created using 'openssl x509 -next_serial' in CA.pl
3360 rather than being initialized to 1.
3363 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
3365 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3366 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3367 [Joe Orton, Steve Henson]
3369 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
3371 [Joe Orton, Steve Henson]
3373 *) Make it possible to have multiple active certificates with the same
3374 subject in the CA index file. This is done only if the keyword
3375 'unique_subject' is set to 'no' in the main CA section (default
3376 if 'CA_default') of the configuration file. The value is saved
3377 with the database itself in a separate index attribute file,
3378 named like the index file with '.attr' appended to the name.
3381 *) X509 verify fixes. Disable broken certificate workarounds when
3382 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
3383 keyUsage extension present. Don't accept CRLs with unhandled critical
3384 extensions: since verify currently doesn't process CRL extensions this
3385 rejects a CRL with *any* critical extensions. Add new verify error codes
3389 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
3390 A clarification of RFC2560 will require the use of OCTET STRINGs and
3391 some implementations cannot handle the current raw format. Since OpenSSL
3392 copies and compares OCSP nonces as opaque blobs without any attempt at
3393 parsing them this should not create any compatibility issues.
3396 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
3397 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
3398 this HMAC (and other) operations are several times slower than OpenSSL
3402 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
3403 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3405 *) Use the correct content when signing type "other".
3408 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
3410 *) Fix various bugs revealed by running the NISCC test suite:
3412 Stop out of bounds reads in the ASN1 code when presented with
3413 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3415 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
3417 If verify callback ignores invalid public key errors don't try to check
3418 certificate signature with the NULL public key.
3422 *) New -ignore_err option in ocsp application to stop the server
3423 exiting on the first error in a request.
3426 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3427 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3431 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3432 extra data after the compression methods not only for TLS 1.0
3433 but also for SSL 3.0 (as required by the specification).
3434 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3436 *) Change X509_certificate_type() to mark the key as exported/exportable
3437 when it's 512 *bits* long, not 512 bytes.
3440 *) Change AES_cbc_encrypt() so it outputs exact multiple of
3441 blocks during encryption.
3444 *) Various fixes to base64 BIO and non blocking I/O. On write
3445 flushes were not handled properly if the BIO retried. On read
3446 data was not being buffered properly and had various logic bugs.
3447 This also affects blocking I/O when the data being decoded is a
3451 *) Various S/MIME bugfixes and compatibility changes:
3452 output correct application/pkcs7 MIME type if
3453 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
3454 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
3455 of files as .eml work). Correctly handle very long lines in MIME
3459 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
3461 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3462 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3463 a protocol version number mismatch like a decryption error
3464 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3467 *) Turn on RSA blinding by default in the default implementation
3468 to avoid a timing attack. Applications that don't want it can call
3469 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3470 They would be ill-advised to do so in most cases.
3471 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3473 *) Change RSA blinding code so that it works when the PRNG is not
3474 seeded (in this case, the secret RSA exponent is abused as
3475 an unpredictable seed -- if it is not unpredictable, there
3476 is no point in blinding anyway). Make RSA blinding thread-safe
3477 by remembering the creator's thread ID in rsa->blinding and
3478 having all other threads use local one-time blinding factors
3479 (this requires more computation than sharing rsa->blinding, but
3480 avoids excessive locking; and if an RSA object is not shared
3481 between threads, blinding will still be very fast).
3484 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
3485 ENGINE as defaults for all supported algorithms irrespective of
3486 the 'flags' parameter. 'flags' is now honoured, so applications
3487 should make sure they are passing it correctly.
3490 *) Target "mingw" now allows native Windows code to be generated in
3491 the Cygwin environment as well as with the MinGW compiler.
3494 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
3496 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3497 via timing by performing a MAC computation even if incorrrect
3498 block cipher padding has been found. This is a countermeasure
3499 against active attacks where the attacker has to distinguish
3500 between bad padding and a MAC verification error. (CVE-2003-0078)
3502 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3503 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3504 Martin Vuagnoux (EPFL, Ilion)]
3506 *) Make the no-err option work as intended. The intention with no-err
3507 is not to have the whole error stack handling routines removed from
3508 libcrypto, it's only intended to remove all the function name and
3509 reason texts, thereby removing some of the footprint that may not
3510 be interesting if those errors aren't displayed anyway.
3512 NOTE: it's still possible for any application or module to have it's
3513 own set of error texts inserted. The routines are there, just not
3514 used by default when no-err is given.
3517 *) Add support for FreeBSD on IA64.
3518 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
3520 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
3521 Kerberos function mit_des_cbc_cksum(). Before this change,
3522 the value returned by DES_cbc_cksum() was like the one from
3523 mit_des_cbc_cksum(), except the bytes were swapped.
3524 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
3526 *) Allow an application to disable the automatic SSL chain building.
3527 Before this a rather primitive chain build was always performed in
3528 ssl3_output_cert_chain(): an application had no way to send the
3529 correct chain if the automatic operation produced an incorrect result.
3531 Now the chain builder is disabled if either:
3533 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
3535 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
3537 The reasoning behind this is that an application would not want the
3538 auto chain building to take place if extra chain certificates are
3539 present and it might also want a means of sending no additional
3540 certificates (for example the chain has two certificates and the
3544 *) Add the possibility to build without the ENGINE framework.
3545 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3547 *) Under Win32 gmtime() can return NULL: check return value in
3548 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
3551 *) DSA routines: under certain error conditions uninitialized BN objects
3552 could be freed. Solution: make sure initialization is performed early
3553 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
3554 Nils Larsch <nla@trustcenter.de> via PR#459)
3557 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
3558 checked on reconnect on the client side, therefore session resumption
3559 could still fail with a "ssl session id is different" error. This
3560 behaviour is masked when SSL_OP_ALL is used due to
3561 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
3562 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3563 followup to PR #377.
3566 *) IA-32 assembler support enhancements: unified ELF targets, support
3567 for SCO/Caldera platforms, fix for Cygwin shared build.
3570 *) Add support for FreeBSD on sparc64. As a consequence, support for
3571 FreeBSD on non-x86 processors is separate from x86 processors on
3572 the config script, much like the NetBSD support.
3573 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
3575 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
3577 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
3580 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
3581 code (06) was taken as the first octet of the session ID and the last
3582 octet was ignored consequently. As a result SSLv2 client side session
3583 caching could not have worked due to the session ID mismatch between
3585 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3589 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
3590 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
3594 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
3595 seems that in spite of existing for more than a year, many application
3596 author have done nothing to provide the necessary callbacks, which
3597 means that this particular engine will not work properly anywhere.
3598 This is a very unfortunate situation which forces us, in the name
3599 of usability, to give the hw_ncipher.c a static lock, which is part
3601 NOTE: This is for the 0.9.7 series ONLY. This hack will never
3602 appear in 0.9.8 or later. We EXPECT application authors to have
3603 dealt properly with this when 0.9.8 is released (unless we actually
3604 make such changes in the libcrypto locking code that changes will
3605 have to be made anyway).
3608 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
3609 octets have been read, EOF or an error occurs. Without this change
3610 some truncated ASN1 structures will not produce an error.
3613 *) Disable Heimdal support, since it hasn't been fully implemented.
3614 Still give the possibility to force the use of Heimdal, but with
3615 warnings and a request that patches get sent to openssl-dev.
3618 *) Add the VC-CE target, introduce the WINCE sysname, and add
3619 INSTALL.WCE and appropriate conditionals to make it build.
3620 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3622 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
3623 cygssl-x.y.z.dll, where x, y and z are the major, minor and
3624 edit numbers of the version.
3625 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3627 *) Introduce safe string copy and catenation functions
3628 (BUF_strlcpy() and BUF_strlcat()).
3629 [Ben Laurie (CHATS) and Richard Levitte]
3631 *) Avoid using fixed-size buffers for one-line DNs.
3632 [Ben Laurie (CHATS)]
3634 *) Add BUF_MEM_grow_clean() to avoid information leakage when
3635 resizing buffers containing secrets, and use where appropriate.
3636 [Ben Laurie (CHATS)]
3638 *) Avoid using fixed size buffers for configuration file location.
3639 [Ben Laurie (CHATS)]
3641 *) Avoid filename truncation for various CA files.
3642 [Ben Laurie (CHATS)]
3644 *) Use sizeof in preference to magic numbers.
3645 [Ben Laurie (CHATS)]
3647 *) Avoid filename truncation in cert requests.
3648 [Ben Laurie (CHATS)]
3650 *) Add assertions to check for (supposedly impossible) buffer
3652 [Ben Laurie (CHATS)]
3654 *) Don't cache truncated DNS entries in the local cache (this could
3655 potentially lead to a spoofing attack).
3656 [Ben Laurie (CHATS)]
3658 *) Fix various buffers to be large enough for hex/decimal
3659 representations in a platform independent manner.
3660 [Ben Laurie (CHATS)]
3662 *) Add CRYPTO_realloc_clean() to avoid information leakage when
3663 resizing buffers containing secrets, and use where appropriate.
3664 [Ben Laurie (CHATS)]
3666 *) Add BIO_indent() to avoid much slightly worrying code to do
3668 [Ben Laurie (CHATS)]
3670 *) Convert sprintf()/BIO_puts() to BIO_printf().
3671 [Ben Laurie (CHATS)]
3673 *) buffer_gets() could terminate with the buffer only half
3675 [Ben Laurie (CHATS)]
3677 *) Add assertions to prevent user-supplied crypto functions from
3678 overflowing internal buffers by having large block sizes, etc.
3679 [Ben Laurie (CHATS)]
3681 *) New OPENSSL_assert() macro (similar to assert(), but enabled
3683 [Ben Laurie (CHATS)]
3685 *) Eliminate unused copy of key in RC4.
3686 [Ben Laurie (CHATS)]
3688 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
3689 [Ben Laurie (CHATS)]
3691 *) Fix off-by-one error in EGD path.
3692 [Ben Laurie (CHATS)]
3694 *) If RANDFILE path is too long, ignore instead of truncating.
3695 [Ben Laurie (CHATS)]
3697 *) Eliminate unused and incorrectly sized X.509 structure
3699 [Ben Laurie (CHATS)]
3701 *) Eliminate unused and dangerous function knumber().
3702 [Ben Laurie (CHATS)]
3704 *) Eliminate unused and dangerous structure, KSSL_ERR.
3705 [Ben Laurie (CHATS)]
3707 *) Protect against overlong session ID context length in an encoded
3708 session object. Since these are local, this does not appear to be
3710 [Ben Laurie (CHATS)]
3712 *) Change from security patch (see 0.9.6e below) that did not affect
3713 the 0.9.6 release series:
3715 Remote buffer overflow in SSL3 protocol - an attacker could
3716 supply an oversized master key in Kerberos-enabled versions.
3718 [Ben Laurie (CHATS)]
3720 *) Change the SSL kerb5 codes to match RFC 2712.
3723 *) Make -nameopt work fully for req and add -reqopt switch.
3724 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
3726 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
3727 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
3729 *) Make sure tests can be performed even if the corresponding algorithms
3730 have been removed entirely. This was also the last step to make
3731 OpenSSL compilable with DJGPP under all reasonable conditions.
3732 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
3734 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
3735 to allow version independent disabling of normally unselected ciphers,
3736 which may be activated as a side-effect of selecting a single cipher.
3738 (E.g., cipher list string "RSA" enables ciphersuites that are left
3739 out of "ALL" because they do not provide symmetric encryption.
3740 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
3741 [Lutz Jaenicke, Bodo Moeller]
3743 *) Add appropriate support for separate platform-dependent build
3744 directories. The recommended way to make a platform-dependent
3745 build directory is the following (tested on Linux), maybe with
3748 # Place yourself outside of the OpenSSL source tree. In
3749 # this example, the environment variable OPENSSL_SOURCE
3750 # is assumed to contain the absolute OpenSSL source directory.
3751 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
3752 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
3753 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
3754 mkdir -p `dirname $F`
3755 ln -s $OPENSSL_SOURCE/$F $F
3758 To be absolutely sure not to disturb the source tree, a "make clean"
3759 is a good thing. If it isn't successfull, don't worry about it,
3760 it probably means the source directory is very clean.
3763 *) Make sure any ENGINE control commands make local copies of string
3764 pointers passed to them whenever necessary. Otherwise it is possible
3765 the caller may have overwritten (or deallocated) the original string
3766 data when a later ENGINE operation tries to use the stored values.
3767 [Götz Babin-Ebell <babinebell@trustcenter.de>]
3769 *) Improve diagnostics in file reading and command-line digests.
3770 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
3772 *) Add AES modes CFB and OFB to the object database. Correct an
3773 error in AES-CFB decryption.
3776 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
3777 allows existing EVP_CIPHER_CTX structures to be reused after
3778 calling EVP_*Final(). This behaviour is used by encryption
3779 BIOs and some applications. This has the side effect that
3780 applications must explicitly clean up cipher contexts with
3781 EVP_CIPHER_CTX_cleanup() or they will leak memory.
3784 *) Check the values of dna and dnb in bn_mul_recursive before calling
3785 bn_mul_comba (a non zero value means the a or b arrays do not contain
3786 n2 elements) and fallback to bn_mul_normal if either is not zero.
3789 *) Fix escaping of non-ASCII characters when using the -subj option
3790 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
3793 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
3794 form for "surname", serialNumber has no short form.
3795 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
3796 therefore remove "mail" short name for "internet 7".
3797 The OID for unique identifiers in X509 certificates is
3798 x500UniqueIdentifier, not uniqueIdentifier.
3799 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
3802 *) Add an "init" command to the ENGINE config module and auto initialize
3803 ENGINEs. Without any "init" command the ENGINE will be initialized
3804 after all ctrl commands have been executed on it. If init=1 the
3805 ENGINE is initailized at that point (ctrls before that point are run
3806 on the uninitialized ENGINE and after on the initialized one). If
3807 init=0 then the ENGINE will not be iniatialized at all.
3810 *) Fix the 'app_verify_callback' interface so that the user-defined
3811 argument is actually passed to the callback: In the
3812 SSL_CTX_set_cert_verify_callback() prototype, the callback
3813 declaration has been changed from
3816 int (*cb)(X509_STORE_CTX *,void *);
3817 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
3818 i=s->ctx->app_verify_callback(&ctx)
3819 has been changed into
3820 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
3822 To update applications using SSL_CTX_set_cert_verify_callback(),
3823 a dummy argument can be added to their callback functions.
3824 [D. K. Smetters <smetters@parc.xerox.com>]
3826 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
3827 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
3829 *) Add and OPENSSL_LOAD_CONF define which will cause
3830 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
3831 This allows older applications to transparently support certain
3832 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
3833 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
3834 load the config file and OPENSSL_add_all_algorithms_conf() which will
3835 always load it have also been added.
3838 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
3839 Adjust NIDs and EVP layer.
3840 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3842 *) Config modules support in openssl utility.
3844 Most commands now load modules from the config file,
3845 though in a few (such as version) this isn't done
3846 because it couldn't be used for anything.
3848 In the case of ca and req the config file used is
3849 the same as the utility itself: that is the -config
3850 command line option can be used to specify an
3854 *) Move default behaviour from OPENSSL_config(). If appname is NULL
3855 use "openssl_conf" if filename is NULL use default openssl config file.
3858 *) Add an argument to OPENSSL_config() to allow the use of an alternative
3859 config section name. Add a new flag to tolerate a missing config file
3860 and move code to CONF_modules_load_file().
3863 *) Support for crypto accelerator cards from Accelerated Encryption
3864 Processing, www.aep.ie. (Use engine 'aep')
3865 The support was copied from 0.9.6c [engine] and adapted/corrected
3866 to work with the new engine framework.
3867 [AEP Inc. and Richard Levitte]
3869 *) Support for SureWare crypto accelerator cards from Baltimore
3870 Technologies. (Use engine 'sureware')
3871 The support was copied from 0.9.6c [engine] and adapted
3872 to work with the new engine framework.
3875 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
3876 make the newer ENGINE framework commands for the CHIL engine work.
3877 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
3879 *) Make it possible to produce shared libraries on ReliantUNIX.
3880 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
3882 *) Add the configuration target debug-linux-ppro.
3883 Make 'openssl rsa' use the general key loading routines
3884 implemented in apps.c, and make those routines able to
3885 handle the key format FORMAT_NETSCAPE and the variant
3887 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3889 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
3890 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3892 *) Add -keyform to rsautl, and document -engine.
3893 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
3895 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
3896 BIO_R_NO_SUCH_FILE error code rather than the generic
3897 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
3900 *) Add new functions
3902 ERR_peek_last_error_line
3903 ERR_peek_last_error_line_data.
3904 These are similar to
3907 ERR_peek_error_line_data,
3908 but report on the latest error recorded rather than the first one
3909 still in the error queue.
3910 [Ben Laurie, Bodo Moeller]
3912 *) default_algorithms option in ENGINE config module. This allows things
3914 default_algorithms = ALL
3915 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
3918 *) Prelminary ENGINE config module.
3921 *) New experimental application configuration code.
3924 *) Change the AES code to follow the same name structure as all other
3925 symmetric ciphers, and behave the same way. Move everything to
3926 the directory crypto/aes, thereby obsoleting crypto/rijndael.
3927 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3929 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
3930 [Ben Laurie and Theo de Raadt]
3932 *) Add option to output public keys in req command.
3933 [Massimiliano Pala madwolf@openca.org]
3935 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
3936 (up to about 10% better than before for P-192 and P-224).
3939 *) New functions/macros
3941 SSL_CTX_set_msg_callback(ctx, cb)
3942 SSL_CTX_set_msg_callback_arg(ctx, arg)
3943 SSL_set_msg_callback(ssl, cb)
3944 SSL_set_msg_callback_arg(ssl, arg)
3946 to request calling a callback function
3948 void cb(int write_p, int version, int content_type,
3949 const void *buf, size_t len, SSL *ssl, void *arg)
3951 whenever a protocol message has been completely received
3952 (write_p == 0) or sent (write_p == 1). Here 'version' is the
3953 protocol version according to which the SSL library interprets
3954 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
3955 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
3956 the content type as defined in the SSL 3.0/TLS 1.0 protocol
3957 specification (change_cipher_spec(20), alert(21), handshake(22)).
3958 'buf' and 'len' point to the actual message, 'ssl' to the
3959 SSL object, and 'arg' is the application-defined value set by
3960 SSL[_CTX]_set_msg_callback_arg().
3962 'openssl s_client' and 'openssl s_server' have new '-msg' options
3963 to enable a callback that displays all protocol messages.
3966 *) Change the shared library support so shared libraries are built as
3967 soon as the corresponding static library is finished, and thereby get
3968 openssl and the test programs linked against the shared library.
3969 This still only happens when the keyword "shard" has been given to
3970 the configuration scripts.
3972 NOTE: shared library support is still an experimental thing, and
3973 backward binary compatibility is still not guaranteed.
3974 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
3976 *) Add support for Subject Information Access extension.
3977 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3979 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
3980 additional bytes when new memory had to be allocated, not just
3981 when reusing an existing buffer.
3984 *) New command line and configuration option 'utf8' for the req command.
3985 This allows field values to be specified as UTF8 strings.
3988 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
3989 runs for the former and machine-readable output for the latter.
3992 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
3993 of the e-mail address in the DN (i.e., it will go into a certificate
3994 extension only). The new configuration file option 'email_in_dn = no'
3995 has the same effect.
3996 [Massimiliano Pala madwolf@openca.org]
3998 *) Change all functions with names starting with des_ to be starting
3999 with DES_ instead. Add wrappers that are compatible with libdes,
4000 but are named _ossl_old_des_*. Finally, add macros that map the
4001 des_* symbols to the corresponding _ossl_old_des_* if libdes
4002 compatibility is desired. If OpenSSL 0.9.6c compatibility is
4003 desired, the des_* symbols will be mapped to DES_*, with one
4006 Since we provide two compatibility mappings, the user needs to
4007 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
4008 compatibility is desired. The default (i.e., when that macro
4009 isn't defined) is OpenSSL 0.9.6c compatibility.
4011 There are also macros that enable and disable the support of old
4012 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
4013 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
4014 are defined, the default will apply: to support the old des routines.
4016 In either case, one must include openssl/des.h to get the correct
4017 definitions. Do not try to just include openssl/des_old.h, that
4020 NOTE: This is a major break of an old API into a new one. Software
4021 authors are encouraged to switch to the DES_ style functions. Some
4022 time in the future, des_old.h and the libdes compatibility functions
4023 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
4024 default), and then completely removed.
4027 *) Test for certificates which contain unsupported critical extensions.
4028 If such a certificate is found during a verify operation it is
4029 rejected by default: this behaviour can be overridden by either
4030 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
4031 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
4032 X509_supported_extension() has also been added which returns 1 if a
4033 particular extension is supported.
4036 *) Modify the behaviour of EVP cipher functions in similar way to digests
4037 to retain compatibility with existing code.
4040 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
4041 compatibility with existing code. In particular the 'ctx' parameter does
4042 not have to be to be initialized before the call to EVP_DigestInit() and
4043 it is tidied up after a call to EVP_DigestFinal(). New function
4044 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
4045 EVP_MD_CTX_copy() changed to not require the destination to be
4046 initialized valid and new function EVP_MD_CTX_copy_ex() added which
4047 requires the destination to be valid.
4049 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
4050 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
4053 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
4054 so that complete 'Handshake' protocol structures are kept in memory
4055 instead of overwriting 'msg_type' and 'length' with 'body' data.
4058 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
4059 [Massimo Santin via Richard Levitte]
4061 *) Major restructuring to the underlying ENGINE code. This includes
4062 reduction of linker bloat, separation of pure "ENGINE" manipulation
4063 (initialisation, etc) from functionality dealing with implementations
4064 of specific crypto iterfaces. This change also introduces integrated
4065 support for symmetric ciphers and digest implementations - so ENGINEs
4066 can now accelerate these by providing EVP_CIPHER and EVP_MD
4067 implementations of their own. This is detailed in crypto/engine/README
4068 as it couldn't be adequately described here. However, there are a few
4069 API changes worth noting - some RSA, DSA, DH, and RAND functions that
4070 were changed in the original introduction of ENGINE code have now
4071 reverted back - the hooking from this code to ENGINE is now a good
4072 deal more passive and at run-time, operations deal directly with
4073 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
4074 dereferencing through an ENGINE pointer any more. Also, the ENGINE
4075 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
4076 they were not being used by the framework as there is no concept of a
4077 BIGNUM_METHOD and they could not be generalised to the new
4078 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
4079 ENGINE_cpy() has been removed as it cannot be consistently defined in
4083 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
4086 *) Change mkdef.pl to sort symbols that get the same entry number,
4087 and make sure the automatically generated functions ERR_load_*
4088 become part of libeay.num as well.
4091 *) New function SSL_renegotiate_pending(). This returns true once
4092 renegotiation has been requested (either SSL_renegotiate() call
4093 or HelloRequest/ClientHello receveived from the peer) and becomes
4094 false once a handshake has been completed.
4095 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
4096 sends a HelloRequest, but does not ensure that a handshake takes
4097 place. SSL_renegotiate_pending() is useful for checking if the
4098 client has followed the request.)
4101 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
4102 By default, clients may request session resumption even during
4103 renegotiation (if session ID contexts permit); with this option,
4104 session resumption is possible only in the first handshake.
4106 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4107 more bits available for options that should not be part of
4108 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4111 *) Add some demos for certificate and certificate request creation.
4114 *) Make maximum certificate chain size accepted from the peer application
4115 settable (SSL*_get/set_max_cert_list()), as proposed by
4116 "Douglas E. Engert" <deengert@anl.gov>.
4119 *) Add support for shared libraries for Unixware-7
4120 (Boyd Lynn Gerber <gerberb@zenez.com>).
4123 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4124 be done prior to destruction. Use this to unload error strings from
4125 ENGINEs that load their own error strings. NB: This adds two new API
4126 functions to "get" and "set" this destroy handler in an ENGINE.
4129 *) Alter all existing ENGINE implementations (except "openssl" and
4130 "openbsd") to dynamically instantiate their own error strings. This
4131 makes them more flexible to be built both as statically-linked ENGINEs
4132 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4133 Also, add stub code to each that makes building them as self-contained
4134 shared-libraries easier (see README.ENGINE).
4137 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4138 implementations into applications that are completely implemented in
4139 self-contained shared-libraries. The "dynamic" ENGINE exposes control
4140 commands that can be used to configure what shared-library to load and
4141 to control aspects of the way it is handled. Also, made an update to
4142 the README.ENGINE file that brings its information up-to-date and
4143 provides some information and instructions on the "dynamic" ENGINE
4144 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
4147 *) Make it possible to unload ranges of ERR strings with a new
4148 "ERR_unload_strings" function.
4151 *) Add a copy() function to EVP_MD.
4154 *) Make EVP_MD routines take a context pointer instead of just the
4155 md_data void pointer.
4158 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
4159 that the digest can only process a single chunk of data
4160 (typically because it is provided by a piece of
4161 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
4162 is only going to provide a single chunk of data, and hence the
4163 framework needn't accumulate the data for oneshot drivers.
4166 *) As with "ERR", make it possible to replace the underlying "ex_data"
4167 functions. This change also alters the storage and management of global
4168 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
4169 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
4170 index counters. The API functions that use this state have been changed
4171 to take a "class_index" rather than pointers to the class's local STACK
4172 and counter, and there is now an API function to dynamically create new
4173 classes. This centralisation allows us to (a) plug a lot of the
4174 thread-safety problems that existed, and (b) makes it possible to clean
4175 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
4176 such data would previously have always leaked in application code and
4177 workarounds were in place to make the memory debugging turn a blind eye
4178 to it. Application code that doesn't use this new function will still
4179 leak as before, but their memory debugging output will announce it now
4180 rather than letting it slide.
4182 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
4183 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
4184 has a return value to indicate success or failure.
4187 *) Make it possible to replace the underlying "ERR" functions such that the
4188 global state (2 LHASH tables and 2 locks) is only used by the "default"
4189 implementation. This change also adds two functions to "get" and "set"
4190 the implementation prior to it being automatically set the first time
4191 any other ERR function takes place. Ie. an application can call "get",
4192 pass the return value to a module it has just loaded, and that module
4193 can call its own "set" function using that value. This means the
4194 module's "ERR" operations will use (and modify) the error state in the
4195 application and not in its own statically linked copy of OpenSSL code.
4198 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
4199 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
4200 the operation, and provides a more encapsulated way for external code
4201 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
4202 to use these functions rather than manually incrementing the counts.
4204 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
4207 *) Add EVP test program.
4210 *) Add symmetric cipher support to ENGINE. Expect the API to change!
4213 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
4214 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
4215 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
4216 These allow a CRL to be built without having to access X509_CRL fields
4217 directly. Modify 'ca' application to use new functions.
4220 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
4221 bug workarounds. Rollback attack detection is a security feature.
4222 The problem will only arise on OpenSSL servers when TLSv1 is not
4223 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
4224 Software authors not wanting to support TLSv1 will have special reasons
4225 for their choice and can explicitly enable this option.
4226 [Bodo Moeller, Lutz Jaenicke]
4228 *) Rationalise EVP so it can be extended: don't include a union of
4229 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
4230 (similar to those existing for EVP_CIPHER_CTX).
4235 EVP_MD_CTX_init(&md); /* new function call */
4236 EVP_DigestInit(&md, EVP_sha1());
4237 EVP_DigestUpdate(&md, in, len);
4238 EVP_DigestFinal(&md, out, NULL);
4239 EVP_MD_CTX_cleanup(&md); /* new function call */
4243 *) Make DES key schedule conform to the usual scheme, as well as
4244 correcting its structure. This means that calls to DES functions
4245 now have to pass a pointer to a des_key_schedule instead of a
4246 plain des_key_schedule (which was actually always a pointer
4249 des_key_schedule ks;
4251 des_set_key_checked(..., &ks);
4252 des_ncbc_encrypt(..., &ks, ...);
4254 (Note that a later change renames 'des_...' into 'DES_...'.)
4257 *) Initial reduction of linker bloat: the use of some functions, such as
4258 PEM causes large amounts of unused functions to be linked in due to
4259 poor organisation. For example pem_all.c contains every PEM function
4260 which has a knock on effect of linking in large amounts of (unused)
4261 ASN1 code. Grouping together similar functions and splitting unrelated
4262 functions prevents this.
4265 *) Cleanup of EVP macros.
4268 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
4269 correct _ecb suffix.
4272 *) Add initial OCSP responder support to ocsp application. The
4273 revocation information is handled using the text based index
4274 use by the ca application. The responder can either handle
4275 requests generated internally, supplied in files (for example
4276 via a CGI script) or using an internal minimal server.
4279 *) Add configuration choices to get zlib compression for TLS.
4282 *) Changes to Kerberos SSL for RFC 2712 compliance:
4283 1. Implemented real KerberosWrapper, instead of just using
4284 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
4285 2. Implemented optional authenticator field of KerberosWrapper.
4287 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
4288 and authenticator structs; see crypto/krb5/.
4290 Generalized Kerberos calls to support multiple Kerberos libraries.
4291 [Vern Staats <staatsvr@asc.hpc.mil>,
4292 Jeffrey Altman <jaltman@columbia.edu>
4293 via Richard Levitte]
4295 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
4296 already does with RSA. testdsa.h now has 'priv_key/pub_key'
4297 values for each of the key sizes rather than having just
4298 parameters (and 'speed' generating keys each time).
4301 *) Speed up EVP routines.
4304 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
4305 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
4306 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
4307 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
4309 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
4310 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
4311 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
4314 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
4316 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
4319 *) Added the OS2-EMX target.
4320 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
4322 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
4323 to support NCONF routines in extension code. New function CONF_set_nconf()
4324 to allow functions which take an NCONF to also handle the old LHASH
4325 structure: this means that the old CONF compatible routines can be
4326 retained (in particular wrt extensions) without having to duplicate the
4327 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
4330 *) Enhance the general user interface with mechanisms for inner control
4331 and with possibilities to have yes/no kind of prompts.
4334 *) Change all calls to low level digest routines in the library and
4335 applications to use EVP. Add missing calls to HMAC_cleanup() and
4336 don't assume HMAC_CTX can be copied using memcpy().
4337 [Verdon Walker <VWalker@novell.com>, Steve Henson]
4339 *) Add the possibility to control engines through control names but with
4340 arbitrary arguments instead of just a string.
4341 Change the key loaders to take a UI_METHOD instead of a callback
4342 function pointer. NOTE: this breaks binary compatibility with earlier
4343 versions of OpenSSL [engine].
4344 Adapt the nCipher code for these new conditions and add a card insertion
4348 *) Enhance the general user interface with mechanisms to better support
4349 dialog box interfaces, application-defined prompts, the possibility
4350 to use defaults (for example default passwords from somewhere else)
4351 and interrupts/cancellations.
4354 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
4355 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
4358 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
4359 tidy up some unnecessarily weird code in 'sk_new()').
4360 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
4362 *) Change the key loading routines for ENGINEs to use the same kind
4363 callback (pem_password_cb) as all other routines that need this
4367 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
4368 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
4369 than this minimum value is recommended.
4372 *) New random seeder for OpenVMS, using the system process statistics
4373 that are easily reachable.
4376 *) Windows apparently can't transparently handle global
4377 variables defined in DLLs. Initialisations such as:
4379 const ASN1_ITEM *it = &ASN1_INTEGER_it;
4381 wont compile. This is used by the any applications that need to
4382 declare their own ASN1 modules. This was fixed by adding the option
4383 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
4384 needed for static libraries under Win32.
4387 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
4388 setting of purpose and trust fields. New X509_STORE trust and
4389 purpose functions and tidy up setting in other SSL functions.
4392 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
4393 structure. These are inherited by X509_STORE_CTX when it is
4394 initialised. This allows various defaults to be set in the
4395 X509_STORE structure (such as flags for CRL checking and custom
4396 purpose or trust settings) for functions which only use X509_STORE_CTX
4397 internally such as S/MIME.
4399 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
4400 trust settings if they are not set in X509_STORE. This allows X509_STORE
4401 purposes and trust (in S/MIME for example) to override any set by default.
4403 Add command line options for CRL checking to smime, s_client and s_server
4407 *) Initial CRL based revocation checking. If the CRL checking flag(s)
4408 are set then the CRL is looked up in the X509_STORE structure and
4409 its validity and signature checked, then if the certificate is found
4410 in the CRL the verify fails with a revoked error.
4412 Various new CRL related callbacks added to X509_STORE_CTX structure.
4414 Command line options added to 'verify' application to support this.
4416 This needs some additional work, such as being able to handle multiple
4417 CRLs with different times, extension based lookup (rather than just
4418 by subject name) and ultimately more complete V2 CRL extension
4422 *) Add a general user interface API (crypto/ui/). This is designed
4423 to replace things like des_read_password and friends (backward
4424 compatibility functions using this new API are provided).
4425 The purpose is to remove prompting functions from the DES code
4426 section as well as provide for prompting through dialog boxes in
4427 a window system and the like.
4430 *) Add "ex_data" support to ENGINE so implementations can add state at a
4431 per-structure level rather than having to store it globally.
4434 *) Make it possible for ENGINE structures to be copied when retrieved by
4435 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
4436 This causes the "original" ENGINE structure to act like a template,
4437 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
4438 operational state can be localised to each ENGINE structure, despite the
4439 fact they all share the same "methods". New ENGINE structures returned in
4440 this case have no functional references and the return value is the single
4441 structural reference. This matches the single structural reference returned
4442 by ENGINE_by_id() normally, when it is incremented on the pre-existing
4446 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
4447 needs to match any other type at all we need to manually clear the
4451 *) Changes to the "openssl engine" utility to include;
4452 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
4453 about an ENGINE's available control commands.
4454 - executing control commands from command line arguments using the
4455 '-pre' and '-post' switches. '-post' is only used if '-t' is
4456 specified and the ENGINE is successfully initialised. The syntax for
4457 the individual commands are colon-separated, for example;
4458 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
4461 *) New dynamic control command support for ENGINEs. ENGINEs can now
4462 declare their own commands (numbers), names (strings), descriptions,
4463 and input types for run-time discovery by calling applications. A
4464 subset of these commands are implicitly classed as "executable"
4465 depending on their input type, and only these can be invoked through
4466 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
4467 can be based on user input, config files, etc). The distinction is
4468 that "executable" commands cannot return anything other than a boolean
4469 result and can only support numeric or string input, whereas some
4470 discoverable commands may only be for direct use through
4471 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
4472 pointers, or other custom uses. The "executable" commands are to
4473 support parameterisations of ENGINE behaviour that can be
4474 unambiguously defined by ENGINEs and used consistently across any
4475 OpenSSL-based application. Commands have been added to all the
4476 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
4477 control over shared-library paths without source code alterations.
4480 *) Changed all ENGINE implementations to dynamically allocate their
4481 ENGINEs rather than declaring them statically. Apart from this being
4482 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
4483 this also allows the implementations to compile without using the
4484 internal engine_int.h header.
4487 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
4488 'const' value. Any code that should be able to modify a RAND_METHOD
4489 should already have non-const pointers to it (ie. they should only
4490 modify their own ones).
4493 *) Made a variety of little tweaks to the ENGINE code.
4494 - "atalla" and "ubsec" string definitions were moved from header files
4495 to C code. "nuron" string definitions were placed in variables
4496 rather than hard-coded - allowing parameterisation of these values
4497 later on via ctrl() commands.
4498 - Removed unused "#if 0"'d code.
4499 - Fixed engine list iteration code so it uses ENGINE_free() to release
4500 structural references.
4501 - Constified the RAND_METHOD element of ENGINE structures.
4502 - Constified various get/set functions as appropriate and added
4503 missing functions (including a catch-all ENGINE_cpy that duplicates
4504 all ENGINE values onto a new ENGINE except reference counts/state).
4505 - Removed NULL parameter checks in get/set functions. Setting a method
4506 or function to NULL is a way of cancelling out a previously set
4507 value. Passing a NULL ENGINE parameter is just plain stupid anyway
4508 and doesn't justify the extra error symbols and code.
4509 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
4510 flags from engine_int.h to engine.h.
4511 - Changed prototypes for ENGINE handler functions (init(), finish(),
4512 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
4515 *) Implement binary inversion algorithm for BN_mod_inverse in addition
4516 to the algorithm using long division. The binary algorithm can be
4517 used only if the modulus is odd. On 32-bit systems, it is faster
4518 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
4519 roughly 5-15% for 256-bit moduli), so we use it only for moduli
4520 up to 450 bits. In 64-bit environments, the binary algorithm
4521 appears to be advantageous for much longer moduli; here we use it
4522 for moduli up to 2048 bits.
4525 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
4526 could not support the combine flag in choice fields.
4529 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
4530 extensions from a certificate request to the certificate.
4533 *) Allow multiple 'certopt' and 'nameopt' options to be separated
4534 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
4535 file: this allows the display of the certificate about to be
4536 signed to be customised, to allow certain fields to be included
4537 or excluded and extension details. The old system didn't display
4538 multicharacter strings properly, omitted fields not in the policy
4539 and couldn't display additional details such as extensions.
4542 *) Function EC_POINTs_mul for multiple scalar multiplication
4543 of an arbitrary number of elliptic curve points
4544 \sum scalars[i]*points[i],
4545 optionally including the generator defined for the EC_GROUP:
4546 scalar*generator + \sum scalars[i]*points[i].
4548 EC_POINT_mul is a simple wrapper function for the typical case
4549 that the point list has just one item (besides the optional
4553 *) First EC_METHODs for curves over GF(p):
4555 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
4556 operations and provides various method functions that can also
4557 operate with faster implementations of modular arithmetic.
4559 EC_GFp_mont_method() reuses most functions that are part of
4560 EC_GFp_simple_method, but uses Montgomery arithmetic.
4562 [Bodo Moeller; point addition and point doubling
4563 implementation directly derived from source code provided by
4564 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
4566 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
4567 crypto/ec/ec_lib.c):
4569 Curves are EC_GROUP objects (with an optional group generator)
4570 based on EC_METHODs that are built into the library.
4572 Points are EC_POINT objects based on EC_GROUP objects.
4574 Most of the framework would be able to handle curves over arbitrary
4575 finite fields, but as there are no obvious types for fields other
4576 than GF(p), some functions are limited to that for now.
4579 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
4580 that the file contains a complete HTTP response.
4583 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
4584 change the def and num file printf format specifier from "%-40sXXX"
4585 to "%-39s XXX". The latter will always guarantee a space after the
4586 field while the former will cause them to run together if the field
4587 is 40 of more characters long.
4590 *) Constify the cipher and digest 'method' functions and structures
4591 and modify related functions to take constant EVP_MD and EVP_CIPHER
4595 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
4596 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
4599 *) Modify EVP_Digest*() routines so they now return values. Although the
4600 internal software routines can never fail additional hardware versions
4604 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
4606 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
4607 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
4610 ERR_R_NESTED_ASN1_ERROR
4612 ERR_R_MISSING_ASN1_EOS
4613 were 4 .. 9, conflicting with
4614 ERR_LIB_RSA (= ERR_R_RSA_LIB)
4616 ERR_LIB_PEM (= ERR_R_PEM_LIB).
4617 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
4619 Add new error code 'ERR_R_INTERNAL_ERROR'.
4622 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
4626 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
4627 sets the subject name for a new request or supersedes the
4628 subject name in a given request. Formats that can be parsed are
4629 'CN=Some Name, OU=myOU, C=IT'
4631 'CN=Some Name/OU=myOU/C=IT'.
4633 Add options '-batch' and '-verbose' to 'openssl req'.
4634 [Massimiliano Pala <madwolf@hackmasters.net>]
4636 *) Introduce the possibility to access global variables through
4637 functions on platform were that's the best way to handle exporting
4638 global variables in shared libraries. To enable this functionality,
4639 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
4640 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
4641 is normally done by Configure or something similar).
4643 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
4644 in the source file (foo.c) like this:
4646 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
4647 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
4649 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
4650 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
4652 OPENSSL_DECLARE_GLOBAL(int,foo);
4653 #define foo OPENSSL_GLOBAL_REF(foo)
4654 OPENSSL_DECLARE_GLOBAL(double,bar);
4655 #define bar OPENSSL_GLOBAL_REF(bar)
4657 The #defines are very important, and therefore so is including the
4658 header file everywhere where the defined globals are used.
4660 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
4661 of ASN.1 items, but that structure is a bit different.
4663 The largest change is in util/mkdef.pl which has been enhanced with
4664 better and easier to understand logic to choose which symbols should
4665 go into the Windows .def files as well as a number of fixes and code
4666 cleanup (among others, algorithm keywords are now sorted
4667 lexicographically to avoid constant rewrites).
4670 *) In BN_div() keep a copy of the sign of 'num' before writing the
4671 result to 'rm' because if rm==num the value will be overwritten
4672 and produce the wrong result if 'num' is negative: this caused
4673 problems with BN_mod() and BN_nnmod().
4676 *) Function OCSP_request_verify(). This checks the signature on an
4677 OCSP request and verifies the signer certificate. The signer
4678 certificate is just checked for a generic purpose and OCSP request
4682 *) Add OCSP_check_validity() function to check the validity of OCSP
4683 responses. OCSP responses are prepared in real time and may only
4684 be a few seconds old. Simply checking that the current time lies
4685 between thisUpdate and nextUpdate max reject otherwise valid responses
4686 caused by either OCSP responder or client clock inaccuracy. Instead
4687 we allow thisUpdate and nextUpdate to fall within a certain period of
4688 the current time. The age of the response can also optionally be
4689 checked. Two new options -validity_period and -status_age added to
4693 *) If signature or public key algorithm is unrecognized print out its
4694 OID rather that just UNKNOWN.
4697 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
4698 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
4699 ID to be generated from the issuer certificate alone which can then be
4700 passed to OCSP_id_issuer_cmp().
4703 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
4704 ASN1 modules to export functions returning ASN1_ITEM pointers
4705 instead of the ASN1_ITEM structures themselves. This adds several
4706 new macros which allow the underlying ASN1 function/structure to
4707 be accessed transparently. As a result code should not use ASN1_ITEM
4708 references directly (such as &X509_it) but instead use the relevant
4709 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
4710 use of the new ASN1 code on platforms where exporting structures
4711 is problematical (for example in shared libraries) but exporting
4712 functions returning pointers to structures is not.
4715 *) Add support for overriding the generation of SSL/TLS session IDs.
4716 These callbacks can be registered either in an SSL_CTX or per SSL.
4717 The purpose of this is to allow applications to control, if they wish,
4718 the arbitrary values chosen for use as session IDs, particularly as it
4719 can be useful for session caching in multiple-server environments. A
4720 command-line switch for testing this (and any client code that wishes
4721 to use such a feature) has been added to "s_server".
4722 [Geoff Thorpe, Lutz Jaenicke]
4724 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
4725 of the form '#if defined(...) || defined(...) || ...' and
4726 '#if !defined(...) && !defined(...) && ...'. This also avoids
4727 the growing number of special cases it was previously handling.
4730 *) Make all configuration macros available for application by making
4731 sure they are available in opensslconf.h, by giving them names starting
4732 with "OPENSSL_" to avoid conflicts with other packages and by making
4733 sure e_os2.h will cover all platform-specific cases together with
4735 Additionally, it is now possible to define configuration/platform-
4736 specific names (called "system identities"). In the C code, these
4737 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
4738 macro with the name beginning with "OPENSSL_SYS_", which is determined
4739 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
4743 *) New option -set_serial to 'req' and 'x509' this allows the serial
4744 number to use to be specified on the command line. Previously self
4745 signed certificates were hard coded with serial number 0 and the
4746 CA options of 'x509' had to use a serial number in a file which was
4750 *) New options to 'ca' utility to support V2 CRL entry extensions.
4751 Currently CRL reason, invalidity date and hold instruction are
4752 supported. Add new CRL extensions to V3 code and some new objects.
4755 *) New function EVP_CIPHER_CTX_set_padding() this is used to
4756 disable standard block padding (aka PKCS#5 padding) in the EVP
4757 API, which was previously mandatory. This means that the data is
4758 not padded in any way and so the total length much be a multiple
4759 of the block size, otherwise an error occurs.
4762 *) Initial (incomplete) OCSP SSL support.
4765 *) New function OCSP_parse_url(). This splits up a URL into its host,
4766 port and path components: primarily to parse OCSP URLs. New -url
4767 option to ocsp utility.
4770 *) New nonce behavior. The return value of OCSP_check_nonce() now
4771 reflects the various checks performed. Applications can decide
4772 whether to tolerate certain situations such as an absent nonce
4773 in a response when one was present in a request: the ocsp application
4774 just prints out a warning. New function OCSP_add1_basic_nonce()
4775 this is to allow responders to include a nonce in a response even if
4776 the request is nonce-less.
4779 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
4780 skipped when using openssl x509 multiple times on a single input file,
4781 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
4784 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
4785 set string type: to handle setting ASN1_TIME structures. Fix ca
4786 utility to correctly initialize revocation date of CRLs.
4789 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
4790 the clients preferred ciphersuites and rather use its own preferences.
4791 Should help to work around M$ SGC (Server Gated Cryptography) bug in
4792 Internet Explorer by ensuring unchanged hash method during stepup.
4793 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
4796 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
4797 to aes and add a new 'exist' option to print out symbols that don't
4801 *) Additional options to ocsp utility to allow flags to be set and
4802 additional certificates supplied.
4805 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
4806 OCSP client a number of certificate to only verify the response
4810 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
4811 handle the new API. Currently only ECB, CBC modes supported. Add new
4814 Add TLS AES ciphersuites as described in RFC3268, "Advanced
4815 Encryption Standard (AES) Ciphersuites for Transport Layer
4816 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
4817 not enabled by default and were not part of the "ALL" ciphersuite
4818 alias because they were not yet official; they could be
4819 explicitly requested by specifying the "AESdraft" ciphersuite
4820 group alias. In the final release of OpenSSL 0.9.7, the group
4821 alias is called "AES" and is part of "ALL".)
4822 [Ben Laurie, Steve Henson, Bodo Moeller]
4824 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
4825 request to response.
4828 *) Functions for OCSP responders. OCSP_request_onereq_count(),
4829 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
4830 extract information from a certificate request. OCSP_response_create()
4831 creates a response and optionally adds a basic response structure.
4832 OCSP_basic_add1_status() adds a complete single response to a basic
4833 response and returns the OCSP_SINGLERESP structure just added (to allow
4834 extensions to be included for example). OCSP_basic_add1_cert() adds a
4835 certificate to a basic response and OCSP_basic_sign() signs a basic
4836 response with various flags. New helper functions ASN1_TIME_check()
4837 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
4838 (converts ASN1_TIME to GeneralizedTime).
4841 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
4842 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
4843 structure from a certificate. X509_pubkey_digest() digests the public_key
4844 contents: this is used in various key identifiers.
4847 *) Make sk_sort() tolerate a NULL argument.
4848 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
4850 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
4851 passed by the function are trusted implicitly. If any of them signed the
4852 response then it is assumed to be valid and is not verified.
4855 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
4856 to data. This was previously part of the PKCS7 ASN1 code. This
4857 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
4858 [Steve Henson, reported by Kenneth R. Robinette
4859 <support@securenetterm.com>]
4861 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
4862 routines: without these tracing memory leaks is very painful.
4863 Fix leaks in PKCS12 and PKCS7 routines.
4866 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
4867 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
4868 effectively meant GeneralizedTime would never be used. Now it
4869 is initialised to -1 but X509_time_adj() now has to check the value
4870 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
4871 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
4872 [Steve Henson, reported by Kenneth R. Robinette
4873 <support@securenetterm.com>]
4875 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
4876 result in a zero length in the ASN1_INTEGER structure which was
4877 not consistent with the structure when d2i_ASN1_INTEGER() was used
4878 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
4879 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
4880 where it did not print out a minus for negative ASN1_INTEGER.
4883 *) Add summary printout to ocsp utility. The various functions which
4884 convert status values to strings have been renamed to:
4885 OCSP_response_status_str(), OCSP_cert_status_str() and
4886 OCSP_crl_reason_str() and are no longer static. New options
4887 to verify nonce values and to disable verification. OCSP response
4888 printout format cleaned up.
4891 *) Add additional OCSP certificate checks. These are those specified
4892 in RFC2560. This consists of two separate checks: the CA of the
4893 certificate being checked must either be the OCSP signer certificate
4894 or the issuer of the OCSP signer certificate. In the latter case the
4895 OCSP signer certificate must contain the OCSP signing extended key
4896 usage. This check is performed by attempting to match the OCSP
4897 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
4898 in the OCSP_CERTID structures of the response.
4901 *) Initial OCSP certificate verification added to OCSP_basic_verify()
4902 and related routines. This uses the standard OpenSSL certificate
4903 verify routines to perform initial checks (just CA validity) and
4904 to obtain the certificate chain. Then additional checks will be
4905 performed on the chain. Currently the root CA is checked to see
4906 if it is explicitly trusted for OCSP signing. This is used to set
4907 a root CA as a global signing root: that is any certificate that
4908 chains to that CA is an acceptable OCSP signing certificate.
4911 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
4912 extensions from a separate configuration file.
4913 As when reading extensions from the main configuration file,
4914 the '-extensions ...' option may be used for specifying the
4916 [Massimiliano Pala <madwolf@comune.modena.it>]
4918 *) New OCSP utility. Allows OCSP requests to be generated or
4919 read. The request can be sent to a responder and the output
4920 parsed, outputed or printed in text form. Not complete yet:
4921 still needs to check the OCSP response validity.
4924 *) New subcommands for 'openssl ca':
4925 'openssl ca -status <serial>' prints the status of the cert with
4926 the given serial number (according to the index file).
4927 'openssl ca -updatedb' updates the expiry status of certificates
4929 [Massimiliano Pala <madwolf@comune.modena.it>]
4931 *) New '-newreq-nodes' command option to CA.pl. This is like
4932 '-newreq', but calls 'openssl req' with the '-nodes' option
4933 so that the resulting key is not encrypted.
4934 [Damien Miller <djm@mindrot.org>]
4936 *) New configuration for the GNU Hurd.
4937 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
4939 *) Initial code to implement OCSP basic response verify. This
4940 is currently incomplete. Currently just finds the signer's
4941 certificate and verifies the signature on the response.
4944 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
4945 value of OPENSSLDIR. This is available via the new '-d' option
4946 to 'openssl version', and is also included in 'openssl version -a'.
4949 *) Allowing defining memory allocation callbacks that will be given
4950 file name and line number information in additional arguments
4951 (a const char* and an int). The basic functionality remains, as
4952 well as the original possibility to just replace malloc(),
4953 realloc() and free() by functions that do not know about these
4954 additional arguments. To register and find out the current
4955 settings for extended allocation functions, the following
4956 functions are provided:
4958 CRYPTO_set_mem_ex_functions
4959 CRYPTO_set_locked_mem_ex_functions
4960 CRYPTO_get_mem_ex_functions
4961 CRYPTO_get_locked_mem_ex_functions
4963 These work the same way as CRYPTO_set_mem_functions and friends.
4964 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
4965 extended allocation function is enabled.
4966 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
4967 a conventional allocation function is enabled.
4968 [Richard Levitte, Bodo Moeller]
4970 *) Finish off removing the remaining LHASH function pointer casts.
4971 There should no longer be any prototype-casting required when using
4972 the LHASH abstraction, and any casts that remain are "bugs". See
4973 the callback types and macros at the head of lhash.h for details
4974 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
4977 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
4978 If /dev/[u]random devices are not available or do not return enough
4979 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
4981 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
4982 /etc/entropy will be queried once each in this sequence, quering stops
4983 when enough entropy was collected without querying more sockets.
4986 *) Change the Unix RAND_poll() variant to be able to poll several
4987 random devices, as specified by DEVRANDOM, until a sufficient amount
4988 of data has been collected. We spend at most 10 ms on each file
4989 (select timeout) and read in non-blocking mode. DEVRANDOM now
4990 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
4991 (previously it was just the string "/dev/urandom"), so on typical
4992 platforms the 10 ms delay will never occur.
4993 Also separate out the Unix variant to its own file, rand_unix.c.
4994 For VMS, there's a currently-empty rand_vms.c.
4997 *) Move OCSP client related routines to ocsp_cl.c. These
4998 provide utility functions which an application needing
4999 to issue a request to an OCSP responder and analyse the
5000 response will typically need: as opposed to those which an
5001 OCSP responder itself would need which will be added later.
5003 OCSP_request_sign() signs an OCSP request with an API similar
5004 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
5005 response. OCSP_response_get1_basic() extracts basic response
5006 from response. OCSP_resp_find_status(): finds and extracts status
5007 information from an OCSP_CERTID structure (which will be created
5008 when the request structure is built). These are built from lower
5009 level functions which work on OCSP_SINGLERESP structures but
5010 wont normally be used unless the application wishes to examine
5011 extensions in the OCSP response for example.
5013 Replace nonce routines with a pair of functions.
5014 OCSP_request_add1_nonce() adds a nonce value and optionally
5015 generates a random value. OCSP_check_nonce() checks the
5016 validity of the nonce in an OCSP response.
5019 *) Change function OCSP_request_add() to OCSP_request_add0_id().
5020 This doesn't copy the supplied OCSP_CERTID and avoids the
5021 need to free up the newly created id. Change return type
5022 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
5023 This can then be used to add extensions to the request.
5024 Deleted OCSP_request_new(), since most of its functionality
5025 is now in OCSP_REQUEST_new() (and the case insensitive name
5026 clash) apart from the ability to set the request name which
5027 will be added elsewhere.
5030 *) Update OCSP API. Remove obsolete extensions argument from
5031 various functions. Extensions are now handled using the new
5032 OCSP extension code. New simple OCSP HTTP function which
5033 can be used to send requests and parse the response.
5036 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
5037 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
5038 uses the special reorder version of SET OF to sort the attributes
5039 and reorder them to match the encoded order. This resolves a long
5040 standing problem: a verify on a PKCS7 structure just after signing
5041 it used to fail because the attribute order did not match the
5042 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
5043 it uses the received order. This is necessary to tolerate some broken
5044 software that does not order SET OF. This is handled by encoding
5045 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
5046 to produce the required SET OF.
5049 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
5050 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
5051 files to get correct declarations of the ASN.1 item variables.
5054 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
5055 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
5056 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
5057 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
5058 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
5059 ASN1_ITEM and no wrapper functions.
5062 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
5063 replace the old function pointer based I/O routines. Change most of
5064 the *_d2i_bio() and *_d2i_fp() functions to use these.
5067 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
5068 lines, recognice more "algorithms" that can be deselected, and make
5069 it complain about algorithm deselection that isn't recognised.
5072 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
5073 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
5074 to use new functions. Add NO_ASN1_OLD which can be set to remove
5075 some old style ASN1 functions: this can be used to determine if old
5076 code will still work when these eventually go away.
5079 *) New extension functions for OCSP structures, these follow the
5080 same conventions as certificates and CRLs.
5083 *) New function X509V3_add1_i2d(). This automatically encodes and
5084 adds an extension. Its behaviour can be customised with various
5085 flags to append, replace or delete. Various wrappers added for
5086 certifcates and CRLs.
5089 *) Fix to avoid calling the underlying ASN1 print routine when
5090 an extension cannot be parsed. Correct a typo in the
5091 OCSP_SERVICELOC extension. Tidy up print OCSP format.
5094 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
5095 entries for variables.
5098 *) Add functionality to apps/openssl.c for detecting locking
5099 problems: As the program is single-threaded, all we have
5100 to do is register a locking callback using an array for
5101 storing which locks are currently held by the program.
5104 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
5105 SSL_get_ex_data_X509_STORE_idx(), which is used in
5106 ssl_verify_cert_chain() and thus can be called at any time
5107 during TLS/SSL handshakes so that thread-safety is essential.
5108 Unfortunately, the ex_data design is not at all suited
5109 for multi-threaded use, so it probably should be abolished.
5112 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5113 [Broadcom, tweaked and integrated by Geoff Thorpe]
5115 *) Move common extension printing code to new function
5116 X509V3_print_extensions(). Reorganise OCSP print routines and
5117 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5120 *) New function X509_signature_print() to remove duplication in some
5124 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5125 set (this was treated exactly the same as SET OF previously). This
5126 is used to reorder the STACK representing the structure to match the
5127 encoding. This will be used to get round a problem where a PKCS7
5128 structure which was signed could not be verified because the STACK
5129 order did not reflect the encoded order.
5132 *) Reimplement the OCSP ASN1 module using the new code.
5135 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5136 for its ASN1 operations. The old style function pointers still exist
5137 for now but they will eventually go away.
5140 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
5141 completely replaces the old ASN1 functionality with a table driven
5142 encoder and decoder which interprets an ASN1_ITEM structure describing
5143 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
5144 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
5145 has also been converted to the new form.
5148 *) Change BN_mod_exp_recp so that negative moduli are tolerated
5149 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
5150 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
5151 for negative moduli.
5154 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
5155 of not touching the result's sign bit.
5158 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
5162 *) Changed the LHASH code to use prototypes for callbacks, and created
5163 macros to declare and implement thin (optionally static) functions
5164 that provide type-safety and avoid function pointer casting for the
5165 type-specific callbacks.
5168 *) Added Kerberos Cipher Suites to be used with TLS, as written in
5170 [Veers Staats <staatsvr@asc.hpc.mil>,
5171 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
5173 *) Reformat the FAQ so the different questions and answers can be divided
5174 in sections depending on the subject.
5177 *) Have the zlib compression code load ZLIB.DLL dynamically under
5181 *) New function BN_mod_sqrt for computing square roots modulo a prime
5182 (using the probabilistic Tonelli-Shanks algorithm unless
5183 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
5184 be handled deterministically).
5185 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5187 *) Make BN_mod_inverse faster by explicitly handling small quotients
5188 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
5189 512 bits], about 30% for larger ones [1024 or 2048 bits].)
5192 *) New function BN_kronecker.
5195 *) Fix BN_gcd so that it works on negative inputs; the result is
5196 positive unless both parameters are zero.
5197 Previously something reasonably close to an infinite loop was
5198 possible because numbers could be growing instead of shrinking
5199 in the implementation of Euclid's algorithm.
5202 *) Fix BN_is_word() and BN_is_one() macros to take into account the
5203 sign of the number in question.
5205 Fix BN_is_word(a,w) to work correctly for w == 0.
5207 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
5208 because its test if the absolute value of 'a' equals 'w'.
5209 Note that BN_abs_is_word does *not* handle w == 0 reliably;
5210 it exists mostly for use in the implementations of BN_is_zero(),
5211 BN_is_one(), and BN_is_word().
5214 *) New function BN_swap.
5217 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
5218 the exponentiation functions are more likely to produce reasonable
5219 results on negative inputs.
5222 *) Change BN_mod_mul so that the result is always non-negative.
5223 Previously, it could be negative if one of the factors was negative;
5224 I don't think anyone really wanted that behaviour.
5227 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
5228 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
5229 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
5230 and add new functions:
5239 BN_mod_lshift1_quick
5243 These functions always generate non-negative results.
5245 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
5246 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
5248 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
5249 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
5250 be reduced modulo m.
5251 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5254 The following entry accidentily appeared in the CHANGES file
5255 distributed with OpenSSL 0.9.7. The modifications described in
5256 it do *not* apply to OpenSSL 0.9.7.
5258 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5259 was actually never needed) and in BN_mul(). The removal in BN_mul()
5260 required a small change in bn_mul_part_recursive() and the addition
5261 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5262 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5263 bn_sub_words() and bn_add_words() except they take arrays with
5268 *) In 'openssl passwd', verify passwords read from the terminal
5269 unless the '-salt' option is used (which usually means that
5270 verification would just waste user's time since the resulting
5271 hash is going to be compared with some given password hash)
5272 or the new '-noverify' option is used.
5274 This is an incompatible change, but it does not affect
5275 non-interactive use of 'openssl passwd' (passwords on the command
5276 line, '-stdin' option, '-in ...' option) and thus should not
5280 *) Remove all references to RSAref, since there's no more need for it.
5283 *) Make DSO load along a path given through an environment variable
5284 (SHLIB_PATH) with shl_load().
5287 *) Constify the ENGINE code as a result of BIGNUM constification.
5288 Also constify the RSA code and most things related to it. In a
5289 few places, most notable in the depth of the ASN.1 code, ugly
5290 casts back to non-const were required (to be solved at a later
5294 *) Make it so the openssl application has all engines loaded by default.
5297 *) Constify the BIGNUM routines a little more.
5300 *) Add the following functions:
5302 ENGINE_load_cswift()
5304 ENGINE_load_atalla()
5306 ENGINE_load_builtin_engines()
5308 That way, an application can itself choose if external engines that
5309 are built-in in OpenSSL shall ever be used or not. The benefit is
5310 that applications won't have to be linked with libdl or other dso
5311 libraries unless it's really needed.
5313 Changed 'openssl engine' to load all engines on demand.
5314 Changed the engine header files to avoid the duplication of some
5315 declarations (they differed!).
5318 *) 'openssl engine' can now list capabilities.
5321 *) Better error reporting in 'openssl engine'.
5324 *) Never call load_dh_param(NULL) in s_server.
5327 *) Add engine application. It can currently list engines by name and
5328 identity, and test if they are actually available.
5331 *) Improve RPM specification file by forcing symbolic linking and making
5332 sure the installed documentation is also owned by root.root.
5333 [Damien Miller <djm@mindrot.org>]
5335 *) Give the OpenSSL applications more possibilities to make use of
5336 keys (public as well as private) handled by engines.
5339 *) Add OCSP code that comes from CertCo.
5342 *) Add VMS support for the Rijndael code.
5345 *) Added untested support for Nuron crypto accelerator.
5348 *) Add support for external cryptographic devices. This code was
5349 previously distributed separately as the "engine" branch.
5350 [Geoff Thorpe, Richard Levitte]
5352 *) Rework the filename-translation in the DSO code. It is now possible to
5353 have far greater control over how a "name" is turned into a filename
5354 depending on the operating environment and any oddities about the
5355 different shared library filenames on each system.
5358 *) Support threads on FreeBSD-elf in Configure.
5361 *) Fix for SHA1 assembly problem with MASM: it produces
5362 warnings about corrupt line number information when assembling
5363 with debugging information. This is caused by the overlapping
5365 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
5368 NCONF_get_number() has no error checking at all. As a replacement,
5369 NCONF_get_number_e() is defined (_e for "error checking") and is
5370 promoted strongly. The old NCONF_get_number is kept around for
5371 binary backward compatibility.
5372 Make it possible for methods to load from something other than a BIO,
5373 by providing a function pointer that is given a name instead of a BIO.
5374 For example, this could be used to load configuration data from an
5378 *) Fix for non blocking accept BIOs. Added new I/O special reason
5379 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
5380 with non blocking I/O was not possible because no retry code was
5381 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
5385 *) Added the beginnings of Rijndael support.
5388 *) Fix for bug in DirectoryString mask setting. Add support for
5389 X509_NAME_print_ex() in 'req' and X509_print_ex() function
5390 to allow certificate printing to more controllable, additional
5391 'certopt' option to 'x509' to allow new printing options to be
5395 *) Clean old EAY MD5 hack from e_os.h.
5398 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
5400 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5401 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5402 [Joe Orton, Steve Henson]
5404 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
5406 *) Fix additional bug revealed by the NISCC test suite:
5408 Stop bug triggering large recursion when presented with
5409 certain ASN.1 tags (CVE-2003-0851)
5412 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
5414 *) Fix various bugs revealed by running the NISCC test suite:
5416 Stop out of bounds reads in the ASN1 code when presented with
5417 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5419 If verify callback ignores invalid public key errors don't try to check
5420 certificate signature with the NULL public key.
5424 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5425 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5429 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5430 extra data after the compression methods not only for TLS 1.0
5431 but also for SSL 3.0 (as required by the specification).
5432 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5434 *) Change X509_certificate_type() to mark the key as exported/exportable
5435 when it's 512 *bits* long, not 512 bytes.
5438 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
5440 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5441 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5442 a protocol version number mismatch like a decryption error
5443 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5446 *) Turn on RSA blinding by default in the default implementation
5447 to avoid a timing attack. Applications that don't want it can call
5448 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5449 They would be ill-advised to do so in most cases.
5450 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5452 *) Change RSA blinding code so that it works when the PRNG is not
5453 seeded (in this case, the secret RSA exponent is abused as
5454 an unpredictable seed -- if it is not unpredictable, there
5455 is no point in blinding anyway). Make RSA blinding thread-safe
5456 by remembering the creator's thread ID in rsa->blinding and
5457 having all other threads use local one-time blinding factors
5458 (this requires more computation than sharing rsa->blinding, but
5459 avoids excessive locking; and if an RSA object is not shared
5460 between threads, blinding will still be very fast).
5463 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
5465 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5466 via timing by performing a MAC computation even if incorrrect
5467 block cipher padding has been found. This is a countermeasure
5468 against active attacks where the attacker has to distinguish
5469 between bad padding and a MAC verification error. (CVE-2003-0078)
5471 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5472 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5473 Martin Vuagnoux (EPFL, Ilion)]
5475 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
5477 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
5478 memory from it's contents. This is done with a counter that will
5479 place alternating values in each byte. This can be used to solve
5480 two issues: 1) the removal of calls to memset() by highly optimizing
5481 compilers, and 2) cleansing with other values than 0, since those can
5482 be read through on certain media, for example a swap space on disk.
5485 *) Bugfix: client side session caching did not work with external caching,
5486 because the session->cipher setting was not restored when reloading
5487 from the external cache. This problem was masked, when
5488 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
5489 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
5492 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
5493 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
5494 [Zeev Lieber <zeev-l@yahoo.com>]
5496 *) Undo an undocumented change introduced in 0.9.6e which caused
5497 repeated calls to OpenSSL_add_all_ciphers() and
5498 OpenSSL_add_all_digests() to be ignored, even after calling
5502 *) Change the default configuration reader to deal with last line not
5503 being properly terminated.
5506 *) Change X509_NAME_cmp() so it applies the special rules on handling
5507 DN values that are of type PrintableString, as well as RDNs of type
5508 emailAddress where the value has the type ia5String.
5509 [stefank@valicert.com via Richard Levitte]
5511 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
5512 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
5513 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
5514 the bitwise-OR of the two for use by the majority of applications
5515 wanting this behaviour, and update the docs. The documented
5516 behaviour and actual behaviour were inconsistent and had been
5517 changing anyway, so this is more a bug-fix than a behavioural
5519 [Geoff Thorpe, diagnosed by Nadav Har'El]
5521 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
5522 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
5525 *) Fix initialization code race conditions in
5526 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
5527 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
5528 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
5529 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
5530 ssl2_get_cipher_by_char(),
5531 ssl3_get_cipher_by_char().
5532 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
5534 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
5535 the cached sessions are flushed, as the remove_cb() might use ex_data
5536 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
5537 (see [openssl.org #212]).
5538 [Geoff Thorpe, Lutz Jaenicke]
5540 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
5541 length, instead of the encoding length to d2i_ASN1_OBJECT.
5544 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
5546 *) [In 0.9.6g-engine release:]
5547 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
5548 [Lynn Gazis <lgazis@rainbow.com>]
5550 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
5552 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
5553 and get fix the header length calculation.
5554 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
5555 Alon Kantor <alonk@checkpoint.com> (and others),
5558 *) Use proper error handling instead of 'assertions' in buffer
5559 overflow checks added in 0.9.6e. This prevents DoS (the
5560 assertions could call abort()).
5561 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
5563 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
5565 *) Add various sanity checks to asn1_get_length() to reject
5566 the ASN1 length bytes if they exceed sizeof(long), will appear
5567 negative or the content length exceeds the length of the
5569 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5571 *) Fix cipher selection routines: ciphers without encryption had no flags
5572 for the cipher strength set and where therefore not handled correctly
5573 by the selection routines (PR #130).
5576 *) Fix EVP_dsa_sha macro.
5580 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
5581 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
5582 that was added in OpenSSL 0.9.6d.
5584 As the countermeasure turned out to be incompatible with some
5585 broken SSL implementations, the new option is part of SSL_OP_ALL.
5586 SSL_OP_ALL is usually employed when compatibility with weird SSL
5587 implementations is desired (e.g. '-bugs' option to 's_client' and
5588 's_server'), so the new option is automatically set in many
5592 *) Changes in security patch:
5594 Changes marked "(CHATS)" were sponsored by the Defense Advanced
5595 Research Projects Agency (DARPA) and Air Force Research Laboratory,
5596 Air Force Materiel Command, USAF, under agreement number
5599 *) Add various sanity checks to asn1_get_length() to reject
5600 the ASN1 length bytes if they exceed sizeof(long), will appear
5601 negative or the content length exceeds the length of the
5602 supplied buffer. (CVE-2002-0659)
5603 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5605 *) Assertions for various potential buffer overflows, not known to
5607 [Ben Laurie (CHATS)]
5609 *) Various temporary buffers to hold ASCII versions of integers were
5610 too small for 64 bit platforms. (CVE-2002-0655)
5611 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
5613 *) Remote buffer overflow in SSL3 protocol - an attacker could
5614 supply an oversized session ID to a client. (CVE-2002-0656)
5615 [Ben Laurie (CHATS)]
5617 *) Remote buffer overflow in SSL2 protocol - an attacker could
5618 supply an oversized client master key. (CVE-2002-0656)
5619 [Ben Laurie (CHATS)]
5621 Changes between 0.9.6c and 0.9.6d [9 May 2002]
5623 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
5624 encoded as NULL) with id-dsa-with-sha1.
5625 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
5627 *) Check various X509_...() return values in apps/req.c.
5628 [Nils Larsch <nla@trustcenter.de>]
5630 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
5631 an end-of-file condition would erronously be flagged, when the CRLF
5632 was just at the end of a processed block. The bug was discovered when
5633 processing data through a buffering memory BIO handing the data to a
5634 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
5635 <ptsekov@syntrex.com> and Nedelcho Stanev.
5638 *) Implement a countermeasure against a vulnerability recently found
5639 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
5640 before application data chunks to avoid the use of known IVs
5641 with data potentially chosen by the attacker.
5644 *) Fix length checks in ssl3_get_client_hello().
5647 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
5648 to prevent ssl3_read_internal() from incorrectly assuming that
5649 ssl3_read_bytes() found application data while handshake
5650 processing was enabled when in fact s->s3->in_read_app_data was
5651 merely automatically cleared during the initial handshake.
5652 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
5654 *) Fix object definitions for Private and Enterprise: they were not
5655 recognized in their shortname (=lowercase) representation. Extend
5656 obj_dat.pl to issue an error when using undefined keywords instead
5657 of silently ignoring the problem (Svenning Sorensen
5658 <sss@sss.dnsalias.net>).
5661 *) Fix DH_generate_parameters() so that it works for 'non-standard'
5662 generators, i.e. generators other than 2 and 5. (Previously, the
5663 code did not properly initialise the 'add' and 'rem' values to
5664 BN_generate_prime().)
5666 In the new general case, we do not insist that 'generator' is
5667 actually a primitive root: This requirement is rather pointless;
5668 a generator of the order-q subgroup is just as good, if not
5672 *) Map new X509 verification errors to alerts. Discovered and submitted by
5673 Tom Wu <tom@arcot.com>.
5676 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
5677 returning non-zero before the data has been completely received
5678 when using non-blocking I/O.
5679 [Bodo Moeller; problem pointed out by John Hughes]
5681 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
5682 [Ben Laurie, Lutz Jaenicke]
5684 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
5685 Yoram Zahavi <YoramZ@gilian.com>).
5688 *) Add information about CygWin 1.3 and on, and preserve proper
5689 configuration for the versions before that.
5690 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5692 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
5693 check whether we deal with a copy of a session and do not delete from
5694 the cache in this case. Problem reported by "Izhar Shoshani Levi"
5695 <izhar@checkpoint.com>.
5698 *) Do not store session data into the internal session cache, if it
5699 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
5700 flag is set). Proposed by Aslam <aslam@funk.com>.
5703 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
5707 *) [In 0.9.6d-engine release:]
5708 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5709 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5711 *) Add the configuration target linux-s390x.
5712 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
5714 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
5715 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
5716 variable as an indication that a ClientHello message has been
5717 received. As the flag value will be lost between multiple
5718 invocations of ssl3_accept when using non-blocking I/O, the
5719 function may not be aware that a handshake has actually taken
5720 place, thus preventing a new session from being added to the
5723 To avoid this problem, we now set s->new_session to 2 instead of
5724 using a local variable.
5725 [Lutz Jaenicke, Bodo Moeller]
5727 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
5728 if the SSL_R_LENGTH_MISMATCH error is detected.
5729 [Geoff Thorpe, Bodo Moeller]
5731 *) New 'shared_ldflag' column in Configure platform table.
5734 *) Fix EVP_CIPHER_mode macro.
5735 ["Dan S. Camper" <dan@bti.net>]
5737 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
5738 type, we must throw them away by setting rr->length to 0.
5739 [D P Chang <dpc@qualys.com>]
5741 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
5743 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
5744 <Dominikus.Scherkl@biodata.com>. (The previous implementation
5745 worked incorrectly for those cases where range = 10..._2 and
5746 3*range is two bits longer than range.)
5749 *) Only add signing time to PKCS7 structures if it is not already
5753 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
5754 OBJ_ld_ce should be OBJ_id_ce.
5755 Also some ip-pda OIDs in crypto/objects/objects.txt were
5756 incorrect (cf. RFC 3039).
5757 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
5759 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
5760 returns early because it has nothing to do.
5761 [Andy Schneider <andy.schneider@bjss.co.uk>]
5763 *) [In 0.9.6c-engine release:]
5764 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
5765 [Andy Schneider <andy.schneider@bjss.co.uk>]
5767 *) [In 0.9.6c-engine release:]
5768 Add support for Cryptographic Appliance's keyserver technology.
5769 (Use engine 'keyclient')
5770 [Cryptographic Appliances and Geoff Thorpe]
5772 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
5773 is called via tools/c89.sh because arguments have to be
5774 rearranged (all '-L' options must appear before the first object
5776 [Richard Shapiro <rshapiro@abinitio.com>]
5778 *) [In 0.9.6c-engine release:]
5779 Add support for Broadcom crypto accelerator cards, backported
5781 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
5783 *) [In 0.9.6c-engine release:]
5784 Add support for SureWare crypto accelerator cards from
5785 Baltimore Technologies. (Use engine 'sureware')
5786 [Baltimore Technologies and Mark Cox]
5788 *) [In 0.9.6c-engine release:]
5789 Add support for crypto accelerator cards from Accelerated
5790 Encryption Processing, www.aep.ie. (Use engine 'aep')
5791 [AEP Inc. and Mark Cox]
5793 *) Add a configuration entry for gcc on UnixWare.
5794 [Gary Benson <gbenson@redhat.com>]
5796 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
5797 messages are stored in a single piece (fixed-length part and
5798 variable-length part combined) and fix various bugs found on the way.
5801 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
5802 instead. BIO_gethostbyname() does not know what timeouts are
5803 appropriate, so entries would stay in cache even when they have
5805 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
5807 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
5808 faced with a pathologically small ClientHello fragment that does
5809 not contain client_version: Instead of aborting with an error,
5810 simply choose the highest available protocol version (i.e.,
5811 TLS 1.0 unless it is disabled). In practice, ClientHello
5812 messages are never sent like this, but this change gives us
5813 strictly correct behaviour at least for TLS.
5816 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
5817 never resets s->method to s->ctx->method when called from within
5818 one of the SSL handshake functions.
5819 [Bodo Moeller; problem pointed out by Niko Baric]
5821 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
5822 (sent using the client's version number) if client_version is
5823 smaller than the protocol version in use. Also change
5824 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
5825 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
5826 the client will at least see that alert.
5829 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
5833 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
5834 client receives HelloRequest while in a handshake.
5835 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
5837 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
5838 should end in 'break', not 'goto end' which circuments various
5839 cleanups done in state SSL_ST_OK. But session related stuff
5840 must be disabled for SSL_ST_OK in the case that we just sent a
5843 Also avoid some overhead by not calling ssl_init_wbio_buffer()
5844 before just sending a HelloRequest.
5845 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
5847 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
5848 reveal whether illegal block cipher padding was found or a MAC
5849 verification error occured. (Neither SSLerr() codes nor alerts
5850 are directly visible to potential attackers, but the information
5851 may leak via logfiles.)
5853 Similar changes are not required for the SSL 2.0 implementation
5854 because the number of padding bytes is sent in clear for SSL 2.0,
5855 and the extra bytes are just ignored. However ssl/s2_pkt.c
5856 failed to verify that the purported number of padding bytes is in
5860 *) Add OpenUNIX-8 support including shared libraries
5861 (Boyd Lynn Gerber <gerberb@zenez.com>).
5864 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
5865 'wristwatch attack' using huge encoding parameters (cf.
5866 James H. Manger's CRYPTO 2001 paper). Note that the
5867 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
5868 encoding parameters and hence was not vulnerable.
5871 *) BN_sqr() bug fix.
5872 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
5874 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
5875 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
5876 followed by modular reduction.
5877 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
5879 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
5880 equivalent based on BN_pseudo_rand() instead of BN_rand().
5883 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
5884 This function was broken, as the check for a new client hello message
5885 to handle SGC did not allow these large messages.
5886 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
5889 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
5892 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
5893 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
5896 *) Rework the configuration and shared library support for Tru64 Unix.
5897 The configuration part makes use of modern compiler features and
5898 still retains old compiler behavior for those that run older versions
5899 of the OS. The shared library support part includes a variant that
5900 uses the RPATH feature, and is available through the special
5901 configuration target "alpha-cc-rpath", which will never be selected
5903 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
5905 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
5906 with the same message size as in ssl3_get_certificate_request().
5907 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
5908 messages might inadvertently be reject as too long.
5909 [Petr Lampa <lampa@fee.vutbr.cz>]
5911 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
5914 *) Modified SSL library such that the verify_callback that has been set
5915 specificly for an SSL object with SSL_set_verify() is actually being
5916 used. Before the change, a verify_callback set with this function was
5917 ignored and the verify_callback() set in the SSL_CTX at the time of
5918 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
5919 to allow the necessary settings.
5922 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
5923 explicitly to NULL, as at least on Solaris 8 this seems not always to be
5924 done automatically (in contradiction to the requirements of the C
5925 standard). This made problems when used from OpenSSH.
5928 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
5929 dh->length and always used
5931 BN_rand_range(priv_key, dh->p).
5933 BN_rand_range() is not necessary for Diffie-Hellman, and this
5934 specific range makes Diffie-Hellman unnecessarily inefficient if
5935 dh->length (recommended exponent length) is much smaller than the
5936 length of dh->p. We could use BN_rand_range() if the order of
5937 the subgroup was stored in the DH structure, but we only have
5942 BN_rand(priv_key, l, ...)
5944 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
5950 RSA_eay_public_encrypt
5951 RSA_eay_private_decrypt
5952 RSA_eay_private_encrypt (signing)
5953 RSA_eay_public_decrypt (signature verification)
5955 (default implementations for RSA_public_encrypt,
5956 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
5957 always reject numbers >= n.
5960 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
5961 to synchronize access to 'locking_thread'. This is necessary on
5962 systems where access to 'locking_thread' (an 'unsigned long'
5963 variable) is not atomic.
5966 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
5967 *before* setting the 'crypto_lock_rand' flag. The previous code had
5968 a race condition if 0 is a valid thread ID.
5969 [Travis Vitek <vitek@roguewave.com>]
5971 *) Add support for shared libraries under Irix.
5972 [Albert Chin-A-Young <china@thewrittenword.com>]
5974 *) Add configuration option to build on Linux on both big-endian and
5976 [Ralf Baechle <ralf@uni-koblenz.de>]
5978 *) Add the possibility to create shared libraries on HP-UX.
5981 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
5983 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
5984 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
5985 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
5986 PRNG state recovery was possible based on the output of
5987 one PRNG request appropriately sized to gain knowledge on
5988 'md' followed by enough consecutive 1-byte PRNG requests
5989 to traverse all of 'state'.
5991 1. When updating 'md_local' (the current thread's copy of 'md')
5992 during PRNG output generation, hash all of the previous
5993 'md_local' value, not just the half used for PRNG output.
5995 2. Make the number of bytes from 'state' included into the hash
5996 independent from the number of PRNG bytes requested.
5998 The first measure alone would be sufficient to avoid
5999 Markku-Juhani's attack. (Actually it had never occurred
6000 to me that the half of 'md_local' used for chaining was the
6001 half from which PRNG output bytes were taken -- I had always
6002 assumed that the secret half would be used.) The second
6003 measure makes sure that additional data from 'state' is never
6004 mixed into 'md_local' in small portions; this heuristically
6005 further strengthens the PRNG.
6008 *) Fix crypto/bn/asm/mips3.s.
6011 *) When only the key is given to "enc", the IV is undefined. Print out
6012 an error message in this case.
6015 *) Handle special case when X509_NAME is empty in X509 printing routines.
6018 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
6019 positive and less than q.
6022 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
6023 used: it isn't thread safe and the add_lock_callback should handle
6025 [Paul Rose <Paul.Rose@bridge.com>]
6027 *) Verify that incoming data obeys the block size in
6028 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
6032 [Ulf Möller, Bodo Möller]
6034 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
6035 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
6036 when fixing the server behaviour for backwards-compatible 'client
6037 hello' messages. (Note that the attack is impractical against
6038 SSL 3.0 and TLS 1.0 anyway because length and version checking
6039 means that the probability of guessing a valid ciphertext is
6040 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
6043 Before 0.9.5, the countermeasure (hide the error by generating a
6044 random 'decryption result') did not work properly because
6045 ERR_clear_error() was missing, meaning that SSL_get_error() would
6046 detect the supposedly ignored error.
6048 Both problems are now fixed.
6051 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
6052 (previously it was 1024).
6055 *) Fix for compatibility mode trust settings: ignore trust settings
6056 unless some valid trust or reject settings are present.
6059 *) Fix for blowfish EVP: its a variable length cipher.
6062 *) Fix various bugs related to DSA S/MIME verification. Handle missing
6063 parameters in DSA public key structures and return an error in the
6064 DSA routines if parameters are absent.
6067 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
6068 in the current directory if neither $RANDFILE nor $HOME was set.
6069 RAND_file_name() in 0.9.6a returned NULL in this case. This has
6070 caused some confusion to Windows users who haven't defined $HOME.
6071 Thus RAND_file_name() is changed again: e_os.h can define a
6072 DEFAULT_HOME, which will be used if $HOME is not set.
6073 For Windows, we use "C:"; on other platforms, we still require
6074 environment variables.
6076 *) Move 'if (!initialized) RAND_poll()' into regions protected by
6077 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
6078 having multiple threads call RAND_poll() concurrently.
6081 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
6082 combination of a flag and a thread ID variable.
6083 Otherwise while one thread is in ssleay_rand_bytes (which sets the
6084 flag), *other* threads can enter ssleay_add_bytes without obeying
6085 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
6086 that they do not hold after the first thread unsets add_do_not_lock).
6089 *) Change bctest again: '-x' expressions are not available in all
6093 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
6095 *) Fix a couple of memory leaks in PKCS7_dataDecode()
6096 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
6098 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
6099 the default extension for executables, if any. Also, make the perl
6100 scripts that use symlink() to test if it really exists and use "cp"
6101 if it doesn't. All this made OpenSSL compilable and installable in
6105 *) Fix for asn1_GetSequence() for indefinite length constructed data.
6106 If SEQUENCE is length is indefinite just set c->slen to the total
6107 amount of data available.
6108 [Steve Henson, reported by shige@FreeBSD.org]
6109 [This change does not apply to 0.9.7.]
6111 *) Change bctest to avoid here-documents inside command substitution
6112 (workaround for FreeBSD /bin/sh bug).
6113 For compatibility with Ultrix, avoid shell functions (introduced
6114 in the bctest version that searches along $PATH).
6117 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6118 with des_encrypt() defined on some operating systems, like Solaris
6122 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6123 On the Importance of Eliminating Errors in Cryptographic
6124 Computations, J. Cryptology 14 (2001) 2, 101-119,
6125 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6128 *) MIPS assembler BIGNUM division bug fix.
6131 *) Disabled incorrect Alpha assembler code.
6134 *) Fix PKCS#7 decode routines so they correctly update the length
6135 after reading an EOC for the EXPLICIT tag.
6137 [This change does not apply to 0.9.7.]
6139 *) Fix bug in PKCS#12 key generation routines. This was triggered
6140 if a 3DES key was generated with a 0 initial byte. Include
6141 PKCS12_BROKEN_KEYGEN compilation option to retain the old
6142 (but broken) behaviour.
6145 *) Enhance bctest to search for a working bc along $PATH and print
6147 [Tim Rice <tim@multitalents.net> via Richard Levitte]
6149 *) Fix memory leaks in err.c: free err_data string if necessary;
6150 don't write to the wrong index in ERR_set_error_data.
6153 *) Implement ssl23_peek (analogous to ssl23_read), which previously
6157 *) Replace rdtsc with _emit statements for VC++ version 5.
6158 [Jeremy Cooper <jeremy@baymoo.org>]
6160 *) Make it possible to reuse SSLv2 sessions.
6163 *) In copy_email() check for >= 0 as a return value for
6164 X509_NAME_get_index_by_NID() since 0 is a valid index.
6165 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
6167 *) Avoid coredump with unsupported or invalid public keys by checking if
6168 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
6169 PKCS7_verify() fails with non detached data.
6172 *) Don't use getenv in library functions when run as setuid/setgid.
6173 New function OPENSSL_issetugid().
6176 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
6177 due to incorrect handling of multi-threading:
6179 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
6181 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
6183 3. Count how many times MemCheck_off() has been called so that
6184 nested use can be treated correctly. This also avoids
6185 inband-signalling in the previous code (which relied on the
6186 assumption that thread ID 0 is impossible).
6189 *) Add "-rand" option also to s_client and s_server.
6192 *) Fix CPU detection on Irix 6.x.
6193 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
6194 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6196 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
6199 [This change does not apply to 0.9.7.]
6201 *) Use the cached encoding of an X509_NAME structure rather than
6202 copying it. This is apparently the reason for the libsafe "errors"
6203 but the code is actually correct.
6206 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
6207 Bleichenbacher's DSA attack.
6208 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
6209 to be set and top=0 forces the highest bit to be set; top=-1 is new
6210 and leaves the highest bit random.
6211 [Ulf Moeller, Bodo Moeller]
6213 *) In the NCONF_...-based implementations for CONF_... queries
6214 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
6215 a temporary CONF structure with the data component set to NULL
6216 (which gives segmentation faults in lh_retrieve).
6217 Instead, use NULL for the CONF pointer in CONF_get_string and
6218 CONF_get_number (which may use environment variables) and directly
6219 return NULL from CONF_get_section.
6222 *) Fix potential buffer overrun for EBCDIC.
6225 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
6226 keyUsage if basicConstraints absent for a CA.
6229 *) Make SMIME_write_PKCS7() write mail header values with a format that
6230 is more generally accepted (no spaces before the semicolon), since
6231 some programs can't parse those values properly otherwise. Also make
6232 sure BIO's that break lines after each write do not create invalid
6236 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
6237 macros previously used would not encode an empty SEQUENCE OF
6238 and break the signature.
6240 [This change does not apply to 0.9.7.]
6242 *) Zero the premaster secret after deriving the master secret in
6246 *) Add some EVP_add_digest_alias registrations (as found in
6247 OpenSSL_add_all_digests()) to SSL_library_init()
6248 aka OpenSSL_add_ssl_algorithms(). This provides improved
6249 compatibility with peers using X.509 certificates
6250 with unconventional AlgorithmIdentifier OIDs.
6253 *) Fix for Irix with NO_ASM.
6254 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6256 *) ./config script fixes.
6257 [Ulf Moeller, Richard Levitte]
6259 *) Fix 'openssl passwd -1'.
6262 *) Change PKCS12_key_gen_asc() so it can cope with non null
6263 terminated strings whose length is passed in the passlen
6264 parameter, for example from PEM callbacks. This was done
6265 by adding an extra length parameter to asc2uni().
6266 [Steve Henson, reported by <oddissey@samsung.co.kr>]
6268 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
6269 call failed, free the DSA structure.
6272 *) Fix to uni2asc() to cope with zero length Unicode strings.
6273 These are present in some PKCS#12 files.
6276 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
6277 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
6278 when writing a 32767 byte record.
6279 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
6281 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
6282 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
6284 (RSA objects have a reference count access to which is protected
6285 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
6286 so they are meant to be shared between threads.)
6287 [Bodo Moeller, Geoff Thorpe; original patch submitted by
6288 "Reddie, Steven" <Steven.Reddie@ca.com>]
6290 *) Fix a deadlock in CRYPTO_mem_leaks().
6293 *) Use better test patterns in bntest.
6296 *) rand_win.c fix for Borland C.
6299 *) BN_rshift bugfix for n == 0.
6302 *) Add a 'bctest' script that checks for some known 'bc' bugs
6303 so that 'make test' does not abort just because 'bc' is broken.
6306 *) Store verify_result within SSL_SESSION also for client side to
6307 avoid potential security hole. (Re-used sessions on the client side
6308 always resulted in verify_result==X509_V_OK, not using the original
6309 result of the server certificate verification.)
6312 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
6313 SSL3_RT_APPLICATION_DATA, return 0.
6314 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
6318 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
6319 releases, have been re-implemented by renaming the previous
6320 implementations of ssl2_read and ssl3_read to ssl2_read_internal
6321 and ssl3_read_internal, respectively, and adding 'peek' parameters
6322 to them. The new ssl[23]_{read,peek} functions are calls to
6323 ssl[23]_read_internal with the 'peek' flag set appropriately.
6324 A 'peek' parameter has also been added to ssl3_read_bytes, which
6325 does the actual work for ssl3_read_internal.
6328 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
6329 the method-specific "init()" handler. Also clean up ex_data after
6330 calling the method-specific "finish()" handler. Previously, this was
6331 happening the other way round.
6334 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
6335 The previous value, 12, was not always sufficient for BN_mod_exp().
6338 *) Make sure that shared libraries get the internal name engine with
6339 the full version number and not just 0. This should mark the
6340 shared libraries as not backward compatible. Of course, this should
6341 be changed again when we can guarantee backward binary compatibility.
6344 *) Fix typo in get_cert_by_subject() in by_dir.c
6345 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
6347 *) Rework the system to generate shared libraries:
6349 - Make note of the expected extension for the shared libraries and
6350 if there is a need for symbolic links from for example libcrypto.so.0
6351 to libcrypto.so.0.9.7. There is extended info in Configure for
6354 - Make as few rebuilds of the shared libraries as possible.
6356 - Still avoid linking the OpenSSL programs with the shared libraries.
6358 - When installing, install the shared libraries separately from the
6362 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
6364 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
6365 and not in SSL_clear because the latter is also used by the
6366 accept/connect functions; previously, the settings made by
6367 SSL_set_read_ahead would be lost during the handshake.
6368 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
6370 *) Correct util/mkdef.pl to be selective about disabled algorithms.
6371 Previously, it would create entries for disableed algorithms no
6375 *) Added several new manual pages for SSL_* function.
6378 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
6380 *) In ssl23_get_client_hello, generate an error message when faced
6381 with an initial SSL 3.0/TLS record that is too small to contain the
6382 first two bytes of the ClientHello message, i.e. client_version.
6383 (Note that this is a pathologic case that probably has never happened
6384 in real life.) The previous approach was to use the version number
6385 from the record header as a substitute; but our protocol choice
6386 should not depend on that one because it is not authenticated
6387 by the Finished messages.
6390 *) More robust randomness gathering functions for Windows.
6391 [Jeffrey Altman <jaltman@columbia.edu>]
6393 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
6394 not set then we don't setup the error code for issuer check errors
6395 to avoid possibly overwriting other errors which the callback does
6396 handle. If an application does set the flag then we assume it knows
6397 what it is doing and can handle the new informational codes
6401 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
6402 a general "ANY" type, as such it should be able to decode anything
6403 including tagged types. However it didn't check the class so it would
6404 wrongly interpret tagged types in the same way as their universal
6405 counterpart and unknown types were just rejected. Changed so that the
6406 tagged and unknown types are handled in the same way as a SEQUENCE:
6407 that is the encoding is stored intact. There is also a new type
6408 "V_ASN1_OTHER" which is used when the class is not universal, in this
6409 case we have no idea what the actual type is so we just lump them all
6413 *) On VMS, stdout may very well lead to a file that is written to
6414 in a record-oriented fashion. That means that every write() will
6415 write a separate record, which will be read separately by the
6416 programs trying to read from it. This can be very confusing.
6418 The solution is to put a BIO filter in the way that will buffer
6419 text until a linefeed is reached, and then write everything a
6420 line at a time, so every record written will be an actual line,
6421 not chunks of lines and not (usually doesn't happen, but I've
6422 seen it once) several lines in one record. BIO_f_linebuffer() is
6425 Currently, it's a VMS-only method, because that's where it has
6426 been tested well enough.
6429 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
6430 it can return incorrect results.
6431 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
6432 but it was in 0.9.6-beta[12].)
6435 *) Disable the check for content being present when verifying detached
6436 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
6437 include zero length content when signing messages.
6440 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
6441 BIO_ctrl (for BIO pairs).
6444 *) Add DSO method for VMS.
6447 *) Bug fix: Montgomery multiplication could produce results with the
6451 *) Add RPM specification openssl.spec and modify it to build three
6452 packages. The default package contains applications, application
6453 documentation and run-time libraries. The devel package contains
6454 include files, static libraries and function documentation. The
6455 doc package contains the contents of the doc directory. The original
6456 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
6459 *) Add a large number of documentation files for many SSL routines.
6460 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6462 *) Add a configuration entry for Sony News 4.
6463 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
6465 *) Don't set the two most significant bits to one when generating a
6466 random number < q in the DSA library.
6469 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
6470 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
6471 the underlying transport is blocking) if a handshake took place.
6472 (The default behaviour is needed by applications such as s_client
6473 and s_server that use select() to determine when to use SSL_read;
6474 but for applications that know in advance when to expect data, it
6475 just makes things more complicated.)
6478 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
6482 *) Add a few more EBCDIC conditionals that make `req' and `x509'
6483 work better on such systems.
6484 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6486 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
6487 Update PKCS12_parse() so it copies the friendlyName and the
6488 keyid to the certificates aux info.
6491 *) Fix bug in PKCS7_verify() which caused an infinite loop
6492 if there was more than one signature.
6493 [Sven Uszpelkat <su@celocom.de>]
6495 *) Major change in util/mkdef.pl to include extra information
6496 about each symbol, as well as presentig variables as well
6497 as functions. This change means that there's n more need
6498 to rebuild the .num files when some algorithms are excluded.
6501 *) Allow the verify time to be set by an application,
6502 rather than always using the current time.
6505 *) Phase 2 verify code reorganisation. The certificate
6506 verify code now looks up an issuer certificate by a
6507 number of criteria: subject name, authority key id
6508 and key usage. It also verifies self signed certificates
6509 by the same criteria. The main comparison function is
6510 X509_check_issued() which performs these checks.
6512 Lot of changes were necessary in order to support this
6513 without completely rewriting the lookup code.
6515 Authority and subject key identifier are now cached.
6517 The LHASH 'certs' is X509_STORE has now been replaced
6518 by a STACK_OF(X509_OBJECT). This is mainly because an
6519 LHASH can't store or retrieve multiple objects with
6520 the same hash value.
6522 As a result various functions (which were all internal
6523 use only) have changed to handle the new X509_STORE
6524 structure. This will break anything that messed round
6525 with X509_STORE internally.
6527 The functions X509_STORE_add_cert() now checks for an
6528 exact match, rather than just subject name.
6530 The X509_STORE API doesn't directly support the retrieval
6531 of multiple certificates matching a given criteria, however
6532 this can be worked round by performing a lookup first
6533 (which will fill the cache with candidate certificates)
6534 and then examining the cache for matches. This is probably
6535 the best we can do without throwing out X509_LOOKUP
6536 entirely (maybe later...).
6538 The X509_VERIFY_CTX structure has been enhanced considerably.
6540 All certificate lookup operations now go via a get_issuer()
6541 callback. Although this currently uses an X509_STORE it
6542 can be replaced by custom lookups. This is a simple way
6543 to bypass the X509_STORE hackery necessary to make this
6544 work and makes it possible to use more efficient techniques
6545 in future. A very simple version which uses a simple
6546 STACK for its trusted certificate store is also provided
6547 using X509_STORE_CTX_trusted_stack().
6549 The verify_cb() and verify() callbacks now have equivalents
6550 in the X509_STORE_CTX structure.
6552 X509_STORE_CTX also has a 'flags' field which can be used
6553 to customise the verify behaviour.
6556 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
6557 excludes S/MIME capabilities.
6560 *) When a certificate request is read in keep a copy of the
6561 original encoding of the signed data and use it when outputing
6562 again. Signatures then use the original encoding rather than
6563 a decoded, encoded version which may cause problems if the
6564 request is improperly encoded.
6567 *) For consistency with other BIO_puts implementations, call
6568 buffer_write(b, ...) directly in buffer_puts instead of calling
6571 In BIO_puts, increment b->num_write as in BIO_write.
6572 [Peter.Sylvester@EdelWeb.fr]
6574 *) Fix BN_mul_word for the case where the word is 0. (We have to use
6575 BN_zero, we may not return a BIGNUM with an array consisting of
6579 *) Avoid calling abort() from within the library when problems are
6580 detected, except if preprocessor symbols have been defined
6581 (such as REF_CHECK, BN_DEBUG etc.).
6584 *) New openssl application 'rsautl'. This utility can be
6585 used for low level RSA operations. DER public key
6586 BIO/fp routines also added.
6589 *) New Configure entry and patches for compiling on QNX 4.
6590 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
6592 *) A demo state-machine implementation was sponsored by
6593 Nuron (http://www.nuron.com/) and is now available in
6594 demos/state_machine.
6597 *) New options added to the 'dgst' utility for signature
6598 generation and verification.
6601 *) Unrecognized PKCS#7 content types are now handled via a
6602 catch all ASN1_TYPE structure. This allows unsupported
6603 types to be stored as a "blob" and an application can
6604 encode and decode it manually.
6607 *) Fix various signed/unsigned issues to make a_strex.c
6609 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
6611 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
6612 length if passed a buffer. ASN1_INTEGER_to_BN failed
6613 if passed a NULL BN and its argument was negative.
6614 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
6616 *) Modification to PKCS#7 encoding routines to output definite
6617 length encoding. Since currently the whole structures are in
6618 memory there's not real point in using indefinite length
6619 constructed encoding. However if OpenSSL is compiled with
6620 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
6623 *) Added BIO_vprintf() and BIO_vsnprintf().
6626 *) Added more prefixes to parse for in the the strings written
6627 through a logging bio, to cover all the levels that are available
6628 through syslog. The prefixes are now:
6630 PANIC, EMERG, EMR => LOG_EMERG
6631 ALERT, ALR => LOG_ALERT
6632 CRIT, CRI => LOG_CRIT
6633 ERROR, ERR => LOG_ERR
6634 WARNING, WARN, WAR => LOG_WARNING
6635 NOTICE, NOTE, NOT => LOG_NOTICE
6636 INFO, INF => LOG_INFO
6637 DEBUG, DBG => LOG_DEBUG
6639 and as before, if none of those prefixes are present at the
6640 beginning of the string, LOG_ERR is chosen.
6642 On Win32, the LOG_* levels are mapped according to this:
6644 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
6645 LOG_WARNING => EVENTLOG_WARNING_TYPE
6646 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
6650 *) Made it possible to reconfigure with just the configuration
6651 argument "reconf" or "reconfigure". The command line arguments
6652 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
6653 and are retrieved from there when reconfiguring.
6657 [Assar Westerlund <assar@sics.se>, Richard Levitte]
6659 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
6662 *) The obj_dat.pl script was messing up the sorting of object
6663 names. The reason was that it compared the quoted version
6664 of strings as a result "OCSP" > "OCSP Signing" because
6665 " > SPACE. Changed script to store unquoted versions of
6666 names and add quotes on output. It was also omitting some
6667 names from the lookup table if they were given a default
6668 value (that is if SN is missing it is given the same
6669 value as LN and vice versa), these are now added on the
6670 grounds that if an object has a name we should be able to
6671 look it up. Finally added warning output when duplicate
6672 short or long names are found.
6675 *) Changes needed for Tandem NSK.
6676 [Scott Uroff <scott@xypro.com>]
6678 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
6679 RSA_padding_check_SSLv23(), special padding was never detected
6680 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
6681 version rollback attacks was not effective.
6683 In s23_clnt.c, don't use special rollback-attack detection padding
6684 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
6685 client; similarly, in s23_srvr.c, don't do the rollback check if
6686 SSL 2.0 is the only protocol enabled in the server.
6689 *) Make it possible to get hexdumps of unprintable data with 'openssl
6690 asn1parse'. By implication, the functions ASN1_parse_dump() and
6691 BIO_dump_indent() are added.
6694 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
6695 these print out strings and name structures based on various
6696 flags including RFC2253 support and proper handling of
6697 multibyte characters. Added options to the 'x509' utility
6698 to allow the various flags to be set.
6701 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
6702 Also change the functions X509_cmp_current_time() and
6703 X509_gmtime_adj() work with an ASN1_TIME structure,
6704 this will enable certificates using GeneralizedTime in validity
6705 dates to be checked.
6708 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
6709 negative public key encodings) on by default,
6710 NO_NEG_PUBKEY_BUG can be set to disable it.
6713 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
6714 content octets. An i2c_ASN1_OBJECT is unnecessary because
6715 the encoding can be trivially obtained from the structure.
6718 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
6719 not read locks (CRYPTO_r_[un]lock).
6722 *) A first attempt at creating official support for shared
6723 libraries through configuration. I've kept it so the
6724 default is static libraries only, and the OpenSSL programs
6725 are always statically linked for now, but there are
6726 preparations for dynamic linking in place.
6727 This has been tested on Linux and Tru64.
6730 *) Randomness polling function for Win9x, as described in:
6731 Peter Gutmann, Software Generation of Practically Strong
6735 *) Fix so PRNG is seeded in req if using an already existing
6739 *) New options to smime application. -inform and -outform
6740 allow alternative formats for the S/MIME message including
6741 PEM and DER. The -content option allows the content to be
6742 specified separately. This should allow things like Netscape
6743 form signing output easier to verify.
6746 *) Fix the ASN1 encoding of tags using the 'long form'.
6749 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
6750 STRING types. These convert content octets to and from the
6751 underlying type. The actual tag and length octets are
6752 already assumed to have been read in and checked. These
6753 are needed because all other string types have virtually
6754 identical handling apart from the tag. By having versions
6755 of the ASN1 functions that just operate on content octets
6756 IMPLICIT tagging can be handled properly. It also allows
6757 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
6758 and ASN1_INTEGER are identical apart from the tag.
6761 *) Change the handling of OID objects as follows:
6763 - New object identifiers are inserted in objects.txt, following
6764 the syntax given in objects.README.
6765 - objects.pl is used to process obj_mac.num and create a new
6767 - obj_dat.pl is used to create a new obj_dat.h, using the data in
6770 This is currently kind of a hack, and the perl code in objects.pl
6771 isn't very elegant, but it works as I intended. The simplest way
6772 to check that it worked correctly is to look in obj_dat.h and
6773 check the array nid_objs and make sure the objects haven't moved
6774 around (this is important!). Additions are OK, as well as
6775 consistent name changes.
6778 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
6781 *) Addition of the command line parameter '-rand file' to 'openssl req'.
6782 The given file adds to whatever has already been seeded into the
6783 random pool through the RANDFILE configuration file option or
6784 environment variable, or the default random state file.
6787 *) mkstack.pl now sorts each macro group into lexical order.
6788 Previously the output order depended on the order the files
6789 appeared in the directory, resulting in needless rewriting
6793 *) Patches to make OpenSSL compile under Win32 again. Mostly
6794 work arounds for the VC++ problem that it treats func() as
6795 func(void). Also stripped out the parts of mkdef.pl that
6796 added extra typesafe functions: these no longer exist.
6799 *) Reorganisation of the stack code. The macros are now all
6800 collected in safestack.h . Each macro is defined in terms of
6801 a "stack macro" of the form SKM_<name>(type, a, b). The
6802 DEBUG_SAFESTACK is now handled in terms of function casts,
6803 this has the advantage of retaining type safety without the
6804 use of additional functions. If DEBUG_SAFESTACK is not defined
6805 then the non typesafe macros are used instead. Also modified the
6806 mkstack.pl script to handle the new form. Needs testing to see
6807 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
6808 the default if no major problems. Similar behaviour for ASN1_SET_OF
6809 and PKCS12_STACK_OF.
6812 *) When some versions of IIS use the 'NET' form of private key the
6813 key derivation algorithm is different. Normally MD5(password) is
6814 used as a 128 bit RC4 key. In the modified case
6815 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
6816 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
6817 as the old Netscape_RSA functions except they have an additional
6818 'sgckey' parameter which uses the modified algorithm. Also added
6819 an -sgckey command line option to the rsa utility. Thanks to
6820 Adrian Peck <bertie@ncipher.com> for posting details of the modified
6821 algorithm to openssl-dev.
6824 *) The evp_local.h macros were using 'c.##kname' which resulted in
6825 invalid expansion on some systems (SCO 5.0.5 for example).
6826 Corrected to 'c.kname'.
6827 [Phillip Porch <root@theporch.com>]
6829 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
6830 a STACK of email addresses from a certificate or request, these look
6831 in the subject name and the subject alternative name extensions and
6832 omit any duplicate addresses.
6835 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
6836 This makes DSA verification about 2 % faster.
6839 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
6840 (meaning that now 2^5 values will be precomputed, which is only 4 KB
6841 plus overhead for 1024 bit moduli).
6842 This makes exponentiations about 0.5 % faster for 1024 bit
6843 exponents (as measured by "openssl speed rsa2048").
6846 *) Rename memory handling macros to avoid conflicts with other
6848 Malloc => OPENSSL_malloc
6849 Malloc_locked => OPENSSL_malloc_locked
6850 Realloc => OPENSSL_realloc
6851 Free => OPENSSL_free
6854 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
6855 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
6858 *) CygWin32 support.
6859 [John Jarvie <jjarvie@newsguy.com>]
6861 *) The type-safe stack code has been rejigged. It is now only compiled
6862 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
6863 by default all type-specific stack functions are "#define"d back to
6864 standard stack functions. This results in more streamlined output
6865 but retains the type-safety checking possibilities of the original
6869 *) The STACK code has been cleaned up, and certain type declarations
6870 that didn't make a lot of sense have been brought in line. This has
6871 also involved a cleanup of sorts in safestack.h to more correctly
6872 map type-safe stack functions onto their plain stack counterparts.
6873 This work has also resulted in a variety of "const"ifications of
6874 lots of the code, especially "_cmp" operations which should normally
6875 be prototyped with "const" parameters anyway.
6878 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
6879 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
6880 (The PRNG state consists of two parts, the large pool 'state' and 'md',
6881 where all of 'md' is used each time the PRNG is used, but 'state'
6882 is used only indexed by a cyclic counter. As entropy may not be
6883 well distributed from the beginning, 'md' is important as a
6884 chaining variable. However, the output function chains only half
6885 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
6886 all of 'md', and seeding with STATE_SIZE dummy bytes will result
6887 in all of 'state' being rewritten, with the new values depending
6888 on virtually all of 'md'. This overcomes the 80 bit limitation.)
6891 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
6892 the handshake is continued after ssl_verify_cert_chain();
6893 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
6894 can lead to 'unexplainable' connection aborts later.
6895 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
6897 *) Major EVP API cipher revision.
6898 Add hooks for extra EVP features. This allows various cipher
6899 parameters to be set in the EVP interface. Support added for variable
6900 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
6901 setting of RC2 and RC5 parameters.
6903 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
6906 Remove lots of duplicated code from the EVP library. For example *every*
6907 cipher init() function handles the 'iv' in the same way according to the
6908 cipher mode. They also all do nothing if the 'key' parameter is NULL and
6909 for CFB and OFB modes they zero ctx->num.
6911 New functionality allows removal of S/MIME code RC2 hack.
6913 Most of the routines have the same form and so can be declared in terms
6916 By shifting this to the top level EVP_CipherInit() it can be removed from
6917 all individual ciphers. If the cipher wants to handle IVs or keys
6918 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
6921 Change lots of functions like EVP_EncryptUpdate() to now return a
6922 value: although software versions of the algorithms cannot fail
6923 any installed hardware versions can.
6926 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
6927 this option is set, tolerate broken clients that send the negotiated
6928 protocol version number instead of the requested protocol version
6932 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
6933 i.e. non-zero for export ciphersuites, zero otherwise.
6934 Previous versions had this flag inverted, inconsistent with
6935 rsa_tmp_cb (..._TMP_RSA_CB).
6936 [Bodo Moeller; problem reported by Amit Chopra]
6938 *) Add missing DSA library text string. Work around for some IIS
6939 key files with invalid SEQUENCE encoding.
6942 *) Add a document (doc/standards.txt) that list all kinds of standards
6943 and so on that are implemented in OpenSSL.
6946 *) Enhance c_rehash script. Old version would mishandle certificates
6947 with the same subject name hash and wouldn't handle CRLs at all.
6948 Added -fingerprint option to crl utility, to support new c_rehash
6952 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
6955 *) Fix for SSL server purpose checking. Server checking was
6956 rejecting certificates which had extended key usage present
6957 but no ssl client purpose.
6958 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
6960 *) Make PKCS#12 code work with no password. The PKCS#12 spec
6961 is a little unclear about how a blank password is handled.
6962 Since the password in encoded as a BMPString with terminating
6963 double NULL a zero length password would end up as just the
6964 double NULL. However no password at all is different and is
6965 handled differently in the PKCS#12 key generation code. NS
6966 treats a blank password as zero length. MSIE treats it as no
6967 password on export: but it will try both on import. We now do
6968 the same: PKCS12_parse() tries zero length and no password if
6969 the password is set to "" or NULL (NULL is now a valid password:
6970 it wasn't before) as does the pkcs12 application.
6973 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
6974 perror when PEM_read_bio_X509_REQ fails, the error message must
6975 be obtained from the error queue.
6978 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
6979 it in ERR_remove_state if appropriate, and change ERR_get_state
6980 accordingly to avoid race conditions (this is necessary because
6981 thread_hash is no longer constant once set).
6984 *) Bugfix for linux-elf makefile.one.
6987 *) RSA_get_default_method() will now cause a default
6988 RSA_METHOD to be chosen if one doesn't exist already.
6989 Previously this was only set during a call to RSA_new()
6990 or RSA_new_method(NULL) meaning it was possible for
6991 RSA_get_default_method() to return NULL.
6994 *) Added native name translation to the existing DSO code
6995 that will convert (if the flag to do so is set) filenames
6996 that are sufficiently small and have no path information
6997 into a canonical native form. Eg. "blah" converted to
6998 "libblah.so" or "blah.dll" etc.
7001 *) New function ERR_error_string_n(e, buf, len) which is like
7002 ERR_error_string(e, buf), but writes at most 'len' bytes
7003 including the 0 terminator. For ERR_error_string_n, 'buf'
7005 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
7007 *) CONF library reworked to become more general. A new CONF
7008 configuration file reader "class" is implemented as well as a
7009 new functions (NCONF_*, for "New CONF") to handle it. The now
7010 old CONF_* functions are still there, but are reimplemented to
7011 work in terms of the new functions. Also, a set of functions
7012 to handle the internal storage of the configuration data is
7013 provided to make it easier to write new configuration file
7014 reader "classes" (I can definitely see something reading a
7015 configuration file in XML format, for example), called _CONF_*,
7016 or "the configuration storage API"...
7018 The new configuration file reading functions are:
7020 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
7021 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
7023 NCONF_default, NCONF_WIN32
7025 NCONF_dump_fp, NCONF_dump_bio
7027 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
7028 NCONF_new creates a new CONF object. This works in the same way
7029 as other interfaces in OpenSSL, like the BIO interface.
7030 NCONF_dump_* dump the internal storage of the configuration file,
7031 which is useful for debugging. All other functions take the same
7032 arguments as the old CONF_* functions wth the exception of the
7033 first that must be a `CONF *' instead of a `LHASH *'.
7035 To make it easer to use the new classes with the old CONF_* functions,
7036 the function CONF_set_default_method is provided.
7039 *) Add '-tls1' option to 'openssl ciphers', which was already
7040 mentioned in the documentation but had not been implemented.
7041 (This option is not yet really useful because even the additional
7042 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
7045 *) Initial DSO code added into libcrypto for letting OpenSSL (and
7046 OpenSSL-based applications) load shared libraries and bind to
7047 them in a portable way.
7048 [Geoff Thorpe, with contributions from Richard Levitte]
7050 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
7052 *) Make sure _lrotl and _lrotr are only used with MSVC.
7054 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
7055 (the default implementation of RAND_status).
7057 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
7058 to '-clrext' (= clear extensions), as intended and documented.
7059 [Bodo Moeller; inconsistency pointed out by Michael Attili
7062 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
7063 was larger than the MD block size.
7064 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
7066 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
7067 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
7068 using the passed key: if the passed key was a private key the result
7069 of X509_print(), for example, would be to print out all the private key
7073 *) des_quad_cksum() byte order bug fix.
7074 [Ulf Möller, using the problem description in krb4-0.9.7, where
7075 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
7077 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
7079 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
7081 *) For easily testing in shell scripts whether some command
7082 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
7083 returns with exit code 0 iff no command of the given name is available.
7084 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
7085 the output goes to stdout and nothing is printed to stderr.
7086 Additional arguments are always ignored.
7088 Since for each cipher there is a command of the same name,
7089 the 'no-cipher' compilation switches can be tested this way.
7091 ('openssl no-XXX' is not able to detect pseudo-commands such
7092 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
7095 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
7098 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
7099 is set; it will be thrown away anyway because each handshake creates
7101 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
7102 to parameters -- in previous versions (since OpenSSL 0.9.3) the
7103 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
7104 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7107 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7108 'Q' and 'R' lose their special meanings (quit/renegotiate).
7109 This is part of what -quiet does; unlike -quiet, -ign_eof
7110 does not suppress any output.
7113 *) Add compatibility options to the purpose and trust code. The
7114 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7115 accepts a certificate or CA, this was the previous behaviour,
7116 with all the associated security issues.
7118 X509_TRUST_COMPAT is the old trust behaviour: only and
7119 automatically trust self signed roots in certificate store. A
7120 new trust setting X509_TRUST_DEFAULT is used to specify that
7121 a purpose has no associated trust setting and it should instead
7122 use the value in the default purpose.
7125 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7126 and fix a memory leak.
7129 *) In util/mkerr.pl (which implements 'make errors'), preserve
7130 reason strings from the previous version of the .c file, as
7131 the default to have only downcase letters (and digits) in
7132 automatically generated reasons codes is not always appropriate.
7135 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7136 using strerror. Previously, ERR_reason_error_string() returned
7137 library names as reason strings for SYSerr; but SYSerr is a special
7138 case where small numbers are errno values, not library numbers.
7141 *) Add '-dsaparam' option to 'openssl dhparam' application. This
7142 converts DSA parameters into DH parameters. (When creating parameters,
7143 DSA_generate_parameters is used.)
7146 *) Include 'length' (recommended exponent length) in C code generated
7147 by 'openssl dhparam -C'.
7150 *) The second argument to set_label in perlasm was already being used
7151 so couldn't be used as a "file scope" flag. Moved to third argument
7155 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
7156 instead of RAND_bytes for encryption IVs and salts.
7159 *) Include RAND_status() into RAND_METHOD instead of implementing
7160 it only for md_rand.c Otherwise replacing the PRNG by calling
7161 RAND_set_rand_method would be impossible.
7164 *) Don't let DSA_generate_key() enter an infinite loop if the random
7165 number generation fails.
7168 *) New 'rand' application for creating pseudo-random output.
7171 *) Added configuration support for Linux/IA64
7172 [Rolf Haberrecker <rolf@suse.de>]
7174 *) Assembler module support for Mingw32.
7177 *) Shared library support for HPUX (in shlib/).
7178 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
7180 *) Shared library support for Solaris gcc.
7181 [Lutz Behnke <behnke@trustcenter.de>]
7183 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
7185 *) PKCS7_encrypt() was adding text MIME headers twice because they
7186 were added manually and by SMIME_crlf_copy().
7189 *) In bntest.c don't call BN_rand with zero bits argument.
7190 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
7192 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
7193 case was implemented. This caused BN_div_recp() to fail occasionally.
7196 *) Add an optional second argument to the set_label() in the perl
7197 assembly language builder. If this argument exists and is set
7198 to 1 it signals that the assembler should use a symbol whose
7199 scope is the entire file, not just the current function. This
7200 is needed with MASM which uses the format label:: for this scope.
7201 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
7203 *) Change the ASN1 types so they are typedefs by default. Before
7204 almost all types were #define'd to ASN1_STRING which was causing
7205 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
7209 *) Change names of new functions to the new get1/get0 naming
7210 convention: After 'get1', the caller owns a reference count
7211 and has to call ..._free; 'get0' returns a pointer to some
7212 data structure without incrementing reference counters.
7213 (Some of the existing 'get' functions increment a reference
7214 counter, some don't.)
7215 Similarly, 'set1' and 'add1' functions increase reference
7216 counters or duplicate objects.
7219 *) Allow for the possibility of temp RSA key generation failure:
7220 the code used to assume it always worked and crashed on failure.
7223 *) Fix potential buffer overrun problem in BIO_printf().
7224 [Ulf Möller, using public domain code by Patrick Powell; problem
7225 pointed out by David Sacerdote <das33@cornell.edu>]
7227 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
7228 RAND_egd() and RAND_status(). In the command line application,
7229 the EGD socket can be specified like a seed file using RANDFILE
7233 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
7234 Some CAs (e.g. Verisign) distribute certificates in this form.
7237 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
7238 list to exclude them. This means that no special compilation option
7239 is needed to use anonymous DH: it just needs to be included in the
7243 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
7244 EVP_MD_type. The old functionality is available in a new macro called
7245 EVP_MD_md(). Change code that uses it and update docs.
7248 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
7249 where the 'void *' argument is replaced by a function pointer argument.
7250 Previously 'void *' was abused to point to functions, which works on
7251 many platforms, but is not correct. As these functions are usually
7252 called by macros defined in OpenSSL header files, most source code
7253 should work without changes.
7256 *) <openssl/opensslconf.h> (which is created by Configure) now contains
7257 sections with information on -D... compiler switches used for
7258 compiling the library so that applications can see them. To enable
7259 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
7260 must be defined. E.g.,
7261 #define OPENSSL_ALGORITHM_DEFINES
7262 #include <openssl/opensslconf.h>
7263 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
7264 [Richard Levitte, Ulf and Bodo Möller]
7266 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
7270 *) Change the 'other' type in certificate aux info to a STACK_OF
7271 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
7272 the required ASN1 format: arbitrary types determined by an OID.
7275 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
7276 argument to 'req'. This is not because the function is newer or
7277 better than others it just uses the work 'NEW' in the certificate
7278 request header lines. Some software needs this.
7281 *) Reorganise password command line arguments: now passwords can be
7282 obtained from various sources. Delete the PEM_cb function and make
7283 it the default behaviour: i.e. if the callback is NULL and the
7284 usrdata argument is not NULL interpret it as a null terminated pass
7285 phrase. If usrdata and the callback are NULL then the pass phrase
7286 is prompted for as usual.
7289 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
7290 the support is automatically enabled. The resulting binaries will
7291 autodetect the card and use it if present.
7292 [Ben Laurie and Compaq Inc.]
7294 *) Work around for Netscape hang bug. This sends certificate request
7295 and server done in one record. Since this is perfectly legal in the
7296 SSL/TLS protocol it isn't a "bug" option and is on by default. See
7297 the bugs/SSLv3 entry for more info.
7300 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
7303 *) Add -rand argument to smime and pkcs12 applications and read/write
7307 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
7310 *) Add command line password options to the remaining applications.
7313 *) Bug fix for BN_div_recp() for numerators with an even number of
7317 *) More tests in bntest.c, and changed test_bn output.
7320 *) ./config recognizes MacOS X now.
7323 *) Bug fix for BN_div() when the first words of num and divsor are
7324 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
7327 *) Add support for various broken PKCS#8 formats, and command line
7328 options to produce them.
7331 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
7332 get temporary BIGNUMs from a BN_CTX.
7335 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
7339 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
7340 include a #define from the old name to the new. The original intent
7341 was that statically linked binaries could for example just call
7342 SSLeay_add_all_ciphers() to just add ciphers to the table and not
7343 link with digests. This never worked becayse SSLeay_add_all_digests()
7344 and SSLeay_add_all_ciphers() were in the same source file so calling
7345 one would link with the other. They are now in separate source files.
7348 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
7351 *) Use a less unusual form of the Miller-Rabin primality test (it used
7352 a binary algorithm for exponentiation integrated into the Miller-Rabin
7353 loop, our standard modexp algorithms are faster).
7356 *) Support for the EBCDIC character set completed.
7357 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
7359 *) Source code cleanups: use const where appropriate, eliminate casts,
7360 use void * instead of char * in lhash.
7363 *) Bugfix: ssl3_send_server_key_exchange was not restartable
7364 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
7365 this the server could overwrite ephemeral keys that the client
7369 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
7370 using 50 iterations of the Rabin-Miller test.
7372 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
7373 iterations of the Rabin-Miller test as required by the appendix
7374 to FIPS PUB 186[-1]) instead of DSA_is_prime.
7375 As BN_is_prime_fasttest includes trial division, DSA parameter
7376 generation becomes much faster.
7378 This implies a change for the callback functions in DSA_is_prime
7379 and DSA_generate_parameters: The callback function is called once
7380 for each positive witness in the Rabin-Miller test, not just
7381 occasionally in the inner loop; and the parameters to the
7382 callback function now provide an iteration count for the outer
7383 loop rather than for the current invocation of the inner loop.
7384 DSA_generate_parameters additionally can call the callback
7385 function with an 'iteration count' of -1, meaning that a
7386 candidate has passed the trial division test (when q is generated
7387 from an application-provided seed, trial division is skipped).
7390 *) New function BN_is_prime_fasttest that optionally does trial
7391 division before starting the Rabin-Miller test and has
7392 an additional BN_CTX * argument (whereas BN_is_prime always
7393 has to allocate at least one BN_CTX).
7394 'callback(1, -1, cb_arg)' is called when a number has passed the
7395 trial division stage.
7398 *) Fix for bug in CRL encoding. The validity dates weren't being handled
7402 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
7405 *) New function BN_pseudo_rand().
7408 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
7409 bignum version of BN_from_montgomery() with the working code from
7410 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
7414 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
7415 made it impossible to use the same SSL_SESSION data structure in
7416 SSL2 clients in multiple threads.
7419 *) The return value of RAND_load_file() no longer counts bytes obtained
7420 by stat(). RAND_load_file(..., -1) is new and uses the complete file
7421 to seed the PRNG (previously an explicit byte count was required).
7422 [Ulf Möller, Bodo Möller]
7424 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
7425 used (char *) instead of (void *) and had casts all over the place.
7428 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
7431 *) Retain source code compatibility for BN_prime_checks macro:
7432 BN_is_prime(..., BN_prime_checks, ...) now uses
7433 BN_prime_checks_for_size to determine the appropriate number of
7434 Rabin-Miller iterations.
7437 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
7438 DH_CHECK_P_NOT_SAFE_PRIME.
7439 (Check if this is true? OpenPGP calls them "strong".)
7442 *) Merge the functionality of "dh" and "gendh" programs into a new program
7443 "dhparam". The old programs are retained for now but will handle DH keys
7444 (instead of parameters) in future.
7447 *) Make the ciphers, s_server and s_client programs check the return values
7448 when a new cipher list is set.
7451 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
7452 ciphers. Before when the 56bit ciphers were enabled the sorting was
7455 The syntax for the cipher sorting has been extended to support sorting by
7456 cipher-strength (using the strength_bits hard coded in the tables).
7457 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
7459 Fix a bug in the cipher-command parser: when supplying a cipher command
7460 string with an "undefined" symbol (neither command nor alphanumeric
7461 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
7462 an error is flagged.
7464 Due to the strength-sorting extension, the code of the
7465 ssl_create_cipher_list() function was completely rearranged. I hope that
7466 the readability was also increased :-)
7467 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7469 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
7470 for the first serial number and places 2 in the serial number file. This
7471 avoids problems when the root CA is created with serial number zero and
7472 the first user certificate has the same issuer name and serial number
7476 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
7477 the new code. Add documentation for this stuff.
7480 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
7481 X509_*() to X509at_*() on the grounds that they don't handle X509
7482 structures and behave in an analagous way to the X509v3 functions:
7483 they shouldn't be called directly but wrapper functions should be used
7486 So we also now have some wrapper functions that call the X509at functions
7487 when passed certificate requests. (TO DO: similar things can be done with
7488 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
7489 things. Some of these need some d2i or i2d and print functionality
7490 because they handle more complex structures.)
7493 *) Add missing #ifndefs that caused missing symbols when building libssl
7494 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
7495 NO_RSA in ssl/s2*.c.
7496 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
7498 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
7499 has a return value which indicates the quality of the random data
7500 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
7501 error queue. New function RAND_pseudo_bytes() generates output that is
7502 guaranteed to be unique but not unpredictable. RAND_add is like
7503 RAND_seed, but takes an extra argument for an entropy estimate
7504 (RAND_seed always assumes full entropy).
7507 *) Do more iterations of Rabin-Miller probable prime test (specifically,
7508 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
7509 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
7510 in crypto/bn/bn_prime.c for the complete table). This guarantees a
7511 false-positive rate of at most 2^-80 for random input.
7514 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
7517 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
7518 in the 0.9.5 release), this returns the chain
7519 from an X509_CTX structure with a dup of the stack and all
7520 the X509 reference counts upped: so the stack will exist
7521 after X509_CTX_cleanup() has been called. Modify pkcs12.c
7524 Also make SSL_SESSION_print() print out the verify return
7528 *) Add manpage for the pkcs12 command. Also change the default
7529 behaviour so MAC iteration counts are used unless the new
7530 -nomaciter option is used. This improves file security and
7531 only older versions of MSIE (4.0 for example) need it.
7534 *) Honor the no-xxx Configure options when creating .DEF files.
7537 *) Add PKCS#10 attributes to field table: challengePassword,
7538 unstructuredName and unstructuredAddress. These are taken from
7539 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
7540 international characters are used.
7542 More changes to X509_ATTRIBUTE code: allow the setting of types
7543 based on strings. Remove the 'loc' parameter when adding
7544 attributes because these will be a SET OF encoding which is sorted
7548 *) Initial changes to the 'req' utility to allow request generation
7549 automation. This will allow an application to just generate a template
7550 file containing all the field values and have req construct the
7553 Initial support for X509_ATTRIBUTE handling. Stacks of these are
7554 used all over the place including certificate requests and PKCS#7
7555 structures. They are currently handled manually where necessary with
7556 some primitive wrappers for PKCS#7. The new functions behave in a
7557 manner analogous to the X509 extension functions: they allow
7558 attributes to be looked up by NID and added.
7560 Later something similar to the X509V3 code would be desirable to
7561 automatically handle the encoding, decoding and printing of the
7562 more complex types. The string types like challengePassword can
7563 be handled by the string table functions.
7565 Also modified the multi byte string table handling. Now there is
7566 a 'global mask' which masks out certain types. The table itself
7567 can use the flag STABLE_NO_MASK to ignore the mask setting: this
7568 is useful when for example there is only one permissible type
7569 (as in countryName) and using the mask might result in no valid
7573 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
7574 SSL_get_peer_finished to allow applications to obtain the latest
7575 Finished messages sent to the peer or expected from the peer,
7576 respectively. (SSL_get_peer_finished is usually the Finished message
7577 actually received from the peer, otherwise the protocol will be aborted.)
7579 As the Finished message are message digests of the complete handshake
7580 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
7581 be used for external authentication procedures when the authentication
7582 provided by SSL/TLS is not desired or is not enough.
7585 *) Enhanced support for Alpha Linux is added. Now ./config checks if
7586 the host supports BWX extension and if Compaq C is present on the
7587 $PATH. Just exploiting of the BWX extension results in 20-30%
7588 performance kick for some algorithms, e.g. DES and RC4 to mention
7589 a couple. Compaq C in turn generates ~20% faster code for MD5 and
7593 *) Add support for MS "fast SGC". This is arguably a violation of the
7594 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
7595 weak crypto and after checking the certificate is SGC a second one
7596 with strong crypto. MS SGC stops the first handshake after receiving
7597 the server certificate message and sends a second client hello. Since
7598 a server will typically do all the time consuming operations before
7599 expecting any further messages from the client (server key exchange
7600 is the most expensive) there is little difference between the two.
7602 To get OpenSSL to support MS SGC we have to permit a second client
7603 hello message after we have sent server done. In addition we have to
7604 reset the MAC if we do get this second client hello.
7607 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
7608 if a DER encoded private key is RSA or DSA traditional format. Changed
7609 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
7610 format DER encoded private key. Newer code should use PKCS#8 format which
7611 has the key type encoded in the ASN1 structure. Added DER private key
7612 support to pkcs8 application.
7615 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
7616 ciphersuites has been selected (as required by the SSL 3/TLS 1
7617 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
7618 is set, we interpret this as a request to violate the specification
7619 (the worst that can happen is a handshake failure, and 'correct'
7620 behaviour would result in a handshake failure anyway).
7623 *) In SSL_CTX_add_session, take into account that there might be multiple
7624 SSL_SESSION structures with the same session ID (e.g. when two threads
7625 concurrently obtain them from an external cache).
7626 The internal cache can handle only one SSL_SESSION with a given ID,
7627 so if there's a conflict, we now throw out the old one to achieve
7631 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
7632 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
7633 some routines that use cipher OIDs: some ciphers do not have OIDs
7634 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
7638 *) Simplify the trust setting structure and code. Now we just have
7639 two sequences of OIDs for trusted and rejected settings. These will
7640 typically have values the same as the extended key usage extension
7641 and any application specific purposes.
7643 The trust checking code now has a default behaviour: it will just
7644 check for an object with the same NID as the passed id. Functions can
7645 be provided to override either the default behaviour or the behaviour
7646 for a given id. SSL client, server and email already have functions
7647 in place for compatibility: they check the NID and also return "trusted"
7648 if the certificate is self signed.
7651 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
7652 traditional format into an EVP_PKEY structure.
7655 *) Add a password callback function PEM_cb() which either prompts for
7656 a password if usr_data is NULL or otherwise assumes it is a null
7657 terminated password. Allow passwords to be passed on command line
7658 environment or config files in a few more utilities.
7661 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
7662 keys. Add some short names for PKCS#8 PBE algorithms and allow them
7663 to be specified on the command line for the pkcs8 and pkcs12 utilities.
7664 Update documentation.
7667 *) Support for ASN1 "NULL" type. This could be handled before by using
7668 ASN1_TYPE but there wasn't any function that would try to read a NULL
7669 and produce an error if it couldn't. For compatibility we also have
7670 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
7671 don't allocate anything because they don't need to.
7674 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
7676 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
7678 *) Rebuild of the memory allocation routines used by OpenSSL code and
7679 possibly others as well. The purpose is to make an interface that
7680 provide hooks so anyone can build a separate set of allocation and
7681 deallocation routines to be used by OpenSSL, for example memory
7682 pool implementations, or something else, which was previously hard
7683 since Malloc(), Realloc() and Free() were defined as macros having
7684 the values malloc, realloc and free, respectively (except for Win32
7685 compilations). The same is provided for memory debugging code.
7686 OpenSSL already comes with functionality to find memory leaks, but
7687 this gives people a chance to debug other memory problems.
7689 With these changes, a new set of functions and macros have appeared:
7691 CRYPTO_set_mem_debug_functions() [F]
7692 CRYPTO_get_mem_debug_functions() [F]
7693 CRYPTO_dbg_set_options() [F]
7694 CRYPTO_dbg_get_options() [F]
7695 CRYPTO_malloc_debug_init() [M]
7697 The memory debug functions are NULL by default, unless the library
7698 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
7699 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
7700 gives the standard debugging functions that come with OpenSSL) or
7701 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
7702 provided by the library user) must be used. When the standard
7703 debugging functions are used, CRYPTO_dbg_set_options can be used to
7704 request additional information:
7705 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
7706 the CRYPTO_MDEBUG_xxx macro when compiling the library.
7708 Also, things like CRYPTO_set_mem_functions will always give the
7709 expected result (the new set of functions is used for allocation
7710 and deallocation) at all times, regardless of platform and compiler
7713 To finish it up, some functions that were never use in any other
7714 way than through macros have a new API and new semantic:
7717 CRYPTO_dbg_realloc()
7720 All macros of value have retained their old syntax.
7721 [Richard Levitte and Bodo Moeller]
7723 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
7724 ordering of SMIMECapabilities wasn't in "strength order" and there
7725 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
7729 *) Some ASN1 types with illegal zero length encoding (INTEGER,
7730 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
7731 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
7733 *) Merge in my S/MIME library for OpenSSL. This provides a simple
7734 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
7735 functionality to handle multipart/signed properly) and a utility
7736 called 'smime' to call all this stuff. This is based on code I
7737 originally wrote for Celo who have kindly allowed it to be
7738 included in OpenSSL.
7741 *) Add variants des_set_key_checked and des_set_key_unchecked of
7742 des_set_key (aka des_key_sched). Global variable des_check_key
7743 decides which of these is called by des_set_key; this way
7744 des_check_key behaves as it always did, but applications and
7745 the library itself, which was buggy for des_check_key == 1,
7746 have a cleaner way to pick the version they need.
7749 *) New function PKCS12_newpass() which changes the password of a
7753 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
7754 dynamic mix. In both cases the ids can be used as an index into the
7755 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
7756 functions so they accept a list of the field values and the
7757 application doesn't need to directly manipulate the X509_TRUST
7761 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
7765 *) Modify the way the V3 extension code looks up extensions. This now
7766 works in a similar way to the object code: we have some "standard"
7767 extensions in a static table which is searched with OBJ_bsearch()
7768 and the application can add dynamic ones if needed. The file
7769 crypto/x509v3/ext_dat.h now has the info: this file needs to be
7770 updated whenever a new extension is added to the core code and kept
7771 in ext_nid order. There is a simple program 'tabtest.c' which checks
7772 this. New extensions are not added too often so this file can readily
7773 be maintained manually.
7775 There are two big advantages in doing things this way. The extensions
7776 can be looked up immediately and no longer need to be "added" using
7777 X509V3_add_standard_extensions(): this function now does nothing.
7778 [Side note: I get *lots* of email saying the extension code doesn't
7779 work because people forget to call this function]
7780 Also no dynamic allocation is done unless new extensions are added:
7781 so if we don't add custom extensions there is no need to call
7782 X509V3_EXT_cleanup().
7785 *) Modify enc utility's salting as follows: make salting the default. Add a
7786 magic header, so unsalted files fail gracefully instead of just decrypting
7787 to garbage. This is because not salting is a big security hole, so people
7788 should be discouraged from doing it.
7791 *) Fixes and enhancements to the 'x509' utility. It allowed a message
7792 digest to be passed on the command line but it only used this
7793 parameter when signing a certificate. Modified so all relevant
7794 operations are affected by the digest parameter including the
7795 -fingerprint and -x509toreq options. Also -x509toreq choked if a
7796 DSA key was used because it didn't fix the digest.
7799 *) Initial certificate chain verify code. Currently tests the untrusted
7800 certificates for consistency with the verify purpose (which is set
7801 when the X509_STORE_CTX structure is set up) and checks the pathlength.
7803 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
7804 this is because it will reject chains with invalid extensions whereas
7805 every previous version of OpenSSL and SSLeay made no checks at all.
7807 Trust code: checks the root CA for the relevant trust settings. Trust
7808 settings have an initial value consistent with the verify purpose: e.g.
7809 if the verify purpose is for SSL client use it expects the CA to be
7810 trusted for SSL client use. However the default value can be changed to
7811 permit custom trust settings: one example of this would be to only trust
7812 certificates from a specific "secure" set of CAs.
7814 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
7815 which should be used for version portability: especially since the
7816 verify structure is likely to change more often now.
7818 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
7819 to set them. If not set then assume SSL clients will verify SSL servers
7822 Two new options to the verify program: -untrusted allows a set of
7823 untrusted certificates to be passed in and -purpose which sets the
7824 intended purpose of the certificate. If a purpose is set then the
7825 new chain verify code is used to check extension consistency.
7828 *) Support for the authority information access extension.
7831 *) Modify RSA and DSA PEM read routines to transparently handle
7832 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
7833 public keys in a format compatible with certificate
7834 SubjectPublicKeyInfo structures. Unfortunately there were already
7835 functions called *_PublicKey_* which used various odd formats so
7836 these are retained for compatibility: however the DSA variants were
7837 never in a public release so they have been deleted. Changed dsa/rsa
7838 utilities to handle the new format: note no releases ever handled public
7839 keys so we should be OK.
7841 The primary motivation for this change is to avoid the same fiasco
7842 that dogs private keys: there are several incompatible private key
7843 formats some of which are standard and some OpenSSL specific and
7844 require various evil hacks to allow partial transparent handling and
7845 even then it doesn't work with DER formats. Given the option anything
7846 other than PKCS#8 should be dumped: but the other formats have to
7847 stay in the name of compatibility.
7849 With public keys and the benefit of hindsight one standard format
7850 is used which works with EVP_PKEY, RSA or DSA structures: though
7851 it clearly returns an error if you try to read the wrong kind of key.
7853 Added a -pubkey option to the 'x509' utility to output the public key.
7854 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
7855 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
7856 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
7857 that do the same as the EVP_PKEY_assign_*() except they up the
7858 reference count of the added key (they don't "swallow" the
7862 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
7863 CRLs would fail if the file contained no certificates or no CRLs:
7864 added a new function to read in both types and return the number
7865 read: this means that if none are read it will be an error. The
7866 DER versions of the certificate and CRL reader would always fail
7867 because it isn't possible to mix certificates and CRLs in DER format
7868 without choking one or the other routine. Changed this to just read
7869 a certificate: this is the best we can do. Also modified the code
7870 in apps/verify.c to take notice of return codes: it was previously
7871 attempting to read in certificates from NULL pointers and ignoring
7872 any errors: this is one reason why the cert and CRL reader seemed
7873 to work. It doesn't check return codes from the default certificate
7874 routines: these may well fail if the certificates aren't installed.
7877 *) Code to support otherName option in GeneralName.
7880 *) First update to verify code. Change the verify utility
7881 so it warns if it is passed a self signed certificate:
7882 for consistency with the normal behaviour. X509_verify
7883 has been modified to it will now verify a self signed
7884 certificate if *exactly* the same certificate appears
7885 in the store: it was previously impossible to trust a
7886 single self signed certificate. This means that:
7887 openssl verify ss.pem
7888 now gives a warning about a self signed certificate but
7889 openssl verify -CAfile ss.pem ss.pem
7893 *) For servers, store verify_result in SSL_SESSION data structure
7894 (and add it to external session representation).
7895 This is needed when client certificate verifications fails,
7896 but an application-provided verification callback (set by
7897 SSL_CTX_set_cert_verify_callback) allows accepting the session
7898 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
7899 but returns 1): When the session is reused, we have to set
7900 ssl->verify_result to the appropriate error code to avoid
7902 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
7904 *) Fix a bug in the new PKCS#7 code: it didn't consider the
7905 case in PKCS7_dataInit() where the signed PKCS7 structure
7906 didn't contain any existing data because it was being created.
7907 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
7909 *) Add a salt to the key derivation routines in enc.c. This
7910 forms the first 8 bytes of the encrypted file. Also add a
7911 -S option to allow a salt to be input on the command line.
7914 *) New function X509_cmp(). Oddly enough there wasn't a function
7915 to compare two certificates. We do this by working out the SHA1
7916 hash and comparing that. X509_cmp() will be needed by the trust
7920 *) SSL_get1_session() is like SSL_get_session(), but increments
7921 the reference count in the SSL_SESSION returned.
7922 [Geoff Thorpe <geoff@eu.c2.net>]
7924 *) Fix for 'req': it was adding a null to request attributes.
7925 Also change the X509_LOOKUP and X509_INFO code to handle
7926 certificate auxiliary information.
7929 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
7933 *) Add the possibility to add extra information to the memory leak
7934 detecting output, to form tracebacks, showing from where each
7935 allocation was originated: CRYPTO_push_info("constant string") adds
7936 the string plus current file name and line number to a per-thread
7937 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
7938 is like calling CYRPTO_pop_info() until the stack is empty.
7939 Also updated memory leak detection code to be multi-thread-safe.
7942 *) Add options -text and -noout to pkcs7 utility and delete the
7943 encryption options which never did anything. Update docs.
7946 *) Add options to some of the utilities to allow the pass phrase
7947 to be included on either the command line (not recommended on
7948 OSes like Unix) or read from the environment. Update the
7949 manpages and fix a few bugs.
7952 *) Add a few manpages for some of the openssl commands.
7955 *) Fix the -revoke option in ca. It was freeing up memory twice,
7956 leaking and not finding already revoked certificates.
7959 *) Extensive changes to support certificate auxiliary information.
7960 This involves the use of X509_CERT_AUX structure and X509_AUX
7961 functions. An X509_AUX function such as PEM_read_X509_AUX()
7962 can still read in a certificate file in the usual way but it
7963 will also read in any additional "auxiliary information". By
7964 doing things this way a fair degree of compatibility can be
7965 retained: existing certificates can have this information added
7966 using the new 'x509' options.
7968 Current auxiliary information includes an "alias" and some trust
7969 settings. The trust settings will ultimately be used in enhanced
7970 certificate chain verification routines: currently a certificate
7971 can only be trusted if it is self signed and then it is trusted
7975 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
7976 The problem was that one of the replacement routines had not been working
7977 since SSLeay releases. For now the offending routine has been replaced
7978 with non-optimised assembler. Even so, this now gives around 95%
7979 performance improvement for 1024 bit RSA signs.
7982 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
7983 handling. Most clients have the effective key size in bits equal to
7984 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
7985 A few however don't do this and instead use the size of the decrypted key
7986 to determine the RC2 key length and the AlgorithmIdentifier to determine
7987 the effective key length. In this case the effective key length can still
7988 be 40 bits but the key length can be 168 bits for example. This is fixed
7989 by manually forcing an RC2 key into the EVP_PKEY structure because the
7990 EVP code can't currently handle unusual RC2 key sizes: it always assumes
7991 the key length and effective key length are equal.
7994 *) Add a bunch of functions that should simplify the creation of
7995 X509_NAME structures. Now you should be able to do:
7996 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
7997 and have it automatically work out the correct field type and fill in
7998 the structures. The more adventurous can try:
7999 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
8000 and it will (hopefully) work out the correct multibyte encoding.
8003 *) Change the 'req' utility to use the new field handling and multibyte
8004 copy routines. Before the DN field creation was handled in an ad hoc
8005 way in req, ca, and x509 which was rather broken and didn't support
8006 BMPStrings or UTF8Strings. Since some software doesn't implement
8007 BMPStrings or UTF8Strings yet, they can be enabled using the config file
8008 using the dirstring_type option. See the new comment in the default
8009 openssl.cnf for more info.
8012 *) Make crypto/rand/md_rand.c more robust:
8013 - Assure unique random numbers after fork().
8014 - Make sure that concurrent threads access the global counter and
8015 md serializably so that we never lose entropy in them
8016 or use exactly the same state in multiple threads.
8017 Access to the large state is not always serializable because
8018 the additional locking could be a performance killer, and
8019 md should be large enough anyway.
8022 *) New file apps/app_rand.c with commonly needed functionality
8023 for handling the random seed file.
8025 Use the random seed file in some applications that previously did not:
8027 dsaparam -genkey (which also ignored its '-rand' option),
8030 x509 (when signing).
8031 Except on systems with /dev/urandom, it is crucial to have a random
8032 seed file at least for key creation, DSA signing, and for DH exchanges;
8033 for RSA signatures we could do without one.
8035 gendh and gendsa (unlike genrsa) used to read only the first byte
8036 of each file listed in the '-rand' option. The function as previously
8037 found in genrsa is now in app_rand.c and is used by all programs
8038 that support '-rand'.
8041 *) In RAND_write_file, use mode 0600 for creating files;
8042 don't just chmod when it may be too late.
8045 *) Report an error from X509_STORE_load_locations
8046 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
8049 *) New function ASN1_mbstring_copy() this copies a string in either
8050 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
8051 into an ASN1_STRING type. A mask of permissible types is passed
8052 and it chooses the "minimal" type to use or an error if not type
8056 *) Add function equivalents to the various macros in asn1.h. The old
8057 macros are retained with an M_ prefix. Code inside the library can
8058 use the M_ macros. External code (including the openssl utility)
8059 should *NOT* in order to be "shared library friendly".
8062 *) Add various functions that can check a certificate's extensions
8063 to see if it usable for various purposes such as SSL client,
8064 server or S/MIME and CAs of these types. This is currently
8065 VERY EXPERIMENTAL but will ultimately be used for certificate chain
8066 verification. Also added a -purpose flag to x509 utility to
8067 print out all the purposes.
8070 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
8074 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
8075 for, obtain and decode and extension and obtain its critical flag.
8076 This allows all the necessary extension code to be handled in a
8077 single function call.
8080 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
8081 platforms. See crypto/rc4/rc4_enc.c for further details.
8084 *) New -noout option to asn1parse. This causes no output to be produced
8085 its main use is when combined with -strparse and -out to extract data
8086 from a file (which may not be in ASN.1 format).
8089 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
8090 when producing the local key id.
8091 [Richard Levitte <levitte@stacken.kth.se>]
8093 *) New option -dhparam in s_server. This allows a DH parameter file to be
8094 stated explicitly. If it is not stated then it tries the first server
8095 certificate file. The previous behaviour hard coded the filename
8099 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
8100 a public key to be input or output. For example:
8101 openssl rsa -in key.pem -pubout -out pubkey.pem
8102 Also added necessary DSA public key functions to handle this.
8105 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
8106 in the message. This was handled by allowing
8107 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8108 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8110 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8111 to the end of the strings whereas this didn't. This would cause problems
8112 if strings read with d2i_ASN1_bytes() were later modified.
8113 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8115 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8116 data and it contains EOF it will end up returning an error. This is
8117 caused by input 46 bytes long. The cause is due to the way base64
8118 BIOs find the start of base64 encoded data. They do this by trying a
8119 trial decode on each line until they find one that works. When they
8120 do a flag is set and it starts again knowing it can pass all the
8121 data directly through the decoder. Unfortunately it doesn't reset
8122 the context it uses. This means that if EOF is reached an attempt
8123 is made to pass two EOFs through the context and this causes the
8124 resulting error. This can also cause other problems as well. As is
8125 usual with these problems it takes *ages* to find and the fix is
8126 trivial: move one line.
8127 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8129 *) Ugly workaround to get s_client and s_server working under Windows. The
8130 old code wouldn't work because it needed to select() on sockets and the
8131 tty (for keypresses and to see if data could be written). Win32 only
8132 supports select() on sockets so we select() with a 1s timeout on the
8133 sockets and then see if any characters are waiting to be read, if none
8134 are present then we retry, we also assume we can always write data to
8135 the tty. This isn't nice because the code then blocks until we've
8136 received a complete line of data and it is effectively polling the
8137 keyboard at 1s intervals: however it's quite a bit better than not
8138 working at all :-) A dedicated Windows application might handle this
8139 with an event loop for example.
8142 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
8143 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
8144 will be called when RSA_sign() and RSA_verify() are used. This is useful
8145 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
8146 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
8147 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
8148 This necessitated the support of an extra signature type NID_md5_sha1
8149 for SSL signatures and modifications to the SSL library to use it instead
8150 of calling RSA_public_decrypt() and RSA_private_encrypt().
8153 *) Add new -verify -CAfile and -CApath options to the crl program, these
8154 will lookup a CRL issuers certificate and verify the signature in a
8155 similar way to the verify program. Tidy up the crl program so it
8156 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
8157 less strict. It will now permit CRL extensions even if it is not
8158 a V2 CRL: this will allow it to tolerate some broken CRLs.
8161 *) Initialize all non-automatic variables each time one of the openssl
8162 sub-programs is started (this is necessary as they may be started
8163 multiple times from the "OpenSSL>" prompt).
8164 [Lennart Bang, Bodo Moeller]
8166 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
8167 removing all other RSA functionality (this is what NO_RSA does). This
8168 is so (for example) those in the US can disable those operations covered
8169 by the RSA patent while allowing storage and parsing of RSA keys and RSA
8173 *) Non-copying interface to BIO pairs.
8174 (still largely untested)
8177 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
8178 ASCII string. This was handled independently in various places before.
8181 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
8182 UTF8 strings a character at a time.
8185 *) Use client_version from client hello to select the protocol
8186 (s23_srvr.c) and for RSA client key exchange verification
8187 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
8190 *) Add various utility functions to handle SPKACs, these were previously
8191 handled by poking round in the structure internals. Added new function
8192 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
8193 print, verify and generate SPKACs. Based on an original idea from
8194 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
8197 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
8200 *) Allow the config file extension section to be overwritten on the
8201 command line. Based on an original idea from Massimiliano Pala
8202 <madwolf@comune.modena.it>. The new option is called -extensions
8203 and can be applied to ca, req and x509. Also -reqexts to override
8204 the request extensions in req and -crlexts to override the crl extensions
8208 *) Add new feature to the SPKAC handling in ca. Now you can include
8209 the same field multiple times by preceding it by "XXXX." for example:
8212 this is the same syntax as used in the req config file.
8215 *) Allow certificate extensions to be added to certificate requests. These
8216 are specified in a 'req_extensions' option of the req section of the
8217 config file. They can be printed out with the -text option to req but
8218 are otherwise ignored at present.
8221 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
8222 data read consists of only the final block it would not decrypted because
8223 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
8224 A misplaced 'break' also meant the decrypted final block might not be
8225 copied until the next read.
8228 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
8229 a few extra parameters to the DH structure: these will be useful if
8230 for example we want the value of 'q' or implement X9.42 DH.
8233 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
8234 provides hooks that allow the default DSA functions or functions on a
8235 "per key" basis to be replaced. This allows hardware acceleration and
8236 hardware key storage to be handled without major modification to the
8237 library. Also added low level modexp hooks and CRYPTO_EX structure and
8238 associated functions.
8241 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
8242 as "read only": it can't be written to and the buffer it points to will
8243 not be freed. Reading from a read only BIO is much more efficient than
8244 a normal memory BIO. This was added because there are several times when
8245 an area of memory needs to be read from a BIO. The previous method was
8246 to create a memory BIO and write the data to it, this results in two
8247 copies of the data and an O(n^2) reading algorithm. There is a new
8248 function BIO_new_mem_buf() which creates a read only memory BIO from
8249 an area of memory. Also modified the PKCS#7 routines to use read only
8253 *) Bugfix: ssl23_get_client_hello did not work properly when called in
8254 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
8255 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
8256 but a retry condition occured while trying to read the rest.
8259 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
8260 NID_pkcs7_encrypted by default: this was wrong since this should almost
8261 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
8262 the encrypted data type: this is a more sensible place to put it and it
8263 allows the PKCS#12 code to be tidied up that duplicated this
8267 *) Changed obj_dat.pl script so it takes its input and output files on
8268 the command line. This should avoid shell escape redirection problems
8272 *) Initial support for certificate extension requests, these are included
8273 in things like Xenroll certificate requests. Included functions to allow
8274 extensions to be obtained and added.
8277 *) -crlf option to s_client and s_server for sending newlines as
8278 CRLF (as required by many protocols).
8281 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
8283 *) Install libRSAglue.a when OpenSSL is built with RSAref.
8284 [Ralf S. Engelschall]
8286 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
8287 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
8289 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
8293 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
8294 DH parameters/keys (q is lost during that conversion, but the resulting
8295 DH parameters contain its length).
8297 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
8298 much faster than DH_generate_parameters (which creates parameters
8299 where p = 2*q + 1), and also the smaller q makes DH computations
8300 much more efficient (160-bit exponentiation instead of 1024-bit
8301 exponentiation); so this provides a convenient way to support DHE
8302 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
8303 utter importance to use
8304 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8306 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8307 when such DH parameters are used, because otherwise small subgroup
8308 attacks may become possible!
8311 *) Avoid memory leak in i2d_DHparams.
8314 *) Allow the -k option to be used more than once in the enc program:
8315 this allows the same encrypted message to be read by multiple recipients.
8318 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
8319 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
8320 it will always use the numerical form of the OID, even if it has a short
8324 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
8325 method only got called if p,q,dmp1,dmq1,iqmp components were present,
8326 otherwise bn_mod_exp was called. In the case of hardware keys for example
8327 no private key components need be present and it might store extra data
8328 in the RSA structure, which cannot be accessed from bn_mod_exp.
8329 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
8330 private key operations.
8333 *) Added support for SPARC Linux.
8336 *) pem_password_cb function type incompatibly changed from
8337 typedef int pem_password_cb(char *buf, int size, int rwflag);
8339 ....(char *buf, int size, int rwflag, void *userdata);
8340 so that applications can pass data to their callbacks:
8341 The PEM[_ASN1]_{read,write}... functions and macros now take an
8342 additional void * argument, which is just handed through whenever
8343 the password callback is called.
8344 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
8346 New function SSL_CTX_set_default_passwd_cb_userdata.
8348 Compatibility note: As many C implementations push function arguments
8349 onto the stack in reverse order, the new library version is likely to
8350 interoperate with programs that have been compiled with the old
8351 pem_password_cb definition (PEM_whatever takes some data that
8352 happens to be on the stack as its last argument, and the callback
8353 just ignores this garbage); but there is no guarantee whatsoever that
8356 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
8357 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
8358 problems not only on Windows, but also on some Unix platforms.
8359 To avoid problematic command lines, these definitions are now in an
8360 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
8361 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
8364 *) MIPS III/IV assembler module is reimplemented.
8367 *) More DES library cleanups: remove references to srand/rand and
8368 delete an unused file.
8371 *) Add support for the the free Netwide assembler (NASM) under Win32,
8372 since not many people have MASM (ml) and it can be hard to obtain.
8373 This is currently experimental but it seems to work OK and pass all
8374 the tests. Check out INSTALL.W32 for info.
8377 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
8378 without temporary keys kept an extra copy of the server key,
8379 and connections with temporary keys did not free everything in case
8383 *) New function RSA_check_key and new openssl rsa option -check
8384 for verifying the consistency of RSA keys.
8385 [Ulf Moeller, Bodo Moeller]
8387 *) Various changes to make Win32 compile work:
8388 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
8389 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
8390 comparison" warnings.
8391 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
8394 *) Add a debugging option to PKCS#5 v2 key generation function: when
8395 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
8396 derived keys are printed to stderr.
8399 *) Copy the flags in ASN1_STRING_dup().
8400 [Roman E. Pavlov <pre@mo.msk.ru>]
8402 *) The x509 application mishandled signing requests containing DSA
8403 keys when the signing key was also DSA and the parameters didn't match.
8405 It was supposed to omit the parameters when they matched the signing key:
8406 the verifying software was then supposed to automatically use the CA's
8407 parameters if they were absent from the end user certificate.
8409 Omitting parameters is no longer recommended. The test was also
8410 the wrong way round! This was probably due to unusual behaviour in
8411 EVP_cmp_parameters() which returns 1 if the parameters match.
8412 This meant that parameters were omitted when they *didn't* match and
8413 the certificate was useless. Certificates signed with 'ca' didn't have
8415 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
8417 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
8418 The interface is as follows:
8419 Applications can use
8420 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
8421 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
8422 "off" is now the default.
8423 The library internally uses
8424 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
8425 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
8426 to disable memory-checking temporarily.
8428 Some inconsistent states that previously were possible (and were
8429 even the default) are now avoided.
8431 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
8432 with each memory chunk allocated; this is occasionally more helpful
8433 than just having a counter.
8435 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
8437 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
8441 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
8442 which largely parallels "options", but is for changing API behaviour,
8443 whereas "options" are about protocol behaviour.
8444 Initial "mode" flags are:
8446 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
8447 a single record has been written.
8448 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
8449 retries use the same buffer location.
8450 (But all of the contents must be
8454 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
8457 *) Fix problems with no-hmac etc.
8458 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
8460 *) New functions RSA_get_default_method(), RSA_set_method() and
8461 RSA_get_method(). These allows replacement of RSA_METHODs without having
8462 to mess around with the internals of an RSA structure.
8465 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
8466 Also really enable memory leak checks in openssl.c and in some
8468 [Chad C. Mulligan, Bodo Moeller]
8470 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
8471 up the length of negative integers. This has now been simplified to just
8472 store the length when it is first determined and use it later, rather
8473 than trying to keep track of where data is copied and updating it to
8475 [Steve Henson, reported by Brien Wheeler
8476 <bwheeler@authentica-security.com>]
8478 *) Add a new function PKCS7_signatureVerify. This allows the verification
8479 of a PKCS#7 signature but with the signing certificate passed to the
8480 function itself. This contrasts with PKCS7_dataVerify which assumes the
8481 certificate is present in the PKCS#7 structure. This isn't always the
8482 case: certificates can be omitted from a PKCS#7 structure and be
8483 distributed by "out of band" means (such as a certificate database).
8486 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
8487 function prototypes in pem.h, also change util/mkdef.pl to add the
8488 necessary function names.
8491 *) mk1mf.pl (used by Windows builds) did not properly read the
8492 options set by Configure in the top level Makefile, and Configure
8493 was not even able to write more than one option correctly.
8494 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
8497 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
8498 file to be loaded from a BIO or FILE pointer. The BIO version will
8499 for example allow memory BIOs to contain config info.
8502 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
8503 Whoever hopes to achieve shared-library compatibility across versions
8504 must use this, not the compile-time macro.
8505 (Exercise 0.9.4: Which is the minimum library version required by
8507 Note: All this applies only to multi-threaded programs, others don't
8511 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
8512 through a BIO pair triggered the default case, i.e.
8513 SSLerr(...,SSL_R_UNKNOWN_STATE).
8516 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
8517 can use the SSL library even if none of the specific BIOs is
8521 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
8522 for the encoded length.
8523 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
8525 *) Add initial documentation of the X509V3 functions.
8528 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
8529 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
8530 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
8531 secure PKCS#8 private key format with a high iteration count.
8534 *) Fix determination of Perl interpreter: A perl or perl5
8535 _directory_ in $PATH was also accepted as the interpreter.
8536 [Ralf S. Engelschall]
8538 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
8539 wrong with it but it was very old and did things like calling
8540 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
8544 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
8545 to use the new extension code.
8548 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
8549 with macros. This should make it easier to change their form, add extra
8550 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
8554 *) Add to configuration table a new entry that can specify an alternative
8555 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
8556 according to Mark Crispin <MRC@Panda.COM>.
8560 *) DES CBC did not update the IV. Weird.
8563 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
8564 Changing the behaviour of the former might break existing programs --
8565 where IV updating is needed, des_ncbc_encrypt can be used.
8568 *) When bntest is run from "make test" it drives bc to check its
8569 calculations, as well as internally checking them. If an internal check
8570 fails, it needs to cause bc to give a non-zero result or make test carries
8571 on without noticing the failure. Fixed.
8574 *) DES library cleanups.
8577 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
8578 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
8579 ciphers. NOTE: although the key derivation function has been verified
8580 against some published test vectors it has not been extensively tested
8581 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
8585 *) Instead of "mkdir -p", which is not fully portable, use new
8586 Perl script "util/mkdir-p.pl".
8589 *) Rewrite the way password based encryption (PBE) is handled. It used to
8590 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
8591 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
8592 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
8593 the 'parameter' field of the AlgorithmIdentifier is passed to the
8594 underlying key generation function so it must do its own ASN1 parsing.
8595 This has also changed the EVP_PBE_CipherInit() function which now has a
8596 'parameter' argument instead of literal salt and iteration count values
8597 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
8600 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
8601 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
8602 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
8603 KEY" because this clashed with PKCS#8 unencrypted string. Since this
8604 value was just used as a "magic string" and not used directly its
8605 value doesn't matter.
8608 *) Introduce some semblance of const correctness to BN. Shame C doesn't
8612 *) "linux-sparc64" configuration (ultrapenguin).
8613 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
8614 "linux-sparc" configuration.
8615 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
8617 *) config now generates no-xxx options for missing ciphers.
8620 *) Support the EBCDIC character set (work in progress).
8621 File ebcdic.c not yet included because it has a different license.
8622 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8624 *) Support BS2000/OSD-POSIX.
8625 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8627 *) Make callbacks for key generation use void * instead of char *.
8630 *) Make S/MIME samples compile (not yet tested).
8633 *) Additional typesafe stacks.
8636 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
8640 Changes between 0.9.3 and 0.9.3a [29 May 1999]
8642 *) New configuration variant "sco5-gcc".
8644 *) Updated some demos.
8645 [Sean O Riordain, Wade Scholine]
8647 *) Add missing BIO_free at exit of pkcs12 application.
8650 *) Fix memory leak in conf.c.
8653 *) Updates for Win32 to assembler version of MD5.
8656 *) Set #! path to perl in apps/der_chop to where we found it
8657 instead of using a fixed path.
8660 *) SHA library changes for irix64-mips4-cc.
8663 *) Improvements for VMS support.
8667 Changes between 0.9.2b and 0.9.3 [24 May 1999]
8669 *) Bignum library bug fix. IRIX 6 passes "make test" now!
8670 This also avoids the problems with SC4.2 and unpatched SC5.
8671 [Andy Polyakov <appro@fy.chalmers.se>]
8673 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
8674 These are required because of the typesafe stack would otherwise break
8675 existing code. If old code used a structure member which used to be STACK
8676 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
8677 sk_num or sk_value it would produce an error because the num, data members
8678 are not present in STACK_OF. Now it just produces a warning. sk_set
8679 replaces the old method of assigning a value to sk_value
8680 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
8681 that does this will no longer work (and should use sk_set instead) but
8682 this could be regarded as a "questionable" behaviour anyway.
8685 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
8686 correctly handle encrypted S/MIME data.
8689 *) Change type of various DES function arguments from des_cblock
8690 (which means, in function argument declarations, pointer to char)
8691 to des_cblock * (meaning pointer to array with 8 char elements),
8692 which allows the compiler to do more typechecking; it was like
8693 that back in SSLeay, but with lots of ugly casts.
8695 Introduce new type const_des_cblock.
8698 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
8699 problems: find RecipientInfo structure that matches recipient certificate
8700 and initialise the ASN1 structures properly based on passed cipher.
8703 *) Belatedly make the BN tests actually check the results.
8706 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
8707 to and from BNs: it was completely broken. New compilation option
8708 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
8709 key elements as negative integers.
8712 *) Reorganize and speed up MD5.
8713 [Andy Polyakov <appro@fy.chalmers.se>]
8716 [Richard Levitte <richard@levitte.org>]
8718 *) New option -out to asn1parse to allow the parsed structure to be
8719 output to a file. This is most useful when combined with the -strparse
8720 option to examine the output of things like OCTET STRINGS.
8723 *) Make SSL library a little more fool-proof by not requiring any longer
8724 that SSL_set_{accept,connect}_state be called before
8725 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
8726 in many applications because usually everything *appeared* to work as
8727 intended anyway -- now it really works as intended).
8730 *) Move openssl.cnf out of lib/.
8733 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
8734 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
8735 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
8736 [Ralf S. Engelschall]
8738 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
8739 handle PKCS#7 enveloped data properly.
8740 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
8742 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
8743 copying pointers. The cert_st handling is changed by this in
8744 various ways (and thus what used to be known as ctx->default_cert
8745 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
8746 any longer when s->cert does not give us what we need).
8747 ssl_cert_instantiate becomes obsolete by this change.
8748 As soon as we've got the new code right (possibly it already is?),
8749 we have solved a couple of bugs of the earlier code where s->cert
8750 was used as if it could not have been shared with other SSL structures.
8752 Note that using the SSL API in certain dirty ways now will result
8753 in different behaviour than observed with earlier library versions:
8754 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
8755 does not influence s as it used to.
8757 In order to clean up things more thoroughly, inside SSL_SESSION
8758 we don't use CERT any longer, but a new structure SESS_CERT
8759 that holds per-session data (if available); currently, this is
8760 the peer's certificate chain and, for clients, the server's certificate
8761 and temporary key. CERT holds only those values that can have
8762 meaningful defaults in an SSL_CTX.
8765 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
8766 from the internal representation. Various PKCS#7 fixes: remove some
8767 evil casts and set the enc_dig_alg field properly based on the signing
8771 *) Allow PKCS#12 password to be set from the command line or the
8772 environment. Let 'ca' get its config file name from the environment
8773 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
8777 *) Allow certificate policies extension to use an IA5STRING for the
8778 organization field. This is contrary to the PKIX definition but
8779 VeriSign uses it and IE5 only recognises this form. Document 'x509'
8783 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
8784 without disallowing inline assembler and the like for non-pedantic builds.
8787 *) Support Borland C++ builder.
8788 [Janez Jere <jj@void.si>, modified by Ulf Möller]
8793 *) SHA-1 cleanups and performance enhancements.
8794 [Andy Polyakov <appro@fy.chalmers.se>]
8796 *) Sparc v8plus assembler for the bignum library.
8797 [Andy Polyakov <appro@fy.chalmers.se>]
8799 *) Accept any -xxx and +xxx compiler options in Configure.
8802 *) Update HPUX configuration.
8805 *) Add missing sk_<type>_unshift() function to safestack.h
8806 [Ralf S. Engelschall]
8808 *) New function SSL_CTX_use_certificate_chain_file that sets the
8809 "extra_cert"s in addition to the certificate. (This makes sense
8810 only for "PEM" format files, as chains as a whole are not
8814 *) Support verify_depth from the SSL API.
8815 x509_vfy.c had what can be considered an off-by-one-error:
8816 Its depth (which was not part of the external interface)
8817 was actually counting the number of certificates in a chain;
8818 now it really counts the depth.
8821 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
8822 instead of X509err, which often resulted in confusing error
8823 messages since the error codes are not globally unique
8824 (e.g. an alleged error in ssl3_accept when a certificate
8825 didn't match the private key).
8827 *) New function SSL_CTX_set_session_id_context that allows to set a default
8828 value (so that you don't need SSL_set_session_id_context for each
8829 connection using the SSL_CTX).
8832 *) OAEP decoding bug fix.
8835 *) Support INSTALL_PREFIX for package builders, as proposed by
8839 *) New Configure options "threads" and "no-threads". For systems
8840 where the proper compiler options are known (currently Solaris
8841 and Linux), "threads" is the default.
8844 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
8847 *) Install various scripts to $(OPENSSLDIR)/misc, not to
8848 $(INSTALLTOP)/bin -- they shouldn't clutter directories
8849 such as /usr/local/bin.
8852 *) "make linux-shared" to build shared libraries.
8853 [Niels Poppe <niels@netbox.org>]
8855 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
8858 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
8859 extension adding in x509 utility.
8862 *) Remove NOPROTO sections and error code comments.
8865 *) Partial rewrite of the DEF file generator to now parse the ANSI
8869 *) New Configure options --prefix=DIR and --openssldir=DIR.
8872 *) Complete rewrite of the error code script(s). It is all now handled
8873 by one script at the top level which handles error code gathering,
8874 header rewriting and C source file generation. It should be much better
8875 than the old method: it now uses a modified version of Ulf's parser to
8876 read the ANSI prototypes in all header files (thus the old K&R definitions
8877 aren't needed for error creation any more) and do a better job of
8878 translating function codes into names. The old 'ASN1 error code imbedded
8879 in a comment' is no longer necessary and it doesn't use .err files which
8880 have now been deleted. Also the error code call doesn't have to appear all
8881 on one line (which resulted in some large lines...).
8884 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
8887 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
8888 0 (which usually indicates a closed connection), but continue reading.
8891 *) Fix some race conditions.
8894 *) Add support for CRL distribution points extension. Add Certificate
8895 Policies and CRL distribution points documentation.
8898 *) Move the autogenerated header file parts to crypto/opensslconf.h.
8901 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
8902 8 of keying material. Merlin has also confirmed interop with this fix
8903 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
8904 [Merlin Hughes <merlin@baltimore.ie>]
8906 *) Fix lots of warnings.
8907 [Richard Levitte <levitte@stacken.kth.se>]
8909 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
8910 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
8911 [Richard Levitte <levitte@stacken.kth.se>]
8913 *) Fix problems with sizeof(long) == 8.
8914 [Andy Polyakov <appro@fy.chalmers.se>]
8916 *) Change functions to ANSI C.
8919 *) Fix typos in error codes.
8920 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
8922 *) Remove defunct assembler files from Configure.
8925 *) SPARC v8 assembler BIGNUM implementation.
8926 [Andy Polyakov <appro@fy.chalmers.se>]
8928 *) Support for Certificate Policies extension: both print and set.
8929 Various additions to support the r2i method this uses.
8932 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
8933 return a const string when you are expecting an allocated buffer.
8936 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
8937 types DirectoryString and DisplayText.
8940 *) Add code to allow r2i extensions to access the configuration database,
8941 add an LHASH database driver and add several ctx helper functions.
8944 *) Fix an evil bug in bn_expand2() which caused various BN functions to
8945 fail when they extended the size of a BIGNUM.
8948 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
8949 support typesafe stack.
8952 *) Fix typo in SSL_[gs]et_options().
8953 [Nils Frostberg <nils@medcom.se>]
8955 *) Delete various functions and files that belonged to the (now obsolete)
8956 old X509V3 handling code.
8959 *) New Configure option "rsaref".
8962 *) Don't auto-generate pem.h.
8965 *) Introduce type-safe ASN.1 SETs.
8968 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
8969 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
8971 *) Introduce type-safe STACKs. This will almost certainly break lots of code
8972 that links with OpenSSL (well at least cause lots of warnings), but fear
8973 not: the conversion is trivial, and it eliminates loads of evil casts. A
8974 few STACKed things have been converted already. Feel free to convert more.
8975 In the fullness of time, I'll do away with the STACK type altogether.
8978 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
8979 specified in <certfile> by updating the entry in the index.txt file.
8980 This way one no longer has to edit the index.txt file manually for
8981 revoking a certificate. The -revoke option does the gory details now.
8982 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
8984 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
8985 `-text' option at all and this way the `-noout -text' combination was
8986 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
8987 [Ralf S. Engelschall]
8989 *) Make sure a corresponding plain text error message exists for the
8990 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
8991 verify callback function determined that a certificate was revoked.
8992 [Ralf S. Engelschall]
8994 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
8995 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
8996 all available cipers including rc5, which was forgotten until now.
8997 In order to let the testing shell script know which algorithms
8998 are available, a new (up to now undocumented) command
8999 "openssl list-cipher-commands" is used.
9002 *) Bugfix: s_client occasionally would sleep in select() when
9003 it should have checked SSL_pending() first.
9006 *) New functions DSA_do_sign and DSA_do_verify to provide access to
9007 the raw DSA values prior to ASN.1 encoding.
9010 *) Tweaks to Configure
9011 [Niels Poppe <niels@netbox.org>]
9013 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
9017 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
9020 *) New config option to avoid instructions that are illegal on the 80386.
9021 The default code is faster, but requires at least a 486.
9024 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
9025 SSL2_SERVER_VERSION (not used at all) macros, which are now the
9026 same as SSL2_VERSION anyway.
9029 *) New "-showcerts" option for s_client.
9032 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
9033 application. Various cleanups and fixes.
9036 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
9037 modify error routines to work internally. Add error codes and PBE init
9038 to library startup routines.
9041 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
9042 packing functions to asn1 and evp. Changed function names and error
9043 codes along the way.
9046 *) PKCS12 integration: and so it begins... First of several patches to
9047 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
9048 objects to objects.h
9051 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
9052 and display support for Thawte strong extranet extension.
9055 *) Add LinuxPPC support.
9056 [Jeff Dubrule <igor@pobox.org>]
9058 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
9059 bn_div_words in alpha.s.
9060 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
9062 *) Make sure the RSA OAEP test is skipped under -DRSAref because
9063 OAEP isn't supported when OpenSSL is built with RSAref.
9064 [Ulf Moeller <ulf@fitug.de>]
9066 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
9067 so they no longer are missing under -DNOPROTO.
9068 [Soren S. Jorvang <soren@t.dk>]
9071 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
9073 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
9074 doesn't work when the session is reused. Coming soon!
9077 *) Fix a security hole, that allows sessions to be reused in the wrong
9078 context thus bypassing client cert protection! All software that uses
9079 client certs and session caches in multiple contexts NEEDS PATCHING to
9080 allow session reuse! A fuller solution is in the works.
9081 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
9083 *) Some more source tree cleanups (removed obsolete files
9084 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
9085 permission on "config" script to be executable) and a fix for the INSTALL
9087 [Ulf Moeller <ulf@fitug.de>]
9089 *) Remove some legacy and erroneous uses of malloc, free instead of
9091 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
9093 *) Make rsa_oaep_test return non-zero on error.
9094 [Ulf Moeller <ulf@fitug.de>]
9096 *) Add support for native Solaris shared libraries. Configure
9097 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
9098 if someone would make that last step automatic.
9099 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
9101 *) ctx_size was not built with the right compiler during "make links". Fixed.
9104 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
9105 except NULL ciphers". This means the default cipher list will no longer
9106 enable NULL ciphers. They need to be specifically enabled e.g. with
9107 the string "DEFAULT:eNULL".
9110 *) Fix to RSA private encryption routines: if p < q then it would
9111 occasionally produce an invalid result. This will only happen with
9112 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9115 *) Be less restrictive and allow also `perl util/perlpath.pl
9116 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9117 because this way one can also use an interpreter named `perl5' (which is
9118 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9119 installed as `perl').
9120 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9122 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9123 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9125 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9126 advapi32.lib to Win32 build and change the pem test comparision
9127 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9128 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9129 and crypto/des/ede_cbcm_enc.c.
9132 *) DES quad checksum was broken on big-endian architectures. Fixed.
9135 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9136 Win32 test batch file so it (might) work again. The Win32 test batch file
9137 is horrible: I feel ill....
9140 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
9141 in e_os.h. Audit of header files to check ANSI and non ANSI
9142 sections: 10 functions were absent from non ANSI section and not exported
9143 from Windows DLLs. Fixed up libeay.num for new functions.
9146 *) Make `openssl version' output lines consistent.
9147 [Ralf S. Engelschall]
9149 *) Fix Win32 symbol export lists for BIO functions: Added
9150 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
9151 to ms/libeay{16,32}.def.
9152 [Ralf S. Engelschall]
9154 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
9155 fine under Unix and passes some trivial tests I've now added. But the
9156 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
9157 added to make sure no one expects that this stuff really works in the
9158 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
9159 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
9161 [Ralf S. Engelschall]
9163 *) Fix the generation of two part addresses in perl.
9164 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
9166 *) Add config entry for Linux on MIPS.
9167 [John Tobey <jtobey@channel1.com>]
9169 *) Make links whenever Configure is run, unless we are on Windoze.
9172 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
9173 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
9177 *) Add a useful kludge to allow package maintainers to specify compiler and
9178 other platforms details on the command line without having to patch the
9179 Configure script everytime: One now can use ``perl Configure
9180 <id>:<details>'', i.e. platform ids are allowed to have details appended
9181 to them (seperated by colons). This is treated as there would be a static
9182 pre-configured entry in Configure's %table under key <id> with value
9183 <details> and ``perl Configure <id>'' is called. So, when you want to
9184 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
9185 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
9186 now, which overrides the FreeBSD-elf entry on-the-fly.
9187 [Ralf S. Engelschall]
9189 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
9192 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
9193 on the `perl Configure ...' command line. This way one can compile
9194 OpenSSL libraries with Position Independent Code (PIC) which is needed
9195 for linking it into DSOs.
9196 [Ralf S. Engelschall]
9198 *) Remarkably, export ciphers were totally broken and no-one had noticed!
9202 *) Cleaned up the LICENSE document: The official contact for any license
9203 questions now is the OpenSSL core team under openssl-core@openssl.org.
9204 And add a paragraph about the dual-license situation to make sure people
9205 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
9206 to the OpenSSL toolkit.
9207 [Ralf S. Engelschall]
9209 *) General source tree makefile cleanups: Made `making xxx in yyy...'
9210 display consistent in the source tree and replaced `/bin/rm' by `rm'.
9211 Additonally cleaned up the `make links' target: Remove unnecessary
9212 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
9213 to speed processing and no longer clutter the display with confusing
9214 stuff. Instead only the actually done links are displayed.
9215 [Ralf S. Engelschall]
9217 *) Permit null encryption ciphersuites, used for authentication only. It used
9218 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
9219 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
9223 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
9224 signed attributes when verifying signatures (this would break them),
9225 the detached data encoding was wrong and public keys obtained using
9226 X509_get_pubkey() weren't freed.
9229 *) Add text documentation for the BUFFER functions. Also added a work around
9230 to a Win95 console bug. This was triggered by the password read stuff: the
9231 last character typed gets carried over to the next fread(). If you were
9232 generating a new cert request using 'req' for example then the last
9233 character of the passphrase would be CR which would then enter the first
9237 *) Added the new `Includes OpenSSL Cryptography Software' button as
9238 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
9239 button and can be used by applications based on OpenSSL to show the
9240 relationship to the OpenSSL project.
9241 [Ralf S. Engelschall]
9243 *) Remove confusing variables in function signatures in files
9244 ssl/ssl_lib.c and ssl/ssl.h.
9245 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9247 *) Don't install bss_file.c under PREFIX/include/
9248 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9250 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
9251 functions that return function pointers and has support for NT specific
9252 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
9253 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
9254 unsigned to signed types: this was killing the Win32 compile.
9257 *) Add new certificate file to stack functions,
9258 SSL_add_dir_cert_subjects_to_stack() and
9259 SSL_add_file_cert_subjects_to_stack(). These largely supplant
9260 SSL_load_client_CA_file(), and can be used to add multiple certs easily
9261 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
9262 This means that Apache-SSL and similar packages don't have to mess around
9263 to add as many CAs as they want to the preferred list.
9266 *) Experiment with doxygen documentation. Currently only partially applied to
9268 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
9269 openssl.doxy as the configuration file.
9272 *) Get rid of remaining C++-style comments which strict C compilers hate.
9273 [Ralf S. Engelschall, pointed out by Carlos Amengual]
9275 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
9276 compiled in by default: it has problems with large keys.
9279 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
9280 DH private keys and/or callback functions which directly correspond to
9281 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
9282 is needed for applications which have to configure certificates on a
9283 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
9285 For the RSA certificate situation is makes no difference, but
9286 for the DSA certificate situation this fixes the "no shared cipher"
9287 problem where the OpenSSL cipher selection procedure failed because the
9288 temporary keys were not overtaken from the context and the API provided
9289 no way to reconfigure them.
9290 The new functions now let applications reconfigure the stuff and they
9291 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
9292 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
9293 non-public-API function ssl_cert_instantiate() is used as a helper
9294 function and also to reduce code redundancy inside ssl_rsa.c.
9295 [Ralf S. Engelschall]
9297 *) Move s_server -dcert and -dkey options out of the undocumented feature
9298 area because they are useful for the DSA situation and should be
9299 recognized by the users.
9300 [Ralf S. Engelschall]
9302 *) Fix the cipher decision scheme for export ciphers: the export bits are
9303 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
9304 SSL_EXP_MASK. So, the original variable has to be used instead of the
9305 already masked variable.
9306 [Richard Levitte <levitte@stacken.kth.se>]
9308 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
9309 [Richard Levitte <levitte@stacken.kth.se>]
9311 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
9312 from `int' to `unsigned int' because it's a length and initialized by
9313 EVP_DigestFinal() which expects an `unsigned int *'.
9314 [Richard Levitte <levitte@stacken.kth.se>]
9316 *) Don't hard-code path to Perl interpreter on shebang line of Configure
9317 script. Instead use the usual Shell->Perl transition trick.
9318 [Ralf S. Engelschall]
9320 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
9321 (in addition to RSA certificates) to match the behaviour of `openssl dsa
9322 -noout -modulus' as it's already the case for `openssl rsa -noout
9323 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
9324 currently the public key is printed (a decision which was already done by
9325 `openssl dsa -modulus' in the past) which serves a similar purpose.
9326 Additionally the NO_RSA no longer completely removes the whole -modulus
9327 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
9329 [Ralf S. Engelschall]
9331 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
9332 BIO. See the source (crypto/evp/bio_ok.c) for more info.
9333 [Arne Ansper <arne@ats.cyber.ee>]
9335 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
9336 to be added. Now both 'req' and 'ca' can use new objects defined in the
9340 *) Add cool BIO that does syslog (or event log on NT).
9341 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
9343 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
9344 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
9345 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
9346 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
9349 *) Add preliminary config info for new extension code.
9352 *) Make RSA_NO_PADDING really use no padding.
9353 [Ulf Moeller <ulf@fitug.de>]
9355 *) Generate errors when private/public key check is done.
9358 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
9359 for some CRL extensions and new objects added.
9362 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
9363 key usage extension and fuller support for authority key id.
9366 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
9367 padding method for RSA, which is recommended for new applications in PKCS
9368 #1 v2.0 (RFC 2437, October 1998).
9369 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
9370 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
9371 against Bleichbacher's attack on RSA.
9372 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
9375 *) Updates to the new SSL compression code
9376 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9378 *) Fix so that the version number in the master secret, when passed
9379 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
9380 (because the server will not accept higher), that the version number
9381 is 0x03,0x01, not 0x03,0x00
9382 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9384 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
9385 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
9386 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
9389 *) Support for RAW extensions where an arbitrary extension can be
9390 created by including its DER encoding. See apps/openssl.cnf for
9394 *) Make sure latest Perl versions don't interpret some generated C array
9395 code as Perl array code in the crypto/err/err_genc.pl script.
9396 [Lars Weber <3weber@informatik.uni-hamburg.de>]
9398 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
9399 not many people have the assembler. Various Win32 compilation fixes and
9400 update to the INSTALL.W32 file with (hopefully) more accurate Win32
9404 *) Modify configure script 'Configure' to automatically create crypto/date.h
9405 file under Win32 and also build pem.h from pem.org. New script
9406 util/mkfiles.pl to create the MINFO file on environments that can't do a
9407 'make files': perl util/mkfiles.pl >MINFO should work.
9410 *) Major rework of DES function declarations, in the pursuit of correctness
9411 and purity. As a result, many evil casts evaporated, and some weirdness,
9412 too. You may find this causes warnings in your code. Zapping your evil
9413 casts will probably fix them. Mostly.
9416 *) Fix for a typo in asn1.h. Bug fix to object creation script
9417 obj_dat.pl. It considered a zero in an object definition to mean
9418 "end of object": none of the objects in objects.h have any zeros
9419 so it wasn't spotted.
9420 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
9422 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
9423 Masking (CBCM). In the absence of test vectors, the best I have been able
9424 to do is check that the decrypt undoes the encrypt, so far. Send me test
9425 vectors if you have them.
9428 *) Correct calculation of key length for export ciphers (too much space was
9429 allocated for null ciphers). This has not been tested!
9432 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
9433 message is now correct (it understands "crypto" and "ssl" on its
9434 command line). There is also now an "update" option. This will update
9435 the util/ssleay.num and util/libeay.num files with any new functions.
9437 perl util/mkdef.pl crypto ssl update
9438 it will update them.
9441 *) Overhauled the Perl interface (perl/*):
9442 - ported BN stuff to OpenSSL's different BN library
9443 - made the perl/ source tree CVS-aware
9444 - renamed the package from SSLeay to OpenSSL (the files still contain
9445 their history because I've copied them in the repository)
9446 - removed obsolete files (the test scripts will be replaced
9447 by better Test::Harness variants in the future)
9448 [Ralf S. Engelschall]
9450 *) First cut for a very conservative source tree cleanup:
9451 1. merge various obsolete readme texts into doc/ssleay.txt
9452 where we collect the old documents and readme texts.
9453 2. remove the first part of files where I'm already sure that we no
9454 longer need them because of three reasons: either they are just temporary
9455 files which were left by Eric or they are preserved original files where
9456 I've verified that the diff is also available in the CVS via "cvs diff
9457 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
9458 the crypto/md/ stuff).
9459 [Ralf S. Engelschall]
9461 *) More extension code. Incomplete support for subject and issuer alt
9462 name, issuer and authority key id. Change the i2v function parameters
9463 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
9464 what that's for :-) Fix to ASN1 macro which messed up
9465 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
9468 *) Preliminary support for ENUMERATED type. This is largely copied from the
9472 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
9473 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9475 *) Make sure `make rehash' target really finds the `openssl' program.
9476 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9478 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
9479 like to hear about it if this slows down other processors.
9482 *) Add CygWin32 platform information to Configure script.
9483 [Alan Batie <batie@aahz.jf.intel.com>]
9485 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
9486 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
9488 *) New program nseq to manipulate netscape certificate sequences
9491 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
9495 *) Fixes to BN code. Previously the default was to define BN_RECURSION
9496 but the BN code had some problems that would cause failures when
9497 doing certificate verification and some other functions.
9498 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9500 *) Add ASN1 and PEM code to support netscape certificate sequences.
9503 *) Add ASN1 and PEM code to support netscape certificate sequences.
9506 *) Add several PKIX and private extended key usage OIDs.
9509 *) Modify the 'ca' program to handle the new extension code. Modify
9510 openssl.cnf for new extension format, add comments.
9513 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
9514 and add a sample to openssl.cnf so req -x509 now adds appropriate
9518 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
9519 error code, add initial support to X509_print() and x509 application.
9522 *) Takes a deep breath and start addding X509 V3 extension support code. Add
9523 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
9524 stuff is currently isolated and isn't even compiled yet.
9527 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
9528 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
9529 Removed the versions check from X509 routines when loading extensions:
9530 this allows certain broken certificates that don't set the version
9531 properly to be processed.
9534 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
9535 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
9536 can still be regenerated with "make depend".
9539 *) Spelling mistake in C version of CAST-128.
9540 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
9542 *) Changes to the error generation code. The perl script err-code.pl
9543 now reads in the old error codes and retains the old numbers, only
9544 adding new ones if necessary. It also only changes the .err files if new
9545 codes are added. The makefiles have been modified to only insert errors
9546 when needed (to avoid needlessly modifying header files). This is done
9547 by only inserting errors if the .err file is newer than the auto generated
9548 C file. To rebuild all the error codes from scratch (the old behaviour)
9549 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
9550 or delete all the .err files.
9553 *) CAST-128 was incorrectly implemented for short keys. The C version has
9554 been fixed, but is untested. The assembler versions are also fixed, but
9555 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
9556 to regenerate it if needed.
9557 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
9558 Hagino <itojun@kame.net>]
9560 *) File was opened incorrectly in randfile.c.
9561 [Ulf Möller <ulf@fitug.de>]
9563 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
9564 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
9565 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
9566 al: it's just almost always a UTCTime. Note this patch adds new error
9567 codes so do a "make errors" if there are problems.
9570 *) Correct Linux 1 recognition in config.
9571 [Ulf Möller <ulf@fitug.de>]
9573 *) Remove pointless MD5 hash when using DSA keys in ca.
9574 [Anonymous <nobody@replay.com>]
9576 *) Generate an error if given an empty string as a cert directory. Also
9577 generate an error if handed NULL (previously returned 0 to indicate an
9578 error, but didn't set one).
9579 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
9581 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
9584 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
9585 parameters. This was causing a warning which killed off the Win32 compile.
9588 *) Remove C++ style comments from crypto/bn/bn_local.h.
9589 [Neil Costigan <neil.costigan@celocom.com>]
9591 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
9592 based on a text string, looking up short and long names and finally
9593 "dot" format. The "dot" format stuff didn't work. Added new function
9594 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
9595 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
9596 OID is not part of the table.
9599 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
9600 X509_LOOKUP_by_alias().
9603 *) Sort openssl functions by name.
9606 *) Get the gendsa program working (hopefully) and add it to app list. Remove
9607 encryption from sample DSA keys (in case anyone is interested the password
9611 *) Make _all_ *_free functions accept a NULL pointer.
9612 [Frans Heymans <fheymans@isaserver.be>]
9614 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
9616 [Anonymous <nobody@replay.com>]
9618 *) s_server should send the CAfile as acceptable CAs, not its own cert.
9619 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9621 *) Don't blow it for numeric -newkey arguments to apps/req.
9622 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9624 *) Temp key "for export" tests were wrong in s3_srvr.c.
9625 [Anonymous <nobody@replay.com>]
9627 *) Add prototype for temp key callback functions
9628 SSL_CTX_set_tmp_{rsa,dh}_callback().
9631 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
9632 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
9635 *) X509_name_add_entry() freed the wrong thing after an error.
9636 [Arne Ansper <arne@ats.cyber.ee>]
9638 *) rsa_eay.c would attempt to free a NULL context.
9639 [Arne Ansper <arne@ats.cyber.ee>]
9641 *) BIO_s_socket() had a broken should_retry() on Windoze.
9642 [Arne Ansper <arne@ats.cyber.ee>]
9644 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
9645 [Arne Ansper <arne@ats.cyber.ee>]
9647 *) Make sure the already existing X509_STORE->depth variable is initialized
9648 in X509_STORE_new(), but document the fact that this variable is still
9649 unused in the certificate verification process.
9650 [Ralf S. Engelschall]
9652 *) Fix the various library and apps files to free up pkeys obtained from
9653 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
9656 *) Fix reference counting in X509_PUBKEY_get(). This makes
9657 demos/maurice/example2.c work, amongst others, probably.
9658 [Steve Henson and Ben Laurie]
9660 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
9661 `openssl' and second, the shortcut symlinks for the `openssl <command>'
9662 are no longer created. This way we have a single and consistent command
9663 line interface `openssl <command>', similar to `cvs <command>'.
9664 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
9666 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
9667 BIT STRING wrapper always have zero unused bits.
9670 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
9673 *) Make the top-level INSTALL documentation easier to understand.
9676 *) Makefiles updated to exit if an error occurs in a sub-directory
9677 make (including if user presses ^C) [Paul Sutton]
9679 *) Make Montgomery context stuff explicit in RSA data structure.
9682 *) Fix build order of pem and err to allow for generated pem.h.
9685 *) Fix renumbering bug in X509_NAME_delete_entry().
9688 *) Enhanced the err-ins.pl script so it makes the error library number
9689 global and can add a library name. This is needed for external ASN1 and
9690 other error libraries.
9693 *) Fixed sk_insert which never worked properly.
9696 *) Fix ASN1 macros so they can handle indefinite length construted
9697 EXPLICIT tags. Some non standard certificates use these: they can now
9701 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
9702 into a single doc/ssleay.txt bundle. This way the information is still
9703 preserved but no longer messes up this directory. Now it's new room for
9704 the new set of documenation files.
9705 [Ralf S. Engelschall]
9707 *) SETs were incorrectly DER encoded. This was a major pain, because they
9708 shared code with SEQUENCEs, which aren't coded the same. This means that
9709 almost everything to do with SETs or SEQUENCEs has either changed name or
9710 number of arguments.
9711 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
9713 *) Fix test data to work with the above.
9716 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
9717 was already fixed by Eric for 0.9.1 it seems.
9718 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
9720 *) Autodetect FreeBSD3.
9723 *) Fix various bugs in Configure. This affects the following platforms:
9727 unixware-2.0-pentium
9731 *) Eliminate generated files from CVS. Reorder tests to regenerate files
9732 before they are needed.
9735 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
9739 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
9741 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
9742 changed SSLeay to OpenSSL in version strings.
9743 [Ralf S. Engelschall]
9745 *) Some fixups to the top-level documents.
9748 *) Fixed the nasty bug where rsaref.h was not found under compile-time
9749 because the symlink to include/ was missing.
9750 [Ralf S. Engelschall]
9752 *) Incorporated the popular no-RSA/DSA-only patches
9753 which allow to compile a RSA-free SSLeay.
9754 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
9756 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
9757 when "ssleay" is still not found.
9758 [Ralf S. Engelschall]
9760 *) Added more platforms to Configure: Cray T3E, HPUX 11,
9761 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
9763 *) Updated the README file.
9764 [Ralf S. Engelschall]
9766 *) Added various .cvsignore files in the CVS repository subdirs
9767 to make a "cvs update" really silent.
9768 [Ralf S. Engelschall]
9770 *) Recompiled the error-definition header files and added
9771 missing symbols to the Win32 linker tables.
9772 [Ralf S. Engelschall]
9774 *) Cleaned up the top-level documents;
9775 o new files: CHANGES and LICENSE
9776 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
9777 o merged COPYRIGHT into LICENSE
9778 o removed obsolete TODO file
9779 o renamed MICROSOFT to INSTALL.W32
9780 [Ralf S. Engelschall]
9782 *) Removed dummy files from the 0.9.1b source tree:
9783 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
9784 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
9785 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
9786 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
9787 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
9788 [Ralf S. Engelschall]
9790 *) Added various platform portability fixes.
9793 *) The Genesis of the OpenSSL rpject:
9794 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
9795 Young and Tim J. Hudson created while they were working for C2Net until
9797 [The OpenSSL Project]
9800 Changes between 0.9.0b and 0.9.1b [not released]
9802 *) Updated a few CA certificates under certs/
9805 *) Changed some BIGNUM api stuff.
9808 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
9809 DGUX x86, Linux Alpha, etc.
9812 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
9813 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
9817 *) Add -strparse option to asn1pars program which parses nested
9819 [Dr Stephen Henson <shenson@bigfoot.com>]
9821 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
9824 *) DSA fix for "ca" program.
9827 *) Added "-genkey" option to "dsaparam" program.
9830 *) Added RIPE MD160 (rmd160) message digest.
9833 *) Added -a (all) option to "ssleay version" command.
9836 *) Added PLATFORM define which is the id given to Configure.
9839 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
9842 *) Extended the ASN.1 parser routines.
9845 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
9848 *) Added a BN_CTX to the BN library.
9851 *) Fixed the weak key values in DES library
9854 *) Changed API in EVP library for cipher aliases.
9857 *) Added support for RC2/64bit cipher.
9860 *) Converted the lhash library to the crypto/mem.c functions.
9863 *) Added more recognized ASN.1 object ids.
9866 *) Added more RSA padding checks for SSL/TLS.
9869 *) Added BIO proxy/filter functionality.
9872 *) Added extra_certs to SSL_CTX which can be used
9873 send extra CA certificates to the client in the CA cert chain sending
9874 process. It can be configured with SSL_CTX_add_extra_chain_cert().
9877 *) Now Fortezza is denied in the authentication phase because
9878 this is key exchange mechanism is not supported by SSLeay at all.
9881 *) Additional PKCS1 checks.
9884 *) Support the string "TLSv1" for all TLS v1 ciphers.
9887 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
9888 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
9891 *) Fixed a few memory leaks.
9894 *) Fixed various code and comment typos.
9897 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
9898 bytes sent in the client random.
9899 [Edward Bishop <ebishop@spyglass.com>]