5 Changes between 1.0.0q and 1.0.0r [xx XXX xxxx]
7 *) Segmentation fault in ASN1_TYPE_cmp fix
9 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
10 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
11 certificate signature algorithm consistency this can be used to crash any
12 certificate verification operation and exploited in a DoS attack. Any
13 application which performs certificate verification is vulnerable including
14 OpenSSL clients and servers which enable client authentication.
18 *) ASN.1 structure reuse memory corruption fix
20 Reusing a structure in ASN.1 parsing may allow an attacker to cause
21 memory corruption via an invalid write. Such reuse is and has been
22 strongly discouraged and is believed to be rare.
24 Applications that parse structures containing CHOICE or ANY DEFINED BY
25 components may be affected. Certificate parsing (d2i_X509 and related
26 functions) are however not affected. OpenSSL clients and servers are
31 *) PKCS7 NULL pointer dereferences fix
33 The PKCS#7 parsing code does not handle missing outer ContentInfo
34 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
35 missing content and trigger a NULL pointer dereference on parsing.
37 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
38 otherwise parse PKCS#7 structures from untrusted sources are
39 affected. OpenSSL clients and servers are not affected.
41 This issue was reported to OpenSSL by Michal Zalewski (Google).
45 *) DoS via reachable assert in SSLv2 servers fix
47 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
48 servers that both support SSLv2 and enable export cipher suites by sending
49 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
51 This issue was discovered by Sean Burford (Google) and Emilia Käsper
52 (OpenSSL development team).
56 *) Use After Free following d2i_ECPrivatekey error fix
58 A malformed EC private key file consumed via the d2i_ECPrivateKey function
59 could cause a use after free condition. This, in turn, could cause a double
60 free in several private key parsing functions (such as d2i_PrivateKey
61 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
62 for applications that receive EC private keys from untrusted
63 sources. This scenario is considered rare.
65 This issue was discovered by the BoringSSL project and fixed in their
70 *) X509_to_X509_REQ NULL pointer deref fix
72 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
73 the certificate key is invalid. This function is rarely used in practice.
75 This issue was discovered by Brian Carpenter.
79 *) Removed the export ciphers from the DEFAULT ciphers
82 Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
84 *) Build fixes for the Windows and OpenVMS platforms
85 [Matt Caswell and Richard Levitte]
87 Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
89 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
90 message can cause a segmentation fault in OpenSSL due to a NULL pointer
91 dereference. This could lead to a Denial Of Service attack. Thanks to
92 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
96 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
97 dtls1_buffer_record function under certain conditions. In particular this
98 could occur if an attacker sent repeated DTLS records with the same
99 sequence number but for the next epoch. The memory leak could be exploited
100 by an attacker in a Denial of Service attack through memory exhaustion.
101 Thanks to Chris Mueller for reporting this issue.
105 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
106 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
107 method would be set to NULL which could later result in a NULL pointer
108 dereference. Thanks to Frank Schmirler for reporting this issue.
112 *) Abort handshake if server key exchange message is omitted for ephemeral
115 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
116 reporting this issue.
120 *) Remove non-export ephemeral RSA code on client and server. This code
121 violated the TLS standard by allowing the use of temporary RSA keys in
122 non-export ciphersuites and could be used by a server to effectively
123 downgrade the RSA key length used to a value smaller than the server
124 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
125 INRIA or reporting this issue.
129 *) Fixed issue where DH client certificates are accepted without verification.
130 An OpenSSL server will accept a DH certificate for client authentication
131 without the certificate verify message. This effectively allows a client to
132 authenticate without the use of a private key. This only affects servers
133 which trust a client certificate authority which issues certificates
134 containing DH keys: these are extremely rare and hardly ever encountered.
135 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
140 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
141 results on some platforms, including x86_64. This bug occurs at random
142 with a very low probability, and is not known to be exploitable in any
143 way, though its exact impact is difficult to determine. Thanks to Pieter
144 Wuille (Blockstream) who reported this issue and also suggested an initial
145 fix. Further analysis was conducted by the OpenSSL development team and
146 Adam Langley of Google. The final fix was developed by Andy Polyakov of
147 the OpenSSL core team.
151 *) Fix various certificate fingerprint issues.
153 By using non-DER or invalid encodings outside the signed portion of a
154 certificate the fingerprint can be changed without breaking the signature.
155 Although no details of the signed portion of the certificate can be changed
156 this can cause problems with some applications: e.g. those using the
157 certificate fingerprint for blacklists.
159 1. Reject signatures with non zero unused bits.
161 If the BIT STRING containing the signature has non zero unused bits reject
162 the signature. All current signature algorithms require zero unused bits.
164 2. Check certificate algorithm consistency.
166 Check the AlgorithmIdentifier inside TBS matches the one in the
167 certificate signature. NB: this will result in signature failure
168 errors for some broken certificates.
170 Thanks to Konrad Kraszewski from Google for reporting this issue.
172 3. Check DSA/ECDSA signatures use DER.
174 Reencode DSA/ECDSA signatures and compare with the original received
175 signature. Return an error if there is a mismatch.
177 This will reject various cases including garbage after signature
178 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
179 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
180 (negative or with leading zeroes).
182 Further analysis was conducted and fixes were developed by Stephen Henson
183 of the OpenSSL core team.
188 Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
190 *) Session Ticket Memory Leak.
192 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
193 integrity of that ticket is first verified. In the event of a session
194 ticket integrity check failing, OpenSSL will fail to free memory
195 causing a memory leak. By sending a large number of invalid session
196 tickets an attacker could exploit this issue in a Denial Of Service
201 *) Build option no-ssl3 is incomplete.
203 When OpenSSL is configured with "no-ssl3" as a build option, servers
204 could accept and complete a SSL 3.0 handshake, and clients could be
205 configured to send them.
207 [Akamai and the OpenSSL team]
209 *) Add support for TLS_FALLBACK_SCSV.
210 Client applications doing fallback retries should call
211 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
213 [Adam Langley, Bodo Moeller]
215 *) Add additional DigestInfo checks.
217 Reencode DigestInto in DER and check against the original when
218 verifying RSA signature: this will reject any improperly encoded
219 DigestInfo structures.
221 Note: this is a precautionary measure and no attacks are currently known.
225 Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
227 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
228 to a denial of service attack. A malicious server can crash the client
229 with a null pointer dereference (read) by specifying an anonymous (EC)DH
230 ciphersuite and sending carefully crafted handshake messages.
232 Thanks to Felix Gröbert (Google) for discovering and researching this
237 *) By sending carefully crafted DTLS packets an attacker could cause openssl
238 to leak memory. This can be exploited through a Denial of Service attack.
239 Thanks to Adam Langley for discovering and researching this issue.
243 *) An attacker can force openssl to consume large amounts of memory whilst
244 processing DTLS handshake messages. This can be exploited through a
245 Denial of Service attack.
246 Thanks to Adam Langley for discovering and researching this issue.
250 *) An attacker can force an error condition which causes openssl to crash
251 whilst processing DTLS packets due to memory being freed twice. This
252 can be exploited through a Denial of Service attack.
253 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
258 *) If a multithreaded client connects to a malicious server using a resumed
259 session and the server sends an ec point format extension it could write
260 up to 255 bytes to freed memory.
262 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
267 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
268 X509_name_oneline, X509_name_print_ex et al. to leak some information
269 from the stack. Applications may be affected if they echo pretty printing
270 output to the attacker.
272 Thanks to Ivan Fratric (Google) for discovering this issue.
274 [Emilia Käsper, and Steve Henson]
276 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
277 for corner cases. (Certain input points at infinity could lead to
278 bogus results, with non-infinity inputs mapped to infinity too.)
281 Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
283 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
284 handshake can force the use of weak keying material in OpenSSL
285 SSL/TLS clients and servers.
287 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
288 researching this issue. (CVE-2014-0224)
289 [KIKUCHI Masashi, Steve Henson]
291 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
292 OpenSSL DTLS client the code can be made to recurse eventually crashing
295 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
297 [Imre Rad, Steve Henson]
299 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
300 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
301 client or server. This is potentially exploitable to run arbitrary
302 code on a vulnerable client or server.
304 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
305 [Jüri Aedla, Steve Henson]
307 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
308 are subject to a denial of service attack.
310 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
311 this issue. (CVE-2014-3470)
312 [Felix Gröbert, Ivan Fratric, Steve Henson]
314 *) Harmonize version and its documentation. -f flag is used to display
316 [mancha <mancha1@zoho.com>]
318 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
320 [mancha <mancha1@zoho.com>]
322 *) Fix some double frees. These are not thought to be exploitable.
323 [mancha <mancha1@zoho.com>]
325 *) Fix for the attack described in the paper "Recovering OpenSSL
326 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
327 by Yuval Yarom and Naomi Benger. Details can be obtained from:
328 http://eprint.iacr.org/2014/140
330 Thanks to Yuval Yarom and Naomi Benger for discovering this
331 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
332 [Yuval Yarom and Naomi Benger]
334 Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
336 *) Keep original DTLS digest and encryption contexts in retransmission
337 structures so we can use the previous session parameters if they need
338 to be resent. (CVE-2013-6450)
341 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
342 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
343 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
344 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
345 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
346 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
347 [Rob Stradling, Adam Langley]
349 Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
351 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
353 This addresses the flaw in CBC record processing discovered by
354 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
355 at: http://www.isg.rhul.ac.uk/tls/
357 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
358 Security Group at Royal Holloway, University of London
359 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
360 Emilia Käsper for the initial patch.
362 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
364 *) Return an error when checking OCSP signatures when key is NULL.
365 This fixes a DoS attack. (CVE-2013-0166)
368 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
369 the right response is stapled. Also change SSL_get_certificate()
370 so it returns the certificate actually sent.
371 See http://rt.openssl.org/Ticket/Display.html?id=2836.
373 [Rob Stradling <rob.stradling@comodo.com>]
375 *) Fix possible deadlock when decoding public keys.
378 Changes between 1.0.0i and 1.0.0j [10 May 2012]
380 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
383 *) Sanity check record length before skipping explicit IV in DTLS
386 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
387 fuzzing as a service testing platform.
391 *) Initialise tkeylen properly when encrypting CMS messages.
392 Thanks to Solar Designer of Openwall for reporting this issue.
395 Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
397 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
398 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
399 in CRYPTO_realloc_clean.
401 Thanks to Tavis Ormandy, Google Security Team, for discovering this
402 issue and to Adam Langley <agl@chromium.org> for fixing it.
404 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
406 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
408 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
409 in CMS and PKCS7 code. When RSA decryption fails use a random key for
410 content decryption and always return the same error. Note: this attack
411 needs on average 2^20 messages so it only affects automated senders. The
412 old behaviour can be reenabled in the CMS code by setting the
413 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
414 an MMA defence is not necessary.
415 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
416 this issue. (CVE-2012-0884)
419 *) Fix CVE-2011-4619: make sure we really are receiving a
420 client hello before rejecting multiple SGC restarts. Thanks to
421 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
424 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
426 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
427 Thanks to Antonio Martin, Enterprise Secure Access Research and
428 Development, Cisco Systems, Inc. for discovering this bug and
429 preparing a fix. (CVE-2012-0050)
432 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
434 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
435 of the Vaudenay padding oracle attack on CBC mode encryption
436 which enables an efficient plaintext recovery attack against
437 the OpenSSL implementation of DTLS. Their attack exploits timing
438 differences arising during decryption processing. A research
439 paper describing this attack can be found at:
440 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
441 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
442 Security Group at Royal Holloway, University of London
443 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
444 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
445 for preparing the fix. (CVE-2011-4108)
446 [Robin Seggelmann, Michael Tuexen]
448 *) Clear bytes used for block padding of SSL 3.0 records.
450 [Adam Langley (Google)]
452 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
453 Kadianakis <desnacked@gmail.com> for discovering this issue and
454 Adam Langley for preparing the fix. (CVE-2011-4619)
455 [Adam Langley (Google)]
457 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
458 [Andrey Kulikov <amdeich@gmail.com>]
460 *) Prevent malformed RFC3779 data triggering an assertion failure.
461 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
462 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
463 [Rob Austein <sra@hactrn.net>]
465 *) Improved PRNG seeding for VOS.
466 [Paul Green <Paul.Green@stratus.com>]
468 *) Fix ssl_ciph.c set-up race.
469 [Adam Langley (Google)]
471 *) Fix spurious failures in ecdsatest.c.
472 [Emilia Käsper (Google)]
474 *) Fix the BIO_f_buffer() implementation (which was mixing different
475 interpretations of the '..._len' fields).
476 [Adam Langley (Google)]
478 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
479 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
480 threads won't reuse the same blinding coefficients.
482 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
483 lock to call BN_BLINDING_invert_ex, and avoids one use of
484 BN_BLINDING_update for each BN_BLINDING structure (previously,
485 the last update always remained unused).
486 [Emilia Käsper (Google)]
488 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
489 [Bob Buckholz (Google)]
491 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
493 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
494 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
495 [Kaspar Brand <ossl@velox.ch>]
497 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
498 for multi-threaded use of ECDH. (CVE-2011-3210)
499 [Adam Langley (Google)]
501 *) Fix x509_name_ex_d2i memory leak on bad inputs.
504 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
505 signature public key algorithm by using OID xref utilities instead.
506 Before this you could only use some ECC ciphersuites with SHA1 only.
509 *) Add protection against ECDSA timing attacks as mentioned in the paper
510 by Billy Bob Brumley and Nicola Tuveri, see:
512 http://eprint.iacr.org/2011/232.pdf
514 [Billy Bob Brumley and Nicola Tuveri]
516 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
518 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
519 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
521 *) Fix bug in string printing code: if *any* escaping is enabled we must
522 escape the escape character (backslash) or the resulting string is
526 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
528 *) Disable code workaround for ancient and obsolete Netscape browsers
529 and servers: an attacker can use it in a ciphersuite downgrade attack.
530 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
533 *) Fixed J-PAKE implementation error, originally discovered by
534 Sebastien Martini, further info and confirmation from Stefan
535 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
538 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
540 *) Fix extension code to avoid race conditions which can result in a buffer
541 overrun vulnerability: resumed sessions must not be modified as they can
542 be shared by multiple threads. CVE-2010-3864
545 *) Fix WIN32 build system to correctly link an ENGINE directory into
549 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
551 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
553 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
555 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
557 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
558 context. The operation can be customised via the ctrl mechanism in
559 case ENGINEs want to include additional functionality.
562 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
565 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
566 output hashes compatible with older versions of OpenSSL.
567 [Willy Weisz <weisz@vcpc.univie.ac.at>]
569 *) Fix compression algorithm handling: if resuming a session use the
570 compression algorithm of the resumed session instead of determining
571 it from client hello again. Don't allow server to change algorithm.
574 *) Add load_crls() function to apps tidying load_certs() too. Add option
575 to verify utility to allow additional CRLs to be included.
578 *) Update OCSP request code to permit adding custom headers to the request:
579 some responders need this.
582 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
584 [Julia Lawall <julia@diku.dk>]
586 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
587 needlessly dereferenced structures, used obsolete functions and
588 didn't handle all updated verify codes correctly.
591 *) Disable MD2 in the default configuration.
594 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
595 indicate the initial BIO being pushed or popped. This makes it possible
596 to determine whether the BIO is the one explicitly called or as a result
597 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
598 it handles reference counts correctly and doesn't zero out the I/O bio
599 when it is not being explicitly popped. WARNING: applications which
600 included workarounds for the old buggy behaviour will need to be modified
601 or they could free up already freed BIOs.
604 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
605 renaming to all platforms (within the 0.9.8 branch, this was
606 done conditionally on Netware platforms to avoid a name clash).
607 [Guenter <lists@gknw.net>]
609 *) Add ECDHE and PSK support to DTLS.
610 [Michael Tuexen <tuexen@fh-muenster.de>]
612 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
616 *) Add "missing" function EVP_MD_flags() (without this the only way to
617 retrieve a digest flags is by accessing the structure directly. Update
618 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
619 or cipher is registered as in the "from" argument. Print out all
620 registered digests in the dgst usage message instead of manually
621 attempting to work them out.
624 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
625 this allows the use of compression and extensions. Change default cipher
626 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
627 by default unless an application cipher string requests it.
630 *) Alter match criteria in PKCS12_parse(). It used to try to use local
631 key ids to find matching certificates and keys but some PKCS#12 files
632 don't follow the (somewhat unwritten) rules and this strategy fails.
633 Now just gather all certificates together and the first private key
634 then look for the first certificate that matches the key.
637 *) Support use of registered digest and cipher names for dgst and cipher
638 commands instead of having to add each one as a special case. So now
645 openssl dgst -sha256 foo
647 and this works for ENGINE based algorithms too.
651 *) Update Gost ENGINE to support parameter files.
652 [Victor B. Wagner <vitus@cryptocom.ru>]
654 *) Support GeneralizedTime in ca utility.
655 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
657 *) Enhance the hash format used for certificate directory links. The new
658 form uses the canonical encoding (meaning equivalent names will work
659 even if they aren't identical) and uses SHA1 instead of MD5. This form
660 is incompatible with the older format and as a result c_rehash should
661 be used to rebuild symbolic links.
664 *) Make PKCS#8 the default write format for private keys, replacing the
665 traditional format. This form is standardised, more secure and doesn't
666 include an implicit MD5 dependency.
669 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
670 committed to OpenSSL should pass this lot as a minimum.
673 *) Add session ticket override functionality for use by EAP-FAST.
674 [Jouni Malinen <j@w1.fi>]
676 *) Modify HMAC functions to return a value. Since these can be implemented
677 in an ENGINE errors can occur.
680 *) Type-checked OBJ_bsearch_ex.
683 *) Type-checked OBJ_bsearch. Also some constification necessitated
684 by type-checking. Still to come: TXT_DB, bsearch(?),
685 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
689 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
690 seconds to a tm structure directly, instead of going through OS
691 specific date routines. This avoids any issues with OS routines such
692 as the year 2038 bug. New *_adj() functions for ASN1 time structures
693 and X509_time_adj_ex() to cover the extended range. The existing
694 X509_time_adj() is still usable and will no longer have any date issues.
697 *) Delta CRL support. New use deltas option which will attempt to locate
698 and search any appropriate delta CRLs available.
700 This work was sponsored by Google.
703 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
704 code and add additional score elements. Validate alternate CRL paths
705 as part of the CRL checking and indicate a new error "CRL path validation
706 error" in this case. Applications wanting additional details can use
707 the verify callback and check the new "parent" field. If this is not
708 NULL CRL path validation is taking place. Existing applications wont
709 see this because it requires extended CRL support which is off by
712 This work was sponsored by Google.
715 *) Support for freshest CRL extension.
717 This work was sponsored by Google.
720 *) Initial indirect CRL support. Currently only supported in the CRLs
721 passed directly and not via lookup. Process certificate issuer
722 CRL entry extension and lookup CRL entries by bother issuer name
723 and serial number. Check and process CRL issuer entry in IDP extension.
725 This work was sponsored by Google.
728 *) Add support for distinct certificate and CRL paths. The CRL issuer
729 certificate is validated separately in this case. Only enabled if
730 an extended CRL support flag is set: this flag will enable additional
731 CRL functionality in future.
733 This work was sponsored by Google.
736 *) Add support for policy mappings extension.
738 This work was sponsored by Google.
741 *) Fixes to pathlength constraint, self issued certificate handling,
742 policy processing to align with RFC3280 and PKITS tests.
744 This work was sponsored by Google.
747 *) Support for name constraints certificate extension. DN, email, DNS
748 and URI types are currently supported.
750 This work was sponsored by Google.
753 *) To cater for systems that provide a pointer-based thread ID rather
754 than numeric, deprecate the current numeric thread ID mechanism and
755 replace it with a structure and associated callback type. This
756 mechanism allows a numeric "hash" to be extracted from a thread ID in
757 either case, and on platforms where pointers are larger than 'long',
758 mixing is done to help ensure the numeric 'hash' is usable even if it
759 can't be guaranteed unique. The default mechanism is to use "&errno"
760 as a pointer-based thread ID to distinguish between threads.
762 Applications that want to provide their own thread IDs should now use
763 CRYPTO_THREADID_set_callback() to register a callback that will call
764 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
766 Note that ERR_remove_state() is now deprecated, because it is tied
767 to the assumption that thread IDs are numeric. ERR_remove_state(0)
768 to free the current thread's error state should be replaced by
769 ERR_remove_thread_state(NULL).
771 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
772 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
773 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
774 application was previously providing a numeric thread callback that
775 was inappropriate for distinguishing threads, then uniqueness might
776 have been obtained with &errno that happened immediately in the
777 intermediate development versions of OpenSSL; this is no longer the
778 case, the numeric thread callback will now override the automatic use
780 [Geoff Thorpe, with help from Bodo Moeller]
782 *) Initial support for different CRL issuing certificates. This covers a
783 simple case where the self issued certificates in the chain exist and
784 the real CRL issuer is higher in the existing chain.
786 This work was sponsored by Google.
789 *) Removed effectively defunct crypto/store from the build.
792 *) Revamp of STACK to provide stronger type-checking. Still to come:
793 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
794 ASN1_STRING, CONF_VALUE.
797 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
798 RAM on SSL connections. This option can save about 34k per idle SSL.
801 *) Revamp of LHASH to provide stronger type-checking. Still to come:
802 STACK, TXT_DB, bsearch, qsort.
805 *) Initial support for Cryptographic Message Syntax (aka CMS) based
806 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
807 support for data, signedData, compressedData, digestedData and
808 encryptedData, envelopedData types included. Scripts to check against
809 RFC4134 examples draft and interop and consistency checks of many
810 content types and variants.
813 *) Add options to enc utility to support use of zlib compression BIO.
816 *) Extend mk1mf to support importing of options and assembly language
817 files from Configure script, currently only included in VC-WIN32.
818 The assembly language rules can now optionally generate the source
819 files from the associated perl scripts.
822 *) Implement remaining functionality needed to support GOST ciphersuites.
823 Interop testing has been performed using CryptoPro implementations.
824 [Victor B. Wagner <vitus@cryptocom.ru>]
826 *) s390x assembler pack.
829 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
833 *) Implement Opaque PRF Input TLS extension as specified in
834 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
835 official specification yet and no extension type assignment by
836 IANA exists, this extension (for now) will have to be explicitly
837 enabled when building OpenSSL by providing the extension number
838 to use. For example, specify an option
840 -DTLSEXT_TYPE_opaque_prf_input=0x9527
842 to the "config" or "Configure" script to enable the extension,
843 assuming extension number 0x9527 (which is a completely arbitrary
844 and unofficial assignment based on the MD5 hash of the Internet
845 Draft). Note that by doing so, you potentially lose
846 interoperability with other TLS implementations since these might
847 be using the same extension number for other purposes.
849 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
850 opaque PRF input value to use in the handshake. This will create
851 an interal copy of the length-'len' string at 'src', and will
852 return non-zero for success.
854 To get more control and flexibility, provide a callback function
857 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
858 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
862 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
865 Callback function 'cb' will be called in handshakes, and is
866 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
867 Argument 'arg' is for application purposes (the value as given to
868 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
869 be provided to the callback function). The callback function
870 has to return non-zero to report success: usually 1 to use opaque
871 PRF input just if possible, or 2 to enforce use of the opaque PRF
872 input. In the latter case, the library will abort the handshake
873 if opaque PRF input is not successfully negotiated.
875 Arguments 'peerinput' and 'len' given to the callback function
876 will always be NULL and 0 in the case of a client. A server will
877 see the client's opaque PRF input through these variables if
878 available (NULL and 0 otherwise). Note that if the server
879 provides an opaque PRF input, the length must be the same as the
880 length of the client's opaque PRF input.
882 Note that the callback function will only be called when creating
883 a new session (session resumption can resume whatever was
884 previously negotiated), and will not be called in SSL 2.0
885 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
886 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
887 for applications that need to enforce opaque PRF input.
891 *) Update ssl code to support digests other than SHA1+MD5 for handshake
894 [Victor B. Wagner <vitus@cryptocom.ru>]
896 *) Add RFC4507 support to OpenSSL. This includes the corrections in
897 RFC4507bis. The encrypted ticket format is an encrypted encoded
898 SSL_SESSION structure, that way new session features are automatically
901 If a client application caches session in an SSL_SESSION structure
902 support is transparent because tickets are now stored in the encoded
905 The SSL_CTX structure automatically generates keys for ticket
906 protection in servers so again support should be possible
907 with no application modification.
909 If a client or server wishes to disable RFC4507 support then the option
910 SSL_OP_NO_TICKET can be set.
912 Add a TLS extension debugging callback to allow the contents of any client
913 or server extensions to be examined.
915 This work was sponsored by Google.
918 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
919 OpenSSL should now compile cleanly on gcc 4.2
920 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
922 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
923 support including streaming MAC support: this is required for GOST
925 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
927 *) Add option -stream to use PKCS#7 streaming in smime utility. New
928 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
929 to output in BER and PEM format.
932 *) Experimental support for use of HMAC via EVP_PKEY interface. This
933 allows HMAC to be handled via the EVP_DigestSign*() interface. The
934 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
935 ENGINE support for HMAC keys which are unextractable. New -mac and
936 -macopt options to dgst utility.
939 *) New option -sigopt to dgst utility. Update dgst to use
940 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
941 alternative signing paramaters such as X9.31 or PSS in the dgst
945 *) Change ssl_cipher_apply_rule(), the internal function that does
946 the work each time a ciphersuite string requests enabling
947 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
948 removing ("!foo+bar") a class of ciphersuites: Now it maintains
949 the order of disabled ciphersuites such that those ciphersuites
950 that most recently went from enabled to disabled not only stay
951 in order with respect to each other, but also have higher priority
952 than other disabled ciphersuites the next time ciphersuites are
955 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
956 the same ciphersuites as with "HIGH" alone, but in a specific
957 order where the PSK ciphersuites come first (since they are the
958 most recently disabled ciphersuites when "HIGH" is parsed).
960 Also, change ssl_create_cipher_list() (using this new
961 funcionality) such that between otherwise identical
962 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
966 *) Change ssl_create_cipher_list() so that it automatically
967 arranges the ciphersuites in reasonable order before starting
968 to process the rule string. Thus, the definition for "DEFAULT"
969 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
970 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
971 This makes it much easier to arrive at a reasonable default order
972 in applications for which anonymous ciphers are OK (meaning
973 that you can't actually use DEFAULT).
974 [Bodo Moeller; suggested by Victor Duchovni]
976 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
977 processing) into multiple integers instead of setting
978 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
979 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
980 (These masks as well as the individual bit definitions are hidden
981 away into the non-exported interface ssl/ssl_locl.h, so this
982 change to the definition of the SSL_CIPHER structure shouldn't
983 affect applications.) This give us more bits for each of these
984 categories, so there is no longer a need to coagulate AES128 and
985 AES256 into a single algorithm bit, and to coagulate Camellia128
986 and Camellia256 into a single algorithm bit, which has led to all
989 Thus, among other things, the kludge introduced in 0.9.7m and
990 0.9.8e for masking out AES256 independently of AES128 or masking
991 out Camellia256 independently of AES256 is not needed here in 0.9.9.
993 With the change, we also introduce new ciphersuite aliases that
994 so far were missing: "AES128", "AES256", "CAMELLIA128", and
998 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
999 Use the leftmost N bytes of the signature input if the input is
1000 larger than the prime q (with N being the size in bytes of q).
1003 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
1004 it yet and it is largely untested.
1007 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
1010 *) Initial incomplete changes to avoid need for function casts in OpenSSL
1011 some compilers (gcc 4.2 and later) reject their use. Safestack is
1012 reimplemented. Update ASN1 to avoid use of legacy functions.
1015 *) Win32/64 targets are linked with Winsock2.
1018 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
1019 to external functions. This can be used to increase CRL handling
1020 efficiency especially when CRLs are very large by (for example) storing
1021 the CRL revoked certificates in a database.
1024 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
1025 new CRLs added to a directory can be used. New command line option
1026 -verify_return_error to s_client and s_server. This causes real errors
1027 to be returned by the verify callback instead of carrying on no matter
1028 what. This reflects the way a "real world" verify callback would behave.
1031 *) GOST engine, supporting several GOST algorithms and public key formats.
1032 Kindly donated by Cryptocom.
1035 *) Partial support for Issuing Distribution Point CRL extension. CRLs
1036 partitioned by DP are handled but no indirect CRL or reason partitioning
1037 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
1038 selected via a scoring technique which handles IDP and AKID in CRLs.
1041 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
1042 will ultimately be used for all verify operations: this will remove the
1043 X509_STORE dependency on certificate verification and allow alternative
1044 lookup methods. X509_STORE based implementations of these two callbacks.
1047 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
1048 Modify get_crl() to find a valid (unexpired) CRL if possible.
1051 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
1052 this would be called X509_CRL_cmp() but that name is already used by
1053 a function that just compares CRL issuer names. Cache several CRL
1054 extensions in X509_CRL structure and cache CRLDP in X509.
1057 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
1058 this maps equivalent X509_NAME structures into a consistent structure.
1059 Name comparison can then be performed rapidly using memcmp().
1062 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
1066 *) Allow digests to supply their own micalg string for S/MIME type using
1067 the ctrl EVP_MD_CTRL_MICALG.
1070 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
1071 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
1072 ctrl. It can then customise the structure before and/or after signing
1076 *) New function OBJ_add_sigid() to allow application defined signature OIDs
1077 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
1078 to free up any added signature OIDs.
1081 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
1082 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
1083 digest and cipher tables. New options added to openssl utility:
1084 list-message-digest-algorithms and list-cipher-algorithms.
1087 *) Change the array representation of binary polynomials: the list
1088 of degrees of non-zero coefficients is now terminated with -1.
1089 Previously it was terminated with 0, which was also part of the
1090 value; thus, the array representation was not applicable to
1091 polynomials where t^0 has coefficient zero. This change makes
1092 the array representation useful in a more general context.
1095 *) Various modifications and fixes to SSL/TLS cipher string
1096 handling. For ECC, the code now distinguishes between fixed ECDH
1097 with RSA certificates on the one hand and with ECDSA certificates
1098 on the other hand, since these are separate ciphersuites. The
1099 unused code for Fortezza ciphersuites has been removed.
1101 For consistency with EDH, ephemeral ECDH is now called "EECDH"
1102 (not "ECDHE"). For consistency with the code for DH
1103 certificates, use of ECDH certificates is now considered ECDH
1104 authentication, not RSA or ECDSA authentication (the latter is
1105 merely the CA's signing algorithm and not actively used in the
1108 The temporary ciphersuite alias "ECCdraft" is no longer
1109 available, and ECC ciphersuites are no longer excluded from "ALL"
1110 and "DEFAULT". The following aliases now exist for RFC 4492
1111 ciphersuites, most of these by analogy with the DH case:
1113 kECDHr - ECDH cert, signed with RSA
1114 kECDHe - ECDH cert, signed with ECDSA
1115 kECDH - ECDH cert (signed with either RSA or ECDSA)
1116 kEECDH - ephemeral ECDH
1117 ECDH - ECDH cert or ephemeral ECDH
1123 AECDH - anonymous ECDH
1124 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
1128 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
1129 Use correct micalg parameters depending on digest(s) in signed message.
1132 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
1133 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
1136 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
1137 an engine to register a method. Add ENGINE lookups for methods and
1138 functional reference processing.
1141 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
1142 EVP_{Sign,Verify}* which allow an application to customise the signature
1146 *) New -resign option to smime utility. This adds one or more signers
1147 to an existing PKCS#7 signedData structure. Also -md option to use an
1148 alternative message digest algorithm for signing.
1151 *) Tidy up PKCS#7 routines and add new functions to make it easier to
1152 create PKCS7 structures containing multiple signers. Update smime
1153 application to support multiple signers.
1156 *) New -macalg option to pkcs12 utility to allow setting of an alternative
1160 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
1161 Reorganize PBE internals to lookup from a static table using NIDs,
1162 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
1163 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
1164 PRF which will be automatically used with PBES2.
1167 *) Replace the algorithm specific calls to generate keys in "req" with the
1171 *) Update PKCS#7 enveloped data routines to use new API. This is now
1172 supported by any public key method supporting the encrypt operation. A
1173 ctrl is added to allow the public key algorithm to examine or modify
1174 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
1178 *) Add a ctrl to asn1 method to allow a public key algorithm to express
1179 a default digest type to use. In most cases this will be SHA1 but some
1180 algorithms (such as GOST) need to specify an alternative digest. The
1181 return value indicates how strong the prefernce is 1 means optional and
1182 2 is mandatory (that is it is the only supported type). Modify
1183 ASN1_item_sign() to accept a NULL digest argument to indicate it should
1184 use the default md. Update openssl utilities to use the default digest
1185 type for signing if it is not explicitly indicated.
1188 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
1189 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
1190 signing method from the key type. This effectively removes the link
1191 between digests and public key types.
1194 *) Add an OID cross reference table and utility functions. Its purpose is to
1195 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
1196 rsaEncryption. This will allow some of the algorithm specific hackery
1197 needed to use the correct OID to be removed.
1200 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
1201 structures for PKCS7_sign(). They are now set up by the relevant public
1205 *) Add provisional EC pkey method with support for ECDSA and ECDH.
1208 *) Add support for key derivation (agreement) in the API, DH method and
1212 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
1213 public and private key formats. As a side effect these add additional
1214 command line functionality not previously available: DSA signatures can be
1215 generated and verified using pkeyutl and DH key support and generation in
1220 [Oliver Tappe <zooey@hirschkaefer.de>]
1222 *) New make target "install_html_docs" installs HTML renditions of the
1224 [Oliver Tappe <zooey@hirschkaefer.de>]
1226 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
1227 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
1228 support key and parameter generation and add initial key generation
1229 functionality for RSA.
1232 *) Add functions for main EVP_PKEY_method operations. The undocumented
1233 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
1234 EVP_PKEY_{encrypt,decrypt}_old.
1237 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
1238 key API, doesn't do much yet.
1241 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
1242 public key algorithms. New option to openssl utility:
1243 "list-public-key-algorithms" to print out info.
1246 *) Implement the Supported Elliptic Curves Extension for
1247 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1250 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
1251 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
1254 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
1255 utilities such as rsa, dsa, dsaparam etc except they process any key
1259 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1260 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1261 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
1265 *) Initial support for pluggable public key ASN1.
1266 De-spaghettify the public key ASN1 handling. Move public and private
1267 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
1268 algorithm specific handling to a single module within the relevant
1269 algorithm directory. Add functions to allow (near) opaque processing
1270 of public and private key structures.
1273 *) Implement the Supported Point Formats Extension for
1274 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1277 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
1278 for the psk identity [hint] and the psk callback functions to the
1279 SSL_SESSION, SSL and SSL_CTX structure.
1282 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
1286 SSL_CTX_use_psk_identity_hint
1287 SSL_get_psk_identity_hint
1288 SSL_get_psk_identity
1289 SSL_use_psk_identity_hint
1291 [Mika Kousa and Pasi Eronen of Nokia Corporation]
1293 *) Add RFC 3161 compliant time stamp request creation, response generation
1294 and response verification functionality.
1295 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
1297 *) Add initial support for TLS extensions, specifically for the server_name
1298 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1299 have new members for a host name. The SSL data structure has an
1300 additional member SSL_CTX *initial_ctx so that new sessions can be
1301 stored in that context to allow for session resumption, even after the
1302 SSL has been switched to a new SSL_CTX in reaction to a client's
1303 server_name extension.
1305 New functions (subject to change):
1307 SSL_get_servername()
1308 SSL_get_servername_type()
1311 New CTRL codes and macros (subject to change):
1313 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1314 - SSL_CTX_set_tlsext_servername_callback()
1315 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1316 - SSL_CTX_set_tlsext_servername_arg()
1317 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1319 openssl s_client has a new '-servername ...' option.
1321 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1322 '-key2 ...', '-servername_fatal' (subject to change). This allows
1323 testing the HostName extension for a specific single host name ('-cert'
1324 and '-key' remain fallbacks for handshakes without HostName
1325 negotiation). If the unrecogninzed_name alert has to be sent, this by
1326 default is a warning; it becomes fatal with the '-servername_fatal'
1329 [Peter Sylvester, Remy Allais, Christophe Renou]
1331 *) Whirlpool hash implementation is added.
1334 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1335 bn(64,32). Because of instruction set limitations it doesn't have
1336 any negative impact on performance. This was done mostly in order
1337 to make it possible to share assembler modules, such as bn_mul_mont
1338 implementations, between 32- and 64-bit builds without hassle.
1341 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1342 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1346 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1347 dedicated Montgomery multiplication procedure, is introduced.
1348 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1349 "64-bit" performance on certain 32-bit targets.
1352 *) New option SSL_OP_NO_COMP to disable use of compression selectively
1353 in SSL structures. New SSL ctrl to set maximum send fragment size.
1354 Save memory by seeting the I/O buffer sizes dynamically instead of
1355 using the maximum available value.
1358 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
1359 in addition to the text details.
1362 *) Very, very preliminary EXPERIMENTAL support for printing of general
1363 ASN1 structures. This currently produces rather ugly output and doesn't
1364 handle several customised structures at all.
1367 *) Integrated support for PVK file format and some related formats such
1368 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
1369 these in the 'rsa' and 'dsa' utilities.
1372 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
1375 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
1376 place for the (very old) "NETSCAPE" format certificates which are now
1377 handled using new ASN1 code equivalents.
1380 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
1381 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
1382 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
1385 *) Modify CRL distribution points extension code to print out previously
1386 unsupported fields. Enhance extension setting code to allow setting of
1390 *) Add print and set support for Issuing Distribution Point CRL extension.
1393 *) Change 'Configure' script to enable Camellia by default.
1396 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
1398 *) When rejecting SSL/TLS records due to an incorrect version number, never
1399 update s->server with a new major version number. As of
1400 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
1401 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
1402 the previous behavior could result in a read attempt at NULL when
1403 receiving specific incorrect SSL/TLS records once record payload
1404 protection is active. (CVE-2010-0740)
1405 [Bodo Moeller, Adam Langley <agl@chromium.org>]
1407 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
1408 could be crashed if the relevant tables were not present (e.g. chrooted).
1409 [Tomas Hoger <thoger@redhat.com>]
1411 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
1413 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
1414 [Martin Olsson, Neel Mehta]
1416 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
1417 accommodate for stack sorting, always a write lock!).
1420 *) On some versions of WIN32 Heap32Next is very slow. This can cause
1421 excessive delays in the RAND_poll(): over a minute. As a workaround
1422 include a time check in the inner Heap32Next loop too.
1425 *) The code that handled flushing of data in SSL/TLS originally used the
1426 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
1427 the problem outlined in PR#1949. The fix suggested there however can
1428 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
1429 of Apache). So instead simplify the code to flush unconditionally.
1430 This should be fine since flushing with no data to flush is a no op.
1433 *) Handle TLS versions 2.0 and later properly and correctly use the
1434 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
1435 off ancient servers have a habit of sticking around for a while...
1438 *) Modify compression code so it frees up structures without using the
1439 ex_data callbacks. This works around a problem where some applications
1440 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
1441 restarting) then use compression (e.g. SSL with compression) later.
1442 This results in significant per-connection memory leaks and
1443 has caused some security issues including CVE-2008-1678 and
1447 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
1448 change when encrypting or decrypting.
1451 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
1452 connect and renegotiate with servers which do not support RI.
1453 Until RI is more widely deployed this option is enabled by default.
1456 *) Add "missing" ssl ctrls to clear options and mode.
1459 *) If client attempts to renegotiate and doesn't support RI respond with
1460 a no_renegotiation alert as required by RFC5746. Some renegotiating
1461 TLS clients will continue a connection gracefully when they receive
1462 the alert. Unfortunately OpenSSL mishandled this alert and would hang
1463 waiting for a server hello which it will never receive. Now we treat a
1464 received no_renegotiation alert as a fatal error. This is because
1465 applications requesting a renegotiation might well expect it to succeed
1466 and would have no code in place to handle the server denying it so the
1467 only safe thing to do is to terminate the connection.
1470 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
1471 peer supports secure renegotiation and 0 otherwise. Print out peer
1472 renegotiation support in s_client/s_server.
1475 *) Replace the highly broken and deprecated SPKAC certification method with
1476 the updated NID creation version. This should correctly handle UTF8.
1479 *) Implement RFC5746. Re-enable renegotiation but require the extension
1480 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
1481 turns out to be a bad idea. It has been replaced by
1482 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
1483 SSL_CTX_set_options(). This is really not recommended unless you
1484 know what you are doing.
1485 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
1487 *) Fixes to stateless session resumption handling. Use initial_ctx when
1488 issuing and attempting to decrypt tickets in case it has changed during
1489 servername handling. Use a non-zero length session ID when attempting
1490 stateless session resumption: this makes it possible to determine if
1491 a resumption has occurred immediately after receiving server hello
1492 (several places in OpenSSL subtly assume this) instead of later in
1496 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
1497 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
1498 fixes for a few places where the return code is not checked
1500 [Julia Lawall <julia@diku.dk>]
1502 *) Add --strict-warnings option to Configure script to include devteam
1503 warnings in other configurations.
1506 *) Add support for --libdir option and LIBDIR variable in makefiles. This
1507 makes it possible to install openssl libraries in locations which
1508 have names other than "lib", for example "/usr/lib64" which some
1510 [Steve Henson, based on patch from Jeremy Utley]
1512 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
1513 X690 8.9.12 and can produce some misleading textual output of OIDs.
1514 [Steve Henson, reported by Dan Kaminsky]
1516 *) Delete MD2 from algorithm tables. This follows the recommendation in
1517 several standards that it is not used in new applications due to
1518 several cryptographic weaknesses. For binary compatibility reasons
1519 the MD2 API is still compiled in by default.
1522 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
1526 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
1527 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
1529 [Guenter <lists@gknw.net>]
1531 *) Fix the server certificate chain building code to use X509_verify_cert(),
1532 it used to have an ad-hoc builder which was unable to cope with anything
1533 other than a simple chain.
1534 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
1536 *) Don't check self signed certificate signatures in X509_verify_cert()
1537 by default (a flag can override this): it just wastes time without
1538 adding any security. As a useful side effect self signed root CAs
1539 with non-FIPS digests are now usable in FIPS mode.
1542 *) In dtls1_process_out_of_seq_message() the check if the current message
1543 is already buffered was missing. For every new message was memory
1544 allocated, allowing an attacker to perform an denial of service attack
1545 with sending out of seq handshake messages until there is no memory
1546 left. Additionally every future messege was buffered, even if the
1547 sequence number made no sense and would be part of another handshake.
1548 So only messages with sequence numbers less than 10 in advance will be
1549 buffered. (CVE-2009-1378)
1550 [Robin Seggelmann, discovered by Daniel Mentz]
1552 *) Records are buffered if they arrive with a future epoch to be
1553 processed after finishing the corresponding handshake. There is
1554 currently no limitation to this buffer allowing an attacker to perform
1555 a DOS attack with sending records with future epochs until there is no
1556 memory left. This patch adds the pqueue_size() function to detemine
1557 the size of a buffer and limits the record buffer to 100 entries.
1559 [Robin Seggelmann, discovered by Daniel Mentz]
1561 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
1562 parent structure is freed. (CVE-2009-1379)
1565 *) Handle non-blocking I/O properly in SSL_shutdown() call.
1566 [Darryl Miles <darryl-mailinglists@netbauds.net>]
1569 [Ilya O. <vrghost@gmail.com>]
1571 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
1573 *) Disable renegotiation completely - this fixes a severe security
1574 problem (CVE-2009-3555) at the cost of breaking all
1575 renegotiation. Renegotiation can be re-enabled by setting
1576 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
1577 run-time. This is really not recommended unless you know what
1581 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
1583 *) Don't set val to NULL when freeing up structures, it is freed up by
1584 underlying code. If sizeof(void *) > sizeof(long) this can result in
1585 zeroing past the valid field. (CVE-2009-0789)
1586 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
1588 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
1589 checked correctly. This would allow some invalid signed attributes to
1590 appear to verify correctly. (CVE-2009-0591)
1591 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1593 *) Reject UniversalString and BMPString types with invalid lengths. This
1594 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
1595 a legal length. (CVE-2009-0590)
1598 *) Set S/MIME signing as the default purpose rather than setting it
1599 unconditionally. This allows applications to override it at the store
1603 *) Permit restricted recursion of ASN1 strings. This is needed in practice
1604 to handle some structures.
1607 *) Improve efficiency of mem_gets: don't search whole buffer each time
1609 [Jeremy Shapiro <jnshapir@us.ibm.com>]
1611 *) New -hex option for openssl rand.
1614 *) Print out UTF8String and NumericString when parsing ASN1.
1617 *) Support NumericString type for name components.
1620 *) Allow CC in the environment to override the automatically chosen
1621 compiler. Note that nothing is done to ensure flags work with the
1625 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
1627 *) Properly check EVP_VerifyFinal() and similar return values
1629 [Ben Laurie, Bodo Moeller, Google Security Team]
1631 *) Enable TLS extensions by default.
1634 *) Allow the CHIL engine to be loaded, whether the application is
1635 multithreaded or not. (This does not release the developer from the
1636 obligation to set up the dynamic locking callbacks.)
1637 [Sander Temme <sander@temme.net>]
1639 *) Use correct exit code if there is an error in dgst command.
1640 [Steve Henson; problem pointed out by Roland Dirlewanger]
1642 *) Tweak Configure so that you need to say "experimental-jpake" to enable
1643 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
1646 *) Add experimental JPAKE support, including demo authentication in
1647 s_client and s_server.
1650 *) Set the comparison function in v3_addr_canonize().
1651 [Rob Austein <sra@hactrn.net>]
1653 *) Add support for XMPP STARTTLS in s_client.
1654 [Philip Paeps <philip@freebsd.org>]
1656 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
1657 to ensure that even with this option, only ciphersuites in the
1658 server's preference list will be accepted. (Note that the option
1659 applies only when resuming a session, so the earlier behavior was
1660 just about the algorithm choice for symmetric cryptography.)
1663 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
1665 *) Fix NULL pointer dereference if a DTLS server received
1666 ChangeCipherSpec as first record (CVE-2009-1386).
1669 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
1670 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
1673 *) The fix in 0.9.8c that supposedly got rid of unsafe
1674 double-checked locking was incomplete for RSA blinding,
1675 addressing just one layer of what turns out to have been
1676 doubly unsafe triple-checked locking.
1678 So now fix this for real by retiring the MONT_HELPER macro
1679 in crypto/rsa/rsa_eay.c.
1681 [Bodo Moeller; problem pointed out by Marius Schilder]
1683 *) Various precautionary measures:
1685 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
1687 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
1688 (NB: This would require knowledge of the secret session ticket key
1689 to exploit, in which case you'd be SOL either way.)
1691 - Change bn_nist.c so that it will properly handle input BIGNUMs
1692 outside the expected range.
1694 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
1697 [Neel Mehta, Bodo Moeller]
1699 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
1700 the load fails. Useful for distros.
1701 [Ben Laurie and the FreeBSD team]
1703 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
1706 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
1709 *) Expand ENGINE to support engine supplied SSL client certificate functions.
1711 This work was sponsored by Logica.
1714 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
1715 keystores. Support for SSL/TLS client authentication too.
1716 Not compiled unless enable-capieng specified to Configure.
1718 This work was sponsored by Logica.
1721 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
1722 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
1723 attribute creation routines such as certifcate requests and PKCS#12
1727 Changes between 0.9.8g and 0.9.8h [28 May 2008]
1729 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
1730 handshake which could lead to a cilent crash as found using the
1731 Codenomicon TLS test suite (CVE-2008-1672)
1732 [Steve Henson, Mark Cox]
1734 *) Fix double free in TLS server name extensions which could lead to
1735 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
1738 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
1740 Clear the error queue to ensure that error entries left from
1741 older function calls do not interfere with the correct operation.
1742 [Lutz Jaenicke, Erik de Castro Lopo]
1744 *) Remove root CA certificates of commercial CAs:
1746 The OpenSSL project does not recommend any specific CA and does not
1747 have any policy with respect to including or excluding any CA.
1748 Therefore it does not make any sense to ship an arbitrary selection
1749 of root CA certificates with the OpenSSL software.
1752 *) RSA OAEP patches to fix two separate invalid memory reads.
1753 The first one involves inputs when 'lzero' is greater than
1754 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
1755 before the beginning of from). The second one involves inputs where
1756 the 'db' section contains nothing but zeroes (there is a one-byte
1757 invalid read after the end of 'db').
1758 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1760 *) Partial backport from 0.9.9-dev:
1762 Introduce bn_mul_mont (dedicated Montgomery multiplication
1763 procedure) as a candidate for BIGNUM assembler implementation.
1764 While 0.9.9-dev uses assembler for various architectures, only
1765 x86_64 is available by default here in the 0.9.8 branch, and
1766 32-bit x86 is available through a compile-time setting.
1768 To try the 32-bit x86 assembler implementation, use Configure
1769 option "enable-montasm" (which exists only for this backport).
1771 As "enable-montasm" for 32-bit x86 disclaims code stability
1772 anyway, in this constellation we activate additional code
1773 backported from 0.9.9-dev for further performance improvements,
1774 namely BN_from_montgomery_word. (To enable this otherwise,
1775 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
1777 [Andy Polyakov (backport partially by Bodo Moeller)]
1779 *) Add TLS session ticket callback. This allows an application to set
1780 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
1781 values. This is useful for key rollover for example where several key
1782 sets may exist with different names.
1785 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
1786 This was broken until now in 0.9.8 releases, such that the only way
1787 a registered ENGINE could be used (assuming it initialises
1788 successfully on the host) was to explicitly set it as the default
1789 for the relevant algorithms. This is in contradiction with 0.9.7
1790 behaviour and the documentation. With this fix, when an ENGINE is
1791 registered into a given algorithm's table of implementations, the
1792 'uptodate' flag is reset so that auto-discovery will be used next
1793 time a new context for that algorithm attempts to select an
1795 [Ian Lister (tweaked by Geoff Thorpe)]
1797 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
1798 implemention in the following ways:
1800 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
1803 Lack of BER streaming support means one pass streaming processing is
1804 only supported if data is detached: setting the streaming flag is
1805 ignored for embedded content.
1807 CMS support is disabled by default and must be explicitly enabled
1808 with the enable-cms configuration option.
1811 *) Update the GMP engine glue to do direct copies between BIGNUM and
1812 mpz_t when openssl and GMP use the same limb size. Otherwise the
1813 existing "conversion via a text string export" trick is still used.
1814 [Paul Sheer <paulsheer@gmail.com>]
1816 *) Zlib compression BIO. This is a filter BIO which compressed and
1817 uncompresses any data passed through it.
1820 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
1821 RFC3394 compatible AES key wrapping.
1824 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
1825 sets string data without copying. X509_ALGOR_set0() and
1826 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
1827 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
1828 from an X509_ATTRIBUTE structure optionally checking it occurs only
1829 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
1833 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
1834 to get the expected BN_FLG_CONSTTIME behavior.
1835 [Bodo Moeller (Google)]
1839 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
1840 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
1841 - added some more tests to do_tests.pl
1842 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
1843 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
1844 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
1845 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
1846 - various changes to netware.pl to enable gcc-cross builds on Win32
1848 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
1849 - various changes to fix missing prototype warnings
1850 - fixed x86nasm.pl to create correct asm files for NASM COFF output
1851 - added AES, WHIRLPOOL and CPUID assembler code to build files
1852 - added missing AES assembler make rules to mk1mf.pl
1853 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
1854 [Guenter Knauf <eflash@gmx.net>]
1856 *) Implement certificate status request TLS extension defined in RFC3546.
1857 A client can set the appropriate parameters and receive the encoded
1858 OCSP response via a callback. A server can query the supplied parameters
1859 and set the encoded OCSP response in the callback. Add simplified examples
1860 to s_client and s_server.
1863 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
1865 *) Fix various bugs:
1866 + Binary incompatibility of ssl_ctx_st structure
1867 + DTLS interoperation with non-compliant servers
1868 + Don't call get_session_cb() without proposed session
1869 + Fix ia64 assembler code
1870 [Andy Polyakov, Steve Henson]
1872 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
1874 *) DTLS Handshake overhaul. There were longstanding issues with
1875 OpenSSL DTLS implementation, which were making it impossible for
1876 RFC 4347 compliant client to communicate with OpenSSL server.
1877 Unfortunately just fixing these incompatibilities would "cut off"
1878 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
1879 server keeps tolerating non RFC compliant syntax. The opposite is
1880 not true, 0.9.8f client can not communicate with earlier server.
1881 This update even addresses CVE-2007-4995.
1884 *) Changes to avoid need for function casts in OpenSSL: some compilers
1885 (gcc 4.2 and later) reject their use.
1886 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
1889 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1890 RFC4507bis. The encrypted ticket format is an encrypted encoded
1891 SSL_SESSION structure, that way new session features are automatically
1894 If a client application caches session in an SSL_SESSION structure
1895 support is transparent because tickets are now stored in the encoded
1898 The SSL_CTX structure automatically generates keys for ticket
1899 protection in servers so again support should be possible
1900 with no application modification.
1902 If a client or server wishes to disable RFC4507 support then the option
1903 SSL_OP_NO_TICKET can be set.
1905 Add a TLS extension debugging callback to allow the contents of any client
1906 or server extensions to be examined.
1908 This work was sponsored by Google.
1911 *) Add initial support for TLS extensions, specifically for the server_name
1912 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1913 have new members for a host name. The SSL data structure has an
1914 additional member SSL_CTX *initial_ctx so that new sessions can be
1915 stored in that context to allow for session resumption, even after the
1916 SSL has been switched to a new SSL_CTX in reaction to a client's
1917 server_name extension.
1919 New functions (subject to change):
1921 SSL_get_servername()
1922 SSL_get_servername_type()
1925 New CTRL codes and macros (subject to change):
1927 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1928 - SSL_CTX_set_tlsext_servername_callback()
1929 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1930 - SSL_CTX_set_tlsext_servername_arg()
1931 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1933 openssl s_client has a new '-servername ...' option.
1935 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1936 '-key2 ...', '-servername_fatal' (subject to change). This allows
1937 testing the HostName extension for a specific single host name ('-cert'
1938 and '-key' remain fallbacks for handshakes without HostName
1939 negotiation). If the unrecogninzed_name alert has to be sent, this by
1940 default is a warning; it becomes fatal with the '-servername_fatal'
1943 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
1945 *) Add AES and SSE2 assembly language support to VC++ build.
1948 *) Mitigate attack on final subtraction in Montgomery reduction.
1951 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
1952 (which previously caused an internal error).
1955 *) Squeeze another 10% out of IGE mode when in != out.
1958 *) AES IGE mode speedup.
1959 [Dean Gaudet (Google)]
1961 *) Add the Korean symmetric 128-bit cipher SEED (see
1962 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
1963 add SEED ciphersuites from RFC 4162:
1965 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
1966 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
1967 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
1968 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
1970 To minimize changes between patchlevels in the OpenSSL 0.9.8
1971 series, SEED remains excluded from compilation unless OpenSSL
1972 is configured with 'enable-seed'.
1973 [KISA, Bodo Moeller]
1975 *) Mitigate branch prediction attacks, which can be practical if a
1976 single processor is shared, allowing a spy process to extract
1977 information. For detailed background information, see
1978 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
1979 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
1980 and Necessary Software Countermeasures"). The core of the change
1981 are new versions BN_div_no_branch() and
1982 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
1983 respectively, which are slower, but avoid the security-relevant
1984 conditional branches. These are automatically called by BN_div()
1985 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
1986 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
1987 remove a conditional branch.
1989 BN_FLG_CONSTTIME is the new name for the previous
1990 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
1991 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
1992 in the exponent causes BN_mod_exp_mont() to use the alternative
1993 implementation in BN_mod_exp_mont_consttime().) The old name
1994 remains as a deprecated alias.
1996 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
1997 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
1998 constant-time implementations for more than just exponentiation.
1999 Here too the old name is kept as a deprecated alias.
2001 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2002 the BN_BLINDING structure gets an independent copy of the
2003 modulus. This means that the previous "BIGNUM *m" argument to
2004 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2005 essentially becomes "const BIGNUM *m", although we can't actually
2006 change this in the header file before 0.9.9. It allows
2007 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2008 enable BN_FLG_CONSTTIME.
2010 [Matthew D Wood (Intel Corp)]
2012 *) In the SSL/TLS server implementation, be strict about session ID
2013 context matching (which matters if an application uses a single
2014 external cache for different purposes). Previously,
2015 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2016 set. This did ensure strict client verification, but meant that,
2017 with applications using a single external cache for quite
2018 different requirements, clients could circumvent ciphersuite
2019 restrictions for a given session ID context by starting a session
2020 in a different context.
2023 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2024 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2025 authentication-only ciphersuites.
2028 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
2029 not complete and could lead to a possible single byte overflow
2030 (CVE-2007-5135) [Ben Laurie]
2032 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
2034 *) Since AES128 and AES256 (and similarly Camellia128 and
2035 Camellia256) share a single mask bit in the logic of
2036 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2037 kludge to work properly if AES128 is available and AES256 isn't
2038 (or if Camellia128 is available and Camellia256 isn't).
2041 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
2042 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
2043 When a point or a seed is encoded in a BIT STRING, we need to
2044 prevent the removal of trailing zero bits to get the proper DER
2045 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
2046 of a NamedBitList, for which trailing 0 bits need to be removed.)
2049 *) Have SSL/TLS server implementation tolerate "mismatched" record
2050 protocol version while receiving ClientHello even if the
2051 ClientHello is fragmented. (The server can't insist on the
2052 particular protocol version it has chosen before the ServerHello
2053 message has informed the client about his choice.)
2056 *) Add RFC 3779 support.
2057 [Rob Austein for ARIN, Ben Laurie]
2059 *) Load error codes if they are not already present instead of using a
2060 static variable. This allows them to be cleanly unloaded and reloaded.
2061 Improve header file function name parsing.
2064 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
2065 or CAPABILITY handshake as required by RFCs.
2068 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
2070 *) Introduce limits to prevent malicious keys being able to
2071 cause a denial of service. (CVE-2006-2940)
2072 [Steve Henson, Bodo Moeller]
2074 *) Fix ASN.1 parsing of certain invalid structures that can result
2075 in a denial of service. (CVE-2006-2937) [Steve Henson]
2077 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2078 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2080 *) Fix SSL client code which could crash if connecting to a
2081 malicious SSLv2 server. (CVE-2006-4343)
2082 [Tavis Ormandy and Will Drewry, Google Security Team]
2084 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
2085 match only those. Before that, "AES256-SHA" would be interpreted
2086 as a pattern and match "AES128-SHA" too (since AES128-SHA got
2087 the same strength classification in 0.9.7h) as we currently only
2088 have a single AES bit in the ciphersuite description bitmap.
2089 That change, however, also applied to ciphersuite strings such as
2090 "RC4-MD5" that intentionally matched multiple ciphersuites --
2091 namely, SSL 2.0 ciphersuites in addition to the more common ones
2092 from SSL 3.0/TLS 1.0.
2094 So we change the selection algorithm again: Naming an explicit
2095 ciphersuite selects this one ciphersuite, and any other similar
2096 ciphersuite (same bitmap) from *other* protocol versions.
2097 Thus, "RC4-MD5" again will properly select both the SSL 2.0
2098 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
2100 Since SSL 2.0 does not have any ciphersuites for which the
2101 128/256 bit distinction would be relevant, this works for now.
2102 The proper fix will be to use different bits for AES128 and
2103 AES256, which would have avoided the problems from the beginning;
2104 however, bits are scarce, so we can only do this in a new release
2105 (not just a patchlevel) when we can change the SSL_CIPHER
2106 definition to split the single 'unsigned long mask' bitmap into
2107 multiple values to extend the available space.
2111 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2113 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2114 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2116 *) Add AES IGE and biIGE modes.
2119 *) Change the Unix randomness entropy gathering to use poll() when
2120 possible instead of select(), since the latter has some
2121 undesirable limitations.
2122 [Darryl Miles via Richard Levitte and Bodo Moeller]
2124 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2125 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2126 cannot be implicitly activated as part of, e.g., the "AES" alias.
2127 However, please upgrade to OpenSSL 0.9.9[-dev] for
2128 non-experimental use of the ECC ciphersuites to get TLS extension
2129 support, which is required for curve and point format negotiation
2130 to avoid potential handshake problems.
2133 *) Disable rogue ciphersuites:
2135 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2136 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2137 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2139 The latter two were purportedly from
2140 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2143 Also deactivate the remaining ciphersuites from
2144 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2145 unofficial, and the ID has long expired.
2148 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2149 dual-core machines) and other potential thread-safety issues.
2152 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
2153 versions), which is now available for royalty-free use
2154 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
2155 Also, add Camellia TLS ciphersuites from RFC 4132.
2157 To minimize changes between patchlevels in the OpenSSL 0.9.8
2158 series, Camellia remains excluded from compilation unless OpenSSL
2159 is configured with 'enable-camellia'.
2162 *) Disable the padding bug check when compression is in use. The padding
2163 bug check assumes the first packet is of even length, this is not
2164 necessarily true if compresssion is enabled and can result in false
2165 positives causing handshake failure. The actual bug test is ancient
2166 code so it is hoped that implementations will either have fixed it by
2167 now or any which still have the bug do not support compression.
2170 Changes between 0.9.8a and 0.9.8b [04 May 2006]
2172 *) When applying a cipher rule check to see if string match is an explicit
2173 cipher suite and only match that one cipher suite if it is.
2176 *) Link in manifests for VC++ if needed.
2177 [Austin Ziegler <halostatue@gmail.com>]
2179 *) Update support for ECC-based TLS ciphersuites according to
2180 draft-ietf-tls-ecc-12.txt with proposed changes (but without
2181 TLS extensions, which are supported starting with the 0.9.9
2182 branch, not in the OpenSSL 0.9.8 branch).
2185 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
2186 opaque EVP_CIPHER_CTX handling.
2189 *) Fixes and enhancements to zlib compression code. We now only use
2190 "zlib1.dll" and use the default __cdecl calling convention on Win32
2191 to conform with the standards mentioned here:
2192 http://www.zlib.net/DLL_FAQ.txt
2193 Static zlib linking now works on Windows and the new --with-zlib-include
2194 --with-zlib-lib options to Configure can be used to supply the location
2195 of the headers and library. Gracefully handle case where zlib library
2199 *) Several fixes and enhancements to the OID generation code. The old code
2200 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
2201 handle numbers larger than ULONG_MAX, truncated printing and had a
2202 non standard OBJ_obj2txt() behaviour.
2205 *) Add support for building of engines under engine/ as shared libraries
2206 under VC++ build system.
2209 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
2210 Hopefully, we will not see any false combination of paths any more.
2213 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
2215 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2216 (part of SSL_OP_ALL). This option used to disable the
2217 countermeasure against man-in-the-middle protocol-version
2218 rollback in the SSL 2.0 server implementation, which is a bad
2219 idea. (CVE-2005-2969)
2221 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2222 for Information Security, National Institute of Advanced Industrial
2223 Science and Technology [AIST], Japan)]
2225 *) Add two function to clear and return the verify parameter flags.
2228 *) Keep cipherlists sorted in the source instead of sorting them at
2229 runtime, thus removing the need for a lock.
2232 *) Avoid some small subgroup attacks in Diffie-Hellman.
2233 [Nick Mathewson and Ben Laurie]
2235 *) Add functions for well-known primes.
2238 *) Extended Windows CE support.
2239 [Satoshi Nakamura and Andy Polyakov]
2241 *) Initialize SSL_METHOD structures at compile time instead of during
2242 runtime, thus removing the need for a lock.
2245 *) Make PKCS7_decrypt() work even if no certificate is supplied by
2246 attempting to decrypt each encrypted key in turn. Add support to
2250 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
2252 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
2255 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
2258 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
2259 key into the same file any more.
2262 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
2265 *) Add -utf8 command line and config file option to 'ca'.
2266 [Stefan <stf@udoma.org]
2268 *) Removed the macro des_crypt(), as it seems to conflict with some
2269 libraries. Use DES_crypt().
2272 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
2273 involves renaming the source and generated shared-libs for
2274 both. The engines will accept the corrected or legacy ids
2275 ('ncipher' and '4758_cca' respectively) when binding. NB,
2276 this only applies when building 'shared'.
2277 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
2279 *) Add attribute functions to EVP_PKEY structure. Modify
2280 PKCS12_create() to recognize a CSP name attribute and
2281 use it. Make -CSP option work again in pkcs12 utility.
2284 *) Add new functionality to the bn blinding code:
2285 - automatic re-creation of the BN_BLINDING parameters after
2286 a fixed number of uses (currently 32)
2287 - add new function for parameter creation
2288 - introduce flags to control the update behaviour of the
2289 BN_BLINDING parameters
2290 - hide BN_BLINDING structure
2291 Add a second BN_BLINDING slot to the RSA structure to improve
2292 performance when a single RSA object is shared among several
2296 *) Add support for DTLS.
2297 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
2299 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
2300 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
2303 *) Remove buggy and incompletet DH cert support from
2304 ssl/ssl_rsa.c and ssl/s3_both.c
2307 *) Use SHA-1 instead of MD5 as the default digest algorithm for
2308 the apps/openssl applications.
2311 *) Compile clean with "-Wall -Wmissing-prototypes
2312 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
2313 DEBUG_SAFESTACK must also be set.
2316 *) Change ./Configure so that certain algorithms can be disabled by default.
2317 The new counterpiece to "no-xxx" is "enable-xxx".
2319 The patented RC5 and MDC2 algorithms will now be disabled unless
2320 "enable-rc5" and "enable-mdc2", respectively, are specified.
2322 (IDEA remains enabled despite being patented. This is because IDEA
2323 is frequently required for interoperability, and there is no license
2324 fee for non-commercial use. As before, "no-idea" can be used to
2325 avoid this algorithm.)
2329 *) Add processing of proxy certificates (see RFC 3820). This work was
2330 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
2331 EGEE (Enabling Grids for E-science in Europe).
2334 *) RC4 performance overhaul on modern architectures/implementations, such
2335 as Intel P4, IA-64 and AMD64.
2338 *) New utility extract-section.pl. This can be used specify an alternative
2339 section number in a pod file instead of having to treat each file as
2340 a separate case in Makefile. This can be done by adding two lines to the
2343 =for comment openssl_section:XXX
2345 The blank line is mandatory.
2349 *) New arguments -certform, -keyform and -pass for s_client and s_server
2350 to allow alternative format key and certificate files and passphrase
2354 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
2355 update associated structures and add various utility functions.
2357 Add new policy related verify parameters, include policy checking in
2358 standard verify code. Enhance 'smime' application with extra parameters
2359 to support policy checking and print out.
2362 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
2363 Nehemiah processors. These extensions support AES encryption in hardware
2364 as well as RNG (though RNG support is currently disabled).
2365 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
2367 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
2370 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
2371 [Andy Polyakov and a number of other people]
2373 *) Improved PowerPC platform support. Most notably BIGNUM assembler
2374 implementation contributed by IBM.
2375 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
2377 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
2378 exponent rather than 'unsigned long'. There is a corresponding change to
2379 the new 'rsa_keygen' element of the RSA_METHOD structure.
2380 [Jelte Jansen, Geoff Thorpe]
2382 *) Functionality for creating the initial serial number file is now
2383 moved from CA.pl to the 'ca' utility with a new option -create_serial.
2385 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
2386 number file to 1, which is bound to cause problems. To avoid
2387 the problems while respecting compatibility between different 0.9.7
2388 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
2389 CA.pl for serial number initialization. With the new release 0.9.8,
2390 we can fix the problem directly in the 'ca' utility.)
2393 *) Reduced header interdepencies by declaring more opaque objects in
2394 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
2395 give fewer recursive includes, which could break lazy source code - so
2396 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
2397 developers should define this symbol when building and using openssl to
2398 ensure they track the recommended behaviour, interfaces, [etc], but
2399 backwards-compatible behaviour prevails when this isn't defined.
2402 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
2405 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
2406 This will generate a random key of the appropriate length based on the
2407 cipher context. The EVP_CIPHER can provide its own random key generation
2408 routine to support keys of a specific form. This is used in the des and
2409 3des routines to generate a key of the correct parity. Update S/MIME
2410 code to use new functions and hence generate correct parity DES keys.
2411 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
2412 valid (weak or incorrect parity).
2415 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
2416 as looking them up. This is useful when the verified structure may contain
2417 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
2418 present unless the new PKCS7_NO_CRL flag is asserted.
2421 *) Extend ASN1 oid configuration module. It now additionally accepts the
2424 shortName = some long name, 1.2.3.4
2427 *) Reimplemented the BN_CTX implementation. There is now no more static
2428 limitation on the number of variables it can handle nor the depth of the
2429 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
2430 information can now expand as required, and rather than having a single
2431 static array of bignums, BN_CTX now uses a linked-list of such arrays
2432 allowing it to expand on demand whilst maintaining the usefulness of
2433 BN_CTX's "bundling".
2436 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
2437 to allow all RSA operations to function using a single BN_CTX.
2440 *) Preliminary support for certificate policy evaluation and checking. This
2441 is initially intended to pass the tests outlined in "Conformance Testing
2442 of Relying Party Client Certificate Path Processing Logic" v1.07.
2445 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
2446 remained unused and not that useful. A variety of other little bignum
2447 tweaks and fixes have also been made continuing on from the audit (see
2451 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
2452 associated ASN1, EVP and SSL functions and old ASN1 macros.
2455 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
2456 and this should never fail. So the return value from the use of
2457 BN_set_word() (which can fail due to needless expansion) is now deprecated;
2458 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
2461 *) BN_CTX_get() should return zero-valued bignums, providing the same
2462 initialised value as BN_new().
2463 [Geoff Thorpe, suggested by Ulf Möller]
2465 *) Support for inhibitAnyPolicy certificate extension.
2468 *) An audit of the BIGNUM code is underway, for which debugging code is
2469 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
2470 is considered valid when processing BIGNUMs, and causes execution to
2471 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
2472 further steps are taken to deliberately pollute unused data in BIGNUM
2473 structures to try and expose faulty code further on. For now, openssl will
2474 (in its default mode of operation) continue to tolerate the inconsistent
2475 forms that it has tolerated in the past, but authors and packagers should
2476 consider trying openssl and their own applications when compiled with
2477 these debugging symbols defined. It will help highlight potential bugs in
2478 their own code, and will improve the test coverage for OpenSSL itself. At
2479 some point, these tighter rules will become openssl's default to improve
2480 maintainability, though the assert()s and other overheads will remain only
2481 in debugging configurations. See bn.h for more details.
2482 [Geoff Thorpe, Nils Larsch, Ulf Möller]
2484 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
2485 that can only be obtained through BN_CTX_new() (which implicitly
2486 initialises it). The presence of this function only made it possible
2487 to overwrite an existing structure (and cause memory leaks).
2490 *) Because of the callback-based approach for implementing LHASH as a
2491 template type, lh_insert() adds opaque objects to hash-tables and
2492 lh_doall() or lh_doall_arg() are typically used with a destructor callback
2493 to clean up those corresponding objects before destroying the hash table
2494 (and losing the object pointers). So some over-zealous constifications in
2495 LHASH have been relaxed so that lh_insert() does not take (nor store) the
2496 objects as "const" and the lh_doall[_arg] callback wrappers are not
2497 prototyped to have "const" restrictions on the object pointers they are
2498 given (and so aren't required to cast them away any more).
2501 *) The tmdiff.h API was so ugly and minimal that our own timing utility
2502 (speed) prefers to use its own implementation. The two implementations
2503 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
2504 its object type properly exposed (MS_TM) instead of casting to/from "char
2505 *". This may still change yet if someone realises MS_TM and "ms_time_***"
2506 aren't necessarily the greatest nomenclatures - but this is what was used
2507 internally to the implementation so I've used that for now.
2510 *) Ensure that deprecated functions do not get compiled when
2511 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
2512 the self-tests were still using deprecated key-generation functions so
2513 these have been updated also.
2516 *) Reorganise PKCS#7 code to separate the digest location functionality
2517 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
2518 New function PKCS7_set_digest() to set the digest type for PKCS#7
2519 digestedData type. Add additional code to correctly generate the
2520 digestedData type and add support for this type in PKCS7 initialization
2524 *) New function PKCS7_set0_type_other() this initializes a PKCS7
2525 structure of type "other".
2528 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
2529 sure the loop does correctly stop and breaking ("division by zero")
2530 modulus operations are not performed. The (pre-generated) prime
2531 table crypto/bn/bn_prime.h was already correct, but it could not be
2532 re-generated on some platforms because of the "division by zero"
2533 situation in the script.
2534 [Ralf S. Engelschall]
2536 *) Update support for ECC-based TLS ciphersuites according to
2537 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
2538 SHA-1 now is only used for "small" curves (where the
2539 representation of a field element takes up to 24 bytes); for
2540 larger curves, the field element resulting from ECDH is directly
2541 used as premaster secret.
2542 [Douglas Stebila (Sun Microsystems Laboratories)]
2544 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
2545 curve secp160r1 to the tests.
2546 [Douglas Stebila (Sun Microsystems Laboratories)]
2548 *) Add the possibility to load symbols globally with DSO.
2549 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
2551 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
2552 control of the error stack.
2555 *) Add support for STORE in ENGINE.
2558 *) Add the STORE type. The intention is to provide a common interface
2559 to certificate and key stores, be they simple file-based stores, or
2560 HSM-type store, or LDAP stores, or...
2561 NOTE: The code is currently UNTESTED and isn't really used anywhere.
2564 *) Add a generic structure called OPENSSL_ITEM. This can be used to
2565 pass a list of arguments to any function as well as provide a way
2566 for a function to pass data back to the caller.
2569 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
2570 works like BUF_strdup() but can be used to duplicate a portion of
2571 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
2575 *) Add the function sk_find_ex() which works like sk_find(), but will
2576 return an index to an element even if an exact match couldn't be
2577 found. The index is guaranteed to point at the element where the
2578 searched-for key would be inserted to preserve sorting order.
2581 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
2582 takes an extra flags argument for optional functionality. Currently,
2583 the following flags are defined:
2585 OBJ_BSEARCH_VALUE_ON_NOMATCH
2586 This one gets OBJ_bsearch_ex() to return a pointer to the first
2587 element where the comparing function returns a negative or zero
2590 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
2591 This one gets OBJ_bsearch_ex() to return a pointer to the first
2592 element where the comparing function returns zero. This is useful
2593 if there are more than one element where the comparing function
2597 *) Make it possible to create self-signed certificates with 'openssl ca'
2598 in such a way that the self-signed certificate becomes part of the
2599 CA database and uses the same mechanisms for serial number generation
2600 as all other certificate signing. The new flag '-selfsign' enables
2601 this functionality. Adapt CA.sh and CA.pl.in.
2604 *) Add functionality to check the public key of a certificate request
2605 against a given private. This is useful to check that a certificate
2606 request can be signed by that key (self-signing).
2609 *) Make it possible to have multiple active certificates with the same
2610 subject in the CA index file. This is done only if the keyword
2611 'unique_subject' is set to 'no' in the main CA section (default
2612 if 'CA_default') of the configuration file. The value is saved
2613 with the database itself in a separate index attribute file,
2614 named like the index file with '.attr' appended to the name.
2617 *) Generate muti valued AVAs using '+' notation in config files for
2621 *) Support for nameConstraints certificate extension.
2624 *) Support for policyConstraints certificate extension.
2627 *) Support for policyMappings certificate extension.
2630 *) Make sure the default DSA_METHOD implementation only uses its
2631 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
2632 and change its own handlers to be NULL so as to remove unnecessary
2633 indirection. This lets alternative implementations fallback to the
2634 default implementation more easily.
2637 *) Support for directoryName in GeneralName related extensions
2641 *) Make it possible to link applications using Makefile.shared.
2642 Make that possible even when linking against static libraries!
2645 *) Support for single pass processing for S/MIME signing. This now
2646 means that S/MIME signing can be done from a pipe, in addition
2647 cleartext signing (multipart/signed type) is effectively streaming
2648 and the signed data does not need to be all held in memory.
2650 This is done with a new flag PKCS7_STREAM. When this flag is set
2651 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
2652 is done after the data is output (and digests calculated) in
2653 SMIME_write_PKCS7().
2656 *) Add full support for -rpath/-R, both in shared libraries and
2657 applications, at least on the platforms where it's known how
2661 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
2662 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
2663 will now compute a table of multiples of the generator that
2664 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
2665 faster (notably in the case of a single point multiplication,
2666 scalar * generator).
2667 [Nils Larsch, Bodo Moeller]
2669 *) IPv6 support for certificate extensions. The various extensions
2670 which use the IP:a.b.c.d can now take IPv6 addresses using the
2671 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
2675 *) Added an ENGINE that implements RSA by performing private key
2676 exponentiations with the GMP library. The conversions to and from
2677 GMP's mpz_t format aren't optimised nor are any montgomery forms
2678 cached, and on x86 it appears OpenSSL's own performance has caught up.
2679 However there are likely to be other architectures where GMP could
2680 provide a boost. This ENGINE is not built in by default, but it can be
2681 specified at Configure time and should be accompanied by the necessary
2682 linker additions, eg;
2683 ./config -DOPENSSL_USE_GMP -lgmp
2686 *) "openssl engine" will not display ENGINE/DSO load failure errors when
2687 testing availability of engines with "-t" - the old behaviour is
2688 produced by increasing the feature's verbosity with "-tt".
2691 *) ECDSA routines: under certain error conditions uninitialized BN objects
2692 could be freed. Solution: make sure initialization is performed early
2693 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
2697 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
2698 and DH_METHOD (eg. by ENGINE implementations) to override the normal
2699 software implementations. For DSA and DH, parameter generation can
2700 also be overriden by providing the appropriate method callbacks.
2703 *) Change the "progress" mechanism used in key-generation and
2704 primality testing to functions that take a new BN_GENCB pointer in
2705 place of callback/argument pairs. The new API functions have "_ex"
2706 postfixes and the older functions are reimplemented as wrappers for
2707 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
2708 declarations of the old functions to help (graceful) attempts to
2709 migrate to the new functions. Also, the new key-generation API
2710 functions operate on a caller-supplied key-structure and return
2711 success/failure rather than returning a key or NULL - this is to
2712 help make "keygen" another member function of RSA_METHOD etc.
2714 Example for using the new callback interface:
2716 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
2720 BN_GENCB_set(&my_cb, my_callback, my_arg);
2722 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
2723 /* For the meaning of a, b in calls to my_callback(), see the
2724 * documentation of the function that calls the callback.
2725 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
2726 * my_callback should return 1 if it wants BN_is_prime_ex()
2727 * to continue, or 0 to stop.
2732 *) Change the ZLIB compression method to be stateful, and make it
2733 available to TLS with the number defined in
2734 draft-ietf-tls-compression-04.txt.
2737 *) Add the ASN.1 structures and functions for CertificatePair, which
2738 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
2740 CertificatePair ::= SEQUENCE {
2741 forward [0] Certificate OPTIONAL,
2742 reverse [1] Certificate OPTIONAL,
2743 -- at least one of the pair shall be present -- }
2745 Also implement the PEM functions to read and write certificate
2746 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
2748 This needed to be defined, mostly for the sake of the LDAP
2749 attribute crossCertificatePair, but may prove useful elsewhere as
2753 *) Make it possible to inhibit symlinking of shared libraries in
2754 Makefile.shared, for Cygwin's sake.
2757 *) Extend the BIGNUM API by creating a function
2758 void BN_set_negative(BIGNUM *a, int neg);
2759 and a macro that behave like
2760 int BN_is_negative(const BIGNUM *a);
2762 to avoid the need to access 'a->neg' directly in applications.
2765 *) Implement fast modular reduction for pseudo-Mersenne primes
2766 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
2767 EC_GROUP_new_curve_GFp() will now automatically use this
2769 [Nils Larsch <nla@trustcenter.de>]
2771 *) Add new lock type (CRYPTO_LOCK_BN).
2774 *) Change the ENGINE framework to automatically load engines
2775 dynamically from specific directories unless they could be
2776 found to already be built in or loaded. Move all the
2777 current engines except for the cryptodev one to a new
2779 The engines in engines/ are built as shared libraries if
2780 the "shared" options was given to ./Configure or ./config.
2781 Otherwise, they are inserted in libcrypto.a.
2782 /usr/local/ssl/engines is the default directory for dynamic
2783 engines, but that can be overriden at configure time through
2784 the usual use of --prefix and/or --openssldir, and at run
2785 time with the environment variable OPENSSL_ENGINES.
2786 [Geoff Thorpe and Richard Levitte]
2788 *) Add Makefile.shared, a helper makefile to build shared
2789 libraries. Addapt Makefile.org.
2792 *) Add version info to Win32 DLLs.
2793 [Peter 'Luna' Runestig" <peter@runestig.com>]
2795 *) Add new 'medium level' PKCS#12 API. Certificates and keys
2796 can be added using this API to created arbitrary PKCS#12
2797 files while avoiding the low level API.
2799 New options to PKCS12_create(), key or cert can be NULL and
2800 will then be omitted from the output file. The encryption
2801 algorithm NIDs can be set to -1 for no encryption, the mac
2802 iteration count can be set to 0 to omit the mac.
2804 Enhance pkcs12 utility by making the -nokeys and -nocerts
2805 options work when creating a PKCS#12 file. New option -nomac
2806 to omit the mac, NONE can be set for an encryption algorithm.
2807 New code is modified to use the enhanced PKCS12_create()
2808 instead of the low level API.
2811 *) Extend ASN1 encoder to support indefinite length constructed
2812 encoding. This can output sequences tags and octet strings in
2813 this form. Modify pk7_asn1.c to support indefinite length
2814 encoding. This is experimental and needs additional code to
2815 be useful, such as an ASN1 bio and some enhanced streaming
2818 Extend template encode functionality so that tagging is passed
2819 down to the template encoder.
2822 *) Let 'openssl req' fail if an argument to '-newkey' is not
2823 recognized instead of using RSA as a default.
2826 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
2827 As these are not official, they are not included in "ALL";
2828 the "ECCdraft" ciphersuite group alias can be used to select them.
2829 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
2831 *) Add ECDH engine support.
2832 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
2834 *) Add ECDH in new directory crypto/ecdh/.
2835 [Douglas Stebila (Sun Microsystems Laboratories)]
2837 *) Let BN_rand_range() abort with an error after 100 iterations
2838 without success (which indicates a broken PRNG).
2841 *) Change BN_mod_sqrt() so that it verifies that the input value
2842 is really the square of the return value. (Previously,
2843 BN_mod_sqrt would show GIGO behaviour.)
2846 *) Add named elliptic curves over binary fields from X9.62, SECG,
2847 and WAP/WTLS; add OIDs that were still missing.
2849 [Sheueling Chang Shantz and Douglas Stebila
2850 (Sun Microsystems Laboratories)]
2852 *) Extend the EC library for elliptic curves over binary fields
2853 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
2856 EC_GF2m_simple_method
2860 EC_GROUP_new_curve_GF2m
2861 EC_GROUP_set_curve_GF2m
2862 EC_GROUP_get_curve_GF2m
2863 EC_POINT_set_affine_coordinates_GF2m
2864 EC_POINT_get_affine_coordinates_GF2m
2865 EC_POINT_set_compressed_coordinates_GF2m
2867 Point compression for binary fields is disabled by default for
2868 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
2871 As binary polynomials are represented as BIGNUMs, various members
2872 of the EC_GROUP and EC_POINT data structures can be shared
2873 between the implementations for prime fields and binary fields;
2874 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
2875 are essentially identical to their ..._GFp counterparts.
2876 (For simplicity, the '..._GFp' prefix has been dropped from
2877 various internal method names.)
2879 An internal 'field_div' method (similar to 'field_mul' and
2880 'field_sqr') has been added; this is used only for binary fields.
2882 [Sheueling Chang Shantz and Douglas Stebila
2883 (Sun Microsystems Laboratories)]
2885 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
2886 through methods ('mul', 'precompute_mult').
2888 The generic implementations (now internally called 'ec_wNAF_mul'
2889 and 'ec_wNAF_precomputed_mult') remain the default if these
2890 methods are undefined.
2892 [Sheueling Chang Shantz and Douglas Stebila
2893 (Sun Microsystems Laboratories)]
2895 *) New function EC_GROUP_get_degree, which is defined through
2896 EC_METHOD. For curves over prime fields, this returns the bit
2897 length of the modulus.
2899 [Sheueling Chang Shantz and Douglas Stebila
2900 (Sun Microsystems Laboratories)]
2902 *) New functions EC_GROUP_dup, EC_POINT_dup.
2903 (These simply call ..._new and ..._copy).
2905 [Sheueling Chang Shantz and Douglas Stebila
2906 (Sun Microsystems Laboratories)]
2908 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
2909 Polynomials are represented as BIGNUMs (where the sign bit is not
2910 used) in the following functions [macros]:
2913 BN_GF2m_sub [= BN_GF2m_add]
2914 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
2915 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
2916 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
2918 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
2919 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
2920 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
2921 BN_GF2m_cmp [= BN_ucmp]
2923 (Note that only the 'mod' functions are actually for fields GF(2^m).
2924 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
2926 For some functions, an the irreducible polynomial defining a
2927 field can be given as an 'unsigned int[]' with strictly
2928 decreasing elements giving the indices of those bits that are set;
2929 i.e., p[] represents the polynomial
2930 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
2932 p[0] > p[1] > ... > p[k] = 0.
2933 This applies to the following functions:
2938 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
2939 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
2941 BN_GF2m_mod_sqrt_arr
2942 BN_GF2m_mod_solve_quad_arr
2946 Conversion can be performed by the following functions:
2951 bntest.c has additional tests for binary polynomial arithmetic.
2953 Two implementations for BN_GF2m_mod_div() are available.
2954 The default algorithm simply uses BN_GF2m_mod_inv() and
2955 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
2956 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
2957 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
2959 [Sheueling Chang Shantz and Douglas Stebila
2960 (Sun Microsystems Laboratories)]
2962 *) Add new error code 'ERR_R_DISABLED' that can be used when some
2963 functionality is disabled at compile-time.
2964 [Douglas Stebila <douglas.stebila@sun.com>]
2966 *) Change default behaviour of 'openssl asn1parse' so that more
2967 information is visible when viewing, e.g., a certificate:
2969 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
2970 mode the content of non-printable OCTET STRINGs is output in a
2971 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
2972 avoid the appearance of a printable string.
2973 [Nils Larsch <nla@trustcenter.de>]
2975 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
2977 EC_GROUP_set_asn1_flag()
2978 EC_GROUP_get_asn1_flag()
2979 EC_GROUP_set_point_conversion_form()
2980 EC_GROUP_get_point_conversion_form()
2981 These control ASN1 encoding details:
2982 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
2983 has been set to OPENSSL_EC_NAMED_CURVE.
2984 - Points are encoded in uncompressed form by default; options for
2985 asn1_for are as for point2oct, namely
2986 POINT_CONVERSION_COMPRESSED
2987 POINT_CONVERSION_UNCOMPRESSED
2988 POINT_CONVERSION_HYBRID
2990 Also add 'seed' and 'seed_len' members to EC_GROUP with access
2993 EC_GROUP_get0_seed()
2994 EC_GROUP_get_seed_len()
2995 This is used only for ASN1 purposes (so far).
2996 [Nils Larsch <nla@trustcenter.de>]
2998 *) Add 'field_type' member to EC_METHOD, which holds the NID
2999 of the appropriate field type OID. The new function
3000 EC_METHOD_get_field_type() returns this value.
3001 [Nils Larsch <nla@trustcenter.de>]
3006 EC_POINT_point2hex()
3007 EC_POINT_hex2point()
3008 providing useful interfaces to EC_POINT_point2oct() and
3009 EC_POINT_oct2point().
3010 [Nils Larsch <nla@trustcenter.de>]
3012 *) Change internals of the EC library so that the functions
3013 EC_GROUP_set_generator()
3014 EC_GROUP_get_generator()
3015 EC_GROUP_get_order()
3016 EC_GROUP_get_cofactor()
3017 are implemented directly in crypto/ec/ec_lib.c and not dispatched
3018 to methods, which would lead to unnecessary code duplication when
3019 adding different types of curves.
3020 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
3022 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
3023 arithmetic, and such that modified wNAFs are generated
3024 (which avoid length expansion in many cases).
3027 *) Add a function EC_GROUP_check_discriminant() (defined via
3028 EC_METHOD) that verifies that the curve discriminant is non-zero.
3030 Add a function EC_GROUP_check() that makes some sanity tests
3031 on a EC_GROUP, its generator and order. This includes
3032 EC_GROUP_check_discriminant().
3033 [Nils Larsch <nla@trustcenter.de>]
3035 *) Add ECDSA in new directory crypto/ecdsa/.
3037 Add applications 'openssl ecparam' and 'openssl ecdsa'
3038 (these are based on 'openssl dsaparam' and 'openssl dsa').
3040 ECDSA support is also included in various other files across the
3041 library. Most notably,
3042 - 'openssl req' now has a '-newkey ecdsa:file' option;
3043 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
3044 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
3045 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
3046 them suitable for ECDSA where domain parameters must be
3047 extracted before the specific public key;
3048 - ECDSA engine support has been added.
3049 [Nils Larsch <nla@trustcenter.de>]
3051 *) Include some named elliptic curves, and add OIDs from X9.62,
3052 SECG, and WAP/WTLS. Each curve can be obtained from the new
3054 EC_GROUP_new_by_curve_name(),
3055 and the list of available named curves can be obtained with
3056 EC_get_builtin_curves().
3057 Also add a 'curve_name' member to EC_GROUP objects, which can be
3059 EC_GROUP_set_curve_name()
3060 EC_GROUP_get_curve_name()
3061 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
3063 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3064 was actually never needed) and in BN_mul(). The removal in BN_mul()
3065 required a small change in bn_mul_part_recursive() and the addition
3066 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3067 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3068 bn_sub_words() and bn_add_words() except they take arrays with
3072 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
3074 *) Cleanse PEM buffers before freeing them since they may contain
3076 [Benjamin Bennett <ben@psc.edu>]
3078 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3079 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3080 authentication-only ciphersuites.
3083 *) Since AES128 and AES256 share a single mask bit in the logic of
3084 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3085 kludge to work properly if AES128 is available and AES256 isn't.
3088 *) Expand security boundary to match 1.1.1 module.
3091 *) Remove redundant features: hash file source, editing of test vectors
3092 modify fipsld to use external fips_premain.c signature.
3095 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
3096 run algorithm test programs.
3099 *) Make algorithm test programs more tolerant of whitespace.
3102 *) Have SSL/TLS server implementation tolerate "mismatched" record
3103 protocol version while receiving ClientHello even if the
3104 ClientHello is fragmented. (The server can't insist on the
3105 particular protocol version it has chosen before the ServerHello
3106 message has informed the client about his choice.)
3109 *) Load error codes if they are not already present instead of using a
3110 static variable. This allows them to be cleanly unloaded and reloaded.
3113 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3115 *) Introduce limits to prevent malicious keys being able to
3116 cause a denial of service. (CVE-2006-2940)
3117 [Steve Henson, Bodo Moeller]
3119 *) Fix ASN.1 parsing of certain invalid structures that can result
3120 in a denial of service. (CVE-2006-2937) [Steve Henson]
3122 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3123 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3125 *) Fix SSL client code which could crash if connecting to a
3126 malicious SSLv2 server. (CVE-2006-4343)
3127 [Tavis Ormandy and Will Drewry, Google Security Team]
3129 *) Change ciphersuite string processing so that an explicit
3130 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3131 will no longer include "AES128-SHA"), and any other similar
3132 ciphersuite (same bitmap) from *other* protocol versions (so that
3133 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3134 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3135 changes from 0.9.8b and 0.9.8d.
3138 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
3140 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3141 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3143 *) Change the Unix randomness entropy gathering to use poll() when
3144 possible instead of select(), since the latter has some
3145 undesirable limitations.
3146 [Darryl Miles via Richard Levitte and Bodo Moeller]
3148 *) Disable rogue ciphersuites:
3150 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3151 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3152 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3154 The latter two were purportedly from
3155 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3158 Also deactive the remaining ciphersuites from
3159 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3160 unofficial, and the ID has long expired.
3163 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3164 dual-core machines) and other potential thread-safety issues.
3167 Changes between 0.9.7i and 0.9.7j [04 May 2006]
3169 *) Adapt fipsld and the build system to link against the validated FIPS
3170 module in FIPS mode.
3173 *) Fixes for VC++ 2005 build under Windows.
3176 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
3177 from a Windows bash shell such as MSYS. It is autodetected from the
3178 "config" script when run from a VC++ environment. Modify standard VC++
3179 build to use fipscanister.o from the GNU make build.
3182 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
3184 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
3185 The value now differs depending on if you build for FIPS or not.
3186 BEWARE! A program linked with a shared FIPSed libcrypto can't be
3187 safely run with a non-FIPSed libcrypto, as it may crash because of
3188 the difference induced by this change.
3191 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
3193 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3194 (part of SSL_OP_ALL). This option used to disable the
3195 countermeasure against man-in-the-middle protocol-version
3196 rollback in the SSL 2.0 server implementation, which is a bad
3197 idea. (CVE-2005-2969)
3199 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3200 for Information Security, National Institute of Advanced Industrial
3201 Science and Technology [AIST], Japan)]
3203 *) Minimal support for X9.31 signatures and PSS padding modes. This is
3204 mainly for FIPS compliance and not fully integrated at this stage.
3207 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
3208 the exponentiation using a fixed-length exponent. (Otherwise,
3209 the information leaked through timing could expose the secret key
3210 after many signatures; cf. Bleichenbacher's attack on DSA with
3214 *) Make a new fixed-window mod_exp implementation the default for
3215 RSA, DSA, and DH private-key operations so that the sequence of
3216 squares and multiplies and the memory access pattern are
3217 independent of the particular secret key. This will mitigate
3218 cache-timing and potential related attacks.
3220 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
3221 and this is automatically used by BN_mod_exp_mont() if the new flag
3222 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
3223 will use this BN flag for private exponents unless the flag
3224 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
3225 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
3227 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
3229 *) Change the client implementation for SSLv23_method() and
3230 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
3231 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
3232 (Previously, the SSL 2.0 backwards compatible Client Hello
3233 message format would be used even with SSL_OP_NO_SSLv2.)
3236 *) Add support for smime-type MIME parameter in S/MIME messages which some
3240 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
3241 a threadsafe manner. Modify rsa code to use new function and add calls
3242 to dsa and dh code (which had race conditions before).
3245 *) Include the fixed error library code in the C error file definitions
3246 instead of fixing them up at runtime. This keeps the error code
3247 structures constant.
3250 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
3252 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
3255 *) Fixes for newer kerberos headers. NB: the casts are needed because
3256 the 'length' field is signed on one version and unsigned on another
3257 with no (?) obvious way to tell the difference, without these VC++
3258 complains. Also the "definition" of FAR (blank) is no longer included
3259 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
3260 some needed definitions.
3263 *) Undo Cygwin change.
3266 *) Added support for proxy certificates according to RFC 3820.
3267 Because they may be a security thread to unaware applications,
3268 they must be explicitely allowed in run-time. See
3269 docs/HOWTO/proxy_certificates.txt for further information.
3272 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
3274 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
3275 server and client random values. Previously
3276 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
3277 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
3279 This change has negligible security impact because:
3281 1. Server and client random values still have 24 bytes of pseudo random
3284 2. Server and client random values are sent in the clear in the initial
3287 3. The master secret is derived using the premaster secret (48 bytes in
3288 size for static RSA ciphersuites) as well as client server and random
3291 The OpenSSL team would like to thank the UK NISCC for bringing this issue
3294 [Stephen Henson, reported by UK NISCC]
3296 *) Use Windows randomness collection on Cygwin.
3299 *) Fix hang in EGD/PRNGD query when communication socket is closed
3300 prematurely by EGD/PRNGD.
3301 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
3303 *) Prompt for pass phrases when appropriate for PKCS12 input format.
3306 *) Back-port of selected performance improvements from development
3307 branch, as well as improved support for PowerPC platforms.
3310 *) Add lots of checks for memory allocation failure, error codes to indicate
3311 failure and freeing up memory if a failure occurs.
3312 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
3314 *) Add new -passin argument to dgst.
3317 *) Perform some character comparisons of different types in X509_NAME_cmp:
3318 this is needed for some certificates that reencode DNs into UTF8Strings
3319 (in violation of RFC3280) and can't or wont issue name rollover
3323 *) Make an explicit check during certificate validation to see that
3324 the CA setting in each certificate on the chain is correct. As a
3325 side effect always do the following basic checks on extensions,
3326 not just when there's an associated purpose to the check:
3328 - if there is an unhandled critical extension (unless the user
3329 has chosen to ignore this fault)
3330 - if the path length has been exceeded (if one is set at all)
3331 - that certain extensions fit the associated purpose (if one has
3335 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
3337 *) Avoid a race condition when CRLs are checked in a multi threaded
3338 environment. This would happen due to the reordering of the revoked
3339 entries during signature checking and serial number lookup. Now the
3340 encoding is cached and the serial number sort performed under a lock.
3341 Add new STACK function sk_is_sorted().
3344 *) Add Delta CRL to the extension code.
3347 *) Various fixes to s3_pkt.c so alerts are sent properly.
3348 [David Holmes <d.holmes@f5.com>]
3350 *) Reduce the chances of duplicate issuer name and serial numbers (in
3351 violation of RFC3280) using the OpenSSL certificate creation utilities.
3352 This is done by creating a random 64 bit value for the initial serial
3353 number when a serial number file is created or when a self signed
3354 certificate is created using 'openssl req -x509'. The initial serial
3355 number file is created using 'openssl x509 -next_serial' in CA.pl
3356 rather than being initialized to 1.
3359 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
3361 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3362 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3363 [Joe Orton, Steve Henson]
3365 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
3367 [Joe Orton, Steve Henson]
3369 *) Make it possible to have multiple active certificates with the same
3370 subject in the CA index file. This is done only if the keyword
3371 'unique_subject' is set to 'no' in the main CA section (default
3372 if 'CA_default') of the configuration file. The value is saved
3373 with the database itself in a separate index attribute file,
3374 named like the index file with '.attr' appended to the name.
3377 *) X509 verify fixes. Disable broken certificate workarounds when
3378 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
3379 keyUsage extension present. Don't accept CRLs with unhandled critical
3380 extensions: since verify currently doesn't process CRL extensions this
3381 rejects a CRL with *any* critical extensions. Add new verify error codes
3385 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
3386 A clarification of RFC2560 will require the use of OCTET STRINGs and
3387 some implementations cannot handle the current raw format. Since OpenSSL
3388 copies and compares OCSP nonces as opaque blobs without any attempt at
3389 parsing them this should not create any compatibility issues.
3392 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
3393 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
3394 this HMAC (and other) operations are several times slower than OpenSSL
3398 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
3399 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3401 *) Use the correct content when signing type "other".
3404 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
3406 *) Fix various bugs revealed by running the NISCC test suite:
3408 Stop out of bounds reads in the ASN1 code when presented with
3409 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3411 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
3413 If verify callback ignores invalid public key errors don't try to check
3414 certificate signature with the NULL public key.
3418 *) New -ignore_err option in ocsp application to stop the server
3419 exiting on the first error in a request.
3422 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3423 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3427 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3428 extra data after the compression methods not only for TLS 1.0
3429 but also for SSL 3.0 (as required by the specification).
3430 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3432 *) Change X509_certificate_type() to mark the key as exported/exportable
3433 when it's 512 *bits* long, not 512 bytes.
3436 *) Change AES_cbc_encrypt() so it outputs exact multiple of
3437 blocks during encryption.
3440 *) Various fixes to base64 BIO and non blocking I/O. On write
3441 flushes were not handled properly if the BIO retried. On read
3442 data was not being buffered properly and had various logic bugs.
3443 This also affects blocking I/O when the data being decoded is a
3447 *) Various S/MIME bugfixes and compatibility changes:
3448 output correct application/pkcs7 MIME type if
3449 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
3450 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
3451 of files as .eml work). Correctly handle very long lines in MIME
3455 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
3457 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3458 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3459 a protocol version number mismatch like a decryption error
3460 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3463 *) Turn on RSA blinding by default in the default implementation
3464 to avoid a timing attack. Applications that don't want it can call
3465 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3466 They would be ill-advised to do so in most cases.
3467 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3469 *) Change RSA blinding code so that it works when the PRNG is not
3470 seeded (in this case, the secret RSA exponent is abused as
3471 an unpredictable seed -- if it is not unpredictable, there
3472 is no point in blinding anyway). Make RSA blinding thread-safe
3473 by remembering the creator's thread ID in rsa->blinding and
3474 having all other threads use local one-time blinding factors
3475 (this requires more computation than sharing rsa->blinding, but
3476 avoids excessive locking; and if an RSA object is not shared
3477 between threads, blinding will still be very fast).
3480 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
3481 ENGINE as defaults for all supported algorithms irrespective of
3482 the 'flags' parameter. 'flags' is now honoured, so applications
3483 should make sure they are passing it correctly.
3486 *) Target "mingw" now allows native Windows code to be generated in
3487 the Cygwin environment as well as with the MinGW compiler.
3490 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
3492 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3493 via timing by performing a MAC computation even if incorrrect
3494 block cipher padding has been found. This is a countermeasure
3495 against active attacks where the attacker has to distinguish
3496 between bad padding and a MAC verification error. (CVE-2003-0078)
3498 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3499 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3500 Martin Vuagnoux (EPFL, Ilion)]
3502 *) Make the no-err option work as intended. The intention with no-err
3503 is not to have the whole error stack handling routines removed from
3504 libcrypto, it's only intended to remove all the function name and
3505 reason texts, thereby removing some of the footprint that may not
3506 be interesting if those errors aren't displayed anyway.
3508 NOTE: it's still possible for any application or module to have it's
3509 own set of error texts inserted. The routines are there, just not
3510 used by default when no-err is given.
3513 *) Add support for FreeBSD on IA64.
3514 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
3516 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
3517 Kerberos function mit_des_cbc_cksum(). Before this change,
3518 the value returned by DES_cbc_cksum() was like the one from
3519 mit_des_cbc_cksum(), except the bytes were swapped.
3520 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
3522 *) Allow an application to disable the automatic SSL chain building.
3523 Before this a rather primitive chain build was always performed in
3524 ssl3_output_cert_chain(): an application had no way to send the
3525 correct chain if the automatic operation produced an incorrect result.
3527 Now the chain builder is disabled if either:
3529 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
3531 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
3533 The reasoning behind this is that an application would not want the
3534 auto chain building to take place if extra chain certificates are
3535 present and it might also want a means of sending no additional
3536 certificates (for example the chain has two certificates and the
3540 *) Add the possibility to build without the ENGINE framework.
3541 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3543 *) Under Win32 gmtime() can return NULL: check return value in
3544 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
3547 *) DSA routines: under certain error conditions uninitialized BN objects
3548 could be freed. Solution: make sure initialization is performed early
3549 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
3550 Nils Larsch <nla@trustcenter.de> via PR#459)
3553 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
3554 checked on reconnect on the client side, therefore session resumption
3555 could still fail with a "ssl session id is different" error. This
3556 behaviour is masked when SSL_OP_ALL is used due to
3557 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
3558 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3559 followup to PR #377.
3562 *) IA-32 assembler support enhancements: unified ELF targets, support
3563 for SCO/Caldera platforms, fix for Cygwin shared build.
3566 *) Add support for FreeBSD on sparc64. As a consequence, support for
3567 FreeBSD on non-x86 processors is separate from x86 processors on
3568 the config script, much like the NetBSD support.
3569 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
3571 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
3573 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
3576 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
3577 code (06) was taken as the first octet of the session ID and the last
3578 octet was ignored consequently. As a result SSLv2 client side session
3579 caching could not have worked due to the session ID mismatch between
3581 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3585 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
3586 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
3590 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
3591 seems that in spite of existing for more than a year, many application
3592 author have done nothing to provide the necessary callbacks, which
3593 means that this particular engine will not work properly anywhere.
3594 This is a very unfortunate situation which forces us, in the name
3595 of usability, to give the hw_ncipher.c a static lock, which is part
3597 NOTE: This is for the 0.9.7 series ONLY. This hack will never
3598 appear in 0.9.8 or later. We EXPECT application authors to have
3599 dealt properly with this when 0.9.8 is released (unless we actually
3600 make such changes in the libcrypto locking code that changes will
3601 have to be made anyway).
3604 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
3605 octets have been read, EOF or an error occurs. Without this change
3606 some truncated ASN1 structures will not produce an error.
3609 *) Disable Heimdal support, since it hasn't been fully implemented.
3610 Still give the possibility to force the use of Heimdal, but with
3611 warnings and a request that patches get sent to openssl-dev.
3614 *) Add the VC-CE target, introduce the WINCE sysname, and add
3615 INSTALL.WCE and appropriate conditionals to make it build.
3616 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3618 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
3619 cygssl-x.y.z.dll, where x, y and z are the major, minor and
3620 edit numbers of the version.
3621 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3623 *) Introduce safe string copy and catenation functions
3624 (BUF_strlcpy() and BUF_strlcat()).
3625 [Ben Laurie (CHATS) and Richard Levitte]
3627 *) Avoid using fixed-size buffers for one-line DNs.
3628 [Ben Laurie (CHATS)]
3630 *) Add BUF_MEM_grow_clean() to avoid information leakage when
3631 resizing buffers containing secrets, and use where appropriate.
3632 [Ben Laurie (CHATS)]
3634 *) Avoid using fixed size buffers for configuration file location.
3635 [Ben Laurie (CHATS)]
3637 *) Avoid filename truncation for various CA files.
3638 [Ben Laurie (CHATS)]
3640 *) Use sizeof in preference to magic numbers.
3641 [Ben Laurie (CHATS)]
3643 *) Avoid filename truncation in cert requests.
3644 [Ben Laurie (CHATS)]
3646 *) Add assertions to check for (supposedly impossible) buffer
3648 [Ben Laurie (CHATS)]
3650 *) Don't cache truncated DNS entries in the local cache (this could
3651 potentially lead to a spoofing attack).
3652 [Ben Laurie (CHATS)]
3654 *) Fix various buffers to be large enough for hex/decimal
3655 representations in a platform independent manner.
3656 [Ben Laurie (CHATS)]
3658 *) Add CRYPTO_realloc_clean() to avoid information leakage when
3659 resizing buffers containing secrets, and use where appropriate.
3660 [Ben Laurie (CHATS)]
3662 *) Add BIO_indent() to avoid much slightly worrying code to do
3664 [Ben Laurie (CHATS)]
3666 *) Convert sprintf()/BIO_puts() to BIO_printf().
3667 [Ben Laurie (CHATS)]
3669 *) buffer_gets() could terminate with the buffer only half
3671 [Ben Laurie (CHATS)]
3673 *) Add assertions to prevent user-supplied crypto functions from
3674 overflowing internal buffers by having large block sizes, etc.
3675 [Ben Laurie (CHATS)]
3677 *) New OPENSSL_assert() macro (similar to assert(), but enabled
3679 [Ben Laurie (CHATS)]
3681 *) Eliminate unused copy of key in RC4.
3682 [Ben Laurie (CHATS)]
3684 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
3685 [Ben Laurie (CHATS)]
3687 *) Fix off-by-one error in EGD path.
3688 [Ben Laurie (CHATS)]
3690 *) If RANDFILE path is too long, ignore instead of truncating.
3691 [Ben Laurie (CHATS)]
3693 *) Eliminate unused and incorrectly sized X.509 structure
3695 [Ben Laurie (CHATS)]
3697 *) Eliminate unused and dangerous function knumber().
3698 [Ben Laurie (CHATS)]
3700 *) Eliminate unused and dangerous structure, KSSL_ERR.
3701 [Ben Laurie (CHATS)]
3703 *) Protect against overlong session ID context length in an encoded
3704 session object. Since these are local, this does not appear to be
3706 [Ben Laurie (CHATS)]
3708 *) Change from security patch (see 0.9.6e below) that did not affect
3709 the 0.9.6 release series:
3711 Remote buffer overflow in SSL3 protocol - an attacker could
3712 supply an oversized master key in Kerberos-enabled versions.
3714 [Ben Laurie (CHATS)]
3716 *) Change the SSL kerb5 codes to match RFC 2712.
3719 *) Make -nameopt work fully for req and add -reqopt switch.
3720 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
3722 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
3723 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
3725 *) Make sure tests can be performed even if the corresponding algorithms
3726 have been removed entirely. This was also the last step to make
3727 OpenSSL compilable with DJGPP under all reasonable conditions.
3728 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
3730 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
3731 to allow version independent disabling of normally unselected ciphers,
3732 which may be activated as a side-effect of selecting a single cipher.
3734 (E.g., cipher list string "RSA" enables ciphersuites that are left
3735 out of "ALL" because they do not provide symmetric encryption.
3736 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
3737 [Lutz Jaenicke, Bodo Moeller]
3739 *) Add appropriate support for separate platform-dependent build
3740 directories. The recommended way to make a platform-dependent
3741 build directory is the following (tested on Linux), maybe with
3744 # Place yourself outside of the OpenSSL source tree. In
3745 # this example, the environment variable OPENSSL_SOURCE
3746 # is assumed to contain the absolute OpenSSL source directory.
3747 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
3748 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
3749 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
3750 mkdir -p `dirname $F`
3751 ln -s $OPENSSL_SOURCE/$F $F
3754 To be absolutely sure not to disturb the source tree, a "make clean"
3755 is a good thing. If it isn't successfull, don't worry about it,
3756 it probably means the source directory is very clean.
3759 *) Make sure any ENGINE control commands make local copies of string
3760 pointers passed to them whenever necessary. Otherwise it is possible
3761 the caller may have overwritten (or deallocated) the original string
3762 data when a later ENGINE operation tries to use the stored values.
3763 [Götz Babin-Ebell <babinebell@trustcenter.de>]
3765 *) Improve diagnostics in file reading and command-line digests.
3766 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
3768 *) Add AES modes CFB and OFB to the object database. Correct an
3769 error in AES-CFB decryption.
3772 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
3773 allows existing EVP_CIPHER_CTX structures to be reused after
3774 calling EVP_*Final(). This behaviour is used by encryption
3775 BIOs and some applications. This has the side effect that
3776 applications must explicitly clean up cipher contexts with
3777 EVP_CIPHER_CTX_cleanup() or they will leak memory.
3780 *) Check the values of dna and dnb in bn_mul_recursive before calling
3781 bn_mul_comba (a non zero value means the a or b arrays do not contain
3782 n2 elements) and fallback to bn_mul_normal if either is not zero.
3785 *) Fix escaping of non-ASCII characters when using the -subj option
3786 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
3789 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
3790 form for "surname", serialNumber has no short form.
3791 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
3792 therefore remove "mail" short name for "internet 7".
3793 The OID for unique identifiers in X509 certificates is
3794 x500UniqueIdentifier, not uniqueIdentifier.
3795 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
3798 *) Add an "init" command to the ENGINE config module and auto initialize
3799 ENGINEs. Without any "init" command the ENGINE will be initialized
3800 after all ctrl commands have been executed on it. If init=1 the
3801 ENGINE is initailized at that point (ctrls before that point are run
3802 on the uninitialized ENGINE and after on the initialized one). If
3803 init=0 then the ENGINE will not be iniatialized at all.
3806 *) Fix the 'app_verify_callback' interface so that the user-defined
3807 argument is actually passed to the callback: In the
3808 SSL_CTX_set_cert_verify_callback() prototype, the callback
3809 declaration has been changed from
3812 int (*cb)(X509_STORE_CTX *,void *);
3813 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
3814 i=s->ctx->app_verify_callback(&ctx)
3815 has been changed into
3816 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
3818 To update applications using SSL_CTX_set_cert_verify_callback(),
3819 a dummy argument can be added to their callback functions.
3820 [D. K. Smetters <smetters@parc.xerox.com>]
3822 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
3823 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
3825 *) Add and OPENSSL_LOAD_CONF define which will cause
3826 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
3827 This allows older applications to transparently support certain
3828 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
3829 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
3830 load the config file and OPENSSL_add_all_algorithms_conf() which will
3831 always load it have also been added.
3834 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
3835 Adjust NIDs and EVP layer.
3836 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3838 *) Config modules support in openssl utility.
3840 Most commands now load modules from the config file,
3841 though in a few (such as version) this isn't done
3842 because it couldn't be used for anything.
3844 In the case of ca and req the config file used is
3845 the same as the utility itself: that is the -config
3846 command line option can be used to specify an
3850 *) Move default behaviour from OPENSSL_config(). If appname is NULL
3851 use "openssl_conf" if filename is NULL use default openssl config file.
3854 *) Add an argument to OPENSSL_config() to allow the use of an alternative
3855 config section name. Add a new flag to tolerate a missing config file
3856 and move code to CONF_modules_load_file().
3859 *) Support for crypto accelerator cards from Accelerated Encryption
3860 Processing, www.aep.ie. (Use engine 'aep')
3861 The support was copied from 0.9.6c [engine] and adapted/corrected
3862 to work with the new engine framework.
3863 [AEP Inc. and Richard Levitte]
3865 *) Support for SureWare crypto accelerator cards from Baltimore
3866 Technologies. (Use engine 'sureware')
3867 The support was copied from 0.9.6c [engine] and adapted
3868 to work with the new engine framework.
3871 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
3872 make the newer ENGINE framework commands for the CHIL engine work.
3873 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
3875 *) Make it possible to produce shared libraries on ReliantUNIX.
3876 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
3878 *) Add the configuration target debug-linux-ppro.
3879 Make 'openssl rsa' use the general key loading routines
3880 implemented in apps.c, and make those routines able to
3881 handle the key format FORMAT_NETSCAPE and the variant
3883 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3885 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
3886 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3888 *) Add -keyform to rsautl, and document -engine.
3889 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
3891 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
3892 BIO_R_NO_SUCH_FILE error code rather than the generic
3893 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
3896 *) Add new functions
3898 ERR_peek_last_error_line
3899 ERR_peek_last_error_line_data.
3900 These are similar to
3903 ERR_peek_error_line_data,
3904 but report on the latest error recorded rather than the first one
3905 still in the error queue.
3906 [Ben Laurie, Bodo Moeller]
3908 *) default_algorithms option in ENGINE config module. This allows things
3910 default_algorithms = ALL
3911 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
3914 *) Prelminary ENGINE config module.
3917 *) New experimental application configuration code.
3920 *) Change the AES code to follow the same name structure as all other
3921 symmetric ciphers, and behave the same way. Move everything to
3922 the directory crypto/aes, thereby obsoleting crypto/rijndael.
3923 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3925 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
3926 [Ben Laurie and Theo de Raadt]
3928 *) Add option to output public keys in req command.
3929 [Massimiliano Pala madwolf@openca.org]
3931 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
3932 (up to about 10% better than before for P-192 and P-224).
3935 *) New functions/macros
3937 SSL_CTX_set_msg_callback(ctx, cb)
3938 SSL_CTX_set_msg_callback_arg(ctx, arg)
3939 SSL_set_msg_callback(ssl, cb)
3940 SSL_set_msg_callback_arg(ssl, arg)
3942 to request calling a callback function
3944 void cb(int write_p, int version, int content_type,
3945 const void *buf, size_t len, SSL *ssl, void *arg)
3947 whenever a protocol message has been completely received
3948 (write_p == 0) or sent (write_p == 1). Here 'version' is the
3949 protocol version according to which the SSL library interprets
3950 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
3951 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
3952 the content type as defined in the SSL 3.0/TLS 1.0 protocol
3953 specification (change_cipher_spec(20), alert(21), handshake(22)).
3954 'buf' and 'len' point to the actual message, 'ssl' to the
3955 SSL object, and 'arg' is the application-defined value set by
3956 SSL[_CTX]_set_msg_callback_arg().
3958 'openssl s_client' and 'openssl s_server' have new '-msg' options
3959 to enable a callback that displays all protocol messages.
3962 *) Change the shared library support so shared libraries are built as
3963 soon as the corresponding static library is finished, and thereby get
3964 openssl and the test programs linked against the shared library.
3965 This still only happens when the keyword "shard" has been given to
3966 the configuration scripts.
3968 NOTE: shared library support is still an experimental thing, and
3969 backward binary compatibility is still not guaranteed.
3970 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
3972 *) Add support for Subject Information Access extension.
3973 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3975 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
3976 additional bytes when new memory had to be allocated, not just
3977 when reusing an existing buffer.
3980 *) New command line and configuration option 'utf8' for the req command.
3981 This allows field values to be specified as UTF8 strings.
3984 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
3985 runs for the former and machine-readable output for the latter.
3988 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
3989 of the e-mail address in the DN (i.e., it will go into a certificate
3990 extension only). The new configuration file option 'email_in_dn = no'
3991 has the same effect.
3992 [Massimiliano Pala madwolf@openca.org]
3994 *) Change all functions with names starting with des_ to be starting
3995 with DES_ instead. Add wrappers that are compatible with libdes,
3996 but are named _ossl_old_des_*. Finally, add macros that map the
3997 des_* symbols to the corresponding _ossl_old_des_* if libdes
3998 compatibility is desired. If OpenSSL 0.9.6c compatibility is
3999 desired, the des_* symbols will be mapped to DES_*, with one
4002 Since we provide two compatibility mappings, the user needs to
4003 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
4004 compatibility is desired. The default (i.e., when that macro
4005 isn't defined) is OpenSSL 0.9.6c compatibility.
4007 There are also macros that enable and disable the support of old
4008 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
4009 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
4010 are defined, the default will apply: to support the old des routines.
4012 In either case, one must include openssl/des.h to get the correct
4013 definitions. Do not try to just include openssl/des_old.h, that
4016 NOTE: This is a major break of an old API into a new one. Software
4017 authors are encouraged to switch to the DES_ style functions. Some
4018 time in the future, des_old.h and the libdes compatibility functions
4019 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
4020 default), and then completely removed.
4023 *) Test for certificates which contain unsupported critical extensions.
4024 If such a certificate is found during a verify operation it is
4025 rejected by default: this behaviour can be overridden by either
4026 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
4027 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
4028 X509_supported_extension() has also been added which returns 1 if a
4029 particular extension is supported.
4032 *) Modify the behaviour of EVP cipher functions in similar way to digests
4033 to retain compatibility with existing code.
4036 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
4037 compatibility with existing code. In particular the 'ctx' parameter does
4038 not have to be to be initialized before the call to EVP_DigestInit() and
4039 it is tidied up after a call to EVP_DigestFinal(). New function
4040 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
4041 EVP_MD_CTX_copy() changed to not require the destination to be
4042 initialized valid and new function EVP_MD_CTX_copy_ex() added which
4043 requires the destination to be valid.
4045 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
4046 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
4049 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
4050 so that complete 'Handshake' protocol structures are kept in memory
4051 instead of overwriting 'msg_type' and 'length' with 'body' data.
4054 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
4055 [Massimo Santin via Richard Levitte]
4057 *) Major restructuring to the underlying ENGINE code. This includes
4058 reduction of linker bloat, separation of pure "ENGINE" manipulation
4059 (initialisation, etc) from functionality dealing with implementations
4060 of specific crypto iterfaces. This change also introduces integrated
4061 support for symmetric ciphers and digest implementations - so ENGINEs
4062 can now accelerate these by providing EVP_CIPHER and EVP_MD
4063 implementations of their own. This is detailed in crypto/engine/README
4064 as it couldn't be adequately described here. However, there are a few
4065 API changes worth noting - some RSA, DSA, DH, and RAND functions that
4066 were changed in the original introduction of ENGINE code have now
4067 reverted back - the hooking from this code to ENGINE is now a good
4068 deal more passive and at run-time, operations deal directly with
4069 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
4070 dereferencing through an ENGINE pointer any more. Also, the ENGINE
4071 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
4072 they were not being used by the framework as there is no concept of a
4073 BIGNUM_METHOD and they could not be generalised to the new
4074 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
4075 ENGINE_cpy() has been removed as it cannot be consistently defined in
4079 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
4082 *) Change mkdef.pl to sort symbols that get the same entry number,
4083 and make sure the automatically generated functions ERR_load_*
4084 become part of libeay.num as well.
4087 *) New function SSL_renegotiate_pending(). This returns true once
4088 renegotiation has been requested (either SSL_renegotiate() call
4089 or HelloRequest/ClientHello receveived from the peer) and becomes
4090 false once a handshake has been completed.
4091 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
4092 sends a HelloRequest, but does not ensure that a handshake takes
4093 place. SSL_renegotiate_pending() is useful for checking if the
4094 client has followed the request.)
4097 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
4098 By default, clients may request session resumption even during
4099 renegotiation (if session ID contexts permit); with this option,
4100 session resumption is possible only in the first handshake.
4102 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4103 more bits available for options that should not be part of
4104 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4107 *) Add some demos for certificate and certificate request creation.
4110 *) Make maximum certificate chain size accepted from the peer application
4111 settable (SSL*_get/set_max_cert_list()), as proposed by
4112 "Douglas E. Engert" <deengert@anl.gov>.
4115 *) Add support for shared libraries for Unixware-7
4116 (Boyd Lynn Gerber <gerberb@zenez.com>).
4119 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4120 be done prior to destruction. Use this to unload error strings from
4121 ENGINEs that load their own error strings. NB: This adds two new API
4122 functions to "get" and "set" this destroy handler in an ENGINE.
4125 *) Alter all existing ENGINE implementations (except "openssl" and
4126 "openbsd") to dynamically instantiate their own error strings. This
4127 makes them more flexible to be built both as statically-linked ENGINEs
4128 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4129 Also, add stub code to each that makes building them as self-contained
4130 shared-libraries easier (see README.ENGINE).
4133 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4134 implementations into applications that are completely implemented in
4135 self-contained shared-libraries. The "dynamic" ENGINE exposes control
4136 commands that can be used to configure what shared-library to load and
4137 to control aspects of the way it is handled. Also, made an update to
4138 the README.ENGINE file that brings its information up-to-date and
4139 provides some information and instructions on the "dynamic" ENGINE
4140 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
4143 *) Make it possible to unload ranges of ERR strings with a new
4144 "ERR_unload_strings" function.
4147 *) Add a copy() function to EVP_MD.
4150 *) Make EVP_MD routines take a context pointer instead of just the
4151 md_data void pointer.
4154 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
4155 that the digest can only process a single chunk of data
4156 (typically because it is provided by a piece of
4157 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
4158 is only going to provide a single chunk of data, and hence the
4159 framework needn't accumulate the data for oneshot drivers.
4162 *) As with "ERR", make it possible to replace the underlying "ex_data"
4163 functions. This change also alters the storage and management of global
4164 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
4165 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
4166 index counters. The API functions that use this state have been changed
4167 to take a "class_index" rather than pointers to the class's local STACK
4168 and counter, and there is now an API function to dynamically create new
4169 classes. This centralisation allows us to (a) plug a lot of the
4170 thread-safety problems that existed, and (b) makes it possible to clean
4171 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
4172 such data would previously have always leaked in application code and
4173 workarounds were in place to make the memory debugging turn a blind eye
4174 to it. Application code that doesn't use this new function will still
4175 leak as before, but their memory debugging output will announce it now
4176 rather than letting it slide.
4178 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
4179 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
4180 has a return value to indicate success or failure.
4183 *) Make it possible to replace the underlying "ERR" functions such that the
4184 global state (2 LHASH tables and 2 locks) is only used by the "default"
4185 implementation. This change also adds two functions to "get" and "set"
4186 the implementation prior to it being automatically set the first time
4187 any other ERR function takes place. Ie. an application can call "get",
4188 pass the return value to a module it has just loaded, and that module
4189 can call its own "set" function using that value. This means the
4190 module's "ERR" operations will use (and modify) the error state in the
4191 application and not in its own statically linked copy of OpenSSL code.
4194 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
4195 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
4196 the operation, and provides a more encapsulated way for external code
4197 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
4198 to use these functions rather than manually incrementing the counts.
4200 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
4203 *) Add EVP test program.
4206 *) Add symmetric cipher support to ENGINE. Expect the API to change!
4209 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
4210 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
4211 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
4212 These allow a CRL to be built without having to access X509_CRL fields
4213 directly. Modify 'ca' application to use new functions.
4216 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
4217 bug workarounds. Rollback attack detection is a security feature.
4218 The problem will only arise on OpenSSL servers when TLSv1 is not
4219 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
4220 Software authors not wanting to support TLSv1 will have special reasons
4221 for their choice and can explicitly enable this option.
4222 [Bodo Moeller, Lutz Jaenicke]
4224 *) Rationalise EVP so it can be extended: don't include a union of
4225 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
4226 (similar to those existing for EVP_CIPHER_CTX).
4231 EVP_MD_CTX_init(&md); /* new function call */
4232 EVP_DigestInit(&md, EVP_sha1());
4233 EVP_DigestUpdate(&md, in, len);
4234 EVP_DigestFinal(&md, out, NULL);
4235 EVP_MD_CTX_cleanup(&md); /* new function call */
4239 *) Make DES key schedule conform to the usual scheme, as well as
4240 correcting its structure. This means that calls to DES functions
4241 now have to pass a pointer to a des_key_schedule instead of a
4242 plain des_key_schedule (which was actually always a pointer
4245 des_key_schedule ks;
4247 des_set_key_checked(..., &ks);
4248 des_ncbc_encrypt(..., &ks, ...);
4250 (Note that a later change renames 'des_...' into 'DES_...'.)
4253 *) Initial reduction of linker bloat: the use of some functions, such as
4254 PEM causes large amounts of unused functions to be linked in due to
4255 poor organisation. For example pem_all.c contains every PEM function
4256 which has a knock on effect of linking in large amounts of (unused)
4257 ASN1 code. Grouping together similar functions and splitting unrelated
4258 functions prevents this.
4261 *) Cleanup of EVP macros.
4264 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
4265 correct _ecb suffix.
4268 *) Add initial OCSP responder support to ocsp application. The
4269 revocation information is handled using the text based index
4270 use by the ca application. The responder can either handle
4271 requests generated internally, supplied in files (for example
4272 via a CGI script) or using an internal minimal server.
4275 *) Add configuration choices to get zlib compression for TLS.
4278 *) Changes to Kerberos SSL for RFC 2712 compliance:
4279 1. Implemented real KerberosWrapper, instead of just using
4280 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
4281 2. Implemented optional authenticator field of KerberosWrapper.
4283 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
4284 and authenticator structs; see crypto/krb5/.
4286 Generalized Kerberos calls to support multiple Kerberos libraries.
4287 [Vern Staats <staatsvr@asc.hpc.mil>,
4288 Jeffrey Altman <jaltman@columbia.edu>
4289 via Richard Levitte]
4291 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
4292 already does with RSA. testdsa.h now has 'priv_key/pub_key'
4293 values for each of the key sizes rather than having just
4294 parameters (and 'speed' generating keys each time).
4297 *) Speed up EVP routines.
4300 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
4301 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
4302 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
4303 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
4305 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
4306 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
4307 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
4310 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
4312 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
4315 *) Added the OS2-EMX target.
4316 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
4318 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
4319 to support NCONF routines in extension code. New function CONF_set_nconf()
4320 to allow functions which take an NCONF to also handle the old LHASH
4321 structure: this means that the old CONF compatible routines can be
4322 retained (in particular wrt extensions) without having to duplicate the
4323 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
4326 *) Enhance the general user interface with mechanisms for inner control
4327 and with possibilities to have yes/no kind of prompts.
4330 *) Change all calls to low level digest routines in the library and
4331 applications to use EVP. Add missing calls to HMAC_cleanup() and
4332 don't assume HMAC_CTX can be copied using memcpy().
4333 [Verdon Walker <VWalker@novell.com>, Steve Henson]
4335 *) Add the possibility to control engines through control names but with
4336 arbitrary arguments instead of just a string.
4337 Change the key loaders to take a UI_METHOD instead of a callback
4338 function pointer. NOTE: this breaks binary compatibility with earlier
4339 versions of OpenSSL [engine].
4340 Adapt the nCipher code for these new conditions and add a card insertion
4344 *) Enhance the general user interface with mechanisms to better support
4345 dialog box interfaces, application-defined prompts, the possibility
4346 to use defaults (for example default passwords from somewhere else)
4347 and interrupts/cancellations.
4350 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
4351 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
4354 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
4355 tidy up some unnecessarily weird code in 'sk_new()').
4356 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
4358 *) Change the key loading routines for ENGINEs to use the same kind
4359 callback (pem_password_cb) as all other routines that need this
4363 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
4364 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
4365 than this minimum value is recommended.
4368 *) New random seeder for OpenVMS, using the system process statistics
4369 that are easily reachable.
4372 *) Windows apparently can't transparently handle global
4373 variables defined in DLLs. Initialisations such as:
4375 const ASN1_ITEM *it = &ASN1_INTEGER_it;
4377 wont compile. This is used by the any applications that need to
4378 declare their own ASN1 modules. This was fixed by adding the option
4379 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
4380 needed for static libraries under Win32.
4383 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
4384 setting of purpose and trust fields. New X509_STORE trust and
4385 purpose functions and tidy up setting in other SSL functions.
4388 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
4389 structure. These are inherited by X509_STORE_CTX when it is
4390 initialised. This allows various defaults to be set in the
4391 X509_STORE structure (such as flags for CRL checking and custom
4392 purpose or trust settings) for functions which only use X509_STORE_CTX
4393 internally such as S/MIME.
4395 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
4396 trust settings if they are not set in X509_STORE. This allows X509_STORE
4397 purposes and trust (in S/MIME for example) to override any set by default.
4399 Add command line options for CRL checking to smime, s_client and s_server
4403 *) Initial CRL based revocation checking. If the CRL checking flag(s)
4404 are set then the CRL is looked up in the X509_STORE structure and
4405 its validity and signature checked, then if the certificate is found
4406 in the CRL the verify fails with a revoked error.
4408 Various new CRL related callbacks added to X509_STORE_CTX structure.
4410 Command line options added to 'verify' application to support this.
4412 This needs some additional work, such as being able to handle multiple
4413 CRLs with different times, extension based lookup (rather than just
4414 by subject name) and ultimately more complete V2 CRL extension
4418 *) Add a general user interface API (crypto/ui/). This is designed
4419 to replace things like des_read_password and friends (backward
4420 compatibility functions using this new API are provided).
4421 The purpose is to remove prompting functions from the DES code
4422 section as well as provide for prompting through dialog boxes in
4423 a window system and the like.
4426 *) Add "ex_data" support to ENGINE so implementations can add state at a
4427 per-structure level rather than having to store it globally.
4430 *) Make it possible for ENGINE structures to be copied when retrieved by
4431 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
4432 This causes the "original" ENGINE structure to act like a template,
4433 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
4434 operational state can be localised to each ENGINE structure, despite the
4435 fact they all share the same "methods". New ENGINE structures returned in
4436 this case have no functional references and the return value is the single
4437 structural reference. This matches the single structural reference returned
4438 by ENGINE_by_id() normally, when it is incremented on the pre-existing
4442 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
4443 needs to match any other type at all we need to manually clear the
4447 *) Changes to the "openssl engine" utility to include;
4448 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
4449 about an ENGINE's available control commands.
4450 - executing control commands from command line arguments using the
4451 '-pre' and '-post' switches. '-post' is only used if '-t' is
4452 specified and the ENGINE is successfully initialised. The syntax for
4453 the individual commands are colon-separated, for example;
4454 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
4457 *) New dynamic control command support for ENGINEs. ENGINEs can now
4458 declare their own commands (numbers), names (strings), descriptions,
4459 and input types for run-time discovery by calling applications. A
4460 subset of these commands are implicitly classed as "executable"
4461 depending on their input type, and only these can be invoked through
4462 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
4463 can be based on user input, config files, etc). The distinction is
4464 that "executable" commands cannot return anything other than a boolean
4465 result and can only support numeric or string input, whereas some
4466 discoverable commands may only be for direct use through
4467 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
4468 pointers, or other custom uses. The "executable" commands are to
4469 support parameterisations of ENGINE behaviour that can be
4470 unambiguously defined by ENGINEs and used consistently across any
4471 OpenSSL-based application. Commands have been added to all the
4472 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
4473 control over shared-library paths without source code alterations.
4476 *) Changed all ENGINE implementations to dynamically allocate their
4477 ENGINEs rather than declaring them statically. Apart from this being
4478 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
4479 this also allows the implementations to compile without using the
4480 internal engine_int.h header.
4483 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
4484 'const' value. Any code that should be able to modify a RAND_METHOD
4485 should already have non-const pointers to it (ie. they should only
4486 modify their own ones).
4489 *) Made a variety of little tweaks to the ENGINE code.
4490 - "atalla" and "ubsec" string definitions were moved from header files
4491 to C code. "nuron" string definitions were placed in variables
4492 rather than hard-coded - allowing parameterisation of these values
4493 later on via ctrl() commands.
4494 - Removed unused "#if 0"'d code.
4495 - Fixed engine list iteration code so it uses ENGINE_free() to release
4496 structural references.
4497 - Constified the RAND_METHOD element of ENGINE structures.
4498 - Constified various get/set functions as appropriate and added
4499 missing functions (including a catch-all ENGINE_cpy that duplicates
4500 all ENGINE values onto a new ENGINE except reference counts/state).
4501 - Removed NULL parameter checks in get/set functions. Setting a method
4502 or function to NULL is a way of cancelling out a previously set
4503 value. Passing a NULL ENGINE parameter is just plain stupid anyway
4504 and doesn't justify the extra error symbols and code.
4505 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
4506 flags from engine_int.h to engine.h.
4507 - Changed prototypes for ENGINE handler functions (init(), finish(),
4508 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
4511 *) Implement binary inversion algorithm for BN_mod_inverse in addition
4512 to the algorithm using long division. The binary algorithm can be
4513 used only if the modulus is odd. On 32-bit systems, it is faster
4514 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
4515 roughly 5-15% for 256-bit moduli), so we use it only for moduli
4516 up to 450 bits. In 64-bit environments, the binary algorithm
4517 appears to be advantageous for much longer moduli; here we use it
4518 for moduli up to 2048 bits.
4521 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
4522 could not support the combine flag in choice fields.
4525 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
4526 extensions from a certificate request to the certificate.
4529 *) Allow multiple 'certopt' and 'nameopt' options to be separated
4530 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
4531 file: this allows the display of the certificate about to be
4532 signed to be customised, to allow certain fields to be included
4533 or excluded and extension details. The old system didn't display
4534 multicharacter strings properly, omitted fields not in the policy
4535 and couldn't display additional details such as extensions.
4538 *) Function EC_POINTs_mul for multiple scalar multiplication
4539 of an arbitrary number of elliptic curve points
4540 \sum scalars[i]*points[i],
4541 optionally including the generator defined for the EC_GROUP:
4542 scalar*generator + \sum scalars[i]*points[i].
4544 EC_POINT_mul is a simple wrapper function for the typical case
4545 that the point list has just one item (besides the optional
4549 *) First EC_METHODs for curves over GF(p):
4551 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
4552 operations and provides various method functions that can also
4553 operate with faster implementations of modular arithmetic.
4555 EC_GFp_mont_method() reuses most functions that are part of
4556 EC_GFp_simple_method, but uses Montgomery arithmetic.
4558 [Bodo Moeller; point addition and point doubling
4559 implementation directly derived from source code provided by
4560 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
4562 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
4563 crypto/ec/ec_lib.c):
4565 Curves are EC_GROUP objects (with an optional group generator)
4566 based on EC_METHODs that are built into the library.
4568 Points are EC_POINT objects based on EC_GROUP objects.
4570 Most of the framework would be able to handle curves over arbitrary
4571 finite fields, but as there are no obvious types for fields other
4572 than GF(p), some functions are limited to that for now.
4575 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
4576 that the file contains a complete HTTP response.
4579 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
4580 change the def and num file printf format specifier from "%-40sXXX"
4581 to "%-39s XXX". The latter will always guarantee a space after the
4582 field while the former will cause them to run together if the field
4583 is 40 of more characters long.
4586 *) Constify the cipher and digest 'method' functions and structures
4587 and modify related functions to take constant EVP_MD and EVP_CIPHER
4591 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
4592 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
4595 *) Modify EVP_Digest*() routines so they now return values. Although the
4596 internal software routines can never fail additional hardware versions
4600 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
4602 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
4603 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
4606 ERR_R_NESTED_ASN1_ERROR
4608 ERR_R_MISSING_ASN1_EOS
4609 were 4 .. 9, conflicting with
4610 ERR_LIB_RSA (= ERR_R_RSA_LIB)
4612 ERR_LIB_PEM (= ERR_R_PEM_LIB).
4613 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
4615 Add new error code 'ERR_R_INTERNAL_ERROR'.
4618 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
4622 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
4623 sets the subject name for a new request or supersedes the
4624 subject name in a given request. Formats that can be parsed are
4625 'CN=Some Name, OU=myOU, C=IT'
4627 'CN=Some Name/OU=myOU/C=IT'.
4629 Add options '-batch' and '-verbose' to 'openssl req'.
4630 [Massimiliano Pala <madwolf@hackmasters.net>]
4632 *) Introduce the possibility to access global variables through
4633 functions on platform were that's the best way to handle exporting
4634 global variables in shared libraries. To enable this functionality,
4635 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
4636 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
4637 is normally done by Configure or something similar).
4639 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
4640 in the source file (foo.c) like this:
4642 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
4643 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
4645 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
4646 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
4648 OPENSSL_DECLARE_GLOBAL(int,foo);
4649 #define foo OPENSSL_GLOBAL_REF(foo)
4650 OPENSSL_DECLARE_GLOBAL(double,bar);
4651 #define bar OPENSSL_GLOBAL_REF(bar)
4653 The #defines are very important, and therefore so is including the
4654 header file everywhere where the defined globals are used.
4656 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
4657 of ASN.1 items, but that structure is a bit different.
4659 The largest change is in util/mkdef.pl which has been enhanced with
4660 better and easier to understand logic to choose which symbols should
4661 go into the Windows .def files as well as a number of fixes and code
4662 cleanup (among others, algorithm keywords are now sorted
4663 lexicographically to avoid constant rewrites).
4666 *) In BN_div() keep a copy of the sign of 'num' before writing the
4667 result to 'rm' because if rm==num the value will be overwritten
4668 and produce the wrong result if 'num' is negative: this caused
4669 problems with BN_mod() and BN_nnmod().
4672 *) Function OCSP_request_verify(). This checks the signature on an
4673 OCSP request and verifies the signer certificate. The signer
4674 certificate is just checked for a generic purpose and OCSP request
4678 *) Add OCSP_check_validity() function to check the validity of OCSP
4679 responses. OCSP responses are prepared in real time and may only
4680 be a few seconds old. Simply checking that the current time lies
4681 between thisUpdate and nextUpdate max reject otherwise valid responses
4682 caused by either OCSP responder or client clock inaccuracy. Instead
4683 we allow thisUpdate and nextUpdate to fall within a certain period of
4684 the current time. The age of the response can also optionally be
4685 checked. Two new options -validity_period and -status_age added to
4689 *) If signature or public key algorithm is unrecognized print out its
4690 OID rather that just UNKNOWN.
4693 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
4694 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
4695 ID to be generated from the issuer certificate alone which can then be
4696 passed to OCSP_id_issuer_cmp().
4699 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
4700 ASN1 modules to export functions returning ASN1_ITEM pointers
4701 instead of the ASN1_ITEM structures themselves. This adds several
4702 new macros which allow the underlying ASN1 function/structure to
4703 be accessed transparently. As a result code should not use ASN1_ITEM
4704 references directly (such as &X509_it) but instead use the relevant
4705 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
4706 use of the new ASN1 code on platforms where exporting structures
4707 is problematical (for example in shared libraries) but exporting
4708 functions returning pointers to structures is not.
4711 *) Add support for overriding the generation of SSL/TLS session IDs.
4712 These callbacks can be registered either in an SSL_CTX or per SSL.
4713 The purpose of this is to allow applications to control, if they wish,
4714 the arbitrary values chosen for use as session IDs, particularly as it
4715 can be useful for session caching in multiple-server environments. A
4716 command-line switch for testing this (and any client code that wishes
4717 to use such a feature) has been added to "s_server".
4718 [Geoff Thorpe, Lutz Jaenicke]
4720 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
4721 of the form '#if defined(...) || defined(...) || ...' and
4722 '#if !defined(...) && !defined(...) && ...'. This also avoids
4723 the growing number of special cases it was previously handling.
4726 *) Make all configuration macros available for application by making
4727 sure they are available in opensslconf.h, by giving them names starting
4728 with "OPENSSL_" to avoid conflicts with other packages and by making
4729 sure e_os2.h will cover all platform-specific cases together with
4731 Additionally, it is now possible to define configuration/platform-
4732 specific names (called "system identities"). In the C code, these
4733 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
4734 macro with the name beginning with "OPENSSL_SYS_", which is determined
4735 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
4739 *) New option -set_serial to 'req' and 'x509' this allows the serial
4740 number to use to be specified on the command line. Previously self
4741 signed certificates were hard coded with serial number 0 and the
4742 CA options of 'x509' had to use a serial number in a file which was
4746 *) New options to 'ca' utility to support V2 CRL entry extensions.
4747 Currently CRL reason, invalidity date and hold instruction are
4748 supported. Add new CRL extensions to V3 code and some new objects.
4751 *) New function EVP_CIPHER_CTX_set_padding() this is used to
4752 disable standard block padding (aka PKCS#5 padding) in the EVP
4753 API, which was previously mandatory. This means that the data is
4754 not padded in any way and so the total length much be a multiple
4755 of the block size, otherwise an error occurs.
4758 *) Initial (incomplete) OCSP SSL support.
4761 *) New function OCSP_parse_url(). This splits up a URL into its host,
4762 port and path components: primarily to parse OCSP URLs. New -url
4763 option to ocsp utility.
4766 *) New nonce behavior. The return value of OCSP_check_nonce() now
4767 reflects the various checks performed. Applications can decide
4768 whether to tolerate certain situations such as an absent nonce
4769 in a response when one was present in a request: the ocsp application
4770 just prints out a warning. New function OCSP_add1_basic_nonce()
4771 this is to allow responders to include a nonce in a response even if
4772 the request is nonce-less.
4775 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
4776 skipped when using openssl x509 multiple times on a single input file,
4777 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
4780 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
4781 set string type: to handle setting ASN1_TIME structures. Fix ca
4782 utility to correctly initialize revocation date of CRLs.
4785 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
4786 the clients preferred ciphersuites and rather use its own preferences.
4787 Should help to work around M$ SGC (Server Gated Cryptography) bug in
4788 Internet Explorer by ensuring unchanged hash method during stepup.
4789 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
4792 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
4793 to aes and add a new 'exist' option to print out symbols that don't
4797 *) Additional options to ocsp utility to allow flags to be set and
4798 additional certificates supplied.
4801 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
4802 OCSP client a number of certificate to only verify the response
4806 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
4807 handle the new API. Currently only ECB, CBC modes supported. Add new
4810 Add TLS AES ciphersuites as described in RFC3268, "Advanced
4811 Encryption Standard (AES) Ciphersuites for Transport Layer
4812 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
4813 not enabled by default and were not part of the "ALL" ciphersuite
4814 alias because they were not yet official; they could be
4815 explicitly requested by specifying the "AESdraft" ciphersuite
4816 group alias. In the final release of OpenSSL 0.9.7, the group
4817 alias is called "AES" and is part of "ALL".)
4818 [Ben Laurie, Steve Henson, Bodo Moeller]
4820 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
4821 request to response.
4824 *) Functions for OCSP responders. OCSP_request_onereq_count(),
4825 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
4826 extract information from a certificate request. OCSP_response_create()
4827 creates a response and optionally adds a basic response structure.
4828 OCSP_basic_add1_status() adds a complete single response to a basic
4829 response and returns the OCSP_SINGLERESP structure just added (to allow
4830 extensions to be included for example). OCSP_basic_add1_cert() adds a
4831 certificate to a basic response and OCSP_basic_sign() signs a basic
4832 response with various flags. New helper functions ASN1_TIME_check()
4833 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
4834 (converts ASN1_TIME to GeneralizedTime).
4837 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
4838 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
4839 structure from a certificate. X509_pubkey_digest() digests the public_key
4840 contents: this is used in various key identifiers.
4843 *) Make sk_sort() tolerate a NULL argument.
4844 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
4846 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
4847 passed by the function are trusted implicitly. If any of them signed the
4848 response then it is assumed to be valid and is not verified.
4851 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
4852 to data. This was previously part of the PKCS7 ASN1 code. This
4853 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
4854 [Steve Henson, reported by Kenneth R. Robinette
4855 <support@securenetterm.com>]
4857 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
4858 routines: without these tracing memory leaks is very painful.
4859 Fix leaks in PKCS12 and PKCS7 routines.
4862 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
4863 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
4864 effectively meant GeneralizedTime would never be used. Now it
4865 is initialised to -1 but X509_time_adj() now has to check the value
4866 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
4867 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
4868 [Steve Henson, reported by Kenneth R. Robinette
4869 <support@securenetterm.com>]
4871 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
4872 result in a zero length in the ASN1_INTEGER structure which was
4873 not consistent with the structure when d2i_ASN1_INTEGER() was used
4874 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
4875 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
4876 where it did not print out a minus for negative ASN1_INTEGER.
4879 *) Add summary printout to ocsp utility. The various functions which
4880 convert status values to strings have been renamed to:
4881 OCSP_response_status_str(), OCSP_cert_status_str() and
4882 OCSP_crl_reason_str() and are no longer static. New options
4883 to verify nonce values and to disable verification. OCSP response
4884 printout format cleaned up.
4887 *) Add additional OCSP certificate checks. These are those specified
4888 in RFC2560. This consists of two separate checks: the CA of the
4889 certificate being checked must either be the OCSP signer certificate
4890 or the issuer of the OCSP signer certificate. In the latter case the
4891 OCSP signer certificate must contain the OCSP signing extended key
4892 usage. This check is performed by attempting to match the OCSP
4893 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
4894 in the OCSP_CERTID structures of the response.
4897 *) Initial OCSP certificate verification added to OCSP_basic_verify()
4898 and related routines. This uses the standard OpenSSL certificate
4899 verify routines to perform initial checks (just CA validity) and
4900 to obtain the certificate chain. Then additional checks will be
4901 performed on the chain. Currently the root CA is checked to see
4902 if it is explicitly trusted for OCSP signing. This is used to set
4903 a root CA as a global signing root: that is any certificate that
4904 chains to that CA is an acceptable OCSP signing certificate.
4907 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
4908 extensions from a separate configuration file.
4909 As when reading extensions from the main configuration file,
4910 the '-extensions ...' option may be used for specifying the
4912 [Massimiliano Pala <madwolf@comune.modena.it>]
4914 *) New OCSP utility. Allows OCSP requests to be generated or
4915 read. The request can be sent to a responder and the output
4916 parsed, outputed or printed in text form. Not complete yet:
4917 still needs to check the OCSP response validity.
4920 *) New subcommands for 'openssl ca':
4921 'openssl ca -status <serial>' prints the status of the cert with
4922 the given serial number (according to the index file).
4923 'openssl ca -updatedb' updates the expiry status of certificates
4925 [Massimiliano Pala <madwolf@comune.modena.it>]
4927 *) New '-newreq-nodes' command option to CA.pl. This is like
4928 '-newreq', but calls 'openssl req' with the '-nodes' option
4929 so that the resulting key is not encrypted.
4930 [Damien Miller <djm@mindrot.org>]
4932 *) New configuration for the GNU Hurd.
4933 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
4935 *) Initial code to implement OCSP basic response verify. This
4936 is currently incomplete. Currently just finds the signer's
4937 certificate and verifies the signature on the response.
4940 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
4941 value of OPENSSLDIR. This is available via the new '-d' option
4942 to 'openssl version', and is also included in 'openssl version -a'.
4945 *) Allowing defining memory allocation callbacks that will be given
4946 file name and line number information in additional arguments
4947 (a const char* and an int). The basic functionality remains, as
4948 well as the original possibility to just replace malloc(),
4949 realloc() and free() by functions that do not know about these
4950 additional arguments. To register and find out the current
4951 settings for extended allocation functions, the following
4952 functions are provided:
4954 CRYPTO_set_mem_ex_functions
4955 CRYPTO_set_locked_mem_ex_functions
4956 CRYPTO_get_mem_ex_functions
4957 CRYPTO_get_locked_mem_ex_functions
4959 These work the same way as CRYPTO_set_mem_functions and friends.
4960 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
4961 extended allocation function is enabled.
4962 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
4963 a conventional allocation function is enabled.
4964 [Richard Levitte, Bodo Moeller]
4966 *) Finish off removing the remaining LHASH function pointer casts.
4967 There should no longer be any prototype-casting required when using
4968 the LHASH abstraction, and any casts that remain are "bugs". See
4969 the callback types and macros at the head of lhash.h for details
4970 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
4973 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
4974 If /dev/[u]random devices are not available or do not return enough
4975 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
4977 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
4978 /etc/entropy will be queried once each in this sequence, quering stops
4979 when enough entropy was collected without querying more sockets.
4982 *) Change the Unix RAND_poll() variant to be able to poll several
4983 random devices, as specified by DEVRANDOM, until a sufficient amount
4984 of data has been collected. We spend at most 10 ms on each file
4985 (select timeout) and read in non-blocking mode. DEVRANDOM now
4986 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
4987 (previously it was just the string "/dev/urandom"), so on typical
4988 platforms the 10 ms delay will never occur.
4989 Also separate out the Unix variant to its own file, rand_unix.c.
4990 For VMS, there's a currently-empty rand_vms.c.
4993 *) Move OCSP client related routines to ocsp_cl.c. These
4994 provide utility functions which an application needing
4995 to issue a request to an OCSP responder and analyse the
4996 response will typically need: as opposed to those which an
4997 OCSP responder itself would need which will be added later.
4999 OCSP_request_sign() signs an OCSP request with an API similar
5000 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
5001 response. OCSP_response_get1_basic() extracts basic response
5002 from response. OCSP_resp_find_status(): finds and extracts status
5003 information from an OCSP_CERTID structure (which will be created
5004 when the request structure is built). These are built from lower
5005 level functions which work on OCSP_SINGLERESP structures but
5006 wont normally be used unless the application wishes to examine
5007 extensions in the OCSP response for example.
5009 Replace nonce routines with a pair of functions.
5010 OCSP_request_add1_nonce() adds a nonce value and optionally
5011 generates a random value. OCSP_check_nonce() checks the
5012 validity of the nonce in an OCSP response.
5015 *) Change function OCSP_request_add() to OCSP_request_add0_id().
5016 This doesn't copy the supplied OCSP_CERTID and avoids the
5017 need to free up the newly created id. Change return type
5018 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
5019 This can then be used to add extensions to the request.
5020 Deleted OCSP_request_new(), since most of its functionality
5021 is now in OCSP_REQUEST_new() (and the case insensitive name
5022 clash) apart from the ability to set the request name which
5023 will be added elsewhere.
5026 *) Update OCSP API. Remove obsolete extensions argument from
5027 various functions. Extensions are now handled using the new
5028 OCSP extension code. New simple OCSP HTTP function which
5029 can be used to send requests and parse the response.
5032 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
5033 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
5034 uses the special reorder version of SET OF to sort the attributes
5035 and reorder them to match the encoded order. This resolves a long
5036 standing problem: a verify on a PKCS7 structure just after signing
5037 it used to fail because the attribute order did not match the
5038 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
5039 it uses the received order. This is necessary to tolerate some broken
5040 software that does not order SET OF. This is handled by encoding
5041 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
5042 to produce the required SET OF.
5045 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
5046 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
5047 files to get correct declarations of the ASN.1 item variables.
5050 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
5051 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
5052 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
5053 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
5054 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
5055 ASN1_ITEM and no wrapper functions.
5058 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
5059 replace the old function pointer based I/O routines. Change most of
5060 the *_d2i_bio() and *_d2i_fp() functions to use these.
5063 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
5064 lines, recognice more "algorithms" that can be deselected, and make
5065 it complain about algorithm deselection that isn't recognised.
5068 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
5069 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
5070 to use new functions. Add NO_ASN1_OLD which can be set to remove
5071 some old style ASN1 functions: this can be used to determine if old
5072 code will still work when these eventually go away.
5075 *) New extension functions for OCSP structures, these follow the
5076 same conventions as certificates and CRLs.
5079 *) New function X509V3_add1_i2d(). This automatically encodes and
5080 adds an extension. Its behaviour can be customised with various
5081 flags to append, replace or delete. Various wrappers added for
5082 certifcates and CRLs.
5085 *) Fix to avoid calling the underlying ASN1 print routine when
5086 an extension cannot be parsed. Correct a typo in the
5087 OCSP_SERVICELOC extension. Tidy up print OCSP format.
5090 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
5091 entries for variables.
5094 *) Add functionality to apps/openssl.c for detecting locking
5095 problems: As the program is single-threaded, all we have
5096 to do is register a locking callback using an array for
5097 storing which locks are currently held by the program.
5100 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
5101 SSL_get_ex_data_X509_STORE_idx(), which is used in
5102 ssl_verify_cert_chain() and thus can be called at any time
5103 during TLS/SSL handshakes so that thread-safety is essential.
5104 Unfortunately, the ex_data design is not at all suited
5105 for multi-threaded use, so it probably should be abolished.
5108 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5109 [Broadcom, tweaked and integrated by Geoff Thorpe]
5111 *) Move common extension printing code to new function
5112 X509V3_print_extensions(). Reorganise OCSP print routines and
5113 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5116 *) New function X509_signature_print() to remove duplication in some
5120 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5121 set (this was treated exactly the same as SET OF previously). This
5122 is used to reorder the STACK representing the structure to match the
5123 encoding. This will be used to get round a problem where a PKCS7
5124 structure which was signed could not be verified because the STACK
5125 order did not reflect the encoded order.
5128 *) Reimplement the OCSP ASN1 module using the new code.
5131 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5132 for its ASN1 operations. The old style function pointers still exist
5133 for now but they will eventually go away.
5136 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
5137 completely replaces the old ASN1 functionality with a table driven
5138 encoder and decoder which interprets an ASN1_ITEM structure describing
5139 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
5140 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
5141 has also been converted to the new form.
5144 *) Change BN_mod_exp_recp so that negative moduli are tolerated
5145 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
5146 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
5147 for negative moduli.
5150 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
5151 of not touching the result's sign bit.
5154 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
5158 *) Changed the LHASH code to use prototypes for callbacks, and created
5159 macros to declare and implement thin (optionally static) functions
5160 that provide type-safety and avoid function pointer casting for the
5161 type-specific callbacks.
5164 *) Added Kerberos Cipher Suites to be used with TLS, as written in
5166 [Veers Staats <staatsvr@asc.hpc.mil>,
5167 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
5169 *) Reformat the FAQ so the different questions and answers can be divided
5170 in sections depending on the subject.
5173 *) Have the zlib compression code load ZLIB.DLL dynamically under
5177 *) New function BN_mod_sqrt for computing square roots modulo a prime
5178 (using the probabilistic Tonelli-Shanks algorithm unless
5179 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
5180 be handled deterministically).
5181 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5183 *) Make BN_mod_inverse faster by explicitly handling small quotients
5184 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
5185 512 bits], about 30% for larger ones [1024 or 2048 bits].)
5188 *) New function BN_kronecker.
5191 *) Fix BN_gcd so that it works on negative inputs; the result is
5192 positive unless both parameters are zero.
5193 Previously something reasonably close to an infinite loop was
5194 possible because numbers could be growing instead of shrinking
5195 in the implementation of Euclid's algorithm.
5198 *) Fix BN_is_word() and BN_is_one() macros to take into account the
5199 sign of the number in question.
5201 Fix BN_is_word(a,w) to work correctly for w == 0.
5203 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
5204 because its test if the absolute value of 'a' equals 'w'.
5205 Note that BN_abs_is_word does *not* handle w == 0 reliably;
5206 it exists mostly for use in the implementations of BN_is_zero(),
5207 BN_is_one(), and BN_is_word().
5210 *) New function BN_swap.
5213 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
5214 the exponentiation functions are more likely to produce reasonable
5215 results on negative inputs.
5218 *) Change BN_mod_mul so that the result is always non-negative.
5219 Previously, it could be negative if one of the factors was negative;
5220 I don't think anyone really wanted that behaviour.
5223 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
5224 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
5225 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
5226 and add new functions:
5235 BN_mod_lshift1_quick
5239 These functions always generate non-negative results.
5241 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
5242 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
5244 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
5245 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
5246 be reduced modulo m.
5247 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5250 The following entry accidentily appeared in the CHANGES file
5251 distributed with OpenSSL 0.9.7. The modifications described in
5252 it do *not* apply to OpenSSL 0.9.7.
5254 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5255 was actually never needed) and in BN_mul(). The removal in BN_mul()
5256 required a small change in bn_mul_part_recursive() and the addition
5257 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5258 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5259 bn_sub_words() and bn_add_words() except they take arrays with
5264 *) In 'openssl passwd', verify passwords read from the terminal
5265 unless the '-salt' option is used (which usually means that
5266 verification would just waste user's time since the resulting
5267 hash is going to be compared with some given password hash)
5268 or the new '-noverify' option is used.
5270 This is an incompatible change, but it does not affect
5271 non-interactive use of 'openssl passwd' (passwords on the command
5272 line, '-stdin' option, '-in ...' option) and thus should not
5276 *) Remove all references to RSAref, since there's no more need for it.
5279 *) Make DSO load along a path given through an environment variable
5280 (SHLIB_PATH) with shl_load().
5283 *) Constify the ENGINE code as a result of BIGNUM constification.
5284 Also constify the RSA code and most things related to it. In a
5285 few places, most notable in the depth of the ASN.1 code, ugly
5286 casts back to non-const were required (to be solved at a later
5290 *) Make it so the openssl application has all engines loaded by default.
5293 *) Constify the BIGNUM routines a little more.
5296 *) Add the following functions:
5298 ENGINE_load_cswift()
5300 ENGINE_load_atalla()
5302 ENGINE_load_builtin_engines()
5304 That way, an application can itself choose if external engines that
5305 are built-in in OpenSSL shall ever be used or not. The benefit is
5306 that applications won't have to be linked with libdl or other dso
5307 libraries unless it's really needed.
5309 Changed 'openssl engine' to load all engines on demand.
5310 Changed the engine header files to avoid the duplication of some
5311 declarations (they differed!).
5314 *) 'openssl engine' can now list capabilities.
5317 *) Better error reporting in 'openssl engine'.
5320 *) Never call load_dh_param(NULL) in s_server.
5323 *) Add engine application. It can currently list engines by name and
5324 identity, and test if they are actually available.
5327 *) Improve RPM specification file by forcing symbolic linking and making
5328 sure the installed documentation is also owned by root.root.
5329 [Damien Miller <djm@mindrot.org>]
5331 *) Give the OpenSSL applications more possibilities to make use of
5332 keys (public as well as private) handled by engines.
5335 *) Add OCSP code that comes from CertCo.
5338 *) Add VMS support for the Rijndael code.
5341 *) Added untested support for Nuron crypto accelerator.
5344 *) Add support for external cryptographic devices. This code was
5345 previously distributed separately as the "engine" branch.
5346 [Geoff Thorpe, Richard Levitte]
5348 *) Rework the filename-translation in the DSO code. It is now possible to
5349 have far greater control over how a "name" is turned into a filename
5350 depending on the operating environment and any oddities about the
5351 different shared library filenames on each system.
5354 *) Support threads on FreeBSD-elf in Configure.
5357 *) Fix for SHA1 assembly problem with MASM: it produces
5358 warnings about corrupt line number information when assembling
5359 with debugging information. This is caused by the overlapping
5361 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
5364 NCONF_get_number() has no error checking at all. As a replacement,
5365 NCONF_get_number_e() is defined (_e for "error checking") and is
5366 promoted strongly. The old NCONF_get_number is kept around for
5367 binary backward compatibility.
5368 Make it possible for methods to load from something other than a BIO,
5369 by providing a function pointer that is given a name instead of a BIO.
5370 For example, this could be used to load configuration data from an
5374 *) Fix for non blocking accept BIOs. Added new I/O special reason
5375 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
5376 with non blocking I/O was not possible because no retry code was
5377 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
5381 *) Added the beginnings of Rijndael support.
5384 *) Fix for bug in DirectoryString mask setting. Add support for
5385 X509_NAME_print_ex() in 'req' and X509_print_ex() function
5386 to allow certificate printing to more controllable, additional
5387 'certopt' option to 'x509' to allow new printing options to be
5391 *) Clean old EAY MD5 hack from e_os.h.
5394 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
5396 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5397 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5398 [Joe Orton, Steve Henson]
5400 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
5402 *) Fix additional bug revealed by the NISCC test suite:
5404 Stop bug triggering large recursion when presented with
5405 certain ASN.1 tags (CVE-2003-0851)
5408 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
5410 *) Fix various bugs revealed by running the NISCC test suite:
5412 Stop out of bounds reads in the ASN1 code when presented with
5413 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5415 If verify callback ignores invalid public key errors don't try to check
5416 certificate signature with the NULL public key.
5420 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5421 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5425 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5426 extra data after the compression methods not only for TLS 1.0
5427 but also for SSL 3.0 (as required by the specification).
5428 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5430 *) Change X509_certificate_type() to mark the key as exported/exportable
5431 when it's 512 *bits* long, not 512 bytes.
5434 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
5436 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5437 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5438 a protocol version number mismatch like a decryption error
5439 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5442 *) Turn on RSA blinding by default in the default implementation
5443 to avoid a timing attack. Applications that don't want it can call
5444 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5445 They would be ill-advised to do so in most cases.
5446 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5448 *) Change RSA blinding code so that it works when the PRNG is not
5449 seeded (in this case, the secret RSA exponent is abused as
5450 an unpredictable seed -- if it is not unpredictable, there
5451 is no point in blinding anyway). Make RSA blinding thread-safe
5452 by remembering the creator's thread ID in rsa->blinding and
5453 having all other threads use local one-time blinding factors
5454 (this requires more computation than sharing rsa->blinding, but
5455 avoids excessive locking; and if an RSA object is not shared
5456 between threads, blinding will still be very fast).
5459 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
5461 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5462 via timing by performing a MAC computation even if incorrrect
5463 block cipher padding has been found. This is a countermeasure
5464 against active attacks where the attacker has to distinguish
5465 between bad padding and a MAC verification error. (CVE-2003-0078)
5467 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5468 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5469 Martin Vuagnoux (EPFL, Ilion)]
5471 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
5473 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
5474 memory from it's contents. This is done with a counter that will
5475 place alternating values in each byte. This can be used to solve
5476 two issues: 1) the removal of calls to memset() by highly optimizing
5477 compilers, and 2) cleansing with other values than 0, since those can
5478 be read through on certain media, for example a swap space on disk.
5481 *) Bugfix: client side session caching did not work with external caching,
5482 because the session->cipher setting was not restored when reloading
5483 from the external cache. This problem was masked, when
5484 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
5485 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
5488 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
5489 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
5490 [Zeev Lieber <zeev-l@yahoo.com>]
5492 *) Undo an undocumented change introduced in 0.9.6e which caused
5493 repeated calls to OpenSSL_add_all_ciphers() and
5494 OpenSSL_add_all_digests() to be ignored, even after calling
5498 *) Change the default configuration reader to deal with last line not
5499 being properly terminated.
5502 *) Change X509_NAME_cmp() so it applies the special rules on handling
5503 DN values that are of type PrintableString, as well as RDNs of type
5504 emailAddress where the value has the type ia5String.
5505 [stefank@valicert.com via Richard Levitte]
5507 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
5508 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
5509 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
5510 the bitwise-OR of the two for use by the majority of applications
5511 wanting this behaviour, and update the docs. The documented
5512 behaviour and actual behaviour were inconsistent and had been
5513 changing anyway, so this is more a bug-fix than a behavioural
5515 [Geoff Thorpe, diagnosed by Nadav Har'El]
5517 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
5518 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
5521 *) Fix initialization code race conditions in
5522 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
5523 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
5524 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
5525 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
5526 ssl2_get_cipher_by_char(),
5527 ssl3_get_cipher_by_char().
5528 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
5530 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
5531 the cached sessions are flushed, as the remove_cb() might use ex_data
5532 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
5533 (see [openssl.org #212]).
5534 [Geoff Thorpe, Lutz Jaenicke]
5536 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
5537 length, instead of the encoding length to d2i_ASN1_OBJECT.
5540 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
5542 *) [In 0.9.6g-engine release:]
5543 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
5544 [Lynn Gazis <lgazis@rainbow.com>]
5546 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
5548 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
5549 and get fix the header length calculation.
5550 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
5551 Alon Kantor <alonk@checkpoint.com> (and others),
5554 *) Use proper error handling instead of 'assertions' in buffer
5555 overflow checks added in 0.9.6e. This prevents DoS (the
5556 assertions could call abort()).
5557 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
5559 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
5561 *) Add various sanity checks to asn1_get_length() to reject
5562 the ASN1 length bytes if they exceed sizeof(long), will appear
5563 negative or the content length exceeds the length of the
5565 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5567 *) Fix cipher selection routines: ciphers without encryption had no flags
5568 for the cipher strength set and where therefore not handled correctly
5569 by the selection routines (PR #130).
5572 *) Fix EVP_dsa_sha macro.
5576 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
5577 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
5578 that was added in OpenSSL 0.9.6d.
5580 As the countermeasure turned out to be incompatible with some
5581 broken SSL implementations, the new option is part of SSL_OP_ALL.
5582 SSL_OP_ALL is usually employed when compatibility with weird SSL
5583 implementations is desired (e.g. '-bugs' option to 's_client' and
5584 's_server'), so the new option is automatically set in many
5588 *) Changes in security patch:
5590 Changes marked "(CHATS)" were sponsored by the Defense Advanced
5591 Research Projects Agency (DARPA) and Air Force Research Laboratory,
5592 Air Force Materiel Command, USAF, under agreement number
5595 *) Add various sanity checks to asn1_get_length() to reject
5596 the ASN1 length bytes if they exceed sizeof(long), will appear
5597 negative or the content length exceeds the length of the
5598 supplied buffer. (CVE-2002-0659)
5599 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5601 *) Assertions for various potential buffer overflows, not known to
5603 [Ben Laurie (CHATS)]
5605 *) Various temporary buffers to hold ASCII versions of integers were
5606 too small for 64 bit platforms. (CVE-2002-0655)
5607 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
5609 *) Remote buffer overflow in SSL3 protocol - an attacker could
5610 supply an oversized session ID to a client. (CVE-2002-0656)
5611 [Ben Laurie (CHATS)]
5613 *) Remote buffer overflow in SSL2 protocol - an attacker could
5614 supply an oversized client master key. (CVE-2002-0656)
5615 [Ben Laurie (CHATS)]
5617 Changes between 0.9.6c and 0.9.6d [9 May 2002]
5619 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
5620 encoded as NULL) with id-dsa-with-sha1.
5621 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
5623 *) Check various X509_...() return values in apps/req.c.
5624 [Nils Larsch <nla@trustcenter.de>]
5626 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
5627 an end-of-file condition would erronously be flagged, when the CRLF
5628 was just at the end of a processed block. The bug was discovered when
5629 processing data through a buffering memory BIO handing the data to a
5630 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
5631 <ptsekov@syntrex.com> and Nedelcho Stanev.
5634 *) Implement a countermeasure against a vulnerability recently found
5635 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
5636 before application data chunks to avoid the use of known IVs
5637 with data potentially chosen by the attacker.
5640 *) Fix length checks in ssl3_get_client_hello().
5643 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
5644 to prevent ssl3_read_internal() from incorrectly assuming that
5645 ssl3_read_bytes() found application data while handshake
5646 processing was enabled when in fact s->s3->in_read_app_data was
5647 merely automatically cleared during the initial handshake.
5648 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
5650 *) Fix object definitions for Private and Enterprise: they were not
5651 recognized in their shortname (=lowercase) representation. Extend
5652 obj_dat.pl to issue an error when using undefined keywords instead
5653 of silently ignoring the problem (Svenning Sorensen
5654 <sss@sss.dnsalias.net>).
5657 *) Fix DH_generate_parameters() so that it works for 'non-standard'
5658 generators, i.e. generators other than 2 and 5. (Previously, the
5659 code did not properly initialise the 'add' and 'rem' values to
5660 BN_generate_prime().)
5662 In the new general case, we do not insist that 'generator' is
5663 actually a primitive root: This requirement is rather pointless;
5664 a generator of the order-q subgroup is just as good, if not
5668 *) Map new X509 verification errors to alerts. Discovered and submitted by
5669 Tom Wu <tom@arcot.com>.
5672 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
5673 returning non-zero before the data has been completely received
5674 when using non-blocking I/O.
5675 [Bodo Moeller; problem pointed out by John Hughes]
5677 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
5678 [Ben Laurie, Lutz Jaenicke]
5680 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
5681 Yoram Zahavi <YoramZ@gilian.com>).
5684 *) Add information about CygWin 1.3 and on, and preserve proper
5685 configuration for the versions before that.
5686 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5688 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
5689 check whether we deal with a copy of a session and do not delete from
5690 the cache in this case. Problem reported by "Izhar Shoshani Levi"
5691 <izhar@checkpoint.com>.
5694 *) Do not store session data into the internal session cache, if it
5695 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
5696 flag is set). Proposed by Aslam <aslam@funk.com>.
5699 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
5703 *) [In 0.9.6d-engine release:]
5704 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5705 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5707 *) Add the configuration target linux-s390x.
5708 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
5710 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
5711 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
5712 variable as an indication that a ClientHello message has been
5713 received. As the flag value will be lost between multiple
5714 invocations of ssl3_accept when using non-blocking I/O, the
5715 function may not be aware that a handshake has actually taken
5716 place, thus preventing a new session from being added to the
5719 To avoid this problem, we now set s->new_session to 2 instead of
5720 using a local variable.
5721 [Lutz Jaenicke, Bodo Moeller]
5723 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
5724 if the SSL_R_LENGTH_MISMATCH error is detected.
5725 [Geoff Thorpe, Bodo Moeller]
5727 *) New 'shared_ldflag' column in Configure platform table.
5730 *) Fix EVP_CIPHER_mode macro.
5731 ["Dan S. Camper" <dan@bti.net>]
5733 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
5734 type, we must throw them away by setting rr->length to 0.
5735 [D P Chang <dpc@qualys.com>]
5737 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
5739 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
5740 <Dominikus.Scherkl@biodata.com>. (The previous implementation
5741 worked incorrectly for those cases where range = 10..._2 and
5742 3*range is two bits longer than range.)
5745 *) Only add signing time to PKCS7 structures if it is not already
5749 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
5750 OBJ_ld_ce should be OBJ_id_ce.
5751 Also some ip-pda OIDs in crypto/objects/objects.txt were
5752 incorrect (cf. RFC 3039).
5753 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
5755 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
5756 returns early because it has nothing to do.
5757 [Andy Schneider <andy.schneider@bjss.co.uk>]
5759 *) [In 0.9.6c-engine release:]
5760 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
5761 [Andy Schneider <andy.schneider@bjss.co.uk>]
5763 *) [In 0.9.6c-engine release:]
5764 Add support for Cryptographic Appliance's keyserver technology.
5765 (Use engine 'keyclient')
5766 [Cryptographic Appliances and Geoff Thorpe]
5768 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
5769 is called via tools/c89.sh because arguments have to be
5770 rearranged (all '-L' options must appear before the first object
5772 [Richard Shapiro <rshapiro@abinitio.com>]
5774 *) [In 0.9.6c-engine release:]
5775 Add support for Broadcom crypto accelerator cards, backported
5777 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
5779 *) [In 0.9.6c-engine release:]
5780 Add support for SureWare crypto accelerator cards from
5781 Baltimore Technologies. (Use engine 'sureware')
5782 [Baltimore Technologies and Mark Cox]
5784 *) [In 0.9.6c-engine release:]
5785 Add support for crypto accelerator cards from Accelerated
5786 Encryption Processing, www.aep.ie. (Use engine 'aep')
5787 [AEP Inc. and Mark Cox]
5789 *) Add a configuration entry for gcc on UnixWare.
5790 [Gary Benson <gbenson@redhat.com>]
5792 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
5793 messages are stored in a single piece (fixed-length part and
5794 variable-length part combined) and fix various bugs found on the way.
5797 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
5798 instead. BIO_gethostbyname() does not know what timeouts are
5799 appropriate, so entries would stay in cache even when they have
5801 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
5803 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
5804 faced with a pathologically small ClientHello fragment that does
5805 not contain client_version: Instead of aborting with an error,
5806 simply choose the highest available protocol version (i.e.,
5807 TLS 1.0 unless it is disabled). In practice, ClientHello
5808 messages are never sent like this, but this change gives us
5809 strictly correct behaviour at least for TLS.
5812 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
5813 never resets s->method to s->ctx->method when called from within
5814 one of the SSL handshake functions.
5815 [Bodo Moeller; problem pointed out by Niko Baric]
5817 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
5818 (sent using the client's version number) if client_version is
5819 smaller than the protocol version in use. Also change
5820 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
5821 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
5822 the client will at least see that alert.
5825 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
5829 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
5830 client receives HelloRequest while in a handshake.
5831 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
5833 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
5834 should end in 'break', not 'goto end' which circuments various
5835 cleanups done in state SSL_ST_OK. But session related stuff
5836 must be disabled for SSL_ST_OK in the case that we just sent a
5839 Also avoid some overhead by not calling ssl_init_wbio_buffer()
5840 before just sending a HelloRequest.
5841 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
5843 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
5844 reveal whether illegal block cipher padding was found or a MAC
5845 verification error occured. (Neither SSLerr() codes nor alerts
5846 are directly visible to potential attackers, but the information
5847 may leak via logfiles.)
5849 Similar changes are not required for the SSL 2.0 implementation
5850 because the number of padding bytes is sent in clear for SSL 2.0,
5851 and the extra bytes are just ignored. However ssl/s2_pkt.c
5852 failed to verify that the purported number of padding bytes is in
5856 *) Add OpenUNIX-8 support including shared libraries
5857 (Boyd Lynn Gerber <gerberb@zenez.com>).
5860 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
5861 'wristwatch attack' using huge encoding parameters (cf.
5862 James H. Manger's CRYPTO 2001 paper). Note that the
5863 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
5864 encoding parameters and hence was not vulnerable.
5867 *) BN_sqr() bug fix.
5868 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
5870 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
5871 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
5872 followed by modular reduction.
5873 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
5875 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
5876 equivalent based on BN_pseudo_rand() instead of BN_rand().
5879 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
5880 This function was broken, as the check for a new client hello message
5881 to handle SGC did not allow these large messages.
5882 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
5885 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
5888 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
5889 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
5892 *) Rework the configuration and shared library support for Tru64 Unix.
5893 The configuration part makes use of modern compiler features and
5894 still retains old compiler behavior for those that run older versions
5895 of the OS. The shared library support part includes a variant that
5896 uses the RPATH feature, and is available through the special
5897 configuration target "alpha-cc-rpath", which will never be selected
5899 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
5901 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
5902 with the same message size as in ssl3_get_certificate_request().
5903 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
5904 messages might inadvertently be reject as too long.
5905 [Petr Lampa <lampa@fee.vutbr.cz>]
5907 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
5910 *) Modified SSL library such that the verify_callback that has been set
5911 specificly for an SSL object with SSL_set_verify() is actually being
5912 used. Before the change, a verify_callback set with this function was
5913 ignored and the verify_callback() set in the SSL_CTX at the time of
5914 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
5915 to allow the necessary settings.
5918 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
5919 explicitly to NULL, as at least on Solaris 8 this seems not always to be
5920 done automatically (in contradiction to the requirements of the C
5921 standard). This made problems when used from OpenSSH.
5924 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
5925 dh->length and always used
5927 BN_rand_range(priv_key, dh->p).
5929 BN_rand_range() is not necessary for Diffie-Hellman, and this
5930 specific range makes Diffie-Hellman unnecessarily inefficient if
5931 dh->length (recommended exponent length) is much smaller than the
5932 length of dh->p. We could use BN_rand_range() if the order of
5933 the subgroup was stored in the DH structure, but we only have
5938 BN_rand(priv_key, l, ...)
5940 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
5946 RSA_eay_public_encrypt
5947 RSA_eay_private_decrypt
5948 RSA_eay_private_encrypt (signing)
5949 RSA_eay_public_decrypt (signature verification)
5951 (default implementations for RSA_public_encrypt,
5952 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
5953 always reject numbers >= n.
5956 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
5957 to synchronize access to 'locking_thread'. This is necessary on
5958 systems where access to 'locking_thread' (an 'unsigned long'
5959 variable) is not atomic.
5962 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
5963 *before* setting the 'crypto_lock_rand' flag. The previous code had
5964 a race condition if 0 is a valid thread ID.
5965 [Travis Vitek <vitek@roguewave.com>]
5967 *) Add support for shared libraries under Irix.
5968 [Albert Chin-A-Young <china@thewrittenword.com>]
5970 *) Add configuration option to build on Linux on both big-endian and
5972 [Ralf Baechle <ralf@uni-koblenz.de>]
5974 *) Add the possibility to create shared libraries on HP-UX.
5977 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
5979 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
5980 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
5981 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
5982 PRNG state recovery was possible based on the output of
5983 one PRNG request appropriately sized to gain knowledge on
5984 'md' followed by enough consecutive 1-byte PRNG requests
5985 to traverse all of 'state'.
5987 1. When updating 'md_local' (the current thread's copy of 'md')
5988 during PRNG output generation, hash all of the previous
5989 'md_local' value, not just the half used for PRNG output.
5991 2. Make the number of bytes from 'state' included into the hash
5992 independent from the number of PRNG bytes requested.
5994 The first measure alone would be sufficient to avoid
5995 Markku-Juhani's attack. (Actually it had never occurred
5996 to me that the half of 'md_local' used for chaining was the
5997 half from which PRNG output bytes were taken -- I had always
5998 assumed that the secret half would be used.) The second
5999 measure makes sure that additional data from 'state' is never
6000 mixed into 'md_local' in small portions; this heuristically
6001 further strengthens the PRNG.
6004 *) Fix crypto/bn/asm/mips3.s.
6007 *) When only the key is given to "enc", the IV is undefined. Print out
6008 an error message in this case.
6011 *) Handle special case when X509_NAME is empty in X509 printing routines.
6014 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
6015 positive and less than q.
6018 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
6019 used: it isn't thread safe and the add_lock_callback should handle
6021 [Paul Rose <Paul.Rose@bridge.com>]
6023 *) Verify that incoming data obeys the block size in
6024 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
6028 [Ulf Möller, Bodo Möller]
6030 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
6031 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
6032 when fixing the server behaviour for backwards-compatible 'client
6033 hello' messages. (Note that the attack is impractical against
6034 SSL 3.0 and TLS 1.0 anyway because length and version checking
6035 means that the probability of guessing a valid ciphertext is
6036 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
6039 Before 0.9.5, the countermeasure (hide the error by generating a
6040 random 'decryption result') did not work properly because
6041 ERR_clear_error() was missing, meaning that SSL_get_error() would
6042 detect the supposedly ignored error.
6044 Both problems are now fixed.
6047 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
6048 (previously it was 1024).
6051 *) Fix for compatibility mode trust settings: ignore trust settings
6052 unless some valid trust or reject settings are present.
6055 *) Fix for blowfish EVP: its a variable length cipher.
6058 *) Fix various bugs related to DSA S/MIME verification. Handle missing
6059 parameters in DSA public key structures and return an error in the
6060 DSA routines if parameters are absent.
6063 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
6064 in the current directory if neither $RANDFILE nor $HOME was set.
6065 RAND_file_name() in 0.9.6a returned NULL in this case. This has
6066 caused some confusion to Windows users who haven't defined $HOME.
6067 Thus RAND_file_name() is changed again: e_os.h can define a
6068 DEFAULT_HOME, which will be used if $HOME is not set.
6069 For Windows, we use "C:"; on other platforms, we still require
6070 environment variables.
6072 *) Move 'if (!initialized) RAND_poll()' into regions protected by
6073 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
6074 having multiple threads call RAND_poll() concurrently.
6077 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
6078 combination of a flag and a thread ID variable.
6079 Otherwise while one thread is in ssleay_rand_bytes (which sets the
6080 flag), *other* threads can enter ssleay_add_bytes without obeying
6081 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
6082 that they do not hold after the first thread unsets add_do_not_lock).
6085 *) Change bctest again: '-x' expressions are not available in all
6089 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
6091 *) Fix a couple of memory leaks in PKCS7_dataDecode()
6092 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
6094 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
6095 the default extension for executables, if any. Also, make the perl
6096 scripts that use symlink() to test if it really exists and use "cp"
6097 if it doesn't. All this made OpenSSL compilable and installable in
6101 *) Fix for asn1_GetSequence() for indefinite length constructed data.
6102 If SEQUENCE is length is indefinite just set c->slen to the total
6103 amount of data available.
6104 [Steve Henson, reported by shige@FreeBSD.org]
6105 [This change does not apply to 0.9.7.]
6107 *) Change bctest to avoid here-documents inside command substitution
6108 (workaround for FreeBSD /bin/sh bug).
6109 For compatibility with Ultrix, avoid shell functions (introduced
6110 in the bctest version that searches along $PATH).
6113 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6114 with des_encrypt() defined on some operating systems, like Solaris
6118 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6119 On the Importance of Eliminating Errors in Cryptographic
6120 Computations, J. Cryptology 14 (2001) 2, 101-119,
6121 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6124 *) MIPS assembler BIGNUM division bug fix.
6127 *) Disabled incorrect Alpha assembler code.
6130 *) Fix PKCS#7 decode routines so they correctly update the length
6131 after reading an EOC for the EXPLICIT tag.
6133 [This change does not apply to 0.9.7.]
6135 *) Fix bug in PKCS#12 key generation routines. This was triggered
6136 if a 3DES key was generated with a 0 initial byte. Include
6137 PKCS12_BROKEN_KEYGEN compilation option to retain the old
6138 (but broken) behaviour.
6141 *) Enhance bctest to search for a working bc along $PATH and print
6143 [Tim Rice <tim@multitalents.net> via Richard Levitte]
6145 *) Fix memory leaks in err.c: free err_data string if necessary;
6146 don't write to the wrong index in ERR_set_error_data.
6149 *) Implement ssl23_peek (analogous to ssl23_read), which previously
6153 *) Replace rdtsc with _emit statements for VC++ version 5.
6154 [Jeremy Cooper <jeremy@baymoo.org>]
6156 *) Make it possible to reuse SSLv2 sessions.
6159 *) In copy_email() check for >= 0 as a return value for
6160 X509_NAME_get_index_by_NID() since 0 is a valid index.
6161 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
6163 *) Avoid coredump with unsupported or invalid public keys by checking if
6164 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
6165 PKCS7_verify() fails with non detached data.
6168 *) Don't use getenv in library functions when run as setuid/setgid.
6169 New function OPENSSL_issetugid().
6172 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
6173 due to incorrect handling of multi-threading:
6175 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
6177 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
6179 3. Count how many times MemCheck_off() has been called so that
6180 nested use can be treated correctly. This also avoids
6181 inband-signalling in the previous code (which relied on the
6182 assumption that thread ID 0 is impossible).
6185 *) Add "-rand" option also to s_client and s_server.
6188 *) Fix CPU detection on Irix 6.x.
6189 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
6190 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6192 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
6195 [This change does not apply to 0.9.7.]
6197 *) Use the cached encoding of an X509_NAME structure rather than
6198 copying it. This is apparently the reason for the libsafe "errors"
6199 but the code is actually correct.
6202 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
6203 Bleichenbacher's DSA attack.
6204 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
6205 to be set and top=0 forces the highest bit to be set; top=-1 is new
6206 and leaves the highest bit random.
6207 [Ulf Moeller, Bodo Moeller]
6209 *) In the NCONF_...-based implementations for CONF_... queries
6210 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
6211 a temporary CONF structure with the data component set to NULL
6212 (which gives segmentation faults in lh_retrieve).
6213 Instead, use NULL for the CONF pointer in CONF_get_string and
6214 CONF_get_number (which may use environment variables) and directly
6215 return NULL from CONF_get_section.
6218 *) Fix potential buffer overrun for EBCDIC.
6221 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
6222 keyUsage if basicConstraints absent for a CA.
6225 *) Make SMIME_write_PKCS7() write mail header values with a format that
6226 is more generally accepted (no spaces before the semicolon), since
6227 some programs can't parse those values properly otherwise. Also make
6228 sure BIO's that break lines after each write do not create invalid
6232 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
6233 macros previously used would not encode an empty SEQUENCE OF
6234 and break the signature.
6236 [This change does not apply to 0.9.7.]
6238 *) Zero the premaster secret after deriving the master secret in
6242 *) Add some EVP_add_digest_alias registrations (as found in
6243 OpenSSL_add_all_digests()) to SSL_library_init()
6244 aka OpenSSL_add_ssl_algorithms(). This provides improved
6245 compatibility with peers using X.509 certificates
6246 with unconventional AlgorithmIdentifier OIDs.
6249 *) Fix for Irix with NO_ASM.
6250 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6252 *) ./config script fixes.
6253 [Ulf Moeller, Richard Levitte]
6255 *) Fix 'openssl passwd -1'.
6258 *) Change PKCS12_key_gen_asc() so it can cope with non null
6259 terminated strings whose length is passed in the passlen
6260 parameter, for example from PEM callbacks. This was done
6261 by adding an extra length parameter to asc2uni().
6262 [Steve Henson, reported by <oddissey@samsung.co.kr>]
6264 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
6265 call failed, free the DSA structure.
6268 *) Fix to uni2asc() to cope with zero length Unicode strings.
6269 These are present in some PKCS#12 files.
6272 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
6273 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
6274 when writing a 32767 byte record.
6275 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
6277 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
6278 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
6280 (RSA objects have a reference count access to which is protected
6281 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
6282 so they are meant to be shared between threads.)
6283 [Bodo Moeller, Geoff Thorpe; original patch submitted by
6284 "Reddie, Steven" <Steven.Reddie@ca.com>]
6286 *) Fix a deadlock in CRYPTO_mem_leaks().
6289 *) Use better test patterns in bntest.
6292 *) rand_win.c fix for Borland C.
6295 *) BN_rshift bugfix for n == 0.
6298 *) Add a 'bctest' script that checks for some known 'bc' bugs
6299 so that 'make test' does not abort just because 'bc' is broken.
6302 *) Store verify_result within SSL_SESSION also for client side to
6303 avoid potential security hole. (Re-used sessions on the client side
6304 always resulted in verify_result==X509_V_OK, not using the original
6305 result of the server certificate verification.)
6308 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
6309 SSL3_RT_APPLICATION_DATA, return 0.
6310 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
6314 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
6315 releases, have been re-implemented by renaming the previous
6316 implementations of ssl2_read and ssl3_read to ssl2_read_internal
6317 and ssl3_read_internal, respectively, and adding 'peek' parameters
6318 to them. The new ssl[23]_{read,peek} functions are calls to
6319 ssl[23]_read_internal with the 'peek' flag set appropriately.
6320 A 'peek' parameter has also been added to ssl3_read_bytes, which
6321 does the actual work for ssl3_read_internal.
6324 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
6325 the method-specific "init()" handler. Also clean up ex_data after
6326 calling the method-specific "finish()" handler. Previously, this was
6327 happening the other way round.
6330 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
6331 The previous value, 12, was not always sufficient for BN_mod_exp().
6334 *) Make sure that shared libraries get the internal name engine with
6335 the full version number and not just 0. This should mark the
6336 shared libraries as not backward compatible. Of course, this should
6337 be changed again when we can guarantee backward binary compatibility.
6340 *) Fix typo in get_cert_by_subject() in by_dir.c
6341 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
6343 *) Rework the system to generate shared libraries:
6345 - Make note of the expected extension for the shared libraries and
6346 if there is a need for symbolic links from for example libcrypto.so.0
6347 to libcrypto.so.0.9.7. There is extended info in Configure for
6350 - Make as few rebuilds of the shared libraries as possible.
6352 - Still avoid linking the OpenSSL programs with the shared libraries.
6354 - When installing, install the shared libraries separately from the
6358 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
6360 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
6361 and not in SSL_clear because the latter is also used by the
6362 accept/connect functions; previously, the settings made by
6363 SSL_set_read_ahead would be lost during the handshake.
6364 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
6366 *) Correct util/mkdef.pl to be selective about disabled algorithms.
6367 Previously, it would create entries for disableed algorithms no
6371 *) Added several new manual pages for SSL_* function.
6374 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
6376 *) In ssl23_get_client_hello, generate an error message when faced
6377 with an initial SSL 3.0/TLS record that is too small to contain the
6378 first two bytes of the ClientHello message, i.e. client_version.
6379 (Note that this is a pathologic case that probably has never happened
6380 in real life.) The previous approach was to use the version number
6381 from the record header as a substitute; but our protocol choice
6382 should not depend on that one because it is not authenticated
6383 by the Finished messages.
6386 *) More robust randomness gathering functions for Windows.
6387 [Jeffrey Altman <jaltman@columbia.edu>]
6389 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
6390 not set then we don't setup the error code for issuer check errors
6391 to avoid possibly overwriting other errors which the callback does
6392 handle. If an application does set the flag then we assume it knows
6393 what it is doing and can handle the new informational codes
6397 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
6398 a general "ANY" type, as such it should be able to decode anything
6399 including tagged types. However it didn't check the class so it would
6400 wrongly interpret tagged types in the same way as their universal
6401 counterpart and unknown types were just rejected. Changed so that the
6402 tagged and unknown types are handled in the same way as a SEQUENCE:
6403 that is the encoding is stored intact. There is also a new type
6404 "V_ASN1_OTHER" which is used when the class is not universal, in this
6405 case we have no idea what the actual type is so we just lump them all
6409 *) On VMS, stdout may very well lead to a file that is written to
6410 in a record-oriented fashion. That means that every write() will
6411 write a separate record, which will be read separately by the
6412 programs trying to read from it. This can be very confusing.
6414 The solution is to put a BIO filter in the way that will buffer
6415 text until a linefeed is reached, and then write everything a
6416 line at a time, so every record written will be an actual line,
6417 not chunks of lines and not (usually doesn't happen, but I've
6418 seen it once) several lines in one record. BIO_f_linebuffer() is
6421 Currently, it's a VMS-only method, because that's where it has
6422 been tested well enough.
6425 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
6426 it can return incorrect results.
6427 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
6428 but it was in 0.9.6-beta[12].)
6431 *) Disable the check for content being present when verifying detached
6432 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
6433 include zero length content when signing messages.
6436 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
6437 BIO_ctrl (for BIO pairs).
6440 *) Add DSO method for VMS.
6443 *) Bug fix: Montgomery multiplication could produce results with the
6447 *) Add RPM specification openssl.spec and modify it to build three
6448 packages. The default package contains applications, application
6449 documentation and run-time libraries. The devel package contains
6450 include files, static libraries and function documentation. The
6451 doc package contains the contents of the doc directory. The original
6452 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
6455 *) Add a large number of documentation files for many SSL routines.
6456 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6458 *) Add a configuration entry for Sony News 4.
6459 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
6461 *) Don't set the two most significant bits to one when generating a
6462 random number < q in the DSA library.
6465 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
6466 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
6467 the underlying transport is blocking) if a handshake took place.
6468 (The default behaviour is needed by applications such as s_client
6469 and s_server that use select() to determine when to use SSL_read;
6470 but for applications that know in advance when to expect data, it
6471 just makes things more complicated.)
6474 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
6478 *) Add a few more EBCDIC conditionals that make `req' and `x509'
6479 work better on such systems.
6480 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6482 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
6483 Update PKCS12_parse() so it copies the friendlyName and the
6484 keyid to the certificates aux info.
6487 *) Fix bug in PKCS7_verify() which caused an infinite loop
6488 if there was more than one signature.
6489 [Sven Uszpelkat <su@celocom.de>]
6491 *) Major change in util/mkdef.pl to include extra information
6492 about each symbol, as well as presentig variables as well
6493 as functions. This change means that there's n more need
6494 to rebuild the .num files when some algorithms are excluded.
6497 *) Allow the verify time to be set by an application,
6498 rather than always using the current time.
6501 *) Phase 2 verify code reorganisation. The certificate
6502 verify code now looks up an issuer certificate by a
6503 number of criteria: subject name, authority key id
6504 and key usage. It also verifies self signed certificates
6505 by the same criteria. The main comparison function is
6506 X509_check_issued() which performs these checks.
6508 Lot of changes were necessary in order to support this
6509 without completely rewriting the lookup code.
6511 Authority and subject key identifier are now cached.
6513 The LHASH 'certs' is X509_STORE has now been replaced
6514 by a STACK_OF(X509_OBJECT). This is mainly because an
6515 LHASH can't store or retrieve multiple objects with
6516 the same hash value.
6518 As a result various functions (which were all internal
6519 use only) have changed to handle the new X509_STORE
6520 structure. This will break anything that messed round
6521 with X509_STORE internally.
6523 The functions X509_STORE_add_cert() now checks for an
6524 exact match, rather than just subject name.
6526 The X509_STORE API doesn't directly support the retrieval
6527 of multiple certificates matching a given criteria, however
6528 this can be worked round by performing a lookup first
6529 (which will fill the cache with candidate certificates)
6530 and then examining the cache for matches. This is probably
6531 the best we can do without throwing out X509_LOOKUP
6532 entirely (maybe later...).
6534 The X509_VERIFY_CTX structure has been enhanced considerably.
6536 All certificate lookup operations now go via a get_issuer()
6537 callback. Although this currently uses an X509_STORE it
6538 can be replaced by custom lookups. This is a simple way
6539 to bypass the X509_STORE hackery necessary to make this
6540 work and makes it possible to use more efficient techniques
6541 in future. A very simple version which uses a simple
6542 STACK for its trusted certificate store is also provided
6543 using X509_STORE_CTX_trusted_stack().
6545 The verify_cb() and verify() callbacks now have equivalents
6546 in the X509_STORE_CTX structure.
6548 X509_STORE_CTX also has a 'flags' field which can be used
6549 to customise the verify behaviour.
6552 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
6553 excludes S/MIME capabilities.
6556 *) When a certificate request is read in keep a copy of the
6557 original encoding of the signed data and use it when outputing
6558 again. Signatures then use the original encoding rather than
6559 a decoded, encoded version which may cause problems if the
6560 request is improperly encoded.
6563 *) For consistency with other BIO_puts implementations, call
6564 buffer_write(b, ...) directly in buffer_puts instead of calling
6567 In BIO_puts, increment b->num_write as in BIO_write.
6568 [Peter.Sylvester@EdelWeb.fr]
6570 *) Fix BN_mul_word for the case where the word is 0. (We have to use
6571 BN_zero, we may not return a BIGNUM with an array consisting of
6575 *) Avoid calling abort() from within the library when problems are
6576 detected, except if preprocessor symbols have been defined
6577 (such as REF_CHECK, BN_DEBUG etc.).
6580 *) New openssl application 'rsautl'. This utility can be
6581 used for low level RSA operations. DER public key
6582 BIO/fp routines also added.
6585 *) New Configure entry and patches for compiling on QNX 4.
6586 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
6588 *) A demo state-machine implementation was sponsored by
6589 Nuron (http://www.nuron.com/) and is now available in
6590 demos/state_machine.
6593 *) New options added to the 'dgst' utility for signature
6594 generation and verification.
6597 *) Unrecognized PKCS#7 content types are now handled via a
6598 catch all ASN1_TYPE structure. This allows unsupported
6599 types to be stored as a "blob" and an application can
6600 encode and decode it manually.
6603 *) Fix various signed/unsigned issues to make a_strex.c
6605 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
6607 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
6608 length if passed a buffer. ASN1_INTEGER_to_BN failed
6609 if passed a NULL BN and its argument was negative.
6610 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
6612 *) Modification to PKCS#7 encoding routines to output definite
6613 length encoding. Since currently the whole structures are in
6614 memory there's not real point in using indefinite length
6615 constructed encoding. However if OpenSSL is compiled with
6616 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
6619 *) Added BIO_vprintf() and BIO_vsnprintf().
6622 *) Added more prefixes to parse for in the the strings written
6623 through a logging bio, to cover all the levels that are available
6624 through syslog. The prefixes are now:
6626 PANIC, EMERG, EMR => LOG_EMERG
6627 ALERT, ALR => LOG_ALERT
6628 CRIT, CRI => LOG_CRIT
6629 ERROR, ERR => LOG_ERR
6630 WARNING, WARN, WAR => LOG_WARNING
6631 NOTICE, NOTE, NOT => LOG_NOTICE
6632 INFO, INF => LOG_INFO
6633 DEBUG, DBG => LOG_DEBUG
6635 and as before, if none of those prefixes are present at the
6636 beginning of the string, LOG_ERR is chosen.
6638 On Win32, the LOG_* levels are mapped according to this:
6640 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
6641 LOG_WARNING => EVENTLOG_WARNING_TYPE
6642 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
6646 *) Made it possible to reconfigure with just the configuration
6647 argument "reconf" or "reconfigure". The command line arguments
6648 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
6649 and are retrieved from there when reconfiguring.
6653 [Assar Westerlund <assar@sics.se>, Richard Levitte]
6655 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
6658 *) The obj_dat.pl script was messing up the sorting of object
6659 names. The reason was that it compared the quoted version
6660 of strings as a result "OCSP" > "OCSP Signing" because
6661 " > SPACE. Changed script to store unquoted versions of
6662 names and add quotes on output. It was also omitting some
6663 names from the lookup table if they were given a default
6664 value (that is if SN is missing it is given the same
6665 value as LN and vice versa), these are now added on the
6666 grounds that if an object has a name we should be able to
6667 look it up. Finally added warning output when duplicate
6668 short or long names are found.
6671 *) Changes needed for Tandem NSK.
6672 [Scott Uroff <scott@xypro.com>]
6674 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
6675 RSA_padding_check_SSLv23(), special padding was never detected
6676 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
6677 version rollback attacks was not effective.
6679 In s23_clnt.c, don't use special rollback-attack detection padding
6680 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
6681 client; similarly, in s23_srvr.c, don't do the rollback check if
6682 SSL 2.0 is the only protocol enabled in the server.
6685 *) Make it possible to get hexdumps of unprintable data with 'openssl
6686 asn1parse'. By implication, the functions ASN1_parse_dump() and
6687 BIO_dump_indent() are added.
6690 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
6691 these print out strings and name structures based on various
6692 flags including RFC2253 support and proper handling of
6693 multibyte characters. Added options to the 'x509' utility
6694 to allow the various flags to be set.
6697 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
6698 Also change the functions X509_cmp_current_time() and
6699 X509_gmtime_adj() work with an ASN1_TIME structure,
6700 this will enable certificates using GeneralizedTime in validity
6701 dates to be checked.
6704 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
6705 negative public key encodings) on by default,
6706 NO_NEG_PUBKEY_BUG can be set to disable it.
6709 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
6710 content octets. An i2c_ASN1_OBJECT is unnecessary because
6711 the encoding can be trivially obtained from the structure.
6714 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
6715 not read locks (CRYPTO_r_[un]lock).
6718 *) A first attempt at creating official support for shared
6719 libraries through configuration. I've kept it so the
6720 default is static libraries only, and the OpenSSL programs
6721 are always statically linked for now, but there are
6722 preparations for dynamic linking in place.
6723 This has been tested on Linux and Tru64.
6726 *) Randomness polling function for Win9x, as described in:
6727 Peter Gutmann, Software Generation of Practically Strong
6731 *) Fix so PRNG is seeded in req if using an already existing
6735 *) New options to smime application. -inform and -outform
6736 allow alternative formats for the S/MIME message including
6737 PEM and DER. The -content option allows the content to be
6738 specified separately. This should allow things like Netscape
6739 form signing output easier to verify.
6742 *) Fix the ASN1 encoding of tags using the 'long form'.
6745 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
6746 STRING types. These convert content octets to and from the
6747 underlying type. The actual tag and length octets are
6748 already assumed to have been read in and checked. These
6749 are needed because all other string types have virtually
6750 identical handling apart from the tag. By having versions
6751 of the ASN1 functions that just operate on content octets
6752 IMPLICIT tagging can be handled properly. It also allows
6753 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
6754 and ASN1_INTEGER are identical apart from the tag.
6757 *) Change the handling of OID objects as follows:
6759 - New object identifiers are inserted in objects.txt, following
6760 the syntax given in objects.README.
6761 - objects.pl is used to process obj_mac.num and create a new
6763 - obj_dat.pl is used to create a new obj_dat.h, using the data in
6766 This is currently kind of a hack, and the perl code in objects.pl
6767 isn't very elegant, but it works as I intended. The simplest way
6768 to check that it worked correctly is to look in obj_dat.h and
6769 check the array nid_objs and make sure the objects haven't moved
6770 around (this is important!). Additions are OK, as well as
6771 consistent name changes.
6774 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
6777 *) Addition of the command line parameter '-rand file' to 'openssl req'.
6778 The given file adds to whatever has already been seeded into the
6779 random pool through the RANDFILE configuration file option or
6780 environment variable, or the default random state file.
6783 *) mkstack.pl now sorts each macro group into lexical order.
6784 Previously the output order depended on the order the files
6785 appeared in the directory, resulting in needless rewriting
6789 *) Patches to make OpenSSL compile under Win32 again. Mostly
6790 work arounds for the VC++ problem that it treats func() as
6791 func(void). Also stripped out the parts of mkdef.pl that
6792 added extra typesafe functions: these no longer exist.
6795 *) Reorganisation of the stack code. The macros are now all
6796 collected in safestack.h . Each macro is defined in terms of
6797 a "stack macro" of the form SKM_<name>(type, a, b). The
6798 DEBUG_SAFESTACK is now handled in terms of function casts,
6799 this has the advantage of retaining type safety without the
6800 use of additional functions. If DEBUG_SAFESTACK is not defined
6801 then the non typesafe macros are used instead. Also modified the
6802 mkstack.pl script to handle the new form. Needs testing to see
6803 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
6804 the default if no major problems. Similar behaviour for ASN1_SET_OF
6805 and PKCS12_STACK_OF.
6808 *) When some versions of IIS use the 'NET' form of private key the
6809 key derivation algorithm is different. Normally MD5(password) is
6810 used as a 128 bit RC4 key. In the modified case
6811 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
6812 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
6813 as the old Netscape_RSA functions except they have an additional
6814 'sgckey' parameter which uses the modified algorithm. Also added
6815 an -sgckey command line option to the rsa utility. Thanks to
6816 Adrian Peck <bertie@ncipher.com> for posting details of the modified
6817 algorithm to openssl-dev.
6820 *) The evp_local.h macros were using 'c.##kname' which resulted in
6821 invalid expansion on some systems (SCO 5.0.5 for example).
6822 Corrected to 'c.kname'.
6823 [Phillip Porch <root@theporch.com>]
6825 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
6826 a STACK of email addresses from a certificate or request, these look
6827 in the subject name and the subject alternative name extensions and
6828 omit any duplicate addresses.
6831 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
6832 This makes DSA verification about 2 % faster.
6835 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
6836 (meaning that now 2^5 values will be precomputed, which is only 4 KB
6837 plus overhead for 1024 bit moduli).
6838 This makes exponentiations about 0.5 % faster for 1024 bit
6839 exponents (as measured by "openssl speed rsa2048").
6842 *) Rename memory handling macros to avoid conflicts with other
6844 Malloc => OPENSSL_malloc
6845 Malloc_locked => OPENSSL_malloc_locked
6846 Realloc => OPENSSL_realloc
6847 Free => OPENSSL_free
6850 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
6851 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
6854 *) CygWin32 support.
6855 [John Jarvie <jjarvie@newsguy.com>]
6857 *) The type-safe stack code has been rejigged. It is now only compiled
6858 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
6859 by default all type-specific stack functions are "#define"d back to
6860 standard stack functions. This results in more streamlined output
6861 but retains the type-safety checking possibilities of the original
6865 *) The STACK code has been cleaned up, and certain type declarations
6866 that didn't make a lot of sense have been brought in line. This has
6867 also involved a cleanup of sorts in safestack.h to more correctly
6868 map type-safe stack functions onto their plain stack counterparts.
6869 This work has also resulted in a variety of "const"ifications of
6870 lots of the code, especially "_cmp" operations which should normally
6871 be prototyped with "const" parameters anyway.
6874 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
6875 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
6876 (The PRNG state consists of two parts, the large pool 'state' and 'md',
6877 where all of 'md' is used each time the PRNG is used, but 'state'
6878 is used only indexed by a cyclic counter. As entropy may not be
6879 well distributed from the beginning, 'md' is important as a
6880 chaining variable. However, the output function chains only half
6881 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
6882 all of 'md', and seeding with STATE_SIZE dummy bytes will result
6883 in all of 'state' being rewritten, with the new values depending
6884 on virtually all of 'md'. This overcomes the 80 bit limitation.)
6887 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
6888 the handshake is continued after ssl_verify_cert_chain();
6889 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
6890 can lead to 'unexplainable' connection aborts later.
6891 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
6893 *) Major EVP API cipher revision.
6894 Add hooks for extra EVP features. This allows various cipher
6895 parameters to be set in the EVP interface. Support added for variable
6896 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
6897 setting of RC2 and RC5 parameters.
6899 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
6902 Remove lots of duplicated code from the EVP library. For example *every*
6903 cipher init() function handles the 'iv' in the same way according to the
6904 cipher mode. They also all do nothing if the 'key' parameter is NULL and
6905 for CFB and OFB modes they zero ctx->num.
6907 New functionality allows removal of S/MIME code RC2 hack.
6909 Most of the routines have the same form and so can be declared in terms
6912 By shifting this to the top level EVP_CipherInit() it can be removed from
6913 all individual ciphers. If the cipher wants to handle IVs or keys
6914 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
6917 Change lots of functions like EVP_EncryptUpdate() to now return a
6918 value: although software versions of the algorithms cannot fail
6919 any installed hardware versions can.
6922 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
6923 this option is set, tolerate broken clients that send the negotiated
6924 protocol version number instead of the requested protocol version
6928 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
6929 i.e. non-zero for export ciphersuites, zero otherwise.
6930 Previous versions had this flag inverted, inconsistent with
6931 rsa_tmp_cb (..._TMP_RSA_CB).
6932 [Bodo Moeller; problem reported by Amit Chopra]
6934 *) Add missing DSA library text string. Work around for some IIS
6935 key files with invalid SEQUENCE encoding.
6938 *) Add a document (doc/standards.txt) that list all kinds of standards
6939 and so on that are implemented in OpenSSL.
6942 *) Enhance c_rehash script. Old version would mishandle certificates
6943 with the same subject name hash and wouldn't handle CRLs at all.
6944 Added -fingerprint option to crl utility, to support new c_rehash
6948 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
6951 *) Fix for SSL server purpose checking. Server checking was
6952 rejecting certificates which had extended key usage present
6953 but no ssl client purpose.
6954 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
6956 *) Make PKCS#12 code work with no password. The PKCS#12 spec
6957 is a little unclear about how a blank password is handled.
6958 Since the password in encoded as a BMPString with terminating
6959 double NULL a zero length password would end up as just the
6960 double NULL. However no password at all is different and is
6961 handled differently in the PKCS#12 key generation code. NS
6962 treats a blank password as zero length. MSIE treats it as no
6963 password on export: but it will try both on import. We now do
6964 the same: PKCS12_parse() tries zero length and no password if
6965 the password is set to "" or NULL (NULL is now a valid password:
6966 it wasn't before) as does the pkcs12 application.
6969 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
6970 perror when PEM_read_bio_X509_REQ fails, the error message must
6971 be obtained from the error queue.
6974 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
6975 it in ERR_remove_state if appropriate, and change ERR_get_state
6976 accordingly to avoid race conditions (this is necessary because
6977 thread_hash is no longer constant once set).
6980 *) Bugfix for linux-elf makefile.one.
6983 *) RSA_get_default_method() will now cause a default
6984 RSA_METHOD to be chosen if one doesn't exist already.
6985 Previously this was only set during a call to RSA_new()
6986 or RSA_new_method(NULL) meaning it was possible for
6987 RSA_get_default_method() to return NULL.
6990 *) Added native name translation to the existing DSO code
6991 that will convert (if the flag to do so is set) filenames
6992 that are sufficiently small and have no path information
6993 into a canonical native form. Eg. "blah" converted to
6994 "libblah.so" or "blah.dll" etc.
6997 *) New function ERR_error_string_n(e, buf, len) which is like
6998 ERR_error_string(e, buf), but writes at most 'len' bytes
6999 including the 0 terminator. For ERR_error_string_n, 'buf'
7001 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
7003 *) CONF library reworked to become more general. A new CONF
7004 configuration file reader "class" is implemented as well as a
7005 new functions (NCONF_*, for "New CONF") to handle it. The now
7006 old CONF_* functions are still there, but are reimplemented to
7007 work in terms of the new functions. Also, a set of functions
7008 to handle the internal storage of the configuration data is
7009 provided to make it easier to write new configuration file
7010 reader "classes" (I can definitely see something reading a
7011 configuration file in XML format, for example), called _CONF_*,
7012 or "the configuration storage API"...
7014 The new configuration file reading functions are:
7016 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
7017 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
7019 NCONF_default, NCONF_WIN32
7021 NCONF_dump_fp, NCONF_dump_bio
7023 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
7024 NCONF_new creates a new CONF object. This works in the same way
7025 as other interfaces in OpenSSL, like the BIO interface.
7026 NCONF_dump_* dump the internal storage of the configuration file,
7027 which is useful for debugging. All other functions take the same
7028 arguments as the old CONF_* functions wth the exception of the
7029 first that must be a `CONF *' instead of a `LHASH *'.
7031 To make it easer to use the new classes with the old CONF_* functions,
7032 the function CONF_set_default_method is provided.
7035 *) Add '-tls1' option to 'openssl ciphers', which was already
7036 mentioned in the documentation but had not been implemented.
7037 (This option is not yet really useful because even the additional
7038 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
7041 *) Initial DSO code added into libcrypto for letting OpenSSL (and
7042 OpenSSL-based applications) load shared libraries and bind to
7043 them in a portable way.
7044 [Geoff Thorpe, with contributions from Richard Levitte]
7046 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
7048 *) Make sure _lrotl and _lrotr are only used with MSVC.
7050 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
7051 (the default implementation of RAND_status).
7053 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
7054 to '-clrext' (= clear extensions), as intended and documented.
7055 [Bodo Moeller; inconsistency pointed out by Michael Attili
7058 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
7059 was larger than the MD block size.
7060 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
7062 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
7063 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
7064 using the passed key: if the passed key was a private key the result
7065 of X509_print(), for example, would be to print out all the private key
7069 *) des_quad_cksum() byte order bug fix.
7070 [Ulf Möller, using the problem description in krb4-0.9.7, where
7071 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
7073 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
7075 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
7077 *) For easily testing in shell scripts whether some command
7078 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
7079 returns with exit code 0 iff no command of the given name is available.
7080 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
7081 the output goes to stdout and nothing is printed to stderr.
7082 Additional arguments are always ignored.
7084 Since for each cipher there is a command of the same name,
7085 the 'no-cipher' compilation switches can be tested this way.
7087 ('openssl no-XXX' is not able to detect pseudo-commands such
7088 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
7091 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
7094 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
7095 is set; it will be thrown away anyway because each handshake creates
7097 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
7098 to parameters -- in previous versions (since OpenSSL 0.9.3) the
7099 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
7100 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7103 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7104 'Q' and 'R' lose their special meanings (quit/renegotiate).
7105 This is part of what -quiet does; unlike -quiet, -ign_eof
7106 does not suppress any output.
7109 *) Add compatibility options to the purpose and trust code. The
7110 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7111 accepts a certificate or CA, this was the previous behaviour,
7112 with all the associated security issues.
7114 X509_TRUST_COMPAT is the old trust behaviour: only and
7115 automatically trust self signed roots in certificate store. A
7116 new trust setting X509_TRUST_DEFAULT is used to specify that
7117 a purpose has no associated trust setting and it should instead
7118 use the value in the default purpose.
7121 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7122 and fix a memory leak.
7125 *) In util/mkerr.pl (which implements 'make errors'), preserve
7126 reason strings from the previous version of the .c file, as
7127 the default to have only downcase letters (and digits) in
7128 automatically generated reasons codes is not always appropriate.
7131 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7132 using strerror. Previously, ERR_reason_error_string() returned
7133 library names as reason strings for SYSerr; but SYSerr is a special
7134 case where small numbers are errno values, not library numbers.
7137 *) Add '-dsaparam' option to 'openssl dhparam' application. This
7138 converts DSA parameters into DH parameters. (When creating parameters,
7139 DSA_generate_parameters is used.)
7142 *) Include 'length' (recommended exponent length) in C code generated
7143 by 'openssl dhparam -C'.
7146 *) The second argument to set_label in perlasm was already being used
7147 so couldn't be used as a "file scope" flag. Moved to third argument
7151 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
7152 instead of RAND_bytes for encryption IVs and salts.
7155 *) Include RAND_status() into RAND_METHOD instead of implementing
7156 it only for md_rand.c Otherwise replacing the PRNG by calling
7157 RAND_set_rand_method would be impossible.
7160 *) Don't let DSA_generate_key() enter an infinite loop if the random
7161 number generation fails.
7164 *) New 'rand' application for creating pseudo-random output.
7167 *) Added configuration support for Linux/IA64
7168 [Rolf Haberrecker <rolf@suse.de>]
7170 *) Assembler module support for Mingw32.
7173 *) Shared library support for HPUX (in shlib/).
7174 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
7176 *) Shared library support for Solaris gcc.
7177 [Lutz Behnke <behnke@trustcenter.de>]
7179 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
7181 *) PKCS7_encrypt() was adding text MIME headers twice because they
7182 were added manually and by SMIME_crlf_copy().
7185 *) In bntest.c don't call BN_rand with zero bits argument.
7186 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
7188 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
7189 case was implemented. This caused BN_div_recp() to fail occasionally.
7192 *) Add an optional second argument to the set_label() in the perl
7193 assembly language builder. If this argument exists and is set
7194 to 1 it signals that the assembler should use a symbol whose
7195 scope is the entire file, not just the current function. This
7196 is needed with MASM which uses the format label:: for this scope.
7197 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
7199 *) Change the ASN1 types so they are typedefs by default. Before
7200 almost all types were #define'd to ASN1_STRING which was causing
7201 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
7205 *) Change names of new functions to the new get1/get0 naming
7206 convention: After 'get1', the caller owns a reference count
7207 and has to call ..._free; 'get0' returns a pointer to some
7208 data structure without incrementing reference counters.
7209 (Some of the existing 'get' functions increment a reference
7210 counter, some don't.)
7211 Similarly, 'set1' and 'add1' functions increase reference
7212 counters or duplicate objects.
7215 *) Allow for the possibility of temp RSA key generation failure:
7216 the code used to assume it always worked and crashed on failure.
7219 *) Fix potential buffer overrun problem in BIO_printf().
7220 [Ulf Möller, using public domain code by Patrick Powell; problem
7221 pointed out by David Sacerdote <das33@cornell.edu>]
7223 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
7224 RAND_egd() and RAND_status(). In the command line application,
7225 the EGD socket can be specified like a seed file using RANDFILE
7229 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
7230 Some CAs (e.g. Verisign) distribute certificates in this form.
7233 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
7234 list to exclude them. This means that no special compilation option
7235 is needed to use anonymous DH: it just needs to be included in the
7239 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
7240 EVP_MD_type. The old functionality is available in a new macro called
7241 EVP_MD_md(). Change code that uses it and update docs.
7244 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
7245 where the 'void *' argument is replaced by a function pointer argument.
7246 Previously 'void *' was abused to point to functions, which works on
7247 many platforms, but is not correct. As these functions are usually
7248 called by macros defined in OpenSSL header files, most source code
7249 should work without changes.
7252 *) <openssl/opensslconf.h> (which is created by Configure) now contains
7253 sections with information on -D... compiler switches used for
7254 compiling the library so that applications can see them. To enable
7255 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
7256 must be defined. E.g.,
7257 #define OPENSSL_ALGORITHM_DEFINES
7258 #include <openssl/opensslconf.h>
7259 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
7260 [Richard Levitte, Ulf and Bodo Möller]
7262 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
7266 *) Change the 'other' type in certificate aux info to a STACK_OF
7267 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
7268 the required ASN1 format: arbitrary types determined by an OID.
7271 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
7272 argument to 'req'. This is not because the function is newer or
7273 better than others it just uses the work 'NEW' in the certificate
7274 request header lines. Some software needs this.
7277 *) Reorganise password command line arguments: now passwords can be
7278 obtained from various sources. Delete the PEM_cb function and make
7279 it the default behaviour: i.e. if the callback is NULL and the
7280 usrdata argument is not NULL interpret it as a null terminated pass
7281 phrase. If usrdata and the callback are NULL then the pass phrase
7282 is prompted for as usual.
7285 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
7286 the support is automatically enabled. The resulting binaries will
7287 autodetect the card and use it if present.
7288 [Ben Laurie and Compaq Inc.]
7290 *) Work around for Netscape hang bug. This sends certificate request
7291 and server done in one record. Since this is perfectly legal in the
7292 SSL/TLS protocol it isn't a "bug" option and is on by default. See
7293 the bugs/SSLv3 entry for more info.
7296 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
7299 *) Add -rand argument to smime and pkcs12 applications and read/write
7303 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
7306 *) Add command line password options to the remaining applications.
7309 *) Bug fix for BN_div_recp() for numerators with an even number of
7313 *) More tests in bntest.c, and changed test_bn output.
7316 *) ./config recognizes MacOS X now.
7319 *) Bug fix for BN_div() when the first words of num and divsor are
7320 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
7323 *) Add support for various broken PKCS#8 formats, and command line
7324 options to produce them.
7327 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
7328 get temporary BIGNUMs from a BN_CTX.
7331 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
7335 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
7336 include a #define from the old name to the new. The original intent
7337 was that statically linked binaries could for example just call
7338 SSLeay_add_all_ciphers() to just add ciphers to the table and not
7339 link with digests. This never worked becayse SSLeay_add_all_digests()
7340 and SSLeay_add_all_ciphers() were in the same source file so calling
7341 one would link with the other. They are now in separate source files.
7344 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
7347 *) Use a less unusual form of the Miller-Rabin primality test (it used
7348 a binary algorithm for exponentiation integrated into the Miller-Rabin
7349 loop, our standard modexp algorithms are faster).
7352 *) Support for the EBCDIC character set completed.
7353 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
7355 *) Source code cleanups: use const where appropriate, eliminate casts,
7356 use void * instead of char * in lhash.
7359 *) Bugfix: ssl3_send_server_key_exchange was not restartable
7360 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
7361 this the server could overwrite ephemeral keys that the client
7365 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
7366 using 50 iterations of the Rabin-Miller test.
7368 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
7369 iterations of the Rabin-Miller test as required by the appendix
7370 to FIPS PUB 186[-1]) instead of DSA_is_prime.
7371 As BN_is_prime_fasttest includes trial division, DSA parameter
7372 generation becomes much faster.
7374 This implies a change for the callback functions in DSA_is_prime
7375 and DSA_generate_parameters: The callback function is called once
7376 for each positive witness in the Rabin-Miller test, not just
7377 occasionally in the inner loop; and the parameters to the
7378 callback function now provide an iteration count for the outer
7379 loop rather than for the current invocation of the inner loop.
7380 DSA_generate_parameters additionally can call the callback
7381 function with an 'iteration count' of -1, meaning that a
7382 candidate has passed the trial division test (when q is generated
7383 from an application-provided seed, trial division is skipped).
7386 *) New function BN_is_prime_fasttest that optionally does trial
7387 division before starting the Rabin-Miller test and has
7388 an additional BN_CTX * argument (whereas BN_is_prime always
7389 has to allocate at least one BN_CTX).
7390 'callback(1, -1, cb_arg)' is called when a number has passed the
7391 trial division stage.
7394 *) Fix for bug in CRL encoding. The validity dates weren't being handled
7398 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
7401 *) New function BN_pseudo_rand().
7404 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
7405 bignum version of BN_from_montgomery() with the working code from
7406 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
7410 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
7411 made it impossible to use the same SSL_SESSION data structure in
7412 SSL2 clients in multiple threads.
7415 *) The return value of RAND_load_file() no longer counts bytes obtained
7416 by stat(). RAND_load_file(..., -1) is new and uses the complete file
7417 to seed the PRNG (previously an explicit byte count was required).
7418 [Ulf Möller, Bodo Möller]
7420 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
7421 used (char *) instead of (void *) and had casts all over the place.
7424 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
7427 *) Retain source code compatibility for BN_prime_checks macro:
7428 BN_is_prime(..., BN_prime_checks, ...) now uses
7429 BN_prime_checks_for_size to determine the appropriate number of
7430 Rabin-Miller iterations.
7433 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
7434 DH_CHECK_P_NOT_SAFE_PRIME.
7435 (Check if this is true? OpenPGP calls them "strong".)
7438 *) Merge the functionality of "dh" and "gendh" programs into a new program
7439 "dhparam". The old programs are retained for now but will handle DH keys
7440 (instead of parameters) in future.
7443 *) Make the ciphers, s_server and s_client programs check the return values
7444 when a new cipher list is set.
7447 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
7448 ciphers. Before when the 56bit ciphers were enabled the sorting was
7451 The syntax for the cipher sorting has been extended to support sorting by
7452 cipher-strength (using the strength_bits hard coded in the tables).
7453 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
7455 Fix a bug in the cipher-command parser: when supplying a cipher command
7456 string with an "undefined" symbol (neither command nor alphanumeric
7457 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
7458 an error is flagged.
7460 Due to the strength-sorting extension, the code of the
7461 ssl_create_cipher_list() function was completely rearranged. I hope that
7462 the readability was also increased :-)
7463 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7465 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
7466 for the first serial number and places 2 in the serial number file. This
7467 avoids problems when the root CA is created with serial number zero and
7468 the first user certificate has the same issuer name and serial number
7472 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
7473 the new code. Add documentation for this stuff.
7476 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
7477 X509_*() to X509at_*() on the grounds that they don't handle X509
7478 structures and behave in an analagous way to the X509v3 functions:
7479 they shouldn't be called directly but wrapper functions should be used
7482 So we also now have some wrapper functions that call the X509at functions
7483 when passed certificate requests. (TO DO: similar things can be done with
7484 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
7485 things. Some of these need some d2i or i2d and print functionality
7486 because they handle more complex structures.)
7489 *) Add missing #ifndefs that caused missing symbols when building libssl
7490 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
7491 NO_RSA in ssl/s2*.c.
7492 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
7494 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
7495 has a return value which indicates the quality of the random data
7496 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
7497 error queue. New function RAND_pseudo_bytes() generates output that is
7498 guaranteed to be unique but not unpredictable. RAND_add is like
7499 RAND_seed, but takes an extra argument for an entropy estimate
7500 (RAND_seed always assumes full entropy).
7503 *) Do more iterations of Rabin-Miller probable prime test (specifically,
7504 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
7505 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
7506 in crypto/bn/bn_prime.c for the complete table). This guarantees a
7507 false-positive rate of at most 2^-80 for random input.
7510 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
7513 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
7514 in the 0.9.5 release), this returns the chain
7515 from an X509_CTX structure with a dup of the stack and all
7516 the X509 reference counts upped: so the stack will exist
7517 after X509_CTX_cleanup() has been called. Modify pkcs12.c
7520 Also make SSL_SESSION_print() print out the verify return
7524 *) Add manpage for the pkcs12 command. Also change the default
7525 behaviour so MAC iteration counts are used unless the new
7526 -nomaciter option is used. This improves file security and
7527 only older versions of MSIE (4.0 for example) need it.
7530 *) Honor the no-xxx Configure options when creating .DEF files.
7533 *) Add PKCS#10 attributes to field table: challengePassword,
7534 unstructuredName and unstructuredAddress. These are taken from
7535 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
7536 international characters are used.
7538 More changes to X509_ATTRIBUTE code: allow the setting of types
7539 based on strings. Remove the 'loc' parameter when adding
7540 attributes because these will be a SET OF encoding which is sorted
7544 *) Initial changes to the 'req' utility to allow request generation
7545 automation. This will allow an application to just generate a template
7546 file containing all the field values and have req construct the
7549 Initial support for X509_ATTRIBUTE handling. Stacks of these are
7550 used all over the place including certificate requests and PKCS#7
7551 structures. They are currently handled manually where necessary with
7552 some primitive wrappers for PKCS#7. The new functions behave in a
7553 manner analogous to the X509 extension functions: they allow
7554 attributes to be looked up by NID and added.
7556 Later something similar to the X509V3 code would be desirable to
7557 automatically handle the encoding, decoding and printing of the
7558 more complex types. The string types like challengePassword can
7559 be handled by the string table functions.
7561 Also modified the multi byte string table handling. Now there is
7562 a 'global mask' which masks out certain types. The table itself
7563 can use the flag STABLE_NO_MASK to ignore the mask setting: this
7564 is useful when for example there is only one permissible type
7565 (as in countryName) and using the mask might result in no valid
7569 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
7570 SSL_get_peer_finished to allow applications to obtain the latest
7571 Finished messages sent to the peer or expected from the peer,
7572 respectively. (SSL_get_peer_finished is usually the Finished message
7573 actually received from the peer, otherwise the protocol will be aborted.)
7575 As the Finished message are message digests of the complete handshake
7576 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
7577 be used for external authentication procedures when the authentication
7578 provided by SSL/TLS is not desired or is not enough.
7581 *) Enhanced support for Alpha Linux is added. Now ./config checks if
7582 the host supports BWX extension and if Compaq C is present on the
7583 $PATH. Just exploiting of the BWX extension results in 20-30%
7584 performance kick for some algorithms, e.g. DES and RC4 to mention
7585 a couple. Compaq C in turn generates ~20% faster code for MD5 and
7589 *) Add support for MS "fast SGC". This is arguably a violation of the
7590 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
7591 weak crypto and after checking the certificate is SGC a second one
7592 with strong crypto. MS SGC stops the first handshake after receiving
7593 the server certificate message and sends a second client hello. Since
7594 a server will typically do all the time consuming operations before
7595 expecting any further messages from the client (server key exchange
7596 is the most expensive) there is little difference between the two.
7598 To get OpenSSL to support MS SGC we have to permit a second client
7599 hello message after we have sent server done. In addition we have to
7600 reset the MAC if we do get this second client hello.
7603 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
7604 if a DER encoded private key is RSA or DSA traditional format. Changed
7605 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
7606 format DER encoded private key. Newer code should use PKCS#8 format which
7607 has the key type encoded in the ASN1 structure. Added DER private key
7608 support to pkcs8 application.
7611 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
7612 ciphersuites has been selected (as required by the SSL 3/TLS 1
7613 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
7614 is set, we interpret this as a request to violate the specification
7615 (the worst that can happen is a handshake failure, and 'correct'
7616 behaviour would result in a handshake failure anyway).
7619 *) In SSL_CTX_add_session, take into account that there might be multiple
7620 SSL_SESSION structures with the same session ID (e.g. when two threads
7621 concurrently obtain them from an external cache).
7622 The internal cache can handle only one SSL_SESSION with a given ID,
7623 so if there's a conflict, we now throw out the old one to achieve
7627 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
7628 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
7629 some routines that use cipher OIDs: some ciphers do not have OIDs
7630 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
7634 *) Simplify the trust setting structure and code. Now we just have
7635 two sequences of OIDs for trusted and rejected settings. These will
7636 typically have values the same as the extended key usage extension
7637 and any application specific purposes.
7639 The trust checking code now has a default behaviour: it will just
7640 check for an object with the same NID as the passed id. Functions can
7641 be provided to override either the default behaviour or the behaviour
7642 for a given id. SSL client, server and email already have functions
7643 in place for compatibility: they check the NID and also return "trusted"
7644 if the certificate is self signed.
7647 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
7648 traditional format into an EVP_PKEY structure.
7651 *) Add a password callback function PEM_cb() which either prompts for
7652 a password if usr_data is NULL or otherwise assumes it is a null
7653 terminated password. Allow passwords to be passed on command line
7654 environment or config files in a few more utilities.
7657 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
7658 keys. Add some short names for PKCS#8 PBE algorithms and allow them
7659 to be specified on the command line for the pkcs8 and pkcs12 utilities.
7660 Update documentation.
7663 *) Support for ASN1 "NULL" type. This could be handled before by using
7664 ASN1_TYPE but there wasn't any function that would try to read a NULL
7665 and produce an error if it couldn't. For compatibility we also have
7666 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
7667 don't allocate anything because they don't need to.
7670 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
7672 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
7674 *) Rebuild of the memory allocation routines used by OpenSSL code and
7675 possibly others as well. The purpose is to make an interface that
7676 provide hooks so anyone can build a separate set of allocation and
7677 deallocation routines to be used by OpenSSL, for example memory
7678 pool implementations, or something else, which was previously hard
7679 since Malloc(), Realloc() and Free() were defined as macros having
7680 the values malloc, realloc and free, respectively (except for Win32
7681 compilations). The same is provided for memory debugging code.
7682 OpenSSL already comes with functionality to find memory leaks, but
7683 this gives people a chance to debug other memory problems.
7685 With these changes, a new set of functions and macros have appeared:
7687 CRYPTO_set_mem_debug_functions() [F]
7688 CRYPTO_get_mem_debug_functions() [F]
7689 CRYPTO_dbg_set_options() [F]
7690 CRYPTO_dbg_get_options() [F]
7691 CRYPTO_malloc_debug_init() [M]
7693 The memory debug functions are NULL by default, unless the library
7694 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
7695 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
7696 gives the standard debugging functions that come with OpenSSL) or
7697 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
7698 provided by the library user) must be used. When the standard
7699 debugging functions are used, CRYPTO_dbg_set_options can be used to
7700 request additional information:
7701 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
7702 the CRYPTO_MDEBUG_xxx macro when compiling the library.
7704 Also, things like CRYPTO_set_mem_functions will always give the
7705 expected result (the new set of functions is used for allocation
7706 and deallocation) at all times, regardless of platform and compiler
7709 To finish it up, some functions that were never use in any other
7710 way than through macros have a new API and new semantic:
7713 CRYPTO_dbg_realloc()
7716 All macros of value have retained their old syntax.
7717 [Richard Levitte and Bodo Moeller]
7719 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
7720 ordering of SMIMECapabilities wasn't in "strength order" and there
7721 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
7725 *) Some ASN1 types with illegal zero length encoding (INTEGER,
7726 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
7727 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
7729 *) Merge in my S/MIME library for OpenSSL. This provides a simple
7730 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
7731 functionality to handle multipart/signed properly) and a utility
7732 called 'smime' to call all this stuff. This is based on code I
7733 originally wrote for Celo who have kindly allowed it to be
7734 included in OpenSSL.
7737 *) Add variants des_set_key_checked and des_set_key_unchecked of
7738 des_set_key (aka des_key_sched). Global variable des_check_key
7739 decides which of these is called by des_set_key; this way
7740 des_check_key behaves as it always did, but applications and
7741 the library itself, which was buggy for des_check_key == 1,
7742 have a cleaner way to pick the version they need.
7745 *) New function PKCS12_newpass() which changes the password of a
7749 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
7750 dynamic mix. In both cases the ids can be used as an index into the
7751 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
7752 functions so they accept a list of the field values and the
7753 application doesn't need to directly manipulate the X509_TRUST
7757 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
7761 *) Modify the way the V3 extension code looks up extensions. This now
7762 works in a similar way to the object code: we have some "standard"
7763 extensions in a static table which is searched with OBJ_bsearch()
7764 and the application can add dynamic ones if needed. The file
7765 crypto/x509v3/ext_dat.h now has the info: this file needs to be
7766 updated whenever a new extension is added to the core code and kept
7767 in ext_nid order. There is a simple program 'tabtest.c' which checks
7768 this. New extensions are not added too often so this file can readily
7769 be maintained manually.
7771 There are two big advantages in doing things this way. The extensions
7772 can be looked up immediately and no longer need to be "added" using
7773 X509V3_add_standard_extensions(): this function now does nothing.
7774 [Side note: I get *lots* of email saying the extension code doesn't
7775 work because people forget to call this function]
7776 Also no dynamic allocation is done unless new extensions are added:
7777 so if we don't add custom extensions there is no need to call
7778 X509V3_EXT_cleanup().
7781 *) Modify enc utility's salting as follows: make salting the default. Add a
7782 magic header, so unsalted files fail gracefully instead of just decrypting
7783 to garbage. This is because not salting is a big security hole, so people
7784 should be discouraged from doing it.
7787 *) Fixes and enhancements to the 'x509' utility. It allowed a message
7788 digest to be passed on the command line but it only used this
7789 parameter when signing a certificate. Modified so all relevant
7790 operations are affected by the digest parameter including the
7791 -fingerprint and -x509toreq options. Also -x509toreq choked if a
7792 DSA key was used because it didn't fix the digest.
7795 *) Initial certificate chain verify code. Currently tests the untrusted
7796 certificates for consistency with the verify purpose (which is set
7797 when the X509_STORE_CTX structure is set up) and checks the pathlength.
7799 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
7800 this is because it will reject chains with invalid extensions whereas
7801 every previous version of OpenSSL and SSLeay made no checks at all.
7803 Trust code: checks the root CA for the relevant trust settings. Trust
7804 settings have an initial value consistent with the verify purpose: e.g.
7805 if the verify purpose is for SSL client use it expects the CA to be
7806 trusted for SSL client use. However the default value can be changed to
7807 permit custom trust settings: one example of this would be to only trust
7808 certificates from a specific "secure" set of CAs.
7810 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
7811 which should be used for version portability: especially since the
7812 verify structure is likely to change more often now.
7814 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
7815 to set them. If not set then assume SSL clients will verify SSL servers
7818 Two new options to the verify program: -untrusted allows a set of
7819 untrusted certificates to be passed in and -purpose which sets the
7820 intended purpose of the certificate. If a purpose is set then the
7821 new chain verify code is used to check extension consistency.
7824 *) Support for the authority information access extension.
7827 *) Modify RSA and DSA PEM read routines to transparently handle
7828 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
7829 public keys in a format compatible with certificate
7830 SubjectPublicKeyInfo structures. Unfortunately there were already
7831 functions called *_PublicKey_* which used various odd formats so
7832 these are retained for compatibility: however the DSA variants were
7833 never in a public release so they have been deleted. Changed dsa/rsa
7834 utilities to handle the new format: note no releases ever handled public
7835 keys so we should be OK.
7837 The primary motivation for this change is to avoid the same fiasco
7838 that dogs private keys: there are several incompatible private key
7839 formats some of which are standard and some OpenSSL specific and
7840 require various evil hacks to allow partial transparent handling and
7841 even then it doesn't work with DER formats. Given the option anything
7842 other than PKCS#8 should be dumped: but the other formats have to
7843 stay in the name of compatibility.
7845 With public keys and the benefit of hindsight one standard format
7846 is used which works with EVP_PKEY, RSA or DSA structures: though
7847 it clearly returns an error if you try to read the wrong kind of key.
7849 Added a -pubkey option to the 'x509' utility to output the public key.
7850 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
7851 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
7852 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
7853 that do the same as the EVP_PKEY_assign_*() except they up the
7854 reference count of the added key (they don't "swallow" the
7858 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
7859 CRLs would fail if the file contained no certificates or no CRLs:
7860 added a new function to read in both types and return the number
7861 read: this means that if none are read it will be an error. The
7862 DER versions of the certificate and CRL reader would always fail
7863 because it isn't possible to mix certificates and CRLs in DER format
7864 without choking one or the other routine. Changed this to just read
7865 a certificate: this is the best we can do. Also modified the code
7866 in apps/verify.c to take notice of return codes: it was previously
7867 attempting to read in certificates from NULL pointers and ignoring
7868 any errors: this is one reason why the cert and CRL reader seemed
7869 to work. It doesn't check return codes from the default certificate
7870 routines: these may well fail if the certificates aren't installed.
7873 *) Code to support otherName option in GeneralName.
7876 *) First update to verify code. Change the verify utility
7877 so it warns if it is passed a self signed certificate:
7878 for consistency with the normal behaviour. X509_verify
7879 has been modified to it will now verify a self signed
7880 certificate if *exactly* the same certificate appears
7881 in the store: it was previously impossible to trust a
7882 single self signed certificate. This means that:
7883 openssl verify ss.pem
7884 now gives a warning about a self signed certificate but
7885 openssl verify -CAfile ss.pem ss.pem
7889 *) For servers, store verify_result in SSL_SESSION data structure
7890 (and add it to external session representation).
7891 This is needed when client certificate verifications fails,
7892 but an application-provided verification callback (set by
7893 SSL_CTX_set_cert_verify_callback) allows accepting the session
7894 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
7895 but returns 1): When the session is reused, we have to set
7896 ssl->verify_result to the appropriate error code to avoid
7898 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
7900 *) Fix a bug in the new PKCS#7 code: it didn't consider the
7901 case in PKCS7_dataInit() where the signed PKCS7 structure
7902 didn't contain any existing data because it was being created.
7903 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
7905 *) Add a salt to the key derivation routines in enc.c. This
7906 forms the first 8 bytes of the encrypted file. Also add a
7907 -S option to allow a salt to be input on the command line.
7910 *) New function X509_cmp(). Oddly enough there wasn't a function
7911 to compare two certificates. We do this by working out the SHA1
7912 hash and comparing that. X509_cmp() will be needed by the trust
7916 *) SSL_get1_session() is like SSL_get_session(), but increments
7917 the reference count in the SSL_SESSION returned.
7918 [Geoff Thorpe <geoff@eu.c2.net>]
7920 *) Fix for 'req': it was adding a null to request attributes.
7921 Also change the X509_LOOKUP and X509_INFO code to handle
7922 certificate auxiliary information.
7925 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
7929 *) Add the possibility to add extra information to the memory leak
7930 detecting output, to form tracebacks, showing from where each
7931 allocation was originated: CRYPTO_push_info("constant string") adds
7932 the string plus current file name and line number to a per-thread
7933 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
7934 is like calling CYRPTO_pop_info() until the stack is empty.
7935 Also updated memory leak detection code to be multi-thread-safe.
7938 *) Add options -text and -noout to pkcs7 utility and delete the
7939 encryption options which never did anything. Update docs.
7942 *) Add options to some of the utilities to allow the pass phrase
7943 to be included on either the command line (not recommended on
7944 OSes like Unix) or read from the environment. Update the
7945 manpages and fix a few bugs.
7948 *) Add a few manpages for some of the openssl commands.
7951 *) Fix the -revoke option in ca. It was freeing up memory twice,
7952 leaking and not finding already revoked certificates.
7955 *) Extensive changes to support certificate auxiliary information.
7956 This involves the use of X509_CERT_AUX structure and X509_AUX
7957 functions. An X509_AUX function such as PEM_read_X509_AUX()
7958 can still read in a certificate file in the usual way but it
7959 will also read in any additional "auxiliary information". By
7960 doing things this way a fair degree of compatibility can be
7961 retained: existing certificates can have this information added
7962 using the new 'x509' options.
7964 Current auxiliary information includes an "alias" and some trust
7965 settings. The trust settings will ultimately be used in enhanced
7966 certificate chain verification routines: currently a certificate
7967 can only be trusted if it is self signed and then it is trusted
7971 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
7972 The problem was that one of the replacement routines had not been working
7973 since SSLeay releases. For now the offending routine has been replaced
7974 with non-optimised assembler. Even so, this now gives around 95%
7975 performance improvement for 1024 bit RSA signs.
7978 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
7979 handling. Most clients have the effective key size in bits equal to
7980 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
7981 A few however don't do this and instead use the size of the decrypted key
7982 to determine the RC2 key length and the AlgorithmIdentifier to determine
7983 the effective key length. In this case the effective key length can still
7984 be 40 bits but the key length can be 168 bits for example. This is fixed
7985 by manually forcing an RC2 key into the EVP_PKEY structure because the
7986 EVP code can't currently handle unusual RC2 key sizes: it always assumes
7987 the key length and effective key length are equal.
7990 *) Add a bunch of functions that should simplify the creation of
7991 X509_NAME structures. Now you should be able to do:
7992 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
7993 and have it automatically work out the correct field type and fill in
7994 the structures. The more adventurous can try:
7995 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
7996 and it will (hopefully) work out the correct multibyte encoding.
7999 *) Change the 'req' utility to use the new field handling and multibyte
8000 copy routines. Before the DN field creation was handled in an ad hoc
8001 way in req, ca, and x509 which was rather broken and didn't support
8002 BMPStrings or UTF8Strings. Since some software doesn't implement
8003 BMPStrings or UTF8Strings yet, they can be enabled using the config file
8004 using the dirstring_type option. See the new comment in the default
8005 openssl.cnf for more info.
8008 *) Make crypto/rand/md_rand.c more robust:
8009 - Assure unique random numbers after fork().
8010 - Make sure that concurrent threads access the global counter and
8011 md serializably so that we never lose entropy in them
8012 or use exactly the same state in multiple threads.
8013 Access to the large state is not always serializable because
8014 the additional locking could be a performance killer, and
8015 md should be large enough anyway.
8018 *) New file apps/app_rand.c with commonly needed functionality
8019 for handling the random seed file.
8021 Use the random seed file in some applications that previously did not:
8023 dsaparam -genkey (which also ignored its '-rand' option),
8026 x509 (when signing).
8027 Except on systems with /dev/urandom, it is crucial to have a random
8028 seed file at least for key creation, DSA signing, and for DH exchanges;
8029 for RSA signatures we could do without one.
8031 gendh and gendsa (unlike genrsa) used to read only the first byte
8032 of each file listed in the '-rand' option. The function as previously
8033 found in genrsa is now in app_rand.c and is used by all programs
8034 that support '-rand'.
8037 *) In RAND_write_file, use mode 0600 for creating files;
8038 don't just chmod when it may be too late.
8041 *) Report an error from X509_STORE_load_locations
8042 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
8045 *) New function ASN1_mbstring_copy() this copies a string in either
8046 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
8047 into an ASN1_STRING type. A mask of permissible types is passed
8048 and it chooses the "minimal" type to use or an error if not type
8052 *) Add function equivalents to the various macros in asn1.h. The old
8053 macros are retained with an M_ prefix. Code inside the library can
8054 use the M_ macros. External code (including the openssl utility)
8055 should *NOT* in order to be "shared library friendly".
8058 *) Add various functions that can check a certificate's extensions
8059 to see if it usable for various purposes such as SSL client,
8060 server or S/MIME and CAs of these types. This is currently
8061 VERY EXPERIMENTAL but will ultimately be used for certificate chain
8062 verification. Also added a -purpose flag to x509 utility to
8063 print out all the purposes.
8066 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
8070 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
8071 for, obtain and decode and extension and obtain its critical flag.
8072 This allows all the necessary extension code to be handled in a
8073 single function call.
8076 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
8077 platforms. See crypto/rc4/rc4_enc.c for further details.
8080 *) New -noout option to asn1parse. This causes no output to be produced
8081 its main use is when combined with -strparse and -out to extract data
8082 from a file (which may not be in ASN.1 format).
8085 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
8086 when producing the local key id.
8087 [Richard Levitte <levitte@stacken.kth.se>]
8089 *) New option -dhparam in s_server. This allows a DH parameter file to be
8090 stated explicitly. If it is not stated then it tries the first server
8091 certificate file. The previous behaviour hard coded the filename
8095 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
8096 a public key to be input or output. For example:
8097 openssl rsa -in key.pem -pubout -out pubkey.pem
8098 Also added necessary DSA public key functions to handle this.
8101 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
8102 in the message. This was handled by allowing
8103 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8104 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8106 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8107 to the end of the strings whereas this didn't. This would cause problems
8108 if strings read with d2i_ASN1_bytes() were later modified.
8109 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8111 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8112 data and it contains EOF it will end up returning an error. This is
8113 caused by input 46 bytes long. The cause is due to the way base64
8114 BIOs find the start of base64 encoded data. They do this by trying a
8115 trial decode on each line until they find one that works. When they
8116 do a flag is set and it starts again knowing it can pass all the
8117 data directly through the decoder. Unfortunately it doesn't reset
8118 the context it uses. This means that if EOF is reached an attempt
8119 is made to pass two EOFs through the context and this causes the
8120 resulting error. This can also cause other problems as well. As is
8121 usual with these problems it takes *ages* to find and the fix is
8122 trivial: move one line.
8123 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8125 *) Ugly workaround to get s_client and s_server working under Windows. The
8126 old code wouldn't work because it needed to select() on sockets and the
8127 tty (for keypresses and to see if data could be written). Win32 only
8128 supports select() on sockets so we select() with a 1s timeout on the
8129 sockets and then see if any characters are waiting to be read, if none
8130 are present then we retry, we also assume we can always write data to
8131 the tty. This isn't nice because the code then blocks until we've
8132 received a complete line of data and it is effectively polling the
8133 keyboard at 1s intervals: however it's quite a bit better than not
8134 working at all :-) A dedicated Windows application might handle this
8135 with an event loop for example.
8138 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
8139 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
8140 will be called when RSA_sign() and RSA_verify() are used. This is useful
8141 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
8142 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
8143 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
8144 This necessitated the support of an extra signature type NID_md5_sha1
8145 for SSL signatures and modifications to the SSL library to use it instead
8146 of calling RSA_public_decrypt() and RSA_private_encrypt().
8149 *) Add new -verify -CAfile and -CApath options to the crl program, these
8150 will lookup a CRL issuers certificate and verify the signature in a
8151 similar way to the verify program. Tidy up the crl program so it
8152 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
8153 less strict. It will now permit CRL extensions even if it is not
8154 a V2 CRL: this will allow it to tolerate some broken CRLs.
8157 *) Initialize all non-automatic variables each time one of the openssl
8158 sub-programs is started (this is necessary as they may be started
8159 multiple times from the "OpenSSL>" prompt).
8160 [Lennart Bang, Bodo Moeller]
8162 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
8163 removing all other RSA functionality (this is what NO_RSA does). This
8164 is so (for example) those in the US can disable those operations covered
8165 by the RSA patent while allowing storage and parsing of RSA keys and RSA
8169 *) Non-copying interface to BIO pairs.
8170 (still largely untested)
8173 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
8174 ASCII string. This was handled independently in various places before.
8177 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
8178 UTF8 strings a character at a time.
8181 *) Use client_version from client hello to select the protocol
8182 (s23_srvr.c) and for RSA client key exchange verification
8183 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
8186 *) Add various utility functions to handle SPKACs, these were previously
8187 handled by poking round in the structure internals. Added new function
8188 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
8189 print, verify and generate SPKACs. Based on an original idea from
8190 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
8193 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
8196 *) Allow the config file extension section to be overwritten on the
8197 command line. Based on an original idea from Massimiliano Pala
8198 <madwolf@comune.modena.it>. The new option is called -extensions
8199 and can be applied to ca, req and x509. Also -reqexts to override
8200 the request extensions in req and -crlexts to override the crl extensions
8204 *) Add new feature to the SPKAC handling in ca. Now you can include
8205 the same field multiple times by preceding it by "XXXX." for example:
8208 this is the same syntax as used in the req config file.
8211 *) Allow certificate extensions to be added to certificate requests. These
8212 are specified in a 'req_extensions' option of the req section of the
8213 config file. They can be printed out with the -text option to req but
8214 are otherwise ignored at present.
8217 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
8218 data read consists of only the final block it would not decrypted because
8219 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
8220 A misplaced 'break' also meant the decrypted final block might not be
8221 copied until the next read.
8224 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
8225 a few extra parameters to the DH structure: these will be useful if
8226 for example we want the value of 'q' or implement X9.42 DH.
8229 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
8230 provides hooks that allow the default DSA functions or functions on a
8231 "per key" basis to be replaced. This allows hardware acceleration and
8232 hardware key storage to be handled without major modification to the
8233 library. Also added low level modexp hooks and CRYPTO_EX structure and
8234 associated functions.
8237 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
8238 as "read only": it can't be written to and the buffer it points to will
8239 not be freed. Reading from a read only BIO is much more efficient than
8240 a normal memory BIO. This was added because there are several times when
8241 an area of memory needs to be read from a BIO. The previous method was
8242 to create a memory BIO and write the data to it, this results in two
8243 copies of the data and an O(n^2) reading algorithm. There is a new
8244 function BIO_new_mem_buf() which creates a read only memory BIO from
8245 an area of memory. Also modified the PKCS#7 routines to use read only
8249 *) Bugfix: ssl23_get_client_hello did not work properly when called in
8250 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
8251 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
8252 but a retry condition occured while trying to read the rest.
8255 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
8256 NID_pkcs7_encrypted by default: this was wrong since this should almost
8257 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
8258 the encrypted data type: this is a more sensible place to put it and it
8259 allows the PKCS#12 code to be tidied up that duplicated this
8263 *) Changed obj_dat.pl script so it takes its input and output files on
8264 the command line. This should avoid shell escape redirection problems
8268 *) Initial support for certificate extension requests, these are included
8269 in things like Xenroll certificate requests. Included functions to allow
8270 extensions to be obtained and added.
8273 *) -crlf option to s_client and s_server for sending newlines as
8274 CRLF (as required by many protocols).
8277 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
8279 *) Install libRSAglue.a when OpenSSL is built with RSAref.
8280 [Ralf S. Engelschall]
8282 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
8283 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
8285 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
8289 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
8290 DH parameters/keys (q is lost during that conversion, but the resulting
8291 DH parameters contain its length).
8293 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
8294 much faster than DH_generate_parameters (which creates parameters
8295 where p = 2*q + 1), and also the smaller q makes DH computations
8296 much more efficient (160-bit exponentiation instead of 1024-bit
8297 exponentiation); so this provides a convenient way to support DHE
8298 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
8299 utter importance to use
8300 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8302 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8303 when such DH parameters are used, because otherwise small subgroup
8304 attacks may become possible!
8307 *) Avoid memory leak in i2d_DHparams.
8310 *) Allow the -k option to be used more than once in the enc program:
8311 this allows the same encrypted message to be read by multiple recipients.
8314 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
8315 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
8316 it will always use the numerical form of the OID, even if it has a short
8320 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
8321 method only got called if p,q,dmp1,dmq1,iqmp components were present,
8322 otherwise bn_mod_exp was called. In the case of hardware keys for example
8323 no private key components need be present and it might store extra data
8324 in the RSA structure, which cannot be accessed from bn_mod_exp.
8325 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
8326 private key operations.
8329 *) Added support for SPARC Linux.
8332 *) pem_password_cb function type incompatibly changed from
8333 typedef int pem_password_cb(char *buf, int size, int rwflag);
8335 ....(char *buf, int size, int rwflag, void *userdata);
8336 so that applications can pass data to their callbacks:
8337 The PEM[_ASN1]_{read,write}... functions and macros now take an
8338 additional void * argument, which is just handed through whenever
8339 the password callback is called.
8340 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
8342 New function SSL_CTX_set_default_passwd_cb_userdata.
8344 Compatibility note: As many C implementations push function arguments
8345 onto the stack in reverse order, the new library version is likely to
8346 interoperate with programs that have been compiled with the old
8347 pem_password_cb definition (PEM_whatever takes some data that
8348 happens to be on the stack as its last argument, and the callback
8349 just ignores this garbage); but there is no guarantee whatsoever that
8352 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
8353 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
8354 problems not only on Windows, but also on some Unix platforms.
8355 To avoid problematic command lines, these definitions are now in an
8356 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
8357 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
8360 *) MIPS III/IV assembler module is reimplemented.
8363 *) More DES library cleanups: remove references to srand/rand and
8364 delete an unused file.
8367 *) Add support for the the free Netwide assembler (NASM) under Win32,
8368 since not many people have MASM (ml) and it can be hard to obtain.
8369 This is currently experimental but it seems to work OK and pass all
8370 the tests. Check out INSTALL.W32 for info.
8373 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
8374 without temporary keys kept an extra copy of the server key,
8375 and connections with temporary keys did not free everything in case
8379 *) New function RSA_check_key and new openssl rsa option -check
8380 for verifying the consistency of RSA keys.
8381 [Ulf Moeller, Bodo Moeller]
8383 *) Various changes to make Win32 compile work:
8384 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
8385 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
8386 comparison" warnings.
8387 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
8390 *) Add a debugging option to PKCS#5 v2 key generation function: when
8391 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
8392 derived keys are printed to stderr.
8395 *) Copy the flags in ASN1_STRING_dup().
8396 [Roman E. Pavlov <pre@mo.msk.ru>]
8398 *) The x509 application mishandled signing requests containing DSA
8399 keys when the signing key was also DSA and the parameters didn't match.
8401 It was supposed to omit the parameters when they matched the signing key:
8402 the verifying software was then supposed to automatically use the CA's
8403 parameters if they were absent from the end user certificate.
8405 Omitting parameters is no longer recommended. The test was also
8406 the wrong way round! This was probably due to unusual behaviour in
8407 EVP_cmp_parameters() which returns 1 if the parameters match.
8408 This meant that parameters were omitted when they *didn't* match and
8409 the certificate was useless. Certificates signed with 'ca' didn't have
8411 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
8413 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
8414 The interface is as follows:
8415 Applications can use
8416 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
8417 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
8418 "off" is now the default.
8419 The library internally uses
8420 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
8421 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
8422 to disable memory-checking temporarily.
8424 Some inconsistent states that previously were possible (and were
8425 even the default) are now avoided.
8427 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
8428 with each memory chunk allocated; this is occasionally more helpful
8429 than just having a counter.
8431 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
8433 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
8437 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
8438 which largely parallels "options", but is for changing API behaviour,
8439 whereas "options" are about protocol behaviour.
8440 Initial "mode" flags are:
8442 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
8443 a single record has been written.
8444 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
8445 retries use the same buffer location.
8446 (But all of the contents must be
8450 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
8453 *) Fix problems with no-hmac etc.
8454 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
8456 *) New functions RSA_get_default_method(), RSA_set_method() and
8457 RSA_get_method(). These allows replacement of RSA_METHODs without having
8458 to mess around with the internals of an RSA structure.
8461 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
8462 Also really enable memory leak checks in openssl.c and in some
8464 [Chad C. Mulligan, Bodo Moeller]
8466 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
8467 up the length of negative integers. This has now been simplified to just
8468 store the length when it is first determined and use it later, rather
8469 than trying to keep track of where data is copied and updating it to
8471 [Steve Henson, reported by Brien Wheeler
8472 <bwheeler@authentica-security.com>]
8474 *) Add a new function PKCS7_signatureVerify. This allows the verification
8475 of a PKCS#7 signature but with the signing certificate passed to the
8476 function itself. This contrasts with PKCS7_dataVerify which assumes the
8477 certificate is present in the PKCS#7 structure. This isn't always the
8478 case: certificates can be omitted from a PKCS#7 structure and be
8479 distributed by "out of band" means (such as a certificate database).
8482 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
8483 function prototypes in pem.h, also change util/mkdef.pl to add the
8484 necessary function names.
8487 *) mk1mf.pl (used by Windows builds) did not properly read the
8488 options set by Configure in the top level Makefile, and Configure
8489 was not even able to write more than one option correctly.
8490 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
8493 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
8494 file to be loaded from a BIO or FILE pointer. The BIO version will
8495 for example allow memory BIOs to contain config info.
8498 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
8499 Whoever hopes to achieve shared-library compatibility across versions
8500 must use this, not the compile-time macro.
8501 (Exercise 0.9.4: Which is the minimum library version required by
8503 Note: All this applies only to multi-threaded programs, others don't
8507 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
8508 through a BIO pair triggered the default case, i.e.
8509 SSLerr(...,SSL_R_UNKNOWN_STATE).
8512 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
8513 can use the SSL library even if none of the specific BIOs is
8517 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
8518 for the encoded length.
8519 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
8521 *) Add initial documentation of the X509V3 functions.
8524 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
8525 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
8526 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
8527 secure PKCS#8 private key format with a high iteration count.
8530 *) Fix determination of Perl interpreter: A perl or perl5
8531 _directory_ in $PATH was also accepted as the interpreter.
8532 [Ralf S. Engelschall]
8534 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
8535 wrong with it but it was very old and did things like calling
8536 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
8540 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
8541 to use the new extension code.
8544 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
8545 with macros. This should make it easier to change their form, add extra
8546 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
8550 *) Add to configuration table a new entry that can specify an alternative
8551 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
8552 according to Mark Crispin <MRC@Panda.COM>.
8556 *) DES CBC did not update the IV. Weird.
8559 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
8560 Changing the behaviour of the former might break existing programs --
8561 where IV updating is needed, des_ncbc_encrypt can be used.
8564 *) When bntest is run from "make test" it drives bc to check its
8565 calculations, as well as internally checking them. If an internal check
8566 fails, it needs to cause bc to give a non-zero result or make test carries
8567 on without noticing the failure. Fixed.
8570 *) DES library cleanups.
8573 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
8574 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
8575 ciphers. NOTE: although the key derivation function has been verified
8576 against some published test vectors it has not been extensively tested
8577 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
8581 *) Instead of "mkdir -p", which is not fully portable, use new
8582 Perl script "util/mkdir-p.pl".
8585 *) Rewrite the way password based encryption (PBE) is handled. It used to
8586 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
8587 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
8588 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
8589 the 'parameter' field of the AlgorithmIdentifier is passed to the
8590 underlying key generation function so it must do its own ASN1 parsing.
8591 This has also changed the EVP_PBE_CipherInit() function which now has a
8592 'parameter' argument instead of literal salt and iteration count values
8593 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
8596 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
8597 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
8598 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
8599 KEY" because this clashed with PKCS#8 unencrypted string. Since this
8600 value was just used as a "magic string" and not used directly its
8601 value doesn't matter.
8604 *) Introduce some semblance of const correctness to BN. Shame C doesn't
8608 *) "linux-sparc64" configuration (ultrapenguin).
8609 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
8610 "linux-sparc" configuration.
8611 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
8613 *) config now generates no-xxx options for missing ciphers.
8616 *) Support the EBCDIC character set (work in progress).
8617 File ebcdic.c not yet included because it has a different license.
8618 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8620 *) Support BS2000/OSD-POSIX.
8621 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8623 *) Make callbacks for key generation use void * instead of char *.
8626 *) Make S/MIME samples compile (not yet tested).
8629 *) Additional typesafe stacks.
8632 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
8636 Changes between 0.9.3 and 0.9.3a [29 May 1999]
8638 *) New configuration variant "sco5-gcc".
8640 *) Updated some demos.
8641 [Sean O Riordain, Wade Scholine]
8643 *) Add missing BIO_free at exit of pkcs12 application.
8646 *) Fix memory leak in conf.c.
8649 *) Updates for Win32 to assembler version of MD5.
8652 *) Set #! path to perl in apps/der_chop to where we found it
8653 instead of using a fixed path.
8656 *) SHA library changes for irix64-mips4-cc.
8659 *) Improvements for VMS support.
8663 Changes between 0.9.2b and 0.9.3 [24 May 1999]
8665 *) Bignum library bug fix. IRIX 6 passes "make test" now!
8666 This also avoids the problems with SC4.2 and unpatched SC5.
8667 [Andy Polyakov <appro@fy.chalmers.se>]
8669 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
8670 These are required because of the typesafe stack would otherwise break
8671 existing code. If old code used a structure member which used to be STACK
8672 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
8673 sk_num or sk_value it would produce an error because the num, data members
8674 are not present in STACK_OF. Now it just produces a warning. sk_set
8675 replaces the old method of assigning a value to sk_value
8676 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
8677 that does this will no longer work (and should use sk_set instead) but
8678 this could be regarded as a "questionable" behaviour anyway.
8681 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
8682 correctly handle encrypted S/MIME data.
8685 *) Change type of various DES function arguments from des_cblock
8686 (which means, in function argument declarations, pointer to char)
8687 to des_cblock * (meaning pointer to array with 8 char elements),
8688 which allows the compiler to do more typechecking; it was like
8689 that back in SSLeay, but with lots of ugly casts.
8691 Introduce new type const_des_cblock.
8694 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
8695 problems: find RecipientInfo structure that matches recipient certificate
8696 and initialise the ASN1 structures properly based on passed cipher.
8699 *) Belatedly make the BN tests actually check the results.
8702 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
8703 to and from BNs: it was completely broken. New compilation option
8704 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
8705 key elements as negative integers.
8708 *) Reorganize and speed up MD5.
8709 [Andy Polyakov <appro@fy.chalmers.se>]
8712 [Richard Levitte <richard@levitte.org>]
8714 *) New option -out to asn1parse to allow the parsed structure to be
8715 output to a file. This is most useful when combined with the -strparse
8716 option to examine the output of things like OCTET STRINGS.
8719 *) Make SSL library a little more fool-proof by not requiring any longer
8720 that SSL_set_{accept,connect}_state be called before
8721 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
8722 in many applications because usually everything *appeared* to work as
8723 intended anyway -- now it really works as intended).
8726 *) Move openssl.cnf out of lib/.
8729 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
8730 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
8731 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
8732 [Ralf S. Engelschall]
8734 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
8735 handle PKCS#7 enveloped data properly.
8736 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
8738 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
8739 copying pointers. The cert_st handling is changed by this in
8740 various ways (and thus what used to be known as ctx->default_cert
8741 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
8742 any longer when s->cert does not give us what we need).
8743 ssl_cert_instantiate becomes obsolete by this change.
8744 As soon as we've got the new code right (possibly it already is?),
8745 we have solved a couple of bugs of the earlier code where s->cert
8746 was used as if it could not have been shared with other SSL structures.
8748 Note that using the SSL API in certain dirty ways now will result
8749 in different behaviour than observed with earlier library versions:
8750 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
8751 does not influence s as it used to.
8753 In order to clean up things more thoroughly, inside SSL_SESSION
8754 we don't use CERT any longer, but a new structure SESS_CERT
8755 that holds per-session data (if available); currently, this is
8756 the peer's certificate chain and, for clients, the server's certificate
8757 and temporary key. CERT holds only those values that can have
8758 meaningful defaults in an SSL_CTX.
8761 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
8762 from the internal representation. Various PKCS#7 fixes: remove some
8763 evil casts and set the enc_dig_alg field properly based on the signing
8767 *) Allow PKCS#12 password to be set from the command line or the
8768 environment. Let 'ca' get its config file name from the environment
8769 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
8773 *) Allow certificate policies extension to use an IA5STRING for the
8774 organization field. This is contrary to the PKIX definition but
8775 VeriSign uses it and IE5 only recognises this form. Document 'x509'
8779 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
8780 without disallowing inline assembler and the like for non-pedantic builds.
8783 *) Support Borland C++ builder.
8784 [Janez Jere <jj@void.si>, modified by Ulf Möller]
8789 *) SHA-1 cleanups and performance enhancements.
8790 [Andy Polyakov <appro@fy.chalmers.se>]
8792 *) Sparc v8plus assembler for the bignum library.
8793 [Andy Polyakov <appro@fy.chalmers.se>]
8795 *) Accept any -xxx and +xxx compiler options in Configure.
8798 *) Update HPUX configuration.
8801 *) Add missing sk_<type>_unshift() function to safestack.h
8802 [Ralf S. Engelschall]
8804 *) New function SSL_CTX_use_certificate_chain_file that sets the
8805 "extra_cert"s in addition to the certificate. (This makes sense
8806 only for "PEM" format files, as chains as a whole are not
8810 *) Support verify_depth from the SSL API.
8811 x509_vfy.c had what can be considered an off-by-one-error:
8812 Its depth (which was not part of the external interface)
8813 was actually counting the number of certificates in a chain;
8814 now it really counts the depth.
8817 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
8818 instead of X509err, which often resulted in confusing error
8819 messages since the error codes are not globally unique
8820 (e.g. an alleged error in ssl3_accept when a certificate
8821 didn't match the private key).
8823 *) New function SSL_CTX_set_session_id_context that allows to set a default
8824 value (so that you don't need SSL_set_session_id_context for each
8825 connection using the SSL_CTX).
8828 *) OAEP decoding bug fix.
8831 *) Support INSTALL_PREFIX for package builders, as proposed by
8835 *) New Configure options "threads" and "no-threads". For systems
8836 where the proper compiler options are known (currently Solaris
8837 and Linux), "threads" is the default.
8840 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
8843 *) Install various scripts to $(OPENSSLDIR)/misc, not to
8844 $(INSTALLTOP)/bin -- they shouldn't clutter directories
8845 such as /usr/local/bin.
8848 *) "make linux-shared" to build shared libraries.
8849 [Niels Poppe <niels@netbox.org>]
8851 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
8854 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
8855 extension adding in x509 utility.
8858 *) Remove NOPROTO sections and error code comments.
8861 *) Partial rewrite of the DEF file generator to now parse the ANSI
8865 *) New Configure options --prefix=DIR and --openssldir=DIR.
8868 *) Complete rewrite of the error code script(s). It is all now handled
8869 by one script at the top level which handles error code gathering,
8870 header rewriting and C source file generation. It should be much better
8871 than the old method: it now uses a modified version of Ulf's parser to
8872 read the ANSI prototypes in all header files (thus the old K&R definitions
8873 aren't needed for error creation any more) and do a better job of
8874 translating function codes into names. The old 'ASN1 error code imbedded
8875 in a comment' is no longer necessary and it doesn't use .err files which
8876 have now been deleted. Also the error code call doesn't have to appear all
8877 on one line (which resulted in some large lines...).
8880 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
8883 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
8884 0 (which usually indicates a closed connection), but continue reading.
8887 *) Fix some race conditions.
8890 *) Add support for CRL distribution points extension. Add Certificate
8891 Policies and CRL distribution points documentation.
8894 *) Move the autogenerated header file parts to crypto/opensslconf.h.
8897 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
8898 8 of keying material. Merlin has also confirmed interop with this fix
8899 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
8900 [Merlin Hughes <merlin@baltimore.ie>]
8902 *) Fix lots of warnings.
8903 [Richard Levitte <levitte@stacken.kth.se>]
8905 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
8906 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
8907 [Richard Levitte <levitte@stacken.kth.se>]
8909 *) Fix problems with sizeof(long) == 8.
8910 [Andy Polyakov <appro@fy.chalmers.se>]
8912 *) Change functions to ANSI C.
8915 *) Fix typos in error codes.
8916 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
8918 *) Remove defunct assembler files from Configure.
8921 *) SPARC v8 assembler BIGNUM implementation.
8922 [Andy Polyakov <appro@fy.chalmers.se>]
8924 *) Support for Certificate Policies extension: both print and set.
8925 Various additions to support the r2i method this uses.
8928 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
8929 return a const string when you are expecting an allocated buffer.
8932 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
8933 types DirectoryString and DisplayText.
8936 *) Add code to allow r2i extensions to access the configuration database,
8937 add an LHASH database driver and add several ctx helper functions.
8940 *) Fix an evil bug in bn_expand2() which caused various BN functions to
8941 fail when they extended the size of a BIGNUM.
8944 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
8945 support typesafe stack.
8948 *) Fix typo in SSL_[gs]et_options().
8949 [Nils Frostberg <nils@medcom.se>]
8951 *) Delete various functions and files that belonged to the (now obsolete)
8952 old X509V3 handling code.
8955 *) New Configure option "rsaref".
8958 *) Don't auto-generate pem.h.
8961 *) Introduce type-safe ASN.1 SETs.
8964 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
8965 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
8967 *) Introduce type-safe STACKs. This will almost certainly break lots of code
8968 that links with OpenSSL (well at least cause lots of warnings), but fear
8969 not: the conversion is trivial, and it eliminates loads of evil casts. A
8970 few STACKed things have been converted already. Feel free to convert more.
8971 In the fullness of time, I'll do away with the STACK type altogether.
8974 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
8975 specified in <certfile> by updating the entry in the index.txt file.
8976 This way one no longer has to edit the index.txt file manually for
8977 revoking a certificate. The -revoke option does the gory details now.
8978 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
8980 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
8981 `-text' option at all and this way the `-noout -text' combination was
8982 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
8983 [Ralf S. Engelschall]
8985 *) Make sure a corresponding plain text error message exists for the
8986 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
8987 verify callback function determined that a certificate was revoked.
8988 [Ralf S. Engelschall]
8990 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
8991 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
8992 all available cipers including rc5, which was forgotten until now.
8993 In order to let the testing shell script know which algorithms
8994 are available, a new (up to now undocumented) command
8995 "openssl list-cipher-commands" is used.
8998 *) Bugfix: s_client occasionally would sleep in select() when
8999 it should have checked SSL_pending() first.
9002 *) New functions DSA_do_sign and DSA_do_verify to provide access to
9003 the raw DSA values prior to ASN.1 encoding.
9006 *) Tweaks to Configure
9007 [Niels Poppe <niels@netbox.org>]
9009 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
9013 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
9016 *) New config option to avoid instructions that are illegal on the 80386.
9017 The default code is faster, but requires at least a 486.
9020 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
9021 SSL2_SERVER_VERSION (not used at all) macros, which are now the
9022 same as SSL2_VERSION anyway.
9025 *) New "-showcerts" option for s_client.
9028 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
9029 application. Various cleanups and fixes.
9032 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
9033 modify error routines to work internally. Add error codes and PBE init
9034 to library startup routines.
9037 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
9038 packing functions to asn1 and evp. Changed function names and error
9039 codes along the way.
9042 *) PKCS12 integration: and so it begins... First of several patches to
9043 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
9044 objects to objects.h
9047 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
9048 and display support for Thawte strong extranet extension.
9051 *) Add LinuxPPC support.
9052 [Jeff Dubrule <igor@pobox.org>]
9054 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
9055 bn_div_words in alpha.s.
9056 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
9058 *) Make sure the RSA OAEP test is skipped under -DRSAref because
9059 OAEP isn't supported when OpenSSL is built with RSAref.
9060 [Ulf Moeller <ulf@fitug.de>]
9062 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
9063 so they no longer are missing under -DNOPROTO.
9064 [Soren S. Jorvang <soren@t.dk>]
9067 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
9069 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
9070 doesn't work when the session is reused. Coming soon!
9073 *) Fix a security hole, that allows sessions to be reused in the wrong
9074 context thus bypassing client cert protection! All software that uses
9075 client certs and session caches in multiple contexts NEEDS PATCHING to
9076 allow session reuse! A fuller solution is in the works.
9077 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
9079 *) Some more source tree cleanups (removed obsolete files
9080 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
9081 permission on "config" script to be executable) and a fix for the INSTALL
9083 [Ulf Moeller <ulf@fitug.de>]
9085 *) Remove some legacy and erroneous uses of malloc, free instead of
9087 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
9089 *) Make rsa_oaep_test return non-zero on error.
9090 [Ulf Moeller <ulf@fitug.de>]
9092 *) Add support for native Solaris shared libraries. Configure
9093 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
9094 if someone would make that last step automatic.
9095 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
9097 *) ctx_size was not built with the right compiler during "make links". Fixed.
9100 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
9101 except NULL ciphers". This means the default cipher list will no longer
9102 enable NULL ciphers. They need to be specifically enabled e.g. with
9103 the string "DEFAULT:eNULL".
9106 *) Fix to RSA private encryption routines: if p < q then it would
9107 occasionally produce an invalid result. This will only happen with
9108 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9111 *) Be less restrictive and allow also `perl util/perlpath.pl
9112 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9113 because this way one can also use an interpreter named `perl5' (which is
9114 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9115 installed as `perl').
9116 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9118 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9119 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9121 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9122 advapi32.lib to Win32 build and change the pem test comparision
9123 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9124 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9125 and crypto/des/ede_cbcm_enc.c.
9128 *) DES quad checksum was broken on big-endian architectures. Fixed.
9131 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9132 Win32 test batch file so it (might) work again. The Win32 test batch file
9133 is horrible: I feel ill....
9136 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
9137 in e_os.h. Audit of header files to check ANSI and non ANSI
9138 sections: 10 functions were absent from non ANSI section and not exported
9139 from Windows DLLs. Fixed up libeay.num for new functions.
9142 *) Make `openssl version' output lines consistent.
9143 [Ralf S. Engelschall]
9145 *) Fix Win32 symbol export lists for BIO functions: Added
9146 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
9147 to ms/libeay{16,32}.def.
9148 [Ralf S. Engelschall]
9150 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
9151 fine under Unix and passes some trivial tests I've now added. But the
9152 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
9153 added to make sure no one expects that this stuff really works in the
9154 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
9155 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
9157 [Ralf S. Engelschall]
9159 *) Fix the generation of two part addresses in perl.
9160 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
9162 *) Add config entry for Linux on MIPS.
9163 [John Tobey <jtobey@channel1.com>]
9165 *) Make links whenever Configure is run, unless we are on Windoze.
9168 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
9169 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
9173 *) Add a useful kludge to allow package maintainers to specify compiler and
9174 other platforms details on the command line without having to patch the
9175 Configure script everytime: One now can use ``perl Configure
9176 <id>:<details>'', i.e. platform ids are allowed to have details appended
9177 to them (seperated by colons). This is treated as there would be a static
9178 pre-configured entry in Configure's %table under key <id> with value
9179 <details> and ``perl Configure <id>'' is called. So, when you want to
9180 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
9181 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
9182 now, which overrides the FreeBSD-elf entry on-the-fly.
9183 [Ralf S. Engelschall]
9185 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
9188 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
9189 on the `perl Configure ...' command line. This way one can compile
9190 OpenSSL libraries with Position Independent Code (PIC) which is needed
9191 for linking it into DSOs.
9192 [Ralf S. Engelschall]
9194 *) Remarkably, export ciphers were totally broken and no-one had noticed!
9198 *) Cleaned up the LICENSE document: The official contact for any license
9199 questions now is the OpenSSL core team under openssl-core@openssl.org.
9200 And add a paragraph about the dual-license situation to make sure people
9201 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
9202 to the OpenSSL toolkit.
9203 [Ralf S. Engelschall]
9205 *) General source tree makefile cleanups: Made `making xxx in yyy...'
9206 display consistent in the source tree and replaced `/bin/rm' by `rm'.
9207 Additonally cleaned up the `make links' target: Remove unnecessary
9208 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
9209 to speed processing and no longer clutter the display with confusing
9210 stuff. Instead only the actually done links are displayed.
9211 [Ralf S. Engelschall]
9213 *) Permit null encryption ciphersuites, used for authentication only. It used
9214 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
9215 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
9219 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
9220 signed attributes when verifying signatures (this would break them),
9221 the detached data encoding was wrong and public keys obtained using
9222 X509_get_pubkey() weren't freed.
9225 *) Add text documentation for the BUFFER functions. Also added a work around
9226 to a Win95 console bug. This was triggered by the password read stuff: the
9227 last character typed gets carried over to the next fread(). If you were
9228 generating a new cert request using 'req' for example then the last
9229 character of the passphrase would be CR which would then enter the first
9233 *) Added the new `Includes OpenSSL Cryptography Software' button as
9234 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
9235 button and can be used by applications based on OpenSSL to show the
9236 relationship to the OpenSSL project.
9237 [Ralf S. Engelschall]
9239 *) Remove confusing variables in function signatures in files
9240 ssl/ssl_lib.c and ssl/ssl.h.
9241 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9243 *) Don't install bss_file.c under PREFIX/include/
9244 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9246 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
9247 functions that return function pointers and has support for NT specific
9248 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
9249 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
9250 unsigned to signed types: this was killing the Win32 compile.
9253 *) Add new certificate file to stack functions,
9254 SSL_add_dir_cert_subjects_to_stack() and
9255 SSL_add_file_cert_subjects_to_stack(). These largely supplant
9256 SSL_load_client_CA_file(), and can be used to add multiple certs easily
9257 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
9258 This means that Apache-SSL and similar packages don't have to mess around
9259 to add as many CAs as they want to the preferred list.
9262 *) Experiment with doxygen documentation. Currently only partially applied to
9264 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
9265 openssl.doxy as the configuration file.
9268 *) Get rid of remaining C++-style comments which strict C compilers hate.
9269 [Ralf S. Engelschall, pointed out by Carlos Amengual]
9271 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
9272 compiled in by default: it has problems with large keys.
9275 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
9276 DH private keys and/or callback functions which directly correspond to
9277 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
9278 is needed for applications which have to configure certificates on a
9279 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
9281 For the RSA certificate situation is makes no difference, but
9282 for the DSA certificate situation this fixes the "no shared cipher"
9283 problem where the OpenSSL cipher selection procedure failed because the
9284 temporary keys were not overtaken from the context and the API provided
9285 no way to reconfigure them.
9286 The new functions now let applications reconfigure the stuff and they
9287 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
9288 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
9289 non-public-API function ssl_cert_instantiate() is used as a helper
9290 function and also to reduce code redundancy inside ssl_rsa.c.
9291 [Ralf S. Engelschall]
9293 *) Move s_server -dcert and -dkey options out of the undocumented feature
9294 area because they are useful for the DSA situation and should be
9295 recognized by the users.
9296 [Ralf S. Engelschall]
9298 *) Fix the cipher decision scheme for export ciphers: the export bits are
9299 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
9300 SSL_EXP_MASK. So, the original variable has to be used instead of the
9301 already masked variable.
9302 [Richard Levitte <levitte@stacken.kth.se>]
9304 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
9305 [Richard Levitte <levitte@stacken.kth.se>]
9307 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
9308 from `int' to `unsigned int' because it's a length and initialized by
9309 EVP_DigestFinal() which expects an `unsigned int *'.
9310 [Richard Levitte <levitte@stacken.kth.se>]
9312 *) Don't hard-code path to Perl interpreter on shebang line of Configure
9313 script. Instead use the usual Shell->Perl transition trick.
9314 [Ralf S. Engelschall]
9316 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
9317 (in addition to RSA certificates) to match the behaviour of `openssl dsa
9318 -noout -modulus' as it's already the case for `openssl rsa -noout
9319 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
9320 currently the public key is printed (a decision which was already done by
9321 `openssl dsa -modulus' in the past) which serves a similar purpose.
9322 Additionally the NO_RSA no longer completely removes the whole -modulus
9323 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
9325 [Ralf S. Engelschall]
9327 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
9328 BIO. See the source (crypto/evp/bio_ok.c) for more info.
9329 [Arne Ansper <arne@ats.cyber.ee>]
9331 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
9332 to be added. Now both 'req' and 'ca' can use new objects defined in the
9336 *) Add cool BIO that does syslog (or event log on NT).
9337 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
9339 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
9340 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
9341 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
9342 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
9345 *) Add preliminary config info for new extension code.
9348 *) Make RSA_NO_PADDING really use no padding.
9349 [Ulf Moeller <ulf@fitug.de>]
9351 *) Generate errors when private/public key check is done.
9354 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
9355 for some CRL extensions and new objects added.
9358 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
9359 key usage extension and fuller support for authority key id.
9362 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
9363 padding method for RSA, which is recommended for new applications in PKCS
9364 #1 v2.0 (RFC 2437, October 1998).
9365 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
9366 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
9367 against Bleichbacher's attack on RSA.
9368 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
9371 *) Updates to the new SSL compression code
9372 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9374 *) Fix so that the version number in the master secret, when passed
9375 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
9376 (because the server will not accept higher), that the version number
9377 is 0x03,0x01, not 0x03,0x00
9378 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9380 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
9381 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
9382 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
9385 *) Support for RAW extensions where an arbitrary extension can be
9386 created by including its DER encoding. See apps/openssl.cnf for
9390 *) Make sure latest Perl versions don't interpret some generated C array
9391 code as Perl array code in the crypto/err/err_genc.pl script.
9392 [Lars Weber <3weber@informatik.uni-hamburg.de>]
9394 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
9395 not many people have the assembler. Various Win32 compilation fixes and
9396 update to the INSTALL.W32 file with (hopefully) more accurate Win32
9400 *) Modify configure script 'Configure' to automatically create crypto/date.h
9401 file under Win32 and also build pem.h from pem.org. New script
9402 util/mkfiles.pl to create the MINFO file on environments that can't do a
9403 'make files': perl util/mkfiles.pl >MINFO should work.
9406 *) Major rework of DES function declarations, in the pursuit of correctness
9407 and purity. As a result, many evil casts evaporated, and some weirdness,
9408 too. You may find this causes warnings in your code. Zapping your evil
9409 casts will probably fix them. Mostly.
9412 *) Fix for a typo in asn1.h. Bug fix to object creation script
9413 obj_dat.pl. It considered a zero in an object definition to mean
9414 "end of object": none of the objects in objects.h have any zeros
9415 so it wasn't spotted.
9416 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
9418 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
9419 Masking (CBCM). In the absence of test vectors, the best I have been able
9420 to do is check that the decrypt undoes the encrypt, so far. Send me test
9421 vectors if you have them.
9424 *) Correct calculation of key length for export ciphers (too much space was
9425 allocated for null ciphers). This has not been tested!
9428 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
9429 message is now correct (it understands "crypto" and "ssl" on its
9430 command line). There is also now an "update" option. This will update
9431 the util/ssleay.num and util/libeay.num files with any new functions.
9433 perl util/mkdef.pl crypto ssl update
9434 it will update them.
9437 *) Overhauled the Perl interface (perl/*):
9438 - ported BN stuff to OpenSSL's different BN library
9439 - made the perl/ source tree CVS-aware
9440 - renamed the package from SSLeay to OpenSSL (the files still contain
9441 their history because I've copied them in the repository)
9442 - removed obsolete files (the test scripts will be replaced
9443 by better Test::Harness variants in the future)
9444 [Ralf S. Engelschall]
9446 *) First cut for a very conservative source tree cleanup:
9447 1. merge various obsolete readme texts into doc/ssleay.txt
9448 where we collect the old documents and readme texts.
9449 2. remove the first part of files where I'm already sure that we no
9450 longer need them because of three reasons: either they are just temporary
9451 files which were left by Eric or they are preserved original files where
9452 I've verified that the diff is also available in the CVS via "cvs diff
9453 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
9454 the crypto/md/ stuff).
9455 [Ralf S. Engelschall]
9457 *) More extension code. Incomplete support for subject and issuer alt
9458 name, issuer and authority key id. Change the i2v function parameters
9459 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
9460 what that's for :-) Fix to ASN1 macro which messed up
9461 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
9464 *) Preliminary support for ENUMERATED type. This is largely copied from the
9468 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
9469 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9471 *) Make sure `make rehash' target really finds the `openssl' program.
9472 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9474 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
9475 like to hear about it if this slows down other processors.
9478 *) Add CygWin32 platform information to Configure script.
9479 [Alan Batie <batie@aahz.jf.intel.com>]
9481 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
9482 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
9484 *) New program nseq to manipulate netscape certificate sequences
9487 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
9491 *) Fixes to BN code. Previously the default was to define BN_RECURSION
9492 but the BN code had some problems that would cause failures when
9493 doing certificate verification and some other functions.
9494 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9496 *) Add ASN1 and PEM code to support netscape certificate sequences.
9499 *) Add ASN1 and PEM code to support netscape certificate sequences.
9502 *) Add several PKIX and private extended key usage OIDs.
9505 *) Modify the 'ca' program to handle the new extension code. Modify
9506 openssl.cnf for new extension format, add comments.
9509 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
9510 and add a sample to openssl.cnf so req -x509 now adds appropriate
9514 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
9515 error code, add initial support to X509_print() and x509 application.
9518 *) Takes a deep breath and start addding X509 V3 extension support code. Add
9519 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
9520 stuff is currently isolated and isn't even compiled yet.
9523 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
9524 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
9525 Removed the versions check from X509 routines when loading extensions:
9526 this allows certain broken certificates that don't set the version
9527 properly to be processed.
9530 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
9531 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
9532 can still be regenerated with "make depend".
9535 *) Spelling mistake in C version of CAST-128.
9536 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
9538 *) Changes to the error generation code. The perl script err-code.pl
9539 now reads in the old error codes and retains the old numbers, only
9540 adding new ones if necessary. It also only changes the .err files if new
9541 codes are added. The makefiles have been modified to only insert errors
9542 when needed (to avoid needlessly modifying header files). This is done
9543 by only inserting errors if the .err file is newer than the auto generated
9544 C file. To rebuild all the error codes from scratch (the old behaviour)
9545 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
9546 or delete all the .err files.
9549 *) CAST-128 was incorrectly implemented for short keys. The C version has
9550 been fixed, but is untested. The assembler versions are also fixed, but
9551 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
9552 to regenerate it if needed.
9553 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
9554 Hagino <itojun@kame.net>]
9556 *) File was opened incorrectly in randfile.c.
9557 [Ulf Möller <ulf@fitug.de>]
9559 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
9560 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
9561 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
9562 al: it's just almost always a UTCTime. Note this patch adds new error
9563 codes so do a "make errors" if there are problems.
9566 *) Correct Linux 1 recognition in config.
9567 [Ulf Möller <ulf@fitug.de>]
9569 *) Remove pointless MD5 hash when using DSA keys in ca.
9570 [Anonymous <nobody@replay.com>]
9572 *) Generate an error if given an empty string as a cert directory. Also
9573 generate an error if handed NULL (previously returned 0 to indicate an
9574 error, but didn't set one).
9575 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
9577 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
9580 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
9581 parameters. This was causing a warning which killed off the Win32 compile.
9584 *) Remove C++ style comments from crypto/bn/bn_local.h.
9585 [Neil Costigan <neil.costigan@celocom.com>]
9587 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
9588 based on a text string, looking up short and long names and finally
9589 "dot" format. The "dot" format stuff didn't work. Added new function
9590 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
9591 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
9592 OID is not part of the table.
9595 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
9596 X509_LOOKUP_by_alias().
9599 *) Sort openssl functions by name.
9602 *) Get the gendsa program working (hopefully) and add it to app list. Remove
9603 encryption from sample DSA keys (in case anyone is interested the password
9607 *) Make _all_ *_free functions accept a NULL pointer.
9608 [Frans Heymans <fheymans@isaserver.be>]
9610 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
9612 [Anonymous <nobody@replay.com>]
9614 *) s_server should send the CAfile as acceptable CAs, not its own cert.
9615 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9617 *) Don't blow it for numeric -newkey arguments to apps/req.
9618 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9620 *) Temp key "for export" tests were wrong in s3_srvr.c.
9621 [Anonymous <nobody@replay.com>]
9623 *) Add prototype for temp key callback functions
9624 SSL_CTX_set_tmp_{rsa,dh}_callback().
9627 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
9628 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
9631 *) X509_name_add_entry() freed the wrong thing after an error.
9632 [Arne Ansper <arne@ats.cyber.ee>]
9634 *) rsa_eay.c would attempt to free a NULL context.
9635 [Arne Ansper <arne@ats.cyber.ee>]
9637 *) BIO_s_socket() had a broken should_retry() on Windoze.
9638 [Arne Ansper <arne@ats.cyber.ee>]
9640 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
9641 [Arne Ansper <arne@ats.cyber.ee>]
9643 *) Make sure the already existing X509_STORE->depth variable is initialized
9644 in X509_STORE_new(), but document the fact that this variable is still
9645 unused in the certificate verification process.
9646 [Ralf S. Engelschall]
9648 *) Fix the various library and apps files to free up pkeys obtained from
9649 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
9652 *) Fix reference counting in X509_PUBKEY_get(). This makes
9653 demos/maurice/example2.c work, amongst others, probably.
9654 [Steve Henson and Ben Laurie]
9656 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
9657 `openssl' and second, the shortcut symlinks for the `openssl <command>'
9658 are no longer created. This way we have a single and consistent command
9659 line interface `openssl <command>', similar to `cvs <command>'.
9660 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
9662 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
9663 BIT STRING wrapper always have zero unused bits.
9666 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
9669 *) Make the top-level INSTALL documentation easier to understand.
9672 *) Makefiles updated to exit if an error occurs in a sub-directory
9673 make (including if user presses ^C) [Paul Sutton]
9675 *) Make Montgomery context stuff explicit in RSA data structure.
9678 *) Fix build order of pem and err to allow for generated pem.h.
9681 *) Fix renumbering bug in X509_NAME_delete_entry().
9684 *) Enhanced the err-ins.pl script so it makes the error library number
9685 global and can add a library name. This is needed for external ASN1 and
9686 other error libraries.
9689 *) Fixed sk_insert which never worked properly.
9692 *) Fix ASN1 macros so they can handle indefinite length construted
9693 EXPLICIT tags. Some non standard certificates use these: they can now
9697 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
9698 into a single doc/ssleay.txt bundle. This way the information is still
9699 preserved but no longer messes up this directory. Now it's new room for
9700 the new set of documenation files.
9701 [Ralf S. Engelschall]
9703 *) SETs were incorrectly DER encoded. This was a major pain, because they
9704 shared code with SEQUENCEs, which aren't coded the same. This means that
9705 almost everything to do with SETs or SEQUENCEs has either changed name or
9706 number of arguments.
9707 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
9709 *) Fix test data to work with the above.
9712 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
9713 was already fixed by Eric for 0.9.1 it seems.
9714 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
9716 *) Autodetect FreeBSD3.
9719 *) Fix various bugs in Configure. This affects the following platforms:
9723 unixware-2.0-pentium
9727 *) Eliminate generated files from CVS. Reorder tests to regenerate files
9728 before they are needed.
9731 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
9735 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
9737 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
9738 changed SSLeay to OpenSSL in version strings.
9739 [Ralf S. Engelschall]
9741 *) Some fixups to the top-level documents.
9744 *) Fixed the nasty bug where rsaref.h was not found under compile-time
9745 because the symlink to include/ was missing.
9746 [Ralf S. Engelschall]
9748 *) Incorporated the popular no-RSA/DSA-only patches
9749 which allow to compile a RSA-free SSLeay.
9750 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
9752 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
9753 when "ssleay" is still not found.
9754 [Ralf S. Engelschall]
9756 *) Added more platforms to Configure: Cray T3E, HPUX 11,
9757 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
9759 *) Updated the README file.
9760 [Ralf S. Engelschall]
9762 *) Added various .cvsignore files in the CVS repository subdirs
9763 to make a "cvs update" really silent.
9764 [Ralf S. Engelschall]
9766 *) Recompiled the error-definition header files and added
9767 missing symbols to the Win32 linker tables.
9768 [Ralf S. Engelschall]
9770 *) Cleaned up the top-level documents;
9771 o new files: CHANGES and LICENSE
9772 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
9773 o merged COPYRIGHT into LICENSE
9774 o removed obsolete TODO file
9775 o renamed MICROSOFT to INSTALL.W32
9776 [Ralf S. Engelschall]
9778 *) Removed dummy files from the 0.9.1b source tree:
9779 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
9780 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
9781 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
9782 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
9783 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
9784 [Ralf S. Engelschall]
9786 *) Added various platform portability fixes.
9789 *) The Genesis of the OpenSSL rpject:
9790 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
9791 Young and Tim J. Hudson created while they were working for C2Net until
9793 [The OpenSSL Project]
9796 Changes between 0.9.0b and 0.9.1b [not released]
9798 *) Updated a few CA certificates under certs/
9801 *) Changed some BIGNUM api stuff.
9804 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
9805 DGUX x86, Linux Alpha, etc.
9808 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
9809 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
9813 *) Add -strparse option to asn1pars program which parses nested
9815 [Dr Stephen Henson <shenson@bigfoot.com>]
9817 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
9820 *) DSA fix for "ca" program.
9823 *) Added "-genkey" option to "dsaparam" program.
9826 *) Added RIPE MD160 (rmd160) message digest.
9829 *) Added -a (all) option to "ssleay version" command.
9832 *) Added PLATFORM define which is the id given to Configure.
9835 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
9838 *) Extended the ASN.1 parser routines.
9841 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
9844 *) Added a BN_CTX to the BN library.
9847 *) Fixed the weak key values in DES library
9850 *) Changed API in EVP library for cipher aliases.
9853 *) Added support for RC2/64bit cipher.
9856 *) Converted the lhash library to the crypto/mem.c functions.
9859 *) Added more recognized ASN.1 object ids.
9862 *) Added more RSA padding checks for SSL/TLS.
9865 *) Added BIO proxy/filter functionality.
9868 *) Added extra_certs to SSL_CTX which can be used
9869 send extra CA certificates to the client in the CA cert chain sending
9870 process. It can be configured with SSL_CTX_add_extra_chain_cert().
9873 *) Now Fortezza is denied in the authentication phase because
9874 this is key exchange mechanism is not supported by SSLeay at all.
9877 *) Additional PKCS1 checks.
9880 *) Support the string "TLSv1" for all TLS v1 ciphers.
9883 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
9884 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
9887 *) Fixed a few memory leaks.
9890 *) Fixed various code and comment typos.
9893 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
9894 bytes sent in the client random.
9895 [Edward Bishop <ebishop@spyglass.com>]