5 Changes between 1.0.1r and 1.0.1s [xx XXX xxxx]
7 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
8 is by default disabled at build-time. Builds that are not configured with
9 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
10 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
11 will need to explicitly call either of:
13 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
15 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
17 as appropriate. Even if either of those is used, or the application
18 explicitly uses the version-specific SSLv2_method() or its client and
19 server variants, SSLv2 ciphers vulnerable to exhaustive search key
20 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
21 ciphers, and SSLv2 56-bit DES are no longer available.
24 *) Disable SRP fake user seed to address a server memory leak.
26 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
28 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
29 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
30 was changed to ignore the "fake user" SRP seed, even if the seed
33 Users should use SRP_VBASE_get1_by_user instead. Note that in
34 SRP_VBASE_get1_by_user, caller must free the returned value. Note
35 also that even though configuring the SRP seed attempts to hide
36 invalid usernames by continuing the handshake with fake
37 credentials, this behaviour is not constant time and no strong
38 guarantees are made that the handshake is indistinguishable from
43 Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
45 *) Protection for DH small subgroup attacks
47 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
48 switched on by default and cannot be disabled. This could have some
52 *) SSLv2 doesn't block disabled ciphers
54 A malicious client can negotiate SSLv2 ciphers that have been disabled on
55 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
56 been disabled, provided that the SSLv2 protocol was not also disabled via
59 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
60 and Sebastian Schinzel.
64 *) Reject DH handshakes with parameters shorter than 1024 bits.
67 Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
69 *) Certificate verify crash with missing PSS parameter
71 The signature verification routines will crash with a NULL pointer
72 dereference if presented with an ASN.1 signature using the RSA PSS
73 algorithm and absent mask generation function parameter. Since these
74 routines are used to verify certificate signature algorithms this can be
75 used to crash any certificate verification operation and exploited in a
76 DoS attack. Any application which performs certificate verification is
77 vulnerable including OpenSSL clients and servers which enable client
80 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
84 *) X509_ATTRIBUTE memory leak
86 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
87 memory. This structure is used by the PKCS#7 and CMS routines so any
88 application which reads PKCS#7 or CMS data from untrusted sources is
89 affected. SSL/TLS is not affected.
91 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
96 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
97 This changes the decoding behaviour for some invalid messages,
98 though the change is mostly in the more lenient direction, and
99 legacy behaviour is preserved as much as possible.
102 *) In DSA_generate_parameters_ex, if the provided seed is too short,
103 use a random seed, as already documented.
104 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
106 Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
108 *) Alternate chains certificate forgery
110 During certificate verfification, OpenSSL will attempt to find an
111 alternative certificate chain if the first attempt to build such a chain
112 fails. An error in the implementation of this logic can mean that an
113 attacker could cause certain checks on untrusted certificates to be
114 bypassed, such as the CA flag, enabling them to use a valid leaf
115 certificate to act as a CA and "issue" an invalid certificate.
117 This issue was reported to OpenSSL by Adam Langley/David Benjamin
122 *) Race condition handling PSK identify hint
124 If PSK identity hints are received by a multi-threaded client then
125 the values are wrongly updated in the parent SSL_CTX structure. This can
126 result in a race condition potentially leading to a double free of the
131 Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
132 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
133 incompatibility in the handling of HMAC. The previous ABI has now been
136 Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
138 *) Malformed ECParameters causes infinite loop
140 When processing an ECParameters structure OpenSSL enters an infinite loop
141 if the curve specified is over a specially malformed binary polynomial
144 This can be used to perform denial of service against any
145 system which processes public keys, certificate requests or
146 certificates. This includes TLS clients and TLS servers with
147 client authentication enabled.
149 This issue was reported to OpenSSL by Joseph Barr-Pixton.
153 *) Exploitable out-of-bounds read in X509_cmp_time
155 X509_cmp_time does not properly check the length of the ASN1_TIME
156 string and can read a few bytes out of bounds. In addition,
157 X509_cmp_time accepts an arbitrary number of fractional seconds in the
160 An attacker can use this to craft malformed certificates and CRLs of
161 various sizes and potentially cause a segmentation fault, resulting in
162 a DoS on applications that verify certificates or CRLs. TLS clients
163 that verify CRLs are affected. TLS clients and servers with client
164 authentication enabled may be affected if they use custom verification
167 This issue was reported to OpenSSL by Robert Swiecki (Google), and
168 independently by Hanno Böck.
172 *) PKCS7 crash with missing EnvelopedContent
174 The PKCS#7 parsing code does not handle missing inner EncryptedContent
175 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
176 with missing content and trigger a NULL pointer dereference on parsing.
178 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
179 structures from untrusted sources are affected. OpenSSL clients and
180 servers are not affected.
182 This issue was reported to OpenSSL by Michal Zalewski (Google).
186 *) CMS verify infinite loop with unknown hash function
188 When verifying a signedData message the CMS code can enter an infinite loop
189 if presented with an unknown hash function OID. This can be used to perform
190 denial of service against any system which verifies signedData messages using
192 This issue was reported to OpenSSL by Johannes Bauer.
196 *) Race condition handling NewSessionTicket
198 If a NewSessionTicket is received by a multi-threaded client when attempting to
199 reuse a previous ticket then a race condition can occur potentially leading to
200 a double free of the ticket data.
204 *) Reject DH handshakes with parameters shorter than 768 bits.
205 [Kurt Roeckx and Emilia Kasper]
207 *) dhparam: generate 2048-bit parameters by default.
208 [Kurt Roeckx and Emilia Kasper]
210 Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
212 *) Segmentation fault in ASN1_TYPE_cmp fix
214 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
215 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
216 certificate signature algorithm consistency this can be used to crash any
217 certificate verification operation and exploited in a DoS attack. Any
218 application which performs certificate verification is vulnerable including
219 OpenSSL clients and servers which enable client authentication.
223 *) ASN.1 structure reuse memory corruption fix
225 Reusing a structure in ASN.1 parsing may allow an attacker to cause
226 memory corruption via an invalid write. Such reuse is and has been
227 strongly discouraged and is believed to be rare.
229 Applications that parse structures containing CHOICE or ANY DEFINED BY
230 components may be affected. Certificate parsing (d2i_X509 and related
231 functions) are however not affected. OpenSSL clients and servers are
236 *) PKCS7 NULL pointer dereferences fix
238 The PKCS#7 parsing code does not handle missing outer ContentInfo
239 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
240 missing content and trigger a NULL pointer dereference on parsing.
242 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
243 otherwise parse PKCS#7 structures from untrusted sources are
244 affected. OpenSSL clients and servers are not affected.
246 This issue was reported to OpenSSL by Michal Zalewski (Google).
250 *) DoS via reachable assert in SSLv2 servers fix
252 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
253 servers that both support SSLv2 and enable export cipher suites by sending
254 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
256 This issue was discovered by Sean Burford (Google) and Emilia Käsper
257 (OpenSSL development team).
261 *) Use After Free following d2i_ECPrivatekey error fix
263 A malformed EC private key file consumed via the d2i_ECPrivateKey function
264 could cause a use after free condition. This, in turn, could cause a double
265 free in several private key parsing functions (such as d2i_PrivateKey
266 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
267 for applications that receive EC private keys from untrusted
268 sources. This scenario is considered rare.
270 This issue was discovered by the BoringSSL project and fixed in their
275 *) X509_to_X509_REQ NULL pointer deref fix
277 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
278 the certificate key is invalid. This function is rarely used in practice.
280 This issue was discovered by Brian Carpenter.
284 *) Removed the export ciphers from the DEFAULT ciphers
287 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
289 *) Build fixes for the Windows and OpenVMS platforms
290 [Matt Caswell and Richard Levitte]
292 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
294 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
295 message can cause a segmentation fault in OpenSSL due to a NULL pointer
296 dereference. This could lead to a Denial Of Service attack. Thanks to
297 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
301 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
302 dtls1_buffer_record function under certain conditions. In particular this
303 could occur if an attacker sent repeated DTLS records with the same
304 sequence number but for the next epoch. The memory leak could be exploited
305 by an attacker in a Denial of Service attack through memory exhaustion.
306 Thanks to Chris Mueller for reporting this issue.
310 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
311 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
312 method would be set to NULL which could later result in a NULL pointer
313 dereference. Thanks to Frank Schmirler for reporting this issue.
317 *) Abort handshake if server key exchange message is omitted for ephemeral
320 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
321 reporting this issue.
325 *) Remove non-export ephemeral RSA code on client and server. This code
326 violated the TLS standard by allowing the use of temporary RSA keys in
327 non-export ciphersuites and could be used by a server to effectively
328 downgrade the RSA key length used to a value smaller than the server
329 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
330 INRIA or reporting this issue.
334 *) Fixed issue where DH client certificates are accepted without verification.
335 An OpenSSL server will accept a DH certificate for client authentication
336 without the certificate verify message. This effectively allows a client to
337 authenticate without the use of a private key. This only affects servers
338 which trust a client certificate authority which issues certificates
339 containing DH keys: these are extremely rare and hardly ever encountered.
340 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
345 *) Ensure that the session ID context of an SSL is updated when its
346 SSL_CTX is updated via SSL_set_SSL_CTX.
348 The session ID context is typically set from the parent SSL_CTX,
349 and can vary with the CTX.
352 *) Fix various certificate fingerprint issues.
354 By using non-DER or invalid encodings outside the signed portion of a
355 certificate the fingerprint can be changed without breaking the signature.
356 Although no details of the signed portion of the certificate can be changed
357 this can cause problems with some applications: e.g. those using the
358 certificate fingerprint for blacklists.
360 1. Reject signatures with non zero unused bits.
362 If the BIT STRING containing the signature has non zero unused bits reject
363 the signature. All current signature algorithms require zero unused bits.
365 2. Check certificate algorithm consistency.
367 Check the AlgorithmIdentifier inside TBS matches the one in the
368 certificate signature. NB: this will result in signature failure
369 errors for some broken certificates.
371 Thanks to Konrad Kraszewski from Google for reporting this issue.
373 3. Check DSA/ECDSA signatures use DER.
375 Reencode DSA/ECDSA signatures and compare with the original received
376 signature. Return an error if there is a mismatch.
378 This will reject various cases including garbage after signature
379 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
380 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
381 (negative or with leading zeroes).
383 Further analysis was conducted and fixes were developed by Stephen Henson
384 of the OpenSSL core team.
389 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
390 results on some platforms, including x86_64. This bug occurs at random
391 with a very low probability, and is not known to be exploitable in any
392 way, though its exact impact is difficult to determine. Thanks to Pieter
393 Wuille (Blockstream) who reported this issue and also suggested an initial
394 fix. Further analysis was conducted by the OpenSSL development team and
395 Adam Langley of Google. The final fix was developed by Andy Polyakov of
396 the OpenSSL core team.
400 *) Do not resume sessions on the server if the negotiated protocol
401 version does not match the session's version. Resuming with a different
402 version, while not strictly forbidden by the RFC, is of questionable
403 sanity and breaks all known clients.
404 [David Benjamin, Emilia Käsper]
406 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
407 early CCS messages during renegotiation. (Note that because
408 renegotiation is encrypted, this early CCS was not exploitable.)
411 *) Tighten client-side session ticket handling during renegotiation:
412 ensure that the client only accepts a session ticket if the server sends
413 the extension anew in the ServerHello. Previously, a TLS client would
414 reuse the old extension state and thus accept a session ticket if one was
415 announced in the initial ServerHello.
417 Similarly, ensure that the client requires a session ticket if one
418 was advertised in the ServerHello. Previously, a TLS client would
419 ignore a missing NewSessionTicket message.
422 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
426 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
427 sends a carefully crafted handshake message, to cause OpenSSL to fail
428 to free up to 64k of memory causing a memory leak. This could be
429 exploited in a Denial Of Service attack. This issue affects OpenSSL
430 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
431 whether SRTP is used or configured. Implementations of OpenSSL that
432 have been compiled with OPENSSL_NO_SRTP defined are not affected.
434 The fix was developed by the OpenSSL team.
438 *) Session Ticket Memory Leak.
440 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
441 integrity of that ticket is first verified. In the event of a session
442 ticket integrity check failing, OpenSSL will fail to free memory
443 causing a memory leak. By sending a large number of invalid session
444 tickets an attacker could exploit this issue in a Denial Of Service
449 *) Build option no-ssl3 is incomplete.
451 When OpenSSL is configured with "no-ssl3" as a build option, servers
452 could accept and complete a SSL 3.0 handshake, and clients could be
453 configured to send them.
455 [Akamai and the OpenSSL team]
457 *) Add support for TLS_FALLBACK_SCSV.
458 Client applications doing fallback retries should call
459 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
461 [Adam Langley, Bodo Moeller]
463 *) Add additional DigestInfo checks.
465 Reencode DigestInto in DER and check against the original when
466 verifying RSA signature: this will reject any improperly encoded
467 DigestInfo structures.
469 Note: this is a precautionary measure and no attacks are currently known.
473 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
475 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
476 SRP code can be overrun an internal buffer. Add sanity check that
477 g, A, B < N to SRP code.
479 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
480 Group for discovering this issue.
484 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
485 TLS 1.0 instead of higher protocol versions when the ClientHello message
486 is badly fragmented. This allows a man-in-the-middle attacker to force a
487 downgrade to TLS 1.0 even if both the server and the client support a
488 higher protocol version, by modifying the client's TLS records.
490 Thanks to David Benjamin and Adam Langley (Google) for discovering and
491 researching this issue.
495 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
496 to a denial of service attack. A malicious server can crash the client
497 with a null pointer dereference (read) by specifying an anonymous (EC)DH
498 ciphersuite and sending carefully crafted handshake messages.
500 Thanks to Felix Gröbert (Google) for discovering and researching this
505 *) By sending carefully crafted DTLS packets an attacker could cause openssl
506 to leak memory. This can be exploited through a Denial of Service attack.
507 Thanks to Adam Langley for discovering and researching this issue.
511 *) An attacker can force openssl to consume large amounts of memory whilst
512 processing DTLS handshake messages. This can be exploited through a
513 Denial of Service attack.
514 Thanks to Adam Langley for discovering and researching this issue.
518 *) An attacker can force an error condition which causes openssl to crash
519 whilst processing DTLS packets due to memory being freed twice. This
520 can be exploited through a Denial of Service attack.
521 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
526 *) If a multithreaded client connects to a malicious server using a resumed
527 session and the server sends an ec point format extension it could write
528 up to 255 bytes to freed memory.
530 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
535 *) A malicious server can crash an OpenSSL client with a null pointer
536 dereference (read) by specifying an SRP ciphersuite even though it was not
537 properly negotiated with the client. This can be exploited through a
538 Denial of Service attack.
540 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
541 discovering and researching this issue.
545 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
546 X509_name_oneline, X509_name_print_ex et al. to leak some information
547 from the stack. Applications may be affected if they echo pretty printing
548 output to the attacker.
550 Thanks to Ivan Fratric (Google) for discovering this issue.
552 [Emilia Käsper, and Steve Henson]
554 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
555 for corner cases. (Certain input points at infinity could lead to
556 bogus results, with non-infinity inputs mapped to infinity too.)
559 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
561 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
562 handshake can force the use of weak keying material in OpenSSL
563 SSL/TLS clients and servers.
565 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
566 researching this issue. (CVE-2014-0224)
567 [KIKUCHI Masashi, Steve Henson]
569 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
570 OpenSSL DTLS client the code can be made to recurse eventually crashing
573 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
575 [Imre Rad, Steve Henson]
577 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
578 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
579 client or server. This is potentially exploitable to run arbitrary
580 code on a vulnerable client or server.
582 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
583 [Jüri Aedla, Steve Henson]
585 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
586 are subject to a denial of service attack.
588 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
589 this issue. (CVE-2014-3470)
590 [Felix Gröbert, Ivan Fratric, Steve Henson]
592 *) Harmonize version and its documentation. -f flag is used to display
594 [mancha <mancha1@zoho.com>]
596 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
598 [mancha <mancha1@zoho.com>]
600 *) Fix some double frees. These are not thought to be exploitable.
601 [mancha <mancha1@zoho.com>]
603 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
605 *) A missing bounds check in the handling of the TLS heartbeat extension
606 can be used to reveal up to 64k of memory to a connected client or
609 Thanks for Neel Mehta of Google Security for discovering this bug and to
610 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
611 preparing the fix (CVE-2014-0160)
612 [Adam Langley, Bodo Moeller]
614 *) Fix for the attack described in the paper "Recovering OpenSSL
615 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
616 by Yuval Yarom and Naomi Benger. Details can be obtained from:
617 http://eprint.iacr.org/2014/140
619 Thanks to Yuval Yarom and Naomi Benger for discovering this
620 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
621 [Yuval Yarom and Naomi Benger]
623 *) TLS pad extension: draft-agl-tls-padding-03
625 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
626 TLS client Hello record length value would otherwise be > 255 and
627 less that 512 pad with a dummy extension containing zeroes so it
628 is at least 512 bytes long.
630 [Adam Langley, Steve Henson]
632 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
634 *) Fix for TLS record tampering bug. A carefully crafted invalid
635 handshake could crash OpenSSL with a NULL pointer exception.
636 Thanks to Anton Johansson for reporting this issues.
639 *) Keep original DTLS digest and encryption contexts in retransmission
640 structures so we can use the previous session parameters if they need
641 to be resent. (CVE-2013-6450)
644 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
645 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
646 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
647 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
648 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
649 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
650 [Rob Stradling, Adam Langley]
652 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
654 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
655 supporting platforms or when small records were transferred.
656 [Andy Polyakov, Steve Henson]
658 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
660 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
662 This addresses the flaw in CBC record processing discovered by
663 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
664 at: http://www.isg.rhul.ac.uk/tls/
666 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
667 Security Group at Royal Holloway, University of London
668 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
669 Emilia Käsper for the initial patch.
671 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
673 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
674 ciphersuites which can be exploited in a denial of service attack.
675 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
676 and detecting this bug and to Wolfgang Ettlinger
677 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
681 *) Return an error when checking OCSP signatures when key is NULL.
682 This fixes a DoS attack. (CVE-2013-0166)
685 *) Make openssl verify return errors.
686 [Chris Palmer <palmer@google.com> and Ben Laurie]
688 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
689 the right response is stapled. Also change SSL_get_certificate()
690 so it returns the certificate actually sent.
691 See http://rt.openssl.org/Ticket/Display.html?id=2836.
692 [Rob Stradling <rob.stradling@comodo.com>]
694 *) Fix possible deadlock when decoding public keys.
697 *) Don't use TLS 1.0 record version number in initial client hello
701 Changes between 1.0.1b and 1.0.1c [10 May 2012]
703 *) Sanity check record length before skipping explicit IV in TLS
704 1.2, 1.1 and DTLS to fix DoS attack.
706 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
707 fuzzing as a service testing platform.
711 *) Initialise tkeylen properly when encrypting CMS messages.
712 Thanks to Solar Designer of Openwall for reporting this issue.
715 *) In FIPS mode don't try to use composite ciphers as they are not
719 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
721 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
722 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
723 mean any application compiled against OpenSSL 1.0.0 headers setting
724 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
725 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
726 0x10000000L Any application which was previously compiled against
727 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
728 will need to be recompiled as a result. Letting be results in
729 inability to disable specifically TLS 1.1 and in client context,
730 in unlike event, limit maximum offered version to TLS 1.0 [see below].
733 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
734 disable just protocol X, but all protocols above X *if* there are
735 protocols *below* X still enabled. In more practical terms it means
736 that if application wants to disable TLS1.0 in favor of TLS1.1 and
737 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
738 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
742 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
744 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
745 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
746 in CRYPTO_realloc_clean.
748 Thanks to Tavis Ormandy, Google Security Team, for discovering this
749 issue and to Adam Langley <agl@chromium.org> for fixing it.
751 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
753 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
756 *) Workarounds for some broken servers that "hang" if a client hello
757 record length exceeds 255 bytes.
759 1. Do not use record version number > TLS 1.0 in initial client
760 hello: some (but not all) hanging servers will now work.
761 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
762 the number of ciphers sent in the client hello. This should be
763 set to an even number, such as 50, for example by passing:
764 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
765 Most broken servers should now work.
766 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
767 TLS 1.2 client support entirely.
770 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
773 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
775 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
776 STRING form instead of a DigestInfo.
779 *) The format used for MDC2 RSA signatures is inconsistent between EVP
780 and the RSA_sign/RSA_verify functions. This was made more apparent when
781 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
782 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
783 the correct format in RSA_verify so both forms transparently work.
786 *) Some servers which support TLS 1.0 can choke if we initially indicate
787 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
788 encrypted premaster secret. As a workaround use the maximum pemitted
789 client version in client hello, this should keep such servers happy
790 and still work with previous versions of OpenSSL.
793 *) Add support for TLS/DTLS heartbeats.
794 [Robin Seggelmann <seggelmann@fh-muenster.de>]
796 *) Add support for SCTP.
797 [Robin Seggelmann <seggelmann@fh-muenster.de>]
799 *) Improved PRNG seeding for VOS.
800 [Paul Green <Paul.Green@stratus.com>]
802 *) Extensive assembler packs updates, most notably:
804 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
805 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
806 - x86_64: bit-sliced AES implementation;
807 - ARM: NEON support, contemporary platforms optimizations;
808 - s390x: z196 support;
809 - *: GHASH and GF(2^m) multiplication implementations;
813 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
814 (removal of unnecessary code)
815 [Peter Sylvester <peter.sylvester@edelweb.fr>]
817 *) Add TLS key material exporter from RFC 5705.
820 *) Add DTLS-SRTP negotiation from RFC 5764.
823 *) Add Next Protocol Negotiation,
824 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
825 disabled with a no-npn flag to config or Configure. Code donated
827 [Adam Langley <agl@google.com> and Ben Laurie]
829 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
830 NIST-P256, NIST-P521, with constant-time single point multiplication on
831 typical inputs. Compiler support for the nonstandard type __uint128_t is
832 required to use this (present in gcc 4.4 and later, for 64-bit builds).
833 Code made available under Apache License version 2.0.
835 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
836 line to include this in your build of OpenSSL, and run "make depend" (or
837 "make update"). This enables the following EC_METHODs:
839 EC_GFp_nistp224_method()
840 EC_GFp_nistp256_method()
841 EC_GFp_nistp521_method()
843 EC_GROUP_new_by_curve_name() will automatically use these (while
844 EC_GROUP_new_curve_GFp() currently prefers the more flexible
846 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
848 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
849 all platforms. Move ssize_t definition from e_os.h to the public
850 header file e_os2.h as it now appears in public header file cms.h
853 *) New -sigopt option to the ca, req and x509 utilities. Additional
854 signature parameters can be passed using this option and in
858 *) Add RSA PSS signing function. This will generate and set the
859 appropriate AlgorithmIdentifiers for PSS based on those in the
860 corresponding EVP_MD_CTX structure. No application support yet.
863 *) Support for companion algorithm specific ASN1 signing routines.
864 New function ASN1_item_sign_ctx() signs a pre-initialised
865 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
866 the appropriate parameters.
869 *) Add new algorithm specific ASN1 verification initialisation function
870 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
871 handling will be the same no matter what EVP_PKEY_METHOD is used.
872 Add a PSS handler to support verification of PSS signatures: checked
873 against a number of sample certificates.
876 *) Add signature printing for PSS. Add PSS OIDs.
877 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
879 *) Add algorithm specific signature printing. An individual ASN1 method
880 can now print out signatures instead of the standard hex dump.
882 More complex signatures (e.g. PSS) can print out more meaningful
883 information. Include DSA version that prints out the signature
887 *) Password based recipient info support for CMS library: implementing
891 *) Split password based encryption into PBES2 and PBKDF2 functions. This
892 neatly separates the code into cipher and PBE sections and is required
893 for some algorithms that split PBES2 into separate pieces (such as
897 *) Session-handling fixes:
898 - Fix handling of connections that are resuming with a session ID,
899 but also support Session Tickets.
900 - Fix a bug that suppressed issuing of a new ticket if the client
901 presented a ticket with an expired session.
902 - Try to set the ticket lifetime hint to something reasonable.
903 - Make tickets shorter by excluding irrelevant information.
904 - On the client side, don't ignore renewed tickets.
905 [Adam Langley, Bodo Moeller (Google)]
907 *) Fix PSK session representation.
910 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
912 This work was sponsored by Intel.
915 *) Add GCM support to TLS library. Some custom code is needed to split
916 the IV between the fixed (from PRF) and explicit (from TLS record)
917 portions. This adds all GCM ciphersuites supported by RFC5288 and
918 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
919 add a special AESGCM string for GCM only.
922 *) Expand range of ctrls for AES GCM. Permit setting invocation
923 field on decrypt and retrieval of invocation field only on encrypt.
926 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
927 As required by RFC5289 these ciphersuites cannot be used if for
928 versions of TLS earlier than 1.2.
931 *) For FIPS capable OpenSSL interpret a NULL default public key method
932 as unset and return the appopriate default but do *not* set the default.
933 This means we can return the appopriate method in applications that
934 swicth between FIPS and non-FIPS modes.
937 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
938 ENGINE is used then we cannot handle that in the FIPS module so we
939 keep original code iff non-FIPS operations are allowed.
942 *) Add -attime option to openssl utilities.
943 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
945 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
948 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
949 FIPS EC methods unconditionally for now.
952 *) New build option no-ec2m to disable characteristic 2 code.
955 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
956 all cases can be covered as some introduce binary incompatibilities.
959 *) Redirect RSA operations to FIPS module including keygen,
960 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
963 *) Add similar low level API blocking to ciphers.
966 *) Low level digest APIs are not approved in FIPS mode: any attempt
967 to use these will cause a fatal error. Applications that *really* want
968 to use them can use the private_* version instead.
971 *) Redirect cipher operations to FIPS module for FIPS builds.
974 *) Redirect digest operations to FIPS module for FIPS builds.
977 *) Update build system to add "fips" flag which will link in fipscanister.o
978 for static and shared library builds embedding a signature if needed.
981 *) Output TLS supported curves in preference order instead of numerical
982 order. This is currently hardcoded for the highest order curves first.
983 This should be configurable so applications can judge speed vs strength.
986 *) Add TLS v1.2 server support for client authentication.
989 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
993 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
994 FIPS modules versions.
997 *) Add TLS v1.2 client side support for client authentication. Keep cache
998 of handshake records longer as we don't know the hash algorithm to use
999 until after the certificate request message is received.
1002 *) Initial TLS v1.2 client support. Add a default signature algorithms
1003 extension including all the algorithms we support. Parse new signature
1004 format in client key exchange. Relax some ECC signing restrictions for
1005 TLS v1.2 as indicated in RFC5246.
1008 *) Add server support for TLS v1.2 signature algorithms extension. Switch
1009 to new signature format when needed using client digest preference.
1010 All server ciphersuites should now work correctly in TLS v1.2. No client
1011 support yet and no support for client certificates.
1014 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
1015 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
1016 ciphersuites. At present only RSA key exchange ciphersuites work with
1017 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
1018 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
1019 and version checking.
1022 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
1023 with this defined it will not be affected by any changes to ssl internal
1024 structures. Add several utility functions to allow openssl application
1025 to work with OPENSSL_NO_SSL_INTERN defined.
1029 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
1031 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
1034 *) Permit abbreviated handshakes when renegotiating using the function
1035 SSL_renegotiate_abbreviated().
1036 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1038 *) Add call to ENGINE_register_all_complete() to
1039 ENGINE_load_builtin_engines(), so some implementations get used
1040 automatically instead of needing explicit application support.
1043 *) Add support for TLS key exporter as described in RFC5705.
1044 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
1046 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
1047 a few changes are required:
1049 Add SSL_OP_NO_TLSv1_1 flag.
1050 Add TLSv1_1 methods.
1051 Update version checking logic to handle version 1.1.
1052 Add explicit IV handling (ported from DTLS code).
1053 Add command line options to s_client/s_server.
1056 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
1058 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1059 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1060 content decryption and always return the same error. Note: this attack
1061 needs on average 2^20 messages so it only affects automated senders. The
1062 old behaviour can be reenabled in the CMS code by setting the
1063 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1064 an MMA defence is not necessary.
1065 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1066 this issue. (CVE-2012-0884)
1069 *) Fix CVE-2011-4619: make sure we really are receiving a
1070 client hello before rejecting multiple SGC restarts. Thanks to
1071 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1074 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
1076 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1077 Thanks to Antonio Martin, Enterprise Secure Access Research and
1078 Development, Cisco Systems, Inc. for discovering this bug and
1079 preparing a fix. (CVE-2012-0050)
1082 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
1084 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1085 of the Vaudenay padding oracle attack on CBC mode encryption
1086 which enables an efficient plaintext recovery attack against
1087 the OpenSSL implementation of DTLS. Their attack exploits timing
1088 differences arising during decryption processing. A research
1089 paper describing this attack can be found at:
1090 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1091 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1092 Security Group at Royal Holloway, University of London
1093 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1094 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1095 for preparing the fix. (CVE-2011-4108)
1096 [Robin Seggelmann, Michael Tuexen]
1098 *) Clear bytes used for block padding of SSL 3.0 records.
1100 [Adam Langley (Google)]
1102 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1103 Kadianakis <desnacked@gmail.com> for discovering this issue and
1104 Adam Langley for preparing the fix. (CVE-2011-4619)
1105 [Adam Langley (Google)]
1107 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
1108 [Andrey Kulikov <amdeich@gmail.com>]
1110 *) Prevent malformed RFC3779 data triggering an assertion failure.
1111 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1112 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1113 [Rob Austein <sra@hactrn.net>]
1115 *) Improved PRNG seeding for VOS.
1116 [Paul Green <Paul.Green@stratus.com>]
1118 *) Fix ssl_ciph.c set-up race.
1119 [Adam Langley (Google)]
1121 *) Fix spurious failures in ecdsatest.c.
1122 [Emilia Käsper (Google)]
1124 *) Fix the BIO_f_buffer() implementation (which was mixing different
1125 interpretations of the '..._len' fields).
1126 [Adam Langley (Google)]
1128 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1129 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1130 threads won't reuse the same blinding coefficients.
1132 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1133 lock to call BN_BLINDING_invert_ex, and avoids one use of
1134 BN_BLINDING_update for each BN_BLINDING structure (previously,
1135 the last update always remained unused).
1136 [Emilia Käsper (Google)]
1138 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
1139 [Bob Buckholz (Google)]
1141 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
1143 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
1144 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
1145 [Kaspar Brand <ossl@velox.ch>]
1147 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1148 for multi-threaded use of ECDH. (CVE-2011-3210)
1149 [Adam Langley (Google)]
1151 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1154 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
1155 signature public key algorithm by using OID xref utilities instead.
1156 Before this you could only use some ECC ciphersuites with SHA1 only.
1159 *) Add protection against ECDSA timing attacks as mentioned in the paper
1160 by Billy Bob Brumley and Nicola Tuveri, see:
1162 http://eprint.iacr.org/2011/232.pdf
1164 [Billy Bob Brumley and Nicola Tuveri]
1166 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
1168 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1169 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1171 *) Fix bug in string printing code: if *any* escaping is enabled we must
1172 escape the escape character (backslash) or the resulting string is
1176 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
1178 *) Disable code workaround for ancient and obsolete Netscape browsers
1179 and servers: an attacker can use it in a ciphersuite downgrade attack.
1180 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1183 *) Fixed J-PAKE implementation error, originally discovered by
1184 Sebastien Martini, further info and confirmation from Stefan
1185 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1188 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
1190 *) Fix extension code to avoid race conditions which can result in a buffer
1191 overrun vulnerability: resumed sessions must not be modified as they can
1192 be shared by multiple threads. CVE-2010-3864
1195 *) Fix WIN32 build system to correctly link an ENGINE directory into
1199 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
1201 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
1203 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
1205 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
1207 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
1208 context. The operation can be customised via the ctrl mechanism in
1209 case ENGINEs want to include additional functionality.
1212 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
1215 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
1216 output hashes compatible with older versions of OpenSSL.
1217 [Willy Weisz <weisz@vcpc.univie.ac.at>]
1219 *) Fix compression algorithm handling: if resuming a session use the
1220 compression algorithm of the resumed session instead of determining
1221 it from client hello again. Don't allow server to change algorithm.
1224 *) Add load_crls() function to apps tidying load_certs() too. Add option
1225 to verify utility to allow additional CRLs to be included.
1228 *) Update OCSP request code to permit adding custom headers to the request:
1229 some responders need this.
1232 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
1234 [Julia Lawall <julia@diku.dk>]
1236 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
1237 needlessly dereferenced structures, used obsolete functions and
1238 didn't handle all updated verify codes correctly.
1241 *) Disable MD2 in the default configuration.
1244 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
1245 indicate the initial BIO being pushed or popped. This makes it possible
1246 to determine whether the BIO is the one explicitly called or as a result
1247 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
1248 it handles reference counts correctly and doesn't zero out the I/O bio
1249 when it is not being explicitly popped. WARNING: applications which
1250 included workarounds for the old buggy behaviour will need to be modified
1251 or they could free up already freed BIOs.
1254 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
1255 renaming to all platforms (within the 0.9.8 branch, this was
1256 done conditionally on Netware platforms to avoid a name clash).
1257 [Guenter <lists@gknw.net>]
1259 *) Add ECDHE and PSK support to DTLS.
1260 [Michael Tuexen <tuexen@fh-muenster.de>]
1262 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
1266 *) Add "missing" function EVP_MD_flags() (without this the only way to
1267 retrieve a digest flags is by accessing the structure directly. Update
1268 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
1269 or cipher is registered as in the "from" argument. Print out all
1270 registered digests in the dgst usage message instead of manually
1271 attempting to work them out.
1274 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
1275 this allows the use of compression and extensions. Change default cipher
1276 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
1277 by default unless an application cipher string requests it.
1280 *) Alter match criteria in PKCS12_parse(). It used to try to use local
1281 key ids to find matching certificates and keys but some PKCS#12 files
1282 don't follow the (somewhat unwritten) rules and this strategy fails.
1283 Now just gather all certificates together and the first private key
1284 then look for the first certificate that matches the key.
1287 *) Support use of registered digest and cipher names for dgst and cipher
1288 commands instead of having to add each one as a special case. So now
1295 openssl dgst -sha256 foo
1297 and this works for ENGINE based algorithms too.
1301 *) Update Gost ENGINE to support parameter files.
1302 [Victor B. Wagner <vitus@cryptocom.ru>]
1304 *) Support GeneralizedTime in ca utility.
1305 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
1307 *) Enhance the hash format used for certificate directory links. The new
1308 form uses the canonical encoding (meaning equivalent names will work
1309 even if they aren't identical) and uses SHA1 instead of MD5. This form
1310 is incompatible with the older format and as a result c_rehash should
1311 be used to rebuild symbolic links.
1314 *) Make PKCS#8 the default write format for private keys, replacing the
1315 traditional format. This form is standardised, more secure and doesn't
1316 include an implicit MD5 dependency.
1319 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
1320 committed to OpenSSL should pass this lot as a minimum.
1323 *) Add session ticket override functionality for use by EAP-FAST.
1324 [Jouni Malinen <j@w1.fi>]
1326 *) Modify HMAC functions to return a value. Since these can be implemented
1327 in an ENGINE errors can occur.
1330 *) Type-checked OBJ_bsearch_ex.
1333 *) Type-checked OBJ_bsearch. Also some constification necessitated
1334 by type-checking. Still to come: TXT_DB, bsearch(?),
1335 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
1339 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
1340 seconds to a tm structure directly, instead of going through OS
1341 specific date routines. This avoids any issues with OS routines such
1342 as the year 2038 bug. New *_adj() functions for ASN1 time structures
1343 and X509_time_adj_ex() to cover the extended range. The existing
1344 X509_time_adj() is still usable and will no longer have any date issues.
1347 *) Delta CRL support. New use deltas option which will attempt to locate
1348 and search any appropriate delta CRLs available.
1350 This work was sponsored by Google.
1353 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
1354 code and add additional score elements. Validate alternate CRL paths
1355 as part of the CRL checking and indicate a new error "CRL path validation
1356 error" in this case. Applications wanting additional details can use
1357 the verify callback and check the new "parent" field. If this is not
1358 NULL CRL path validation is taking place. Existing applications wont
1359 see this because it requires extended CRL support which is off by
1362 This work was sponsored by Google.
1365 *) Support for freshest CRL extension.
1367 This work was sponsored by Google.
1370 *) Initial indirect CRL support. Currently only supported in the CRLs
1371 passed directly and not via lookup. Process certificate issuer
1372 CRL entry extension and lookup CRL entries by bother issuer name
1373 and serial number. Check and process CRL issuer entry in IDP extension.
1375 This work was sponsored by Google.
1378 *) Add support for distinct certificate and CRL paths. The CRL issuer
1379 certificate is validated separately in this case. Only enabled if
1380 an extended CRL support flag is set: this flag will enable additional
1381 CRL functionality in future.
1383 This work was sponsored by Google.
1386 *) Add support for policy mappings extension.
1388 This work was sponsored by Google.
1391 *) Fixes to pathlength constraint, self issued certificate handling,
1392 policy processing to align with RFC3280 and PKITS tests.
1394 This work was sponsored by Google.
1397 *) Support for name constraints certificate extension. DN, email, DNS
1398 and URI types are currently supported.
1400 This work was sponsored by Google.
1403 *) To cater for systems that provide a pointer-based thread ID rather
1404 than numeric, deprecate the current numeric thread ID mechanism and
1405 replace it with a structure and associated callback type. This
1406 mechanism allows a numeric "hash" to be extracted from a thread ID in
1407 either case, and on platforms where pointers are larger than 'long',
1408 mixing is done to help ensure the numeric 'hash' is usable even if it
1409 can't be guaranteed unique. The default mechanism is to use "&errno"
1410 as a pointer-based thread ID to distinguish between threads.
1412 Applications that want to provide their own thread IDs should now use
1413 CRYPTO_THREADID_set_callback() to register a callback that will call
1414 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
1416 Note that ERR_remove_state() is now deprecated, because it is tied
1417 to the assumption that thread IDs are numeric. ERR_remove_state(0)
1418 to free the current thread's error state should be replaced by
1419 ERR_remove_thread_state(NULL).
1421 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
1422 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
1423 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
1424 application was previously providing a numeric thread callback that
1425 was inappropriate for distinguishing threads, then uniqueness might
1426 have been obtained with &errno that happened immediately in the
1427 intermediate development versions of OpenSSL; this is no longer the
1428 case, the numeric thread callback will now override the automatic use
1430 [Geoff Thorpe, with help from Bodo Moeller]
1432 *) Initial support for different CRL issuing certificates. This covers a
1433 simple case where the self issued certificates in the chain exist and
1434 the real CRL issuer is higher in the existing chain.
1436 This work was sponsored by Google.
1439 *) Removed effectively defunct crypto/store from the build.
1442 *) Revamp of STACK to provide stronger type-checking. Still to come:
1443 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
1444 ASN1_STRING, CONF_VALUE.
1447 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
1448 RAM on SSL connections. This option can save about 34k per idle SSL.
1451 *) Revamp of LHASH to provide stronger type-checking. Still to come:
1452 STACK, TXT_DB, bsearch, qsort.
1455 *) Initial support for Cryptographic Message Syntax (aka CMS) based
1456 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
1457 support for data, signedData, compressedData, digestedData and
1458 encryptedData, envelopedData types included. Scripts to check against
1459 RFC4134 examples draft and interop and consistency checks of many
1460 content types and variants.
1463 *) Add options to enc utility to support use of zlib compression BIO.
1466 *) Extend mk1mf to support importing of options and assembly language
1467 files from Configure script, currently only included in VC-WIN32.
1468 The assembly language rules can now optionally generate the source
1469 files from the associated perl scripts.
1472 *) Implement remaining functionality needed to support GOST ciphersuites.
1473 Interop testing has been performed using CryptoPro implementations.
1474 [Victor B. Wagner <vitus@cryptocom.ru>]
1476 *) s390x assembler pack.
1479 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
1483 *) Implement Opaque PRF Input TLS extension as specified in
1484 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
1485 official specification yet and no extension type assignment by
1486 IANA exists, this extension (for now) will have to be explicitly
1487 enabled when building OpenSSL by providing the extension number
1488 to use. For example, specify an option
1490 -DTLSEXT_TYPE_opaque_prf_input=0x9527
1492 to the "config" or "Configure" script to enable the extension,
1493 assuming extension number 0x9527 (which is a completely arbitrary
1494 and unofficial assignment based on the MD5 hash of the Internet
1495 Draft). Note that by doing so, you potentially lose
1496 interoperability with other TLS implementations since these might
1497 be using the same extension number for other purposes.
1499 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
1500 opaque PRF input value to use in the handshake. This will create
1501 an interal copy of the length-'len' string at 'src', and will
1502 return non-zero for success.
1504 To get more control and flexibility, provide a callback function
1507 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
1508 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
1512 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
1515 Callback function 'cb' will be called in handshakes, and is
1516 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
1517 Argument 'arg' is for application purposes (the value as given to
1518 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
1519 be provided to the callback function). The callback function
1520 has to return non-zero to report success: usually 1 to use opaque
1521 PRF input just if possible, or 2 to enforce use of the opaque PRF
1522 input. In the latter case, the library will abort the handshake
1523 if opaque PRF input is not successfully negotiated.
1525 Arguments 'peerinput' and 'len' given to the callback function
1526 will always be NULL and 0 in the case of a client. A server will
1527 see the client's opaque PRF input through these variables if
1528 available (NULL and 0 otherwise). Note that if the server
1529 provides an opaque PRF input, the length must be the same as the
1530 length of the client's opaque PRF input.
1532 Note that the callback function will only be called when creating
1533 a new session (session resumption can resume whatever was
1534 previously negotiated), and will not be called in SSL 2.0
1535 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
1536 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
1537 for applications that need to enforce opaque PRF input.
1541 *) Update ssl code to support digests other than SHA1+MD5 for handshake
1544 [Victor B. Wagner <vitus@cryptocom.ru>]
1546 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1547 RFC4507bis. The encrypted ticket format is an encrypted encoded
1548 SSL_SESSION structure, that way new session features are automatically
1551 If a client application caches session in an SSL_SESSION structure
1552 support is transparent because tickets are now stored in the encoded
1555 The SSL_CTX structure automatically generates keys for ticket
1556 protection in servers so again support should be possible
1557 with no application modification.
1559 If a client or server wishes to disable RFC4507 support then the option
1560 SSL_OP_NO_TICKET can be set.
1562 Add a TLS extension debugging callback to allow the contents of any client
1563 or server extensions to be examined.
1565 This work was sponsored by Google.
1568 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
1569 OpenSSL should now compile cleanly on gcc 4.2
1570 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
1572 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
1573 support including streaming MAC support: this is required for GOST
1574 ciphersuite support.
1575 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
1577 *) Add option -stream to use PKCS#7 streaming in smime utility. New
1578 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
1579 to output in BER and PEM format.
1582 *) Experimental support for use of HMAC via EVP_PKEY interface. This
1583 allows HMAC to be handled via the EVP_DigestSign*() interface. The
1584 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
1585 ENGINE support for HMAC keys which are unextractable. New -mac and
1586 -macopt options to dgst utility.
1589 *) New option -sigopt to dgst utility. Update dgst to use
1590 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
1591 alternative signing paramaters such as X9.31 or PSS in the dgst
1595 *) Change ssl_cipher_apply_rule(), the internal function that does
1596 the work each time a ciphersuite string requests enabling
1597 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
1598 removing ("!foo+bar") a class of ciphersuites: Now it maintains
1599 the order of disabled ciphersuites such that those ciphersuites
1600 that most recently went from enabled to disabled not only stay
1601 in order with respect to each other, but also have higher priority
1602 than other disabled ciphersuites the next time ciphersuites are
1605 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
1606 the same ciphersuites as with "HIGH" alone, but in a specific
1607 order where the PSK ciphersuites come first (since they are the
1608 most recently disabled ciphersuites when "HIGH" is parsed).
1610 Also, change ssl_create_cipher_list() (using this new
1611 funcionality) such that between otherwise identical
1612 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
1616 *) Change ssl_create_cipher_list() so that it automatically
1617 arranges the ciphersuites in reasonable order before starting
1618 to process the rule string. Thus, the definition for "DEFAULT"
1619 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
1620 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
1621 This makes it much easier to arrive at a reasonable default order
1622 in applications for which anonymous ciphers are OK (meaning
1623 that you can't actually use DEFAULT).
1624 [Bodo Moeller; suggested by Victor Duchovni]
1626 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
1627 processing) into multiple integers instead of setting
1628 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
1629 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
1630 (These masks as well as the individual bit definitions are hidden
1631 away into the non-exported interface ssl/ssl_locl.h, so this
1632 change to the definition of the SSL_CIPHER structure shouldn't
1633 affect applications.) This give us more bits for each of these
1634 categories, so there is no longer a need to coagulate AES128 and
1635 AES256 into a single algorithm bit, and to coagulate Camellia128
1636 and Camellia256 into a single algorithm bit, which has led to all
1639 Thus, among other things, the kludge introduced in 0.9.7m and
1640 0.9.8e for masking out AES256 independently of AES128 or masking
1641 out Camellia256 independently of AES256 is not needed here in 0.9.9.
1643 With the change, we also introduce new ciphersuite aliases that
1644 so far were missing: "AES128", "AES256", "CAMELLIA128", and
1648 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
1649 Use the leftmost N bytes of the signature input if the input is
1650 larger than the prime q (with N being the size in bytes of q).
1653 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
1654 it yet and it is largely untested.
1657 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
1660 *) Initial incomplete changes to avoid need for function casts in OpenSSL
1661 some compilers (gcc 4.2 and later) reject their use. Safestack is
1662 reimplemented. Update ASN1 to avoid use of legacy functions.
1665 *) Win32/64 targets are linked with Winsock2.
1668 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
1669 to external functions. This can be used to increase CRL handling
1670 efficiency especially when CRLs are very large by (for example) storing
1671 the CRL revoked certificates in a database.
1674 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
1675 new CRLs added to a directory can be used. New command line option
1676 -verify_return_error to s_client and s_server. This causes real errors
1677 to be returned by the verify callback instead of carrying on no matter
1678 what. This reflects the way a "real world" verify callback would behave.
1681 *) GOST engine, supporting several GOST algorithms and public key formats.
1682 Kindly donated by Cryptocom.
1685 *) Partial support for Issuing Distribution Point CRL extension. CRLs
1686 partitioned by DP are handled but no indirect CRL or reason partitioning
1687 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
1688 selected via a scoring technique which handles IDP and AKID in CRLs.
1691 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
1692 will ultimately be used for all verify operations: this will remove the
1693 X509_STORE dependency on certificate verification and allow alternative
1694 lookup methods. X509_STORE based implementations of these two callbacks.
1697 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
1698 Modify get_crl() to find a valid (unexpired) CRL if possible.
1701 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
1702 this would be called X509_CRL_cmp() but that name is already used by
1703 a function that just compares CRL issuer names. Cache several CRL
1704 extensions in X509_CRL structure and cache CRLDP in X509.
1707 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
1708 this maps equivalent X509_NAME structures into a consistent structure.
1709 Name comparison can then be performed rapidly using memcmp().
1712 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
1716 *) Allow digests to supply their own micalg string for S/MIME type using
1717 the ctrl EVP_MD_CTRL_MICALG.
1720 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
1721 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
1722 ctrl. It can then customise the structure before and/or after signing
1726 *) New function OBJ_add_sigid() to allow application defined signature OIDs
1727 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
1728 to free up any added signature OIDs.
1731 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
1732 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
1733 digest and cipher tables. New options added to openssl utility:
1734 list-message-digest-algorithms and list-cipher-algorithms.
1737 *) Change the array representation of binary polynomials: the list
1738 of degrees of non-zero coefficients is now terminated with -1.
1739 Previously it was terminated with 0, which was also part of the
1740 value; thus, the array representation was not applicable to
1741 polynomials where t^0 has coefficient zero. This change makes
1742 the array representation useful in a more general context.
1745 *) Various modifications and fixes to SSL/TLS cipher string
1746 handling. For ECC, the code now distinguishes between fixed ECDH
1747 with RSA certificates on the one hand and with ECDSA certificates
1748 on the other hand, since these are separate ciphersuites. The
1749 unused code for Fortezza ciphersuites has been removed.
1751 For consistency with EDH, ephemeral ECDH is now called "EECDH"
1752 (not "ECDHE"). For consistency with the code for DH
1753 certificates, use of ECDH certificates is now considered ECDH
1754 authentication, not RSA or ECDSA authentication (the latter is
1755 merely the CA's signing algorithm and not actively used in the
1758 The temporary ciphersuite alias "ECCdraft" is no longer
1759 available, and ECC ciphersuites are no longer excluded from "ALL"
1760 and "DEFAULT". The following aliases now exist for RFC 4492
1761 ciphersuites, most of these by analogy with the DH case:
1763 kECDHr - ECDH cert, signed with RSA
1764 kECDHe - ECDH cert, signed with ECDSA
1765 kECDH - ECDH cert (signed with either RSA or ECDSA)
1766 kEECDH - ephemeral ECDH
1767 ECDH - ECDH cert or ephemeral ECDH
1773 AECDH - anonymous ECDH
1774 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
1778 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
1779 Use correct micalg parameters depending on digest(s) in signed message.
1782 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
1783 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
1786 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
1787 an engine to register a method. Add ENGINE lookups for methods and
1788 functional reference processing.
1791 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
1792 EVP_{Sign,Verify}* which allow an application to customise the signature
1796 *) New -resign option to smime utility. This adds one or more signers
1797 to an existing PKCS#7 signedData structure. Also -md option to use an
1798 alternative message digest algorithm for signing.
1801 *) Tidy up PKCS#7 routines and add new functions to make it easier to
1802 create PKCS7 structures containing multiple signers. Update smime
1803 application to support multiple signers.
1806 *) New -macalg option to pkcs12 utility to allow setting of an alternative
1810 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
1811 Reorganize PBE internals to lookup from a static table using NIDs,
1812 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
1813 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
1814 PRF which will be automatically used with PBES2.
1817 *) Replace the algorithm specific calls to generate keys in "req" with the
1821 *) Update PKCS#7 enveloped data routines to use new API. This is now
1822 supported by any public key method supporting the encrypt operation. A
1823 ctrl is added to allow the public key algorithm to examine or modify
1824 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
1828 *) Add a ctrl to asn1 method to allow a public key algorithm to express
1829 a default digest type to use. In most cases this will be SHA1 but some
1830 algorithms (such as GOST) need to specify an alternative digest. The
1831 return value indicates how strong the prefernce is 1 means optional and
1832 2 is mandatory (that is it is the only supported type). Modify
1833 ASN1_item_sign() to accept a NULL digest argument to indicate it should
1834 use the default md. Update openssl utilities to use the default digest
1835 type for signing if it is not explicitly indicated.
1838 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
1839 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
1840 signing method from the key type. This effectively removes the link
1841 between digests and public key types.
1844 *) Add an OID cross reference table and utility functions. Its purpose is to
1845 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
1846 rsaEncryption. This will allow some of the algorithm specific hackery
1847 needed to use the correct OID to be removed.
1850 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
1851 structures for PKCS7_sign(). They are now set up by the relevant public
1855 *) Add provisional EC pkey method with support for ECDSA and ECDH.
1858 *) Add support for key derivation (agreement) in the API, DH method and
1862 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
1863 public and private key formats. As a side effect these add additional
1864 command line functionality not previously available: DSA signatures can be
1865 generated and verified using pkeyutl and DH key support and generation in
1870 [Oliver Tappe <zooey@hirschkaefer.de>]
1872 *) New make target "install_html_docs" installs HTML renditions of the
1874 [Oliver Tappe <zooey@hirschkaefer.de>]
1876 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
1877 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
1878 support key and parameter generation and add initial key generation
1879 functionality for RSA.
1882 *) Add functions for main EVP_PKEY_method operations. The undocumented
1883 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
1884 EVP_PKEY_{encrypt,decrypt}_old.
1887 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
1888 key API, doesn't do much yet.
1891 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
1892 public key algorithms. New option to openssl utility:
1893 "list-public-key-algorithms" to print out info.
1896 *) Implement the Supported Elliptic Curves Extension for
1897 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1900 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
1901 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
1904 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
1905 utilities such as rsa, dsa, dsaparam etc except they process any key
1909 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1910 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1911 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
1915 *) Initial support for pluggable public key ASN1.
1916 De-spaghettify the public key ASN1 handling. Move public and private
1917 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
1918 algorithm specific handling to a single module within the relevant
1919 algorithm directory. Add functions to allow (near) opaque processing
1920 of public and private key structures.
1923 *) Implement the Supported Point Formats Extension for
1924 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1927 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
1928 for the psk identity [hint] and the psk callback functions to the
1929 SSL_SESSION, SSL and SSL_CTX structure.
1932 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
1936 SSL_CTX_use_psk_identity_hint
1937 SSL_get_psk_identity_hint
1938 SSL_get_psk_identity
1939 SSL_use_psk_identity_hint
1941 [Mika Kousa and Pasi Eronen of Nokia Corporation]
1943 *) Add RFC 3161 compliant time stamp request creation, response generation
1944 and response verification functionality.
1945 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
1947 *) Add initial support for TLS extensions, specifically for the server_name
1948 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1949 have new members for a host name. The SSL data structure has an
1950 additional member SSL_CTX *initial_ctx so that new sessions can be
1951 stored in that context to allow for session resumption, even after the
1952 SSL has been switched to a new SSL_CTX in reaction to a client's
1953 server_name extension.
1955 New functions (subject to change):
1957 SSL_get_servername()
1958 SSL_get_servername_type()
1961 New CTRL codes and macros (subject to change):
1963 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1964 - SSL_CTX_set_tlsext_servername_callback()
1965 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1966 - SSL_CTX_set_tlsext_servername_arg()
1967 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1969 openssl s_client has a new '-servername ...' option.
1971 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1972 '-key2 ...', '-servername_fatal' (subject to change). This allows
1973 testing the HostName extension for a specific single host name ('-cert'
1974 and '-key' remain fallbacks for handshakes without HostName
1975 negotiation). If the unrecogninzed_name alert has to be sent, this by
1976 default is a warning; it becomes fatal with the '-servername_fatal'
1979 [Peter Sylvester, Remy Allais, Christophe Renou]
1981 *) Whirlpool hash implementation is added.
1984 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1985 bn(64,32). Because of instruction set limitations it doesn't have
1986 any negative impact on performance. This was done mostly in order
1987 to make it possible to share assembler modules, such as bn_mul_mont
1988 implementations, between 32- and 64-bit builds without hassle.
1991 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1992 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1996 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1997 dedicated Montgomery multiplication procedure, is introduced.
1998 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1999 "64-bit" performance on certain 32-bit targets.
2002 *) New option SSL_OP_NO_COMP to disable use of compression selectively
2003 in SSL structures. New SSL ctrl to set maximum send fragment size.
2004 Save memory by seeting the I/O buffer sizes dynamically instead of
2005 using the maximum available value.
2008 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
2009 in addition to the text details.
2012 *) Very, very preliminary EXPERIMENTAL support for printing of general
2013 ASN1 structures. This currently produces rather ugly output and doesn't
2014 handle several customised structures at all.
2017 *) Integrated support for PVK file format and some related formats such
2018 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
2019 these in the 'rsa' and 'dsa' utilities.
2022 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
2025 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
2026 place for the (very old) "NETSCAPE" format certificates which are now
2027 handled using new ASN1 code equivalents.
2030 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
2031 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
2032 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
2035 *) Modify CRL distribution points extension code to print out previously
2036 unsupported fields. Enhance extension setting code to allow setting of
2040 *) Add print and set support for Issuing Distribution Point CRL extension.
2043 *) Change 'Configure' script to enable Camellia by default.
2046 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
2048 *) When rejecting SSL/TLS records due to an incorrect version number, never
2049 update s->server with a new major version number. As of
2050 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
2051 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
2052 the previous behavior could result in a read attempt at NULL when
2053 receiving specific incorrect SSL/TLS records once record payload
2054 protection is active. (CVE-2010-0740)
2055 [Bodo Moeller, Adam Langley <agl@chromium.org>]
2057 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
2058 could be crashed if the relevant tables were not present (e.g. chrooted).
2059 [Tomas Hoger <thoger@redhat.com>]
2061 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
2063 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
2064 [Martin Olsson, Neel Mehta]
2066 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
2067 accommodate for stack sorting, always a write lock!).
2070 *) On some versions of WIN32 Heap32Next is very slow. This can cause
2071 excessive delays in the RAND_poll(): over a minute. As a workaround
2072 include a time check in the inner Heap32Next loop too.
2075 *) The code that handled flushing of data in SSL/TLS originally used the
2076 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
2077 the problem outlined in PR#1949. The fix suggested there however can
2078 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
2079 of Apache). So instead simplify the code to flush unconditionally.
2080 This should be fine since flushing with no data to flush is a no op.
2083 *) Handle TLS versions 2.0 and later properly and correctly use the
2084 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
2085 off ancient servers have a habit of sticking around for a while...
2088 *) Modify compression code so it frees up structures without using the
2089 ex_data callbacks. This works around a problem where some applications
2090 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
2091 restarting) then use compression (e.g. SSL with compression) later.
2092 This results in significant per-connection memory leaks and
2093 has caused some security issues including CVE-2008-1678 and
2097 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
2098 change when encrypting or decrypting.
2101 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
2102 connect and renegotiate with servers which do not support RI.
2103 Until RI is more widely deployed this option is enabled by default.
2106 *) Add "missing" ssl ctrls to clear options and mode.
2109 *) If client attempts to renegotiate and doesn't support RI respond with
2110 a no_renegotiation alert as required by RFC5746. Some renegotiating
2111 TLS clients will continue a connection gracefully when they receive
2112 the alert. Unfortunately OpenSSL mishandled this alert and would hang
2113 waiting for a server hello which it will never receive. Now we treat a
2114 received no_renegotiation alert as a fatal error. This is because
2115 applications requesting a renegotiation might well expect it to succeed
2116 and would have no code in place to handle the server denying it so the
2117 only safe thing to do is to terminate the connection.
2120 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
2121 peer supports secure renegotiation and 0 otherwise. Print out peer
2122 renegotiation support in s_client/s_server.
2125 *) Replace the highly broken and deprecated SPKAC certification method with
2126 the updated NID creation version. This should correctly handle UTF8.
2129 *) Implement RFC5746. Re-enable renegotiation but require the extension
2130 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
2131 turns out to be a bad idea. It has been replaced by
2132 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
2133 SSL_CTX_set_options(). This is really not recommended unless you
2134 know what you are doing.
2135 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
2137 *) Fixes to stateless session resumption handling. Use initial_ctx when
2138 issuing and attempting to decrypt tickets in case it has changed during
2139 servername handling. Use a non-zero length session ID when attempting
2140 stateless session resumption: this makes it possible to determine if
2141 a resumption has occurred immediately after receiving server hello
2142 (several places in OpenSSL subtly assume this) instead of later in
2146 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
2147 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
2148 fixes for a few places where the return code is not checked
2150 [Julia Lawall <julia@diku.dk>]
2152 *) Add --strict-warnings option to Configure script to include devteam
2153 warnings in other configurations.
2156 *) Add support for --libdir option and LIBDIR variable in makefiles. This
2157 makes it possible to install openssl libraries in locations which
2158 have names other than "lib", for example "/usr/lib64" which some
2160 [Steve Henson, based on patch from Jeremy Utley]
2162 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
2163 X690 8.9.12 and can produce some misleading textual output of OIDs.
2164 [Steve Henson, reported by Dan Kaminsky]
2166 *) Delete MD2 from algorithm tables. This follows the recommendation in
2167 several standards that it is not used in new applications due to
2168 several cryptographic weaknesses. For binary compatibility reasons
2169 the MD2 API is still compiled in by default.
2172 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
2176 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
2177 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
2179 [Guenter <lists@gknw.net>]
2181 *) Fix the server certificate chain building code to use X509_verify_cert(),
2182 it used to have an ad-hoc builder which was unable to cope with anything
2183 other than a simple chain.
2184 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
2186 *) Don't check self signed certificate signatures in X509_verify_cert()
2187 by default (a flag can override this): it just wastes time without
2188 adding any security. As a useful side effect self signed root CAs
2189 with non-FIPS digests are now usable in FIPS mode.
2192 *) In dtls1_process_out_of_seq_message() the check if the current message
2193 is already buffered was missing. For every new message was memory
2194 allocated, allowing an attacker to perform an denial of service attack
2195 with sending out of seq handshake messages until there is no memory
2196 left. Additionally every future messege was buffered, even if the
2197 sequence number made no sense and would be part of another handshake.
2198 So only messages with sequence numbers less than 10 in advance will be
2199 buffered. (CVE-2009-1378)
2200 [Robin Seggelmann, discovered by Daniel Mentz]
2202 *) Records are buffered if they arrive with a future epoch to be
2203 processed after finishing the corresponding handshake. There is
2204 currently no limitation to this buffer allowing an attacker to perform
2205 a DOS attack with sending records with future epochs until there is no
2206 memory left. This patch adds the pqueue_size() function to detemine
2207 the size of a buffer and limits the record buffer to 100 entries.
2209 [Robin Seggelmann, discovered by Daniel Mentz]
2211 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
2212 parent structure is freed. (CVE-2009-1379)
2215 *) Handle non-blocking I/O properly in SSL_shutdown() call.
2216 [Darryl Miles <darryl-mailinglists@netbauds.net>]
2219 [Ilya O. <vrghost@gmail.com>]
2221 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
2223 *) Disable renegotiation completely - this fixes a severe security
2224 problem (CVE-2009-3555) at the cost of breaking all
2225 renegotiation. Renegotiation can be re-enabled by setting
2226 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
2227 run-time. This is really not recommended unless you know what
2231 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
2233 *) Don't set val to NULL when freeing up structures, it is freed up by
2234 underlying code. If sizeof(void *) > sizeof(long) this can result in
2235 zeroing past the valid field. (CVE-2009-0789)
2236 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
2238 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
2239 checked correctly. This would allow some invalid signed attributes to
2240 appear to verify correctly. (CVE-2009-0591)
2241 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2243 *) Reject UniversalString and BMPString types with invalid lengths. This
2244 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
2245 a legal length. (CVE-2009-0590)
2248 *) Set S/MIME signing as the default purpose rather than setting it
2249 unconditionally. This allows applications to override it at the store
2253 *) Permit restricted recursion of ASN1 strings. This is needed in practice
2254 to handle some structures.
2257 *) Improve efficiency of mem_gets: don't search whole buffer each time
2259 [Jeremy Shapiro <jnshapir@us.ibm.com>]
2261 *) New -hex option for openssl rand.
2264 *) Print out UTF8String and NumericString when parsing ASN1.
2267 *) Support NumericString type for name components.
2270 *) Allow CC in the environment to override the automatically chosen
2271 compiler. Note that nothing is done to ensure flags work with the
2275 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
2277 *) Properly check EVP_VerifyFinal() and similar return values
2279 [Ben Laurie, Bodo Moeller, Google Security Team]
2281 *) Enable TLS extensions by default.
2284 *) Allow the CHIL engine to be loaded, whether the application is
2285 multithreaded or not. (This does not release the developer from the
2286 obligation to set up the dynamic locking callbacks.)
2287 [Sander Temme <sander@temme.net>]
2289 *) Use correct exit code if there is an error in dgst command.
2290 [Steve Henson; problem pointed out by Roland Dirlewanger]
2292 *) Tweak Configure so that you need to say "experimental-jpake" to enable
2293 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
2296 *) Add experimental JPAKE support, including demo authentication in
2297 s_client and s_server.
2300 *) Set the comparison function in v3_addr_canonize().
2301 [Rob Austein <sra@hactrn.net>]
2303 *) Add support for XMPP STARTTLS in s_client.
2304 [Philip Paeps <philip@freebsd.org>]
2306 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
2307 to ensure that even with this option, only ciphersuites in the
2308 server's preference list will be accepted. (Note that the option
2309 applies only when resuming a session, so the earlier behavior was
2310 just about the algorithm choice for symmetric cryptography.)
2313 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
2315 *) Fix NULL pointer dereference if a DTLS server received
2316 ChangeCipherSpec as first record (CVE-2009-1386).
2319 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
2320 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
2323 *) The fix in 0.9.8c that supposedly got rid of unsafe
2324 double-checked locking was incomplete for RSA blinding,
2325 addressing just one layer of what turns out to have been
2326 doubly unsafe triple-checked locking.
2328 So now fix this for real by retiring the MONT_HELPER macro
2329 in crypto/rsa/rsa_eay.c.
2331 [Bodo Moeller; problem pointed out by Marius Schilder]
2333 *) Various precautionary measures:
2335 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
2337 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
2338 (NB: This would require knowledge of the secret session ticket key
2339 to exploit, in which case you'd be SOL either way.)
2341 - Change bn_nist.c so that it will properly handle input BIGNUMs
2342 outside the expected range.
2344 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
2347 [Neel Mehta, Bodo Moeller]
2349 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
2350 the load fails. Useful for distros.
2351 [Ben Laurie and the FreeBSD team]
2353 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
2356 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
2359 *) Expand ENGINE to support engine supplied SSL client certificate functions.
2361 This work was sponsored by Logica.
2364 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
2365 keystores. Support for SSL/TLS client authentication too.
2366 Not compiled unless enable-capieng specified to Configure.
2368 This work was sponsored by Logica.
2371 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
2372 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
2373 attribute creation routines such as certifcate requests and PKCS#12
2377 Changes between 0.9.8g and 0.9.8h [28 May 2008]
2379 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
2380 handshake which could lead to a cilent crash as found using the
2381 Codenomicon TLS test suite (CVE-2008-1672)
2382 [Steve Henson, Mark Cox]
2384 *) Fix double free in TLS server name extensions which could lead to
2385 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
2388 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
2390 Clear the error queue to ensure that error entries left from
2391 older function calls do not interfere with the correct operation.
2392 [Lutz Jaenicke, Erik de Castro Lopo]
2394 *) Remove root CA certificates of commercial CAs:
2396 The OpenSSL project does not recommend any specific CA and does not
2397 have any policy with respect to including or excluding any CA.
2398 Therefore it does not make any sense to ship an arbitrary selection
2399 of root CA certificates with the OpenSSL software.
2402 *) RSA OAEP patches to fix two separate invalid memory reads.
2403 The first one involves inputs when 'lzero' is greater than
2404 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
2405 before the beginning of from). The second one involves inputs where
2406 the 'db' section contains nothing but zeroes (there is a one-byte
2407 invalid read after the end of 'db').
2408 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2410 *) Partial backport from 0.9.9-dev:
2412 Introduce bn_mul_mont (dedicated Montgomery multiplication
2413 procedure) as a candidate for BIGNUM assembler implementation.
2414 While 0.9.9-dev uses assembler for various architectures, only
2415 x86_64 is available by default here in the 0.9.8 branch, and
2416 32-bit x86 is available through a compile-time setting.
2418 To try the 32-bit x86 assembler implementation, use Configure
2419 option "enable-montasm" (which exists only for this backport).
2421 As "enable-montasm" for 32-bit x86 disclaims code stability
2422 anyway, in this constellation we activate additional code
2423 backported from 0.9.9-dev for further performance improvements,
2424 namely BN_from_montgomery_word. (To enable this otherwise,
2425 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
2427 [Andy Polyakov (backport partially by Bodo Moeller)]
2429 *) Add TLS session ticket callback. This allows an application to set
2430 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
2431 values. This is useful for key rollover for example where several key
2432 sets may exist with different names.
2435 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
2436 This was broken until now in 0.9.8 releases, such that the only way
2437 a registered ENGINE could be used (assuming it initialises
2438 successfully on the host) was to explicitly set it as the default
2439 for the relevant algorithms. This is in contradiction with 0.9.7
2440 behaviour and the documentation. With this fix, when an ENGINE is
2441 registered into a given algorithm's table of implementations, the
2442 'uptodate' flag is reset so that auto-discovery will be used next
2443 time a new context for that algorithm attempts to select an
2445 [Ian Lister (tweaked by Geoff Thorpe)]
2447 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
2448 implemention in the following ways:
2450 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
2453 Lack of BER streaming support means one pass streaming processing is
2454 only supported if data is detached: setting the streaming flag is
2455 ignored for embedded content.
2457 CMS support is disabled by default and must be explicitly enabled
2458 with the enable-cms configuration option.
2461 *) Update the GMP engine glue to do direct copies between BIGNUM and
2462 mpz_t when openssl and GMP use the same limb size. Otherwise the
2463 existing "conversion via a text string export" trick is still used.
2464 [Paul Sheer <paulsheer@gmail.com>]
2466 *) Zlib compression BIO. This is a filter BIO which compressed and
2467 uncompresses any data passed through it.
2470 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
2471 RFC3394 compatible AES key wrapping.
2474 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
2475 sets string data without copying. X509_ALGOR_set0() and
2476 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
2477 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
2478 from an X509_ATTRIBUTE structure optionally checking it occurs only
2479 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
2483 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
2484 to get the expected BN_FLG_CONSTTIME behavior.
2485 [Bodo Moeller (Google)]
2489 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
2490 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
2491 - added some more tests to do_tests.pl
2492 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
2493 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
2494 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
2495 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
2496 - various changes to netware.pl to enable gcc-cross builds on Win32
2498 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
2499 - various changes to fix missing prototype warnings
2500 - fixed x86nasm.pl to create correct asm files for NASM COFF output
2501 - added AES, WHIRLPOOL and CPUID assembler code to build files
2502 - added missing AES assembler make rules to mk1mf.pl
2503 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
2504 [Guenter Knauf <eflash@gmx.net>]
2506 *) Implement certificate status request TLS extension defined in RFC3546.
2507 A client can set the appropriate parameters and receive the encoded
2508 OCSP response via a callback. A server can query the supplied parameters
2509 and set the encoded OCSP response in the callback. Add simplified examples
2510 to s_client and s_server.
2513 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
2515 *) Fix various bugs:
2516 + Binary incompatibility of ssl_ctx_st structure
2517 + DTLS interoperation with non-compliant servers
2518 + Don't call get_session_cb() without proposed session
2519 + Fix ia64 assembler code
2520 [Andy Polyakov, Steve Henson]
2522 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
2524 *) DTLS Handshake overhaul. There were longstanding issues with
2525 OpenSSL DTLS implementation, which were making it impossible for
2526 RFC 4347 compliant client to communicate with OpenSSL server.
2527 Unfortunately just fixing these incompatibilities would "cut off"
2528 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
2529 server keeps tolerating non RFC compliant syntax. The opposite is
2530 not true, 0.9.8f client can not communicate with earlier server.
2531 This update even addresses CVE-2007-4995.
2534 *) Changes to avoid need for function casts in OpenSSL: some compilers
2535 (gcc 4.2 and later) reject their use.
2536 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
2539 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2540 RFC4507bis. The encrypted ticket format is an encrypted encoded
2541 SSL_SESSION structure, that way new session features are automatically
2544 If a client application caches session in an SSL_SESSION structure
2545 support is transparent because tickets are now stored in the encoded
2548 The SSL_CTX structure automatically generates keys for ticket
2549 protection in servers so again support should be possible
2550 with no application modification.
2552 If a client or server wishes to disable RFC4507 support then the option
2553 SSL_OP_NO_TICKET can be set.
2555 Add a TLS extension debugging callback to allow the contents of any client
2556 or server extensions to be examined.
2558 This work was sponsored by Google.
2561 *) Add initial support for TLS extensions, specifically for the server_name
2562 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2563 have new members for a host name. The SSL data structure has an
2564 additional member SSL_CTX *initial_ctx so that new sessions can be
2565 stored in that context to allow for session resumption, even after the
2566 SSL has been switched to a new SSL_CTX in reaction to a client's
2567 server_name extension.
2569 New functions (subject to change):
2571 SSL_get_servername()
2572 SSL_get_servername_type()
2575 New CTRL codes and macros (subject to change):
2577 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2578 - SSL_CTX_set_tlsext_servername_callback()
2579 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2580 - SSL_CTX_set_tlsext_servername_arg()
2581 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2583 openssl s_client has a new '-servername ...' option.
2585 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2586 '-key2 ...', '-servername_fatal' (subject to change). This allows
2587 testing the HostName extension for a specific single host name ('-cert'
2588 and '-key' remain fallbacks for handshakes without HostName
2589 negotiation). If the unrecogninzed_name alert has to be sent, this by
2590 default is a warning; it becomes fatal with the '-servername_fatal'
2593 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
2595 *) Add AES and SSE2 assembly language support to VC++ build.
2598 *) Mitigate attack on final subtraction in Montgomery reduction.
2601 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
2602 (which previously caused an internal error).
2605 *) Squeeze another 10% out of IGE mode when in != out.
2608 *) AES IGE mode speedup.
2609 [Dean Gaudet (Google)]
2611 *) Add the Korean symmetric 128-bit cipher SEED (see
2612 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
2613 add SEED ciphersuites from RFC 4162:
2615 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
2616 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
2617 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
2618 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
2620 To minimize changes between patchlevels in the OpenSSL 0.9.8
2621 series, SEED remains excluded from compilation unless OpenSSL
2622 is configured with 'enable-seed'.
2623 [KISA, Bodo Moeller]
2625 *) Mitigate branch prediction attacks, which can be practical if a
2626 single processor is shared, allowing a spy process to extract
2627 information. For detailed background information, see
2628 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
2629 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
2630 and Necessary Software Countermeasures"). The core of the change
2631 are new versions BN_div_no_branch() and
2632 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
2633 respectively, which are slower, but avoid the security-relevant
2634 conditional branches. These are automatically called by BN_div()
2635 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
2636 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
2637 remove a conditional branch.
2639 BN_FLG_CONSTTIME is the new name for the previous
2640 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
2641 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
2642 in the exponent causes BN_mod_exp_mont() to use the alternative
2643 implementation in BN_mod_exp_mont_consttime().) The old name
2644 remains as a deprecated alias.
2646 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
2647 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
2648 constant-time implementations for more than just exponentiation.
2649 Here too the old name is kept as a deprecated alias.
2651 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2652 the BN_BLINDING structure gets an independent copy of the
2653 modulus. This means that the previous "BIGNUM *m" argument to
2654 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2655 essentially becomes "const BIGNUM *m", although we can't actually
2656 change this in the header file before 0.9.9. It allows
2657 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2658 enable BN_FLG_CONSTTIME.
2660 [Matthew D Wood (Intel Corp)]
2662 *) In the SSL/TLS server implementation, be strict about session ID
2663 context matching (which matters if an application uses a single
2664 external cache for different purposes). Previously,
2665 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2666 set. This did ensure strict client verification, but meant that,
2667 with applications using a single external cache for quite
2668 different requirements, clients could circumvent ciphersuite
2669 restrictions for a given session ID context by starting a session
2670 in a different context.
2673 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2674 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2675 authentication-only ciphersuites.
2678 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
2679 not complete and could lead to a possible single byte overflow
2680 (CVE-2007-5135) [Ben Laurie]
2682 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
2684 *) Since AES128 and AES256 (and similarly Camellia128 and
2685 Camellia256) share a single mask bit in the logic of
2686 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2687 kludge to work properly if AES128 is available and AES256 isn't
2688 (or if Camellia128 is available and Camellia256 isn't).
2691 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
2692 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
2693 When a point or a seed is encoded in a BIT STRING, we need to
2694 prevent the removal of trailing zero bits to get the proper DER
2695 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
2696 of a NamedBitList, for which trailing 0 bits need to be removed.)
2699 *) Have SSL/TLS server implementation tolerate "mismatched" record
2700 protocol version while receiving ClientHello even if the
2701 ClientHello is fragmented. (The server can't insist on the
2702 particular protocol version it has chosen before the ServerHello
2703 message has informed the client about his choice.)
2706 *) Add RFC 3779 support.
2707 [Rob Austein for ARIN, Ben Laurie]
2709 *) Load error codes if they are not already present instead of using a
2710 static variable. This allows them to be cleanly unloaded and reloaded.
2711 Improve header file function name parsing.
2714 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
2715 or CAPABILITY handshake as required by RFCs.
2718 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
2720 *) Introduce limits to prevent malicious keys being able to
2721 cause a denial of service. (CVE-2006-2940)
2722 [Steve Henson, Bodo Moeller]
2724 *) Fix ASN.1 parsing of certain invalid structures that can result
2725 in a denial of service. (CVE-2006-2937) [Steve Henson]
2727 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2728 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2730 *) Fix SSL client code which could crash if connecting to a
2731 malicious SSLv2 server. (CVE-2006-4343)
2732 [Tavis Ormandy and Will Drewry, Google Security Team]
2734 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
2735 match only those. Before that, "AES256-SHA" would be interpreted
2736 as a pattern and match "AES128-SHA" too (since AES128-SHA got
2737 the same strength classification in 0.9.7h) as we currently only
2738 have a single AES bit in the ciphersuite description bitmap.
2739 That change, however, also applied to ciphersuite strings such as
2740 "RC4-MD5" that intentionally matched multiple ciphersuites --
2741 namely, SSL 2.0 ciphersuites in addition to the more common ones
2742 from SSL 3.0/TLS 1.0.
2744 So we change the selection algorithm again: Naming an explicit
2745 ciphersuite selects this one ciphersuite, and any other similar
2746 ciphersuite (same bitmap) from *other* protocol versions.
2747 Thus, "RC4-MD5" again will properly select both the SSL 2.0
2748 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
2750 Since SSL 2.0 does not have any ciphersuites for which the
2751 128/256 bit distinction would be relevant, this works for now.
2752 The proper fix will be to use different bits for AES128 and
2753 AES256, which would have avoided the problems from the beginning;
2754 however, bits are scarce, so we can only do this in a new release
2755 (not just a patchlevel) when we can change the SSL_CIPHER
2756 definition to split the single 'unsigned long mask' bitmap into
2757 multiple values to extend the available space.
2761 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2763 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2764 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2766 *) Add AES IGE and biIGE modes.
2769 *) Change the Unix randomness entropy gathering to use poll() when
2770 possible instead of select(), since the latter has some
2771 undesirable limitations.
2772 [Darryl Miles via Richard Levitte and Bodo Moeller]
2774 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2775 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2776 cannot be implicitly activated as part of, e.g., the "AES" alias.
2777 However, please upgrade to OpenSSL 0.9.9[-dev] for
2778 non-experimental use of the ECC ciphersuites to get TLS extension
2779 support, which is required for curve and point format negotiation
2780 to avoid potential handshake problems.
2783 *) Disable rogue ciphersuites:
2785 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2786 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2787 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2789 The latter two were purportedly from
2790 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2793 Also deactivate the remaining ciphersuites from
2794 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2795 unofficial, and the ID has long expired.
2798 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2799 dual-core machines) and other potential thread-safety issues.
2802 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
2803 versions), which is now available for royalty-free use
2804 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
2805 Also, add Camellia TLS ciphersuites from RFC 4132.
2807 To minimize changes between patchlevels in the OpenSSL 0.9.8
2808 series, Camellia remains excluded from compilation unless OpenSSL
2809 is configured with 'enable-camellia'.
2812 *) Disable the padding bug check when compression is in use. The padding
2813 bug check assumes the first packet is of even length, this is not
2814 necessarily true if compresssion is enabled and can result in false
2815 positives causing handshake failure. The actual bug test is ancient
2816 code so it is hoped that implementations will either have fixed it by
2817 now or any which still have the bug do not support compression.
2820 Changes between 0.9.8a and 0.9.8b [04 May 2006]
2822 *) When applying a cipher rule check to see if string match is an explicit
2823 cipher suite and only match that one cipher suite if it is.
2826 *) Link in manifests for VC++ if needed.
2827 [Austin Ziegler <halostatue@gmail.com>]
2829 *) Update support for ECC-based TLS ciphersuites according to
2830 draft-ietf-tls-ecc-12.txt with proposed changes (but without
2831 TLS extensions, which are supported starting with the 0.9.9
2832 branch, not in the OpenSSL 0.9.8 branch).
2835 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
2836 opaque EVP_CIPHER_CTX handling.
2839 *) Fixes and enhancements to zlib compression code. We now only use
2840 "zlib1.dll" and use the default __cdecl calling convention on Win32
2841 to conform with the standards mentioned here:
2842 http://www.zlib.net/DLL_FAQ.txt
2843 Static zlib linking now works on Windows and the new --with-zlib-include
2844 --with-zlib-lib options to Configure can be used to supply the location
2845 of the headers and library. Gracefully handle case where zlib library
2849 *) Several fixes and enhancements to the OID generation code. The old code
2850 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
2851 handle numbers larger than ULONG_MAX, truncated printing and had a
2852 non standard OBJ_obj2txt() behaviour.
2855 *) Add support for building of engines under engine/ as shared libraries
2856 under VC++ build system.
2859 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
2860 Hopefully, we will not see any false combination of paths any more.
2863 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
2865 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2866 (part of SSL_OP_ALL). This option used to disable the
2867 countermeasure against man-in-the-middle protocol-version
2868 rollback in the SSL 2.0 server implementation, which is a bad
2869 idea. (CVE-2005-2969)
2871 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2872 for Information Security, National Institute of Advanced Industrial
2873 Science and Technology [AIST], Japan)]
2875 *) Add two function to clear and return the verify parameter flags.
2878 *) Keep cipherlists sorted in the source instead of sorting them at
2879 runtime, thus removing the need for a lock.
2882 *) Avoid some small subgroup attacks in Diffie-Hellman.
2883 [Nick Mathewson and Ben Laurie]
2885 *) Add functions for well-known primes.
2888 *) Extended Windows CE support.
2889 [Satoshi Nakamura and Andy Polyakov]
2891 *) Initialize SSL_METHOD structures at compile time instead of during
2892 runtime, thus removing the need for a lock.
2895 *) Make PKCS7_decrypt() work even if no certificate is supplied by
2896 attempting to decrypt each encrypted key in turn. Add support to
2900 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
2902 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
2905 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
2908 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
2909 key into the same file any more.
2912 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
2915 *) Add -utf8 command line and config file option to 'ca'.
2916 [Stefan <stf@udoma.org]
2918 *) Removed the macro des_crypt(), as it seems to conflict with some
2919 libraries. Use DES_crypt().
2922 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
2923 involves renaming the source and generated shared-libs for
2924 both. The engines will accept the corrected or legacy ids
2925 ('ncipher' and '4758_cca' respectively) when binding. NB,
2926 this only applies when building 'shared'.
2927 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
2929 *) Add attribute functions to EVP_PKEY structure. Modify
2930 PKCS12_create() to recognize a CSP name attribute and
2931 use it. Make -CSP option work again in pkcs12 utility.
2934 *) Add new functionality to the bn blinding code:
2935 - automatic re-creation of the BN_BLINDING parameters after
2936 a fixed number of uses (currently 32)
2937 - add new function for parameter creation
2938 - introduce flags to control the update behaviour of the
2939 BN_BLINDING parameters
2940 - hide BN_BLINDING structure
2941 Add a second BN_BLINDING slot to the RSA structure to improve
2942 performance when a single RSA object is shared among several
2946 *) Add support for DTLS.
2947 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
2949 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
2950 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
2953 *) Remove buggy and incompletet DH cert support from
2954 ssl/ssl_rsa.c and ssl/s3_both.c
2957 *) Use SHA-1 instead of MD5 as the default digest algorithm for
2958 the apps/openssl applications.
2961 *) Compile clean with "-Wall -Wmissing-prototypes
2962 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
2963 DEBUG_SAFESTACK must also be set.
2966 *) Change ./Configure so that certain algorithms can be disabled by default.
2967 The new counterpiece to "no-xxx" is "enable-xxx".
2969 The patented RC5 and MDC2 algorithms will now be disabled unless
2970 "enable-rc5" and "enable-mdc2", respectively, are specified.
2972 (IDEA remains enabled despite being patented. This is because IDEA
2973 is frequently required for interoperability, and there is no license
2974 fee for non-commercial use. As before, "no-idea" can be used to
2975 avoid this algorithm.)
2979 *) Add processing of proxy certificates (see RFC 3820). This work was
2980 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
2981 EGEE (Enabling Grids for E-science in Europe).
2984 *) RC4 performance overhaul on modern architectures/implementations, such
2985 as Intel P4, IA-64 and AMD64.
2988 *) New utility extract-section.pl. This can be used specify an alternative
2989 section number in a pod file instead of having to treat each file as
2990 a separate case in Makefile. This can be done by adding two lines to the
2993 =for comment openssl_section:XXX
2995 The blank line is mandatory.
2999 *) New arguments -certform, -keyform and -pass for s_client and s_server
3000 to allow alternative format key and certificate files and passphrase
3004 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
3005 update associated structures and add various utility functions.
3007 Add new policy related verify parameters, include policy checking in
3008 standard verify code. Enhance 'smime' application with extra parameters
3009 to support policy checking and print out.
3012 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
3013 Nehemiah processors. These extensions support AES encryption in hardware
3014 as well as RNG (though RNG support is currently disabled).
3015 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
3017 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
3020 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
3021 [Andy Polyakov and a number of other people]
3023 *) Improved PowerPC platform support. Most notably BIGNUM assembler
3024 implementation contributed by IBM.
3025 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
3027 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
3028 exponent rather than 'unsigned long'. There is a corresponding change to
3029 the new 'rsa_keygen' element of the RSA_METHOD structure.
3030 [Jelte Jansen, Geoff Thorpe]
3032 *) Functionality for creating the initial serial number file is now
3033 moved from CA.pl to the 'ca' utility with a new option -create_serial.
3035 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
3036 number file to 1, which is bound to cause problems. To avoid
3037 the problems while respecting compatibility between different 0.9.7
3038 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
3039 CA.pl for serial number initialization. With the new release 0.9.8,
3040 we can fix the problem directly in the 'ca' utility.)
3043 *) Reduced header interdepencies by declaring more opaque objects in
3044 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
3045 give fewer recursive includes, which could break lazy source code - so
3046 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
3047 developers should define this symbol when building and using openssl to
3048 ensure they track the recommended behaviour, interfaces, [etc], but
3049 backwards-compatible behaviour prevails when this isn't defined.
3052 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
3055 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
3056 This will generate a random key of the appropriate length based on the
3057 cipher context. The EVP_CIPHER can provide its own random key generation
3058 routine to support keys of a specific form. This is used in the des and
3059 3des routines to generate a key of the correct parity. Update S/MIME
3060 code to use new functions and hence generate correct parity DES keys.
3061 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
3062 valid (weak or incorrect parity).
3065 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
3066 as looking them up. This is useful when the verified structure may contain
3067 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
3068 present unless the new PKCS7_NO_CRL flag is asserted.
3071 *) Extend ASN1 oid configuration module. It now additionally accepts the
3074 shortName = some long name, 1.2.3.4
3077 *) Reimplemented the BN_CTX implementation. There is now no more static
3078 limitation on the number of variables it can handle nor the depth of the
3079 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
3080 information can now expand as required, and rather than having a single
3081 static array of bignums, BN_CTX now uses a linked-list of such arrays
3082 allowing it to expand on demand whilst maintaining the usefulness of
3083 BN_CTX's "bundling".
3086 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
3087 to allow all RSA operations to function using a single BN_CTX.
3090 *) Preliminary support for certificate policy evaluation and checking. This
3091 is initially intended to pass the tests outlined in "Conformance Testing
3092 of Relying Party Client Certificate Path Processing Logic" v1.07.
3095 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
3096 remained unused and not that useful. A variety of other little bignum
3097 tweaks and fixes have also been made continuing on from the audit (see
3101 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
3102 associated ASN1, EVP and SSL functions and old ASN1 macros.
3105 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
3106 and this should never fail. So the return value from the use of
3107 BN_set_word() (which can fail due to needless expansion) is now deprecated;
3108 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
3111 *) BN_CTX_get() should return zero-valued bignums, providing the same
3112 initialised value as BN_new().
3113 [Geoff Thorpe, suggested by Ulf Möller]
3115 *) Support for inhibitAnyPolicy certificate extension.
3118 *) An audit of the BIGNUM code is underway, for which debugging code is
3119 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
3120 is considered valid when processing BIGNUMs, and causes execution to
3121 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
3122 further steps are taken to deliberately pollute unused data in BIGNUM
3123 structures to try and expose faulty code further on. For now, openssl will
3124 (in its default mode of operation) continue to tolerate the inconsistent
3125 forms that it has tolerated in the past, but authors and packagers should
3126 consider trying openssl and their own applications when compiled with
3127 these debugging symbols defined. It will help highlight potential bugs in
3128 their own code, and will improve the test coverage for OpenSSL itself. At
3129 some point, these tighter rules will become openssl's default to improve
3130 maintainability, though the assert()s and other overheads will remain only
3131 in debugging configurations. See bn.h for more details.
3132 [Geoff Thorpe, Nils Larsch, Ulf Möller]
3134 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
3135 that can only be obtained through BN_CTX_new() (which implicitly
3136 initialises it). The presence of this function only made it possible
3137 to overwrite an existing structure (and cause memory leaks).
3140 *) Because of the callback-based approach for implementing LHASH as a
3141 template type, lh_insert() adds opaque objects to hash-tables and
3142 lh_doall() or lh_doall_arg() are typically used with a destructor callback
3143 to clean up those corresponding objects before destroying the hash table
3144 (and losing the object pointers). So some over-zealous constifications in
3145 LHASH have been relaxed so that lh_insert() does not take (nor store) the
3146 objects as "const" and the lh_doall[_arg] callback wrappers are not
3147 prototyped to have "const" restrictions on the object pointers they are
3148 given (and so aren't required to cast them away any more).
3151 *) The tmdiff.h API was so ugly and minimal that our own timing utility
3152 (speed) prefers to use its own implementation. The two implementations
3153 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
3154 its object type properly exposed (MS_TM) instead of casting to/from "char
3155 *". This may still change yet if someone realises MS_TM and "ms_time_***"
3156 aren't necessarily the greatest nomenclatures - but this is what was used
3157 internally to the implementation so I've used that for now.
3160 *) Ensure that deprecated functions do not get compiled when
3161 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
3162 the self-tests were still using deprecated key-generation functions so
3163 these have been updated also.
3166 *) Reorganise PKCS#7 code to separate the digest location functionality
3167 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
3168 New function PKCS7_set_digest() to set the digest type for PKCS#7
3169 digestedData type. Add additional code to correctly generate the
3170 digestedData type and add support for this type in PKCS7 initialization
3174 *) New function PKCS7_set0_type_other() this initializes a PKCS7
3175 structure of type "other".
3178 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
3179 sure the loop does correctly stop and breaking ("division by zero")
3180 modulus operations are not performed. The (pre-generated) prime
3181 table crypto/bn/bn_prime.h was already correct, but it could not be
3182 re-generated on some platforms because of the "division by zero"
3183 situation in the script.
3184 [Ralf S. Engelschall]
3186 *) Update support for ECC-based TLS ciphersuites according to
3187 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
3188 SHA-1 now is only used for "small" curves (where the
3189 representation of a field element takes up to 24 bytes); for
3190 larger curves, the field element resulting from ECDH is directly
3191 used as premaster secret.
3192 [Douglas Stebila (Sun Microsystems Laboratories)]
3194 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
3195 curve secp160r1 to the tests.
3196 [Douglas Stebila (Sun Microsystems Laboratories)]
3198 *) Add the possibility to load symbols globally with DSO.
3199 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
3201 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
3202 control of the error stack.
3205 *) Add support for STORE in ENGINE.
3208 *) Add the STORE type. The intention is to provide a common interface
3209 to certificate and key stores, be they simple file-based stores, or
3210 HSM-type store, or LDAP stores, or...
3211 NOTE: The code is currently UNTESTED and isn't really used anywhere.
3214 *) Add a generic structure called OPENSSL_ITEM. This can be used to
3215 pass a list of arguments to any function as well as provide a way
3216 for a function to pass data back to the caller.
3219 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
3220 works like BUF_strdup() but can be used to duplicate a portion of
3221 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
3225 *) Add the function sk_find_ex() which works like sk_find(), but will
3226 return an index to an element even if an exact match couldn't be
3227 found. The index is guaranteed to point at the element where the
3228 searched-for key would be inserted to preserve sorting order.
3231 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
3232 takes an extra flags argument for optional functionality. Currently,
3233 the following flags are defined:
3235 OBJ_BSEARCH_VALUE_ON_NOMATCH
3236 This one gets OBJ_bsearch_ex() to return a pointer to the first
3237 element where the comparing function returns a negative or zero
3240 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
3241 This one gets OBJ_bsearch_ex() to return a pointer to the first
3242 element where the comparing function returns zero. This is useful
3243 if there are more than one element where the comparing function
3247 *) Make it possible to create self-signed certificates with 'openssl ca'
3248 in such a way that the self-signed certificate becomes part of the
3249 CA database and uses the same mechanisms for serial number generation
3250 as all other certificate signing. The new flag '-selfsign' enables
3251 this functionality. Adapt CA.sh and CA.pl.in.
3254 *) Add functionality to check the public key of a certificate request
3255 against a given private. This is useful to check that a certificate
3256 request can be signed by that key (self-signing).
3259 *) Make it possible to have multiple active certificates with the same
3260 subject in the CA index file. This is done only if the keyword
3261 'unique_subject' is set to 'no' in the main CA section (default
3262 if 'CA_default') of the configuration file. The value is saved
3263 with the database itself in a separate index attribute file,
3264 named like the index file with '.attr' appended to the name.
3267 *) Generate muti valued AVAs using '+' notation in config files for
3271 *) Support for nameConstraints certificate extension.
3274 *) Support for policyConstraints certificate extension.
3277 *) Support for policyMappings certificate extension.
3280 *) Make sure the default DSA_METHOD implementation only uses its
3281 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
3282 and change its own handlers to be NULL so as to remove unnecessary
3283 indirection. This lets alternative implementations fallback to the
3284 default implementation more easily.
3287 *) Support for directoryName in GeneralName related extensions
3291 *) Make it possible to link applications using Makefile.shared.
3292 Make that possible even when linking against static libraries!
3295 *) Support for single pass processing for S/MIME signing. This now
3296 means that S/MIME signing can be done from a pipe, in addition
3297 cleartext signing (multipart/signed type) is effectively streaming
3298 and the signed data does not need to be all held in memory.
3300 This is done with a new flag PKCS7_STREAM. When this flag is set
3301 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
3302 is done after the data is output (and digests calculated) in
3303 SMIME_write_PKCS7().
3306 *) Add full support for -rpath/-R, both in shared libraries and
3307 applications, at least on the platforms where it's known how
3311 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
3312 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
3313 will now compute a table of multiples of the generator that
3314 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
3315 faster (notably in the case of a single point multiplication,
3316 scalar * generator).
3317 [Nils Larsch, Bodo Moeller]
3319 *) IPv6 support for certificate extensions. The various extensions
3320 which use the IP:a.b.c.d can now take IPv6 addresses using the
3321 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
3325 *) Added an ENGINE that implements RSA by performing private key
3326 exponentiations with the GMP library. The conversions to and from
3327 GMP's mpz_t format aren't optimised nor are any montgomery forms
3328 cached, and on x86 it appears OpenSSL's own performance has caught up.
3329 However there are likely to be other architectures where GMP could
3330 provide a boost. This ENGINE is not built in by default, but it can be
3331 specified at Configure time and should be accompanied by the necessary
3332 linker additions, eg;
3333 ./config -DOPENSSL_USE_GMP -lgmp
3336 *) "openssl engine" will not display ENGINE/DSO load failure errors when
3337 testing availability of engines with "-t" - the old behaviour is
3338 produced by increasing the feature's verbosity with "-tt".
3341 *) ECDSA routines: under certain error conditions uninitialized BN objects
3342 could be freed. Solution: make sure initialization is performed early
3343 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
3347 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
3348 and DH_METHOD (eg. by ENGINE implementations) to override the normal
3349 software implementations. For DSA and DH, parameter generation can
3350 also be overriden by providing the appropriate method callbacks.
3353 *) Change the "progress" mechanism used in key-generation and
3354 primality testing to functions that take a new BN_GENCB pointer in
3355 place of callback/argument pairs. The new API functions have "_ex"
3356 postfixes and the older functions are reimplemented as wrappers for
3357 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
3358 declarations of the old functions to help (graceful) attempts to
3359 migrate to the new functions. Also, the new key-generation API
3360 functions operate on a caller-supplied key-structure and return
3361 success/failure rather than returning a key or NULL - this is to
3362 help make "keygen" another member function of RSA_METHOD etc.
3364 Example for using the new callback interface:
3366 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
3370 BN_GENCB_set(&my_cb, my_callback, my_arg);
3372 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
3373 /* For the meaning of a, b in calls to my_callback(), see the
3374 * documentation of the function that calls the callback.
3375 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
3376 * my_callback should return 1 if it wants BN_is_prime_ex()
3377 * to continue, or 0 to stop.
3382 *) Change the ZLIB compression method to be stateful, and make it
3383 available to TLS with the number defined in
3384 draft-ietf-tls-compression-04.txt.
3387 *) Add the ASN.1 structures and functions for CertificatePair, which
3388 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
3390 CertificatePair ::= SEQUENCE {
3391 forward [0] Certificate OPTIONAL,
3392 reverse [1] Certificate OPTIONAL,
3393 -- at least one of the pair shall be present -- }
3395 Also implement the PEM functions to read and write certificate
3396 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
3398 This needed to be defined, mostly for the sake of the LDAP
3399 attribute crossCertificatePair, but may prove useful elsewhere as
3403 *) Make it possible to inhibit symlinking of shared libraries in
3404 Makefile.shared, for Cygwin's sake.
3407 *) Extend the BIGNUM API by creating a function
3408 void BN_set_negative(BIGNUM *a, int neg);
3409 and a macro that behave like
3410 int BN_is_negative(const BIGNUM *a);
3412 to avoid the need to access 'a->neg' directly in applications.
3415 *) Implement fast modular reduction for pseudo-Mersenne primes
3416 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
3417 EC_GROUP_new_curve_GFp() will now automatically use this
3419 [Nils Larsch <nla@trustcenter.de>]
3421 *) Add new lock type (CRYPTO_LOCK_BN).
3424 *) Change the ENGINE framework to automatically load engines
3425 dynamically from specific directories unless they could be
3426 found to already be built in or loaded. Move all the
3427 current engines except for the cryptodev one to a new
3429 The engines in engines/ are built as shared libraries if
3430 the "shared" options was given to ./Configure or ./config.
3431 Otherwise, they are inserted in libcrypto.a.
3432 /usr/local/ssl/engines is the default directory for dynamic
3433 engines, but that can be overriden at configure time through
3434 the usual use of --prefix and/or --openssldir, and at run
3435 time with the environment variable OPENSSL_ENGINES.
3436 [Geoff Thorpe and Richard Levitte]
3438 *) Add Makefile.shared, a helper makefile to build shared
3439 libraries. Addapt Makefile.org.
3442 *) Add version info to Win32 DLLs.
3443 [Peter 'Luna' Runestig" <peter@runestig.com>]
3445 *) Add new 'medium level' PKCS#12 API. Certificates and keys
3446 can be added using this API to created arbitrary PKCS#12
3447 files while avoiding the low level API.
3449 New options to PKCS12_create(), key or cert can be NULL and
3450 will then be omitted from the output file. The encryption
3451 algorithm NIDs can be set to -1 for no encryption, the mac
3452 iteration count can be set to 0 to omit the mac.
3454 Enhance pkcs12 utility by making the -nokeys and -nocerts
3455 options work when creating a PKCS#12 file. New option -nomac
3456 to omit the mac, NONE can be set for an encryption algorithm.
3457 New code is modified to use the enhanced PKCS12_create()
3458 instead of the low level API.
3461 *) Extend ASN1 encoder to support indefinite length constructed
3462 encoding. This can output sequences tags and octet strings in
3463 this form. Modify pk7_asn1.c to support indefinite length
3464 encoding. This is experimental and needs additional code to
3465 be useful, such as an ASN1 bio and some enhanced streaming
3468 Extend template encode functionality so that tagging is passed
3469 down to the template encoder.
3472 *) Let 'openssl req' fail if an argument to '-newkey' is not
3473 recognized instead of using RSA as a default.
3476 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
3477 As these are not official, they are not included in "ALL";
3478 the "ECCdraft" ciphersuite group alias can be used to select them.
3479 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
3481 *) Add ECDH engine support.
3482 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
3484 *) Add ECDH in new directory crypto/ecdh/.
3485 [Douglas Stebila (Sun Microsystems Laboratories)]
3487 *) Let BN_rand_range() abort with an error after 100 iterations
3488 without success (which indicates a broken PRNG).
3491 *) Change BN_mod_sqrt() so that it verifies that the input value
3492 is really the square of the return value. (Previously,
3493 BN_mod_sqrt would show GIGO behaviour.)
3496 *) Add named elliptic curves over binary fields from X9.62, SECG,
3497 and WAP/WTLS; add OIDs that were still missing.
3499 [Sheueling Chang Shantz and Douglas Stebila
3500 (Sun Microsystems Laboratories)]
3502 *) Extend the EC library for elliptic curves over binary fields
3503 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
3506 EC_GF2m_simple_method
3510 EC_GROUP_new_curve_GF2m
3511 EC_GROUP_set_curve_GF2m
3512 EC_GROUP_get_curve_GF2m
3513 EC_POINT_set_affine_coordinates_GF2m
3514 EC_POINT_get_affine_coordinates_GF2m
3515 EC_POINT_set_compressed_coordinates_GF2m
3517 Point compression for binary fields is disabled by default for
3518 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
3521 As binary polynomials are represented as BIGNUMs, various members
3522 of the EC_GROUP and EC_POINT data structures can be shared
3523 between the implementations for prime fields and binary fields;
3524 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
3525 are essentially identical to their ..._GFp counterparts.
3526 (For simplicity, the '..._GFp' prefix has been dropped from
3527 various internal method names.)
3529 An internal 'field_div' method (similar to 'field_mul' and
3530 'field_sqr') has been added; this is used only for binary fields.
3532 [Sheueling Chang Shantz and Douglas Stebila
3533 (Sun Microsystems Laboratories)]
3535 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
3536 through methods ('mul', 'precompute_mult').
3538 The generic implementations (now internally called 'ec_wNAF_mul'
3539 and 'ec_wNAF_precomputed_mult') remain the default if these
3540 methods are undefined.
3542 [Sheueling Chang Shantz and Douglas Stebila
3543 (Sun Microsystems Laboratories)]
3545 *) New function EC_GROUP_get_degree, which is defined through
3546 EC_METHOD. For curves over prime fields, this returns the bit
3547 length of the modulus.
3549 [Sheueling Chang Shantz and Douglas Stebila
3550 (Sun Microsystems Laboratories)]
3552 *) New functions EC_GROUP_dup, EC_POINT_dup.
3553 (These simply call ..._new and ..._copy).
3555 [Sheueling Chang Shantz and Douglas Stebila
3556 (Sun Microsystems Laboratories)]
3558 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
3559 Polynomials are represented as BIGNUMs (where the sign bit is not
3560 used) in the following functions [macros]:
3563 BN_GF2m_sub [= BN_GF2m_add]
3564 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
3565 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
3566 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
3568 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
3569 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
3570 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
3571 BN_GF2m_cmp [= BN_ucmp]
3573 (Note that only the 'mod' functions are actually for fields GF(2^m).
3574 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
3576 For some functions, an the irreducible polynomial defining a
3577 field can be given as an 'unsigned int[]' with strictly
3578 decreasing elements giving the indices of those bits that are set;
3579 i.e., p[] represents the polynomial
3580 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
3582 p[0] > p[1] > ... > p[k] = 0.
3583 This applies to the following functions:
3588 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
3589 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
3591 BN_GF2m_mod_sqrt_arr
3592 BN_GF2m_mod_solve_quad_arr
3596 Conversion can be performed by the following functions:
3601 bntest.c has additional tests for binary polynomial arithmetic.
3603 Two implementations for BN_GF2m_mod_div() are available.
3604 The default algorithm simply uses BN_GF2m_mod_inv() and
3605 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
3606 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
3607 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
3609 [Sheueling Chang Shantz and Douglas Stebila
3610 (Sun Microsystems Laboratories)]
3612 *) Add new error code 'ERR_R_DISABLED' that can be used when some
3613 functionality is disabled at compile-time.
3614 [Douglas Stebila <douglas.stebila@sun.com>]
3616 *) Change default behaviour of 'openssl asn1parse' so that more
3617 information is visible when viewing, e.g., a certificate:
3619 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
3620 mode the content of non-printable OCTET STRINGs is output in a
3621 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
3622 avoid the appearance of a printable string.
3623 [Nils Larsch <nla@trustcenter.de>]
3625 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
3627 EC_GROUP_set_asn1_flag()
3628 EC_GROUP_get_asn1_flag()
3629 EC_GROUP_set_point_conversion_form()
3630 EC_GROUP_get_point_conversion_form()
3631 These control ASN1 encoding details:
3632 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
3633 has been set to OPENSSL_EC_NAMED_CURVE.
3634 - Points are encoded in uncompressed form by default; options for
3635 asn1_for are as for point2oct, namely
3636 POINT_CONVERSION_COMPRESSED
3637 POINT_CONVERSION_UNCOMPRESSED
3638 POINT_CONVERSION_HYBRID
3640 Also add 'seed' and 'seed_len' members to EC_GROUP with access
3643 EC_GROUP_get0_seed()
3644 EC_GROUP_get_seed_len()
3645 This is used only for ASN1 purposes (so far).
3646 [Nils Larsch <nla@trustcenter.de>]
3648 *) Add 'field_type' member to EC_METHOD, which holds the NID
3649 of the appropriate field type OID. The new function
3650 EC_METHOD_get_field_type() returns this value.
3651 [Nils Larsch <nla@trustcenter.de>]
3656 EC_POINT_point2hex()
3657 EC_POINT_hex2point()
3658 providing useful interfaces to EC_POINT_point2oct() and
3659 EC_POINT_oct2point().
3660 [Nils Larsch <nla@trustcenter.de>]
3662 *) Change internals of the EC library so that the functions
3663 EC_GROUP_set_generator()
3664 EC_GROUP_get_generator()
3665 EC_GROUP_get_order()
3666 EC_GROUP_get_cofactor()
3667 are implemented directly in crypto/ec/ec_lib.c and not dispatched
3668 to methods, which would lead to unnecessary code duplication when
3669 adding different types of curves.
3670 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
3672 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
3673 arithmetic, and such that modified wNAFs are generated
3674 (which avoid length expansion in many cases).
3677 *) Add a function EC_GROUP_check_discriminant() (defined via
3678 EC_METHOD) that verifies that the curve discriminant is non-zero.
3680 Add a function EC_GROUP_check() that makes some sanity tests
3681 on a EC_GROUP, its generator and order. This includes
3682 EC_GROUP_check_discriminant().
3683 [Nils Larsch <nla@trustcenter.de>]
3685 *) Add ECDSA in new directory crypto/ecdsa/.
3687 Add applications 'openssl ecparam' and 'openssl ecdsa'
3688 (these are based on 'openssl dsaparam' and 'openssl dsa').
3690 ECDSA support is also included in various other files across the
3691 library. Most notably,
3692 - 'openssl req' now has a '-newkey ecdsa:file' option;
3693 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
3694 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
3695 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
3696 them suitable for ECDSA where domain parameters must be
3697 extracted before the specific public key;
3698 - ECDSA engine support has been added.
3699 [Nils Larsch <nla@trustcenter.de>]
3701 *) Include some named elliptic curves, and add OIDs from X9.62,
3702 SECG, and WAP/WTLS. Each curve can be obtained from the new
3704 EC_GROUP_new_by_curve_name(),
3705 and the list of available named curves can be obtained with
3706 EC_get_builtin_curves().
3707 Also add a 'curve_name' member to EC_GROUP objects, which can be
3709 EC_GROUP_set_curve_name()
3710 EC_GROUP_get_curve_name()
3711 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
3713 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3714 was actually never needed) and in BN_mul(). The removal in BN_mul()
3715 required a small change in bn_mul_part_recursive() and the addition
3716 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3717 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3718 bn_sub_words() and bn_add_words() except they take arrays with
3722 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
3724 *) Cleanse PEM buffers before freeing them since they may contain
3726 [Benjamin Bennett <ben@psc.edu>]
3728 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3729 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3730 authentication-only ciphersuites.
3733 *) Since AES128 and AES256 share a single mask bit in the logic of
3734 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3735 kludge to work properly if AES128 is available and AES256 isn't.
3738 *) Expand security boundary to match 1.1.1 module.
3741 *) Remove redundant features: hash file source, editing of test vectors
3742 modify fipsld to use external fips_premain.c signature.
3745 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
3746 run algorithm test programs.
3749 *) Make algorithm test programs more tolerant of whitespace.
3752 *) Have SSL/TLS server implementation tolerate "mismatched" record
3753 protocol version while receiving ClientHello even if the
3754 ClientHello is fragmented. (The server can't insist on the
3755 particular protocol version it has chosen before the ServerHello
3756 message has informed the client about his choice.)
3759 *) Load error codes if they are not already present instead of using a
3760 static variable. This allows them to be cleanly unloaded and reloaded.
3763 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3765 *) Introduce limits to prevent malicious keys being able to
3766 cause a denial of service. (CVE-2006-2940)
3767 [Steve Henson, Bodo Moeller]
3769 *) Fix ASN.1 parsing of certain invalid structures that can result
3770 in a denial of service. (CVE-2006-2937) [Steve Henson]
3772 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3773 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3775 *) Fix SSL client code which could crash if connecting to a
3776 malicious SSLv2 server. (CVE-2006-4343)
3777 [Tavis Ormandy and Will Drewry, Google Security Team]
3779 *) Change ciphersuite string processing so that an explicit
3780 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3781 will no longer include "AES128-SHA"), and any other similar
3782 ciphersuite (same bitmap) from *other* protocol versions (so that
3783 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3784 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3785 changes from 0.9.8b and 0.9.8d.
3788 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
3790 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3791 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3793 *) Change the Unix randomness entropy gathering to use poll() when
3794 possible instead of select(), since the latter has some
3795 undesirable limitations.
3796 [Darryl Miles via Richard Levitte and Bodo Moeller]
3798 *) Disable rogue ciphersuites:
3800 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3801 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3802 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3804 The latter two were purportedly from
3805 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3808 Also deactive the remaining ciphersuites from
3809 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3810 unofficial, and the ID has long expired.
3813 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3814 dual-core machines) and other potential thread-safety issues.
3817 Changes between 0.9.7i and 0.9.7j [04 May 2006]
3819 *) Adapt fipsld and the build system to link against the validated FIPS
3820 module in FIPS mode.
3823 *) Fixes for VC++ 2005 build under Windows.
3826 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
3827 from a Windows bash shell such as MSYS. It is autodetected from the
3828 "config" script when run from a VC++ environment. Modify standard VC++
3829 build to use fipscanister.o from the GNU make build.
3832 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
3834 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
3835 The value now differs depending on if you build for FIPS or not.
3836 BEWARE! A program linked with a shared FIPSed libcrypto can't be
3837 safely run with a non-FIPSed libcrypto, as it may crash because of
3838 the difference induced by this change.
3841 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
3843 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3844 (part of SSL_OP_ALL). This option used to disable the
3845 countermeasure against man-in-the-middle protocol-version
3846 rollback in the SSL 2.0 server implementation, which is a bad
3847 idea. (CVE-2005-2969)
3849 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3850 for Information Security, National Institute of Advanced Industrial
3851 Science and Technology [AIST], Japan)]
3853 *) Minimal support for X9.31 signatures and PSS padding modes. This is
3854 mainly for FIPS compliance and not fully integrated at this stage.
3857 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
3858 the exponentiation using a fixed-length exponent. (Otherwise,
3859 the information leaked through timing could expose the secret key
3860 after many signatures; cf. Bleichenbacher's attack on DSA with
3864 *) Make a new fixed-window mod_exp implementation the default for
3865 RSA, DSA, and DH private-key operations so that the sequence of
3866 squares and multiplies and the memory access pattern are
3867 independent of the particular secret key. This will mitigate
3868 cache-timing and potential related attacks.
3870 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
3871 and this is automatically used by BN_mod_exp_mont() if the new flag
3872 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
3873 will use this BN flag for private exponents unless the flag
3874 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
3875 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
3877 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
3879 *) Change the client implementation for SSLv23_method() and
3880 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
3881 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
3882 (Previously, the SSL 2.0 backwards compatible Client Hello
3883 message format would be used even with SSL_OP_NO_SSLv2.)
3886 *) Add support for smime-type MIME parameter in S/MIME messages which some
3890 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
3891 a threadsafe manner. Modify rsa code to use new function and add calls
3892 to dsa and dh code (which had race conditions before).
3895 *) Include the fixed error library code in the C error file definitions
3896 instead of fixing them up at runtime. This keeps the error code
3897 structures constant.
3900 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
3902 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
3905 *) Fixes for newer kerberos headers. NB: the casts are needed because
3906 the 'length' field is signed on one version and unsigned on another
3907 with no (?) obvious way to tell the difference, without these VC++
3908 complains. Also the "definition" of FAR (blank) is no longer included
3909 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
3910 some needed definitions.
3913 *) Undo Cygwin change.
3916 *) Added support for proxy certificates according to RFC 3820.
3917 Because they may be a security thread to unaware applications,
3918 they must be explicitely allowed in run-time. See
3919 docs/HOWTO/proxy_certificates.txt for further information.
3922 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
3924 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
3925 server and client random values. Previously
3926 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
3927 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
3929 This change has negligible security impact because:
3931 1. Server and client random values still have 24 bytes of pseudo random
3934 2. Server and client random values are sent in the clear in the initial
3937 3. The master secret is derived using the premaster secret (48 bytes in
3938 size for static RSA ciphersuites) as well as client server and random
3941 The OpenSSL team would like to thank the UK NISCC for bringing this issue
3944 [Stephen Henson, reported by UK NISCC]
3946 *) Use Windows randomness collection on Cygwin.
3949 *) Fix hang in EGD/PRNGD query when communication socket is closed
3950 prematurely by EGD/PRNGD.
3951 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
3953 *) Prompt for pass phrases when appropriate for PKCS12 input format.
3956 *) Back-port of selected performance improvements from development
3957 branch, as well as improved support for PowerPC platforms.
3960 *) Add lots of checks for memory allocation failure, error codes to indicate
3961 failure and freeing up memory if a failure occurs.
3962 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
3964 *) Add new -passin argument to dgst.
3967 *) Perform some character comparisons of different types in X509_NAME_cmp:
3968 this is needed for some certificates that reencode DNs into UTF8Strings
3969 (in violation of RFC3280) and can't or wont issue name rollover
3973 *) Make an explicit check during certificate validation to see that
3974 the CA setting in each certificate on the chain is correct. As a
3975 side effect always do the following basic checks on extensions,
3976 not just when there's an associated purpose to the check:
3978 - if there is an unhandled critical extension (unless the user
3979 has chosen to ignore this fault)
3980 - if the path length has been exceeded (if one is set at all)
3981 - that certain extensions fit the associated purpose (if one has
3985 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
3987 *) Avoid a race condition when CRLs are checked in a multi threaded
3988 environment. This would happen due to the reordering of the revoked
3989 entries during signature checking and serial number lookup. Now the
3990 encoding is cached and the serial number sort performed under a lock.
3991 Add new STACK function sk_is_sorted().
3994 *) Add Delta CRL to the extension code.
3997 *) Various fixes to s3_pkt.c so alerts are sent properly.
3998 [David Holmes <d.holmes@f5.com>]
4000 *) Reduce the chances of duplicate issuer name and serial numbers (in
4001 violation of RFC3280) using the OpenSSL certificate creation utilities.
4002 This is done by creating a random 64 bit value for the initial serial
4003 number when a serial number file is created or when a self signed
4004 certificate is created using 'openssl req -x509'. The initial serial
4005 number file is created using 'openssl x509 -next_serial' in CA.pl
4006 rather than being initialized to 1.
4009 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
4011 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4012 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4013 [Joe Orton, Steve Henson]
4015 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
4017 [Joe Orton, Steve Henson]
4019 *) Make it possible to have multiple active certificates with the same
4020 subject in the CA index file. This is done only if the keyword
4021 'unique_subject' is set to 'no' in the main CA section (default
4022 if 'CA_default') of the configuration file. The value is saved
4023 with the database itself in a separate index attribute file,
4024 named like the index file with '.attr' appended to the name.
4027 *) X509 verify fixes. Disable broken certificate workarounds when
4028 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
4029 keyUsage extension present. Don't accept CRLs with unhandled critical
4030 extensions: since verify currently doesn't process CRL extensions this
4031 rejects a CRL with *any* critical extensions. Add new verify error codes
4035 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
4036 A clarification of RFC2560 will require the use of OCTET STRINGs and
4037 some implementations cannot handle the current raw format. Since OpenSSL
4038 copies and compares OCSP nonces as opaque blobs without any attempt at
4039 parsing them this should not create any compatibility issues.
4042 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
4043 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
4044 this HMAC (and other) operations are several times slower than OpenSSL
4048 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
4049 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4051 *) Use the correct content when signing type "other".
4054 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
4056 *) Fix various bugs revealed by running the NISCC test suite:
4058 Stop out of bounds reads in the ASN1 code when presented with
4059 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4061 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
4063 If verify callback ignores invalid public key errors don't try to check
4064 certificate signature with the NULL public key.
4068 *) New -ignore_err option in ocsp application to stop the server
4069 exiting on the first error in a request.
4072 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4073 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4077 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4078 extra data after the compression methods not only for TLS 1.0
4079 but also for SSL 3.0 (as required by the specification).
4080 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4082 *) Change X509_certificate_type() to mark the key as exported/exportable
4083 when it's 512 *bits* long, not 512 bytes.
4086 *) Change AES_cbc_encrypt() so it outputs exact multiple of
4087 blocks during encryption.
4090 *) Various fixes to base64 BIO and non blocking I/O. On write
4091 flushes were not handled properly if the BIO retried. On read
4092 data was not being buffered properly and had various logic bugs.
4093 This also affects blocking I/O when the data being decoded is a
4097 *) Various S/MIME bugfixes and compatibility changes:
4098 output correct application/pkcs7 MIME type if
4099 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
4100 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
4101 of files as .eml work). Correctly handle very long lines in MIME
4105 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
4107 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4108 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4109 a protocol version number mismatch like a decryption error
4110 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4113 *) Turn on RSA blinding by default in the default implementation
4114 to avoid a timing attack. Applications that don't want it can call
4115 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4116 They would be ill-advised to do so in most cases.
4117 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4119 *) Change RSA blinding code so that it works when the PRNG is not
4120 seeded (in this case, the secret RSA exponent is abused as
4121 an unpredictable seed -- if it is not unpredictable, there
4122 is no point in blinding anyway). Make RSA blinding thread-safe
4123 by remembering the creator's thread ID in rsa->blinding and
4124 having all other threads use local one-time blinding factors
4125 (this requires more computation than sharing rsa->blinding, but
4126 avoids excessive locking; and if an RSA object is not shared
4127 between threads, blinding will still be very fast).
4130 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
4131 ENGINE as defaults for all supported algorithms irrespective of
4132 the 'flags' parameter. 'flags' is now honoured, so applications
4133 should make sure they are passing it correctly.
4136 *) Target "mingw" now allows native Windows code to be generated in
4137 the Cygwin environment as well as with the MinGW compiler.
4140 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
4142 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4143 via timing by performing a MAC computation even if incorrrect
4144 block cipher padding has been found. This is a countermeasure
4145 against active attacks where the attacker has to distinguish
4146 between bad padding and a MAC verification error. (CVE-2003-0078)
4148 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4149 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4150 Martin Vuagnoux (EPFL, Ilion)]
4152 *) Make the no-err option work as intended. The intention with no-err
4153 is not to have the whole error stack handling routines removed from
4154 libcrypto, it's only intended to remove all the function name and
4155 reason texts, thereby removing some of the footprint that may not
4156 be interesting if those errors aren't displayed anyway.
4158 NOTE: it's still possible for any application or module to have it's
4159 own set of error texts inserted. The routines are there, just not
4160 used by default when no-err is given.
4163 *) Add support for FreeBSD on IA64.
4164 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
4166 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
4167 Kerberos function mit_des_cbc_cksum(). Before this change,
4168 the value returned by DES_cbc_cksum() was like the one from
4169 mit_des_cbc_cksum(), except the bytes were swapped.
4170 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
4172 *) Allow an application to disable the automatic SSL chain building.
4173 Before this a rather primitive chain build was always performed in
4174 ssl3_output_cert_chain(): an application had no way to send the
4175 correct chain if the automatic operation produced an incorrect result.
4177 Now the chain builder is disabled if either:
4179 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
4181 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
4183 The reasoning behind this is that an application would not want the
4184 auto chain building to take place if extra chain certificates are
4185 present and it might also want a means of sending no additional
4186 certificates (for example the chain has two certificates and the
4190 *) Add the possibility to build without the ENGINE framework.
4191 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4193 *) Under Win32 gmtime() can return NULL: check return value in
4194 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
4197 *) DSA routines: under certain error conditions uninitialized BN objects
4198 could be freed. Solution: make sure initialization is performed early
4199 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
4200 Nils Larsch <nla@trustcenter.de> via PR#459)
4203 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
4204 checked on reconnect on the client side, therefore session resumption
4205 could still fail with a "ssl session id is different" error. This
4206 behaviour is masked when SSL_OP_ALL is used due to
4207 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
4208 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4209 followup to PR #377.
4212 *) IA-32 assembler support enhancements: unified ELF targets, support
4213 for SCO/Caldera platforms, fix for Cygwin shared build.
4216 *) Add support for FreeBSD on sparc64. As a consequence, support for
4217 FreeBSD on non-x86 processors is separate from x86 processors on
4218 the config script, much like the NetBSD support.
4219 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
4221 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
4223 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
4226 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
4227 code (06) was taken as the first octet of the session ID and the last
4228 octet was ignored consequently. As a result SSLv2 client side session
4229 caching could not have worked due to the session ID mismatch between
4231 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4235 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
4236 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
4240 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
4241 seems that in spite of existing for more than a year, many application
4242 author have done nothing to provide the necessary callbacks, which
4243 means that this particular engine will not work properly anywhere.
4244 This is a very unfortunate situation which forces us, in the name
4245 of usability, to give the hw_ncipher.c a static lock, which is part
4247 NOTE: This is for the 0.9.7 series ONLY. This hack will never
4248 appear in 0.9.8 or later. We EXPECT application authors to have
4249 dealt properly with this when 0.9.8 is released (unless we actually
4250 make such changes in the libcrypto locking code that changes will
4251 have to be made anyway).
4254 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
4255 octets have been read, EOF or an error occurs. Without this change
4256 some truncated ASN1 structures will not produce an error.
4259 *) Disable Heimdal support, since it hasn't been fully implemented.
4260 Still give the possibility to force the use of Heimdal, but with
4261 warnings and a request that patches get sent to openssl-dev.
4264 *) Add the VC-CE target, introduce the WINCE sysname, and add
4265 INSTALL.WCE and appropriate conditionals to make it build.
4266 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4268 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
4269 cygssl-x.y.z.dll, where x, y and z are the major, minor and
4270 edit numbers of the version.
4271 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4273 *) Introduce safe string copy and catenation functions
4274 (BUF_strlcpy() and BUF_strlcat()).
4275 [Ben Laurie (CHATS) and Richard Levitte]
4277 *) Avoid using fixed-size buffers for one-line DNs.
4278 [Ben Laurie (CHATS)]
4280 *) Add BUF_MEM_grow_clean() to avoid information leakage when
4281 resizing buffers containing secrets, and use where appropriate.
4282 [Ben Laurie (CHATS)]
4284 *) Avoid using fixed size buffers for configuration file location.
4285 [Ben Laurie (CHATS)]
4287 *) Avoid filename truncation for various CA files.
4288 [Ben Laurie (CHATS)]
4290 *) Use sizeof in preference to magic numbers.
4291 [Ben Laurie (CHATS)]
4293 *) Avoid filename truncation in cert requests.
4294 [Ben Laurie (CHATS)]
4296 *) Add assertions to check for (supposedly impossible) buffer
4298 [Ben Laurie (CHATS)]
4300 *) Don't cache truncated DNS entries in the local cache (this could
4301 potentially lead to a spoofing attack).
4302 [Ben Laurie (CHATS)]
4304 *) Fix various buffers to be large enough for hex/decimal
4305 representations in a platform independent manner.
4306 [Ben Laurie (CHATS)]
4308 *) Add CRYPTO_realloc_clean() to avoid information leakage when
4309 resizing buffers containing secrets, and use where appropriate.
4310 [Ben Laurie (CHATS)]
4312 *) Add BIO_indent() to avoid much slightly worrying code to do
4314 [Ben Laurie (CHATS)]
4316 *) Convert sprintf()/BIO_puts() to BIO_printf().
4317 [Ben Laurie (CHATS)]
4319 *) buffer_gets() could terminate with the buffer only half
4321 [Ben Laurie (CHATS)]
4323 *) Add assertions to prevent user-supplied crypto functions from
4324 overflowing internal buffers by having large block sizes, etc.
4325 [Ben Laurie (CHATS)]
4327 *) New OPENSSL_assert() macro (similar to assert(), but enabled
4329 [Ben Laurie (CHATS)]
4331 *) Eliminate unused copy of key in RC4.
4332 [Ben Laurie (CHATS)]
4334 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
4335 [Ben Laurie (CHATS)]
4337 *) Fix off-by-one error in EGD path.
4338 [Ben Laurie (CHATS)]
4340 *) If RANDFILE path is too long, ignore instead of truncating.
4341 [Ben Laurie (CHATS)]
4343 *) Eliminate unused and incorrectly sized X.509 structure
4345 [Ben Laurie (CHATS)]
4347 *) Eliminate unused and dangerous function knumber().
4348 [Ben Laurie (CHATS)]
4350 *) Eliminate unused and dangerous structure, KSSL_ERR.
4351 [Ben Laurie (CHATS)]
4353 *) Protect against overlong session ID context length in an encoded
4354 session object. Since these are local, this does not appear to be
4356 [Ben Laurie (CHATS)]
4358 *) Change from security patch (see 0.9.6e below) that did not affect
4359 the 0.9.6 release series:
4361 Remote buffer overflow in SSL3 protocol - an attacker could
4362 supply an oversized master key in Kerberos-enabled versions.
4364 [Ben Laurie (CHATS)]
4366 *) Change the SSL kerb5 codes to match RFC 2712.
4369 *) Make -nameopt work fully for req and add -reqopt switch.
4370 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
4372 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
4373 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
4375 *) Make sure tests can be performed even if the corresponding algorithms
4376 have been removed entirely. This was also the last step to make
4377 OpenSSL compilable with DJGPP under all reasonable conditions.
4378 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
4380 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
4381 to allow version independent disabling of normally unselected ciphers,
4382 which may be activated as a side-effect of selecting a single cipher.
4384 (E.g., cipher list string "RSA" enables ciphersuites that are left
4385 out of "ALL" because they do not provide symmetric encryption.
4386 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
4387 [Lutz Jaenicke, Bodo Moeller]
4389 *) Add appropriate support for separate platform-dependent build
4390 directories. The recommended way to make a platform-dependent
4391 build directory is the following (tested on Linux), maybe with
4394 # Place yourself outside of the OpenSSL source tree. In
4395 # this example, the environment variable OPENSSL_SOURCE
4396 # is assumed to contain the absolute OpenSSL source directory.
4397 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
4398 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
4399 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
4400 mkdir -p `dirname $F`
4401 ln -s $OPENSSL_SOURCE/$F $F
4404 To be absolutely sure not to disturb the source tree, a "make clean"
4405 is a good thing. If it isn't successfull, don't worry about it,
4406 it probably means the source directory is very clean.
4409 *) Make sure any ENGINE control commands make local copies of string
4410 pointers passed to them whenever necessary. Otherwise it is possible
4411 the caller may have overwritten (or deallocated) the original string
4412 data when a later ENGINE operation tries to use the stored values.
4413 [Götz Babin-Ebell <babinebell@trustcenter.de>]
4415 *) Improve diagnostics in file reading and command-line digests.
4416 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
4418 *) Add AES modes CFB and OFB to the object database. Correct an
4419 error in AES-CFB decryption.
4422 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
4423 allows existing EVP_CIPHER_CTX structures to be reused after
4424 calling EVP_*Final(). This behaviour is used by encryption
4425 BIOs and some applications. This has the side effect that
4426 applications must explicitly clean up cipher contexts with
4427 EVP_CIPHER_CTX_cleanup() or they will leak memory.
4430 *) Check the values of dna and dnb in bn_mul_recursive before calling
4431 bn_mul_comba (a non zero value means the a or b arrays do not contain
4432 n2 elements) and fallback to bn_mul_normal if either is not zero.
4435 *) Fix escaping of non-ASCII characters when using the -subj option
4436 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
4439 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
4440 form for "surname", serialNumber has no short form.
4441 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
4442 therefore remove "mail" short name for "internet 7".
4443 The OID for unique identifiers in X509 certificates is
4444 x500UniqueIdentifier, not uniqueIdentifier.
4445 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
4448 *) Add an "init" command to the ENGINE config module and auto initialize
4449 ENGINEs. Without any "init" command the ENGINE will be initialized
4450 after all ctrl commands have been executed on it. If init=1 the
4451 ENGINE is initailized at that point (ctrls before that point are run
4452 on the uninitialized ENGINE and after on the initialized one). If
4453 init=0 then the ENGINE will not be iniatialized at all.
4456 *) Fix the 'app_verify_callback' interface so that the user-defined
4457 argument is actually passed to the callback: In the
4458 SSL_CTX_set_cert_verify_callback() prototype, the callback
4459 declaration has been changed from
4462 int (*cb)(X509_STORE_CTX *,void *);
4463 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
4464 i=s->ctx->app_verify_callback(&ctx)
4465 has been changed into
4466 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
4468 To update applications using SSL_CTX_set_cert_verify_callback(),
4469 a dummy argument can be added to their callback functions.
4470 [D. K. Smetters <smetters@parc.xerox.com>]
4472 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
4473 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
4475 *) Add and OPENSSL_LOAD_CONF define which will cause
4476 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
4477 This allows older applications to transparently support certain
4478 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
4479 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
4480 load the config file and OPENSSL_add_all_algorithms_conf() which will
4481 always load it have also been added.
4484 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
4485 Adjust NIDs and EVP layer.
4486 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4488 *) Config modules support in openssl utility.
4490 Most commands now load modules from the config file,
4491 though in a few (such as version) this isn't done
4492 because it couldn't be used for anything.
4494 In the case of ca and req the config file used is
4495 the same as the utility itself: that is the -config
4496 command line option can be used to specify an
4500 *) Move default behaviour from OPENSSL_config(). If appname is NULL
4501 use "openssl_conf" if filename is NULL use default openssl config file.
4504 *) Add an argument to OPENSSL_config() to allow the use of an alternative
4505 config section name. Add a new flag to tolerate a missing config file
4506 and move code to CONF_modules_load_file().
4509 *) Support for crypto accelerator cards from Accelerated Encryption
4510 Processing, www.aep.ie. (Use engine 'aep')
4511 The support was copied from 0.9.6c [engine] and adapted/corrected
4512 to work with the new engine framework.
4513 [AEP Inc. and Richard Levitte]
4515 *) Support for SureWare crypto accelerator cards from Baltimore
4516 Technologies. (Use engine 'sureware')
4517 The support was copied from 0.9.6c [engine] and adapted
4518 to work with the new engine framework.
4521 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
4522 make the newer ENGINE framework commands for the CHIL engine work.
4523 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
4525 *) Make it possible to produce shared libraries on ReliantUNIX.
4526 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
4528 *) Add the configuration target debug-linux-ppro.
4529 Make 'openssl rsa' use the general key loading routines
4530 implemented in apps.c, and make those routines able to
4531 handle the key format FORMAT_NETSCAPE and the variant
4533 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4535 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4536 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4538 *) Add -keyform to rsautl, and document -engine.
4539 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
4541 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
4542 BIO_R_NO_SUCH_FILE error code rather than the generic
4543 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
4546 *) Add new functions
4548 ERR_peek_last_error_line
4549 ERR_peek_last_error_line_data.
4550 These are similar to
4553 ERR_peek_error_line_data,
4554 but report on the latest error recorded rather than the first one
4555 still in the error queue.
4556 [Ben Laurie, Bodo Moeller]
4558 *) default_algorithms option in ENGINE config module. This allows things
4560 default_algorithms = ALL
4561 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
4564 *) Prelminary ENGINE config module.
4567 *) New experimental application configuration code.
4570 *) Change the AES code to follow the same name structure as all other
4571 symmetric ciphers, and behave the same way. Move everything to
4572 the directory crypto/aes, thereby obsoleting crypto/rijndael.
4573 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4575 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
4576 [Ben Laurie and Theo de Raadt]
4578 *) Add option to output public keys in req command.
4579 [Massimiliano Pala madwolf@openca.org]
4581 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
4582 (up to about 10% better than before for P-192 and P-224).
4585 *) New functions/macros
4587 SSL_CTX_set_msg_callback(ctx, cb)
4588 SSL_CTX_set_msg_callback_arg(ctx, arg)
4589 SSL_set_msg_callback(ssl, cb)
4590 SSL_set_msg_callback_arg(ssl, arg)
4592 to request calling a callback function
4594 void cb(int write_p, int version, int content_type,
4595 const void *buf, size_t len, SSL *ssl, void *arg)
4597 whenever a protocol message has been completely received
4598 (write_p == 0) or sent (write_p == 1). Here 'version' is the
4599 protocol version according to which the SSL library interprets
4600 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
4601 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
4602 the content type as defined in the SSL 3.0/TLS 1.0 protocol
4603 specification (change_cipher_spec(20), alert(21), handshake(22)).
4604 'buf' and 'len' point to the actual message, 'ssl' to the
4605 SSL object, and 'arg' is the application-defined value set by
4606 SSL[_CTX]_set_msg_callback_arg().
4608 'openssl s_client' and 'openssl s_server' have new '-msg' options
4609 to enable a callback that displays all protocol messages.
4612 *) Change the shared library support so shared libraries are built as
4613 soon as the corresponding static library is finished, and thereby get
4614 openssl and the test programs linked against the shared library.
4615 This still only happens when the keyword "shard" has been given to
4616 the configuration scripts.
4618 NOTE: shared library support is still an experimental thing, and
4619 backward binary compatibility is still not guaranteed.
4620 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
4622 *) Add support for Subject Information Access extension.
4623 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4625 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
4626 additional bytes when new memory had to be allocated, not just
4627 when reusing an existing buffer.
4630 *) New command line and configuration option 'utf8' for the req command.
4631 This allows field values to be specified as UTF8 strings.
4634 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
4635 runs for the former and machine-readable output for the latter.
4638 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
4639 of the e-mail address in the DN (i.e., it will go into a certificate
4640 extension only). The new configuration file option 'email_in_dn = no'
4641 has the same effect.
4642 [Massimiliano Pala madwolf@openca.org]
4644 *) Change all functions with names starting with des_ to be starting
4645 with DES_ instead. Add wrappers that are compatible with libdes,
4646 but are named _ossl_old_des_*. Finally, add macros that map the
4647 des_* symbols to the corresponding _ossl_old_des_* if libdes
4648 compatibility is desired. If OpenSSL 0.9.6c compatibility is
4649 desired, the des_* symbols will be mapped to DES_*, with one
4652 Since we provide two compatibility mappings, the user needs to
4653 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
4654 compatibility is desired. The default (i.e., when that macro
4655 isn't defined) is OpenSSL 0.9.6c compatibility.
4657 There are also macros that enable and disable the support of old
4658 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
4659 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
4660 are defined, the default will apply: to support the old des routines.
4662 In either case, one must include openssl/des.h to get the correct
4663 definitions. Do not try to just include openssl/des_old.h, that
4666 NOTE: This is a major break of an old API into a new one. Software
4667 authors are encouraged to switch to the DES_ style functions. Some
4668 time in the future, des_old.h and the libdes compatibility functions
4669 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
4670 default), and then completely removed.
4673 *) Test for certificates which contain unsupported critical extensions.
4674 If such a certificate is found during a verify operation it is
4675 rejected by default: this behaviour can be overridden by either
4676 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
4677 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
4678 X509_supported_extension() has also been added which returns 1 if a
4679 particular extension is supported.
4682 *) Modify the behaviour of EVP cipher functions in similar way to digests
4683 to retain compatibility with existing code.
4686 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
4687 compatibility with existing code. In particular the 'ctx' parameter does
4688 not have to be to be initialized before the call to EVP_DigestInit() and
4689 it is tidied up after a call to EVP_DigestFinal(). New function
4690 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
4691 EVP_MD_CTX_copy() changed to not require the destination to be
4692 initialized valid and new function EVP_MD_CTX_copy_ex() added which
4693 requires the destination to be valid.
4695 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
4696 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
4699 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
4700 so that complete 'Handshake' protocol structures are kept in memory
4701 instead of overwriting 'msg_type' and 'length' with 'body' data.
4704 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
4705 [Massimo Santin via Richard Levitte]
4707 *) Major restructuring to the underlying ENGINE code. This includes
4708 reduction of linker bloat, separation of pure "ENGINE" manipulation
4709 (initialisation, etc) from functionality dealing with implementations
4710 of specific crypto iterfaces. This change also introduces integrated
4711 support for symmetric ciphers and digest implementations - so ENGINEs
4712 can now accelerate these by providing EVP_CIPHER and EVP_MD
4713 implementations of their own. This is detailed in crypto/engine/README
4714 as it couldn't be adequately described here. However, there are a few
4715 API changes worth noting - some RSA, DSA, DH, and RAND functions that
4716 were changed in the original introduction of ENGINE code have now
4717 reverted back - the hooking from this code to ENGINE is now a good
4718 deal more passive and at run-time, operations deal directly with
4719 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
4720 dereferencing through an ENGINE pointer any more. Also, the ENGINE
4721 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
4722 they were not being used by the framework as there is no concept of a
4723 BIGNUM_METHOD and they could not be generalised to the new
4724 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
4725 ENGINE_cpy() has been removed as it cannot be consistently defined in
4729 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
4732 *) Change mkdef.pl to sort symbols that get the same entry number,
4733 and make sure the automatically generated functions ERR_load_*
4734 become part of libeay.num as well.
4737 *) New function SSL_renegotiate_pending(). This returns true once
4738 renegotiation has been requested (either SSL_renegotiate() call
4739 or HelloRequest/ClientHello receveived from the peer) and becomes
4740 false once a handshake has been completed.
4741 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
4742 sends a HelloRequest, but does not ensure that a handshake takes
4743 place. SSL_renegotiate_pending() is useful for checking if the
4744 client has followed the request.)
4747 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
4748 By default, clients may request session resumption even during
4749 renegotiation (if session ID contexts permit); with this option,
4750 session resumption is possible only in the first handshake.
4752 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4753 more bits available for options that should not be part of
4754 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4757 *) Add some demos for certificate and certificate request creation.
4760 *) Make maximum certificate chain size accepted from the peer application
4761 settable (SSL*_get/set_max_cert_list()), as proposed by
4762 "Douglas E. Engert" <deengert@anl.gov>.
4765 *) Add support for shared libraries for Unixware-7
4766 (Boyd Lynn Gerber <gerberb@zenez.com>).
4769 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4770 be done prior to destruction. Use this to unload error strings from
4771 ENGINEs that load their own error strings. NB: This adds two new API
4772 functions to "get" and "set" this destroy handler in an ENGINE.
4775 *) Alter all existing ENGINE implementations (except "openssl" and
4776 "openbsd") to dynamically instantiate their own error strings. This
4777 makes them more flexible to be built both as statically-linked ENGINEs
4778 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4779 Also, add stub code to each that makes building them as self-contained
4780 shared-libraries easier (see README.ENGINE).
4783 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4784 implementations into applications that are completely implemented in
4785 self-contained shared-libraries. The "dynamic" ENGINE exposes control
4786 commands that can be used to configure what shared-library to load and
4787 to control aspects of the way it is handled. Also, made an update to
4788 the README.ENGINE file that brings its information up-to-date and
4789 provides some information and instructions on the "dynamic" ENGINE
4790 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
4793 *) Make it possible to unload ranges of ERR strings with a new
4794 "ERR_unload_strings" function.
4797 *) Add a copy() function to EVP_MD.
4800 *) Make EVP_MD routines take a context pointer instead of just the
4801 md_data void pointer.
4804 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
4805 that the digest can only process a single chunk of data
4806 (typically because it is provided by a piece of
4807 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
4808 is only going to provide a single chunk of data, and hence the
4809 framework needn't accumulate the data for oneshot drivers.
4812 *) As with "ERR", make it possible to replace the underlying "ex_data"
4813 functions. This change also alters the storage and management of global
4814 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
4815 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
4816 index counters. The API functions that use this state have been changed
4817 to take a "class_index" rather than pointers to the class's local STACK
4818 and counter, and there is now an API function to dynamically create new
4819 classes. This centralisation allows us to (a) plug a lot of the
4820 thread-safety problems that existed, and (b) makes it possible to clean
4821 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
4822 such data would previously have always leaked in application code and
4823 workarounds were in place to make the memory debugging turn a blind eye
4824 to it. Application code that doesn't use this new function will still
4825 leak as before, but their memory debugging output will announce it now
4826 rather than letting it slide.
4828 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
4829 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
4830 has a return value to indicate success or failure.
4833 *) Make it possible to replace the underlying "ERR" functions such that the
4834 global state (2 LHASH tables and 2 locks) is only used by the "default"
4835 implementation. This change also adds two functions to "get" and "set"
4836 the implementation prior to it being automatically set the first time
4837 any other ERR function takes place. Ie. an application can call "get",
4838 pass the return value to a module it has just loaded, and that module
4839 can call its own "set" function using that value. This means the
4840 module's "ERR" operations will use (and modify) the error state in the
4841 application and not in its own statically linked copy of OpenSSL code.
4844 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
4845 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
4846 the operation, and provides a more encapsulated way for external code
4847 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
4848 to use these functions rather than manually incrementing the counts.
4850 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
4853 *) Add EVP test program.
4856 *) Add symmetric cipher support to ENGINE. Expect the API to change!
4859 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
4860 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
4861 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
4862 These allow a CRL to be built without having to access X509_CRL fields
4863 directly. Modify 'ca' application to use new functions.
4866 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
4867 bug workarounds. Rollback attack detection is a security feature.
4868 The problem will only arise on OpenSSL servers when TLSv1 is not
4869 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
4870 Software authors not wanting to support TLSv1 will have special reasons
4871 for their choice and can explicitly enable this option.
4872 [Bodo Moeller, Lutz Jaenicke]
4874 *) Rationalise EVP so it can be extended: don't include a union of
4875 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
4876 (similar to those existing for EVP_CIPHER_CTX).
4881 EVP_MD_CTX_init(&md); /* new function call */
4882 EVP_DigestInit(&md, EVP_sha1());
4883 EVP_DigestUpdate(&md, in, len);
4884 EVP_DigestFinal(&md, out, NULL);
4885 EVP_MD_CTX_cleanup(&md); /* new function call */
4889 *) Make DES key schedule conform to the usual scheme, as well as
4890 correcting its structure. This means that calls to DES functions
4891 now have to pass a pointer to a des_key_schedule instead of a
4892 plain des_key_schedule (which was actually always a pointer
4895 des_key_schedule ks;
4897 des_set_key_checked(..., &ks);
4898 des_ncbc_encrypt(..., &ks, ...);
4900 (Note that a later change renames 'des_...' into 'DES_...'.)
4903 *) Initial reduction of linker bloat: the use of some functions, such as
4904 PEM causes large amounts of unused functions to be linked in due to
4905 poor organisation. For example pem_all.c contains every PEM function
4906 which has a knock on effect of linking in large amounts of (unused)
4907 ASN1 code. Grouping together similar functions and splitting unrelated
4908 functions prevents this.
4911 *) Cleanup of EVP macros.
4914 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
4915 correct _ecb suffix.
4918 *) Add initial OCSP responder support to ocsp application. The
4919 revocation information is handled using the text based index
4920 use by the ca application. The responder can either handle
4921 requests generated internally, supplied in files (for example
4922 via a CGI script) or using an internal minimal server.
4925 *) Add configuration choices to get zlib compression for TLS.
4928 *) Changes to Kerberos SSL for RFC 2712 compliance:
4929 1. Implemented real KerberosWrapper, instead of just using
4930 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
4931 2. Implemented optional authenticator field of KerberosWrapper.
4933 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
4934 and authenticator structs; see crypto/krb5/.
4936 Generalized Kerberos calls to support multiple Kerberos libraries.
4937 [Vern Staats <staatsvr@asc.hpc.mil>,
4938 Jeffrey Altman <jaltman@columbia.edu>
4939 via Richard Levitte]
4941 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
4942 already does with RSA. testdsa.h now has 'priv_key/pub_key'
4943 values for each of the key sizes rather than having just
4944 parameters (and 'speed' generating keys each time).
4947 *) Speed up EVP routines.
4950 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
4951 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
4952 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
4953 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
4955 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
4956 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
4957 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
4960 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
4962 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
4965 *) Added the OS2-EMX target.
4966 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
4968 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
4969 to support NCONF routines in extension code. New function CONF_set_nconf()
4970 to allow functions which take an NCONF to also handle the old LHASH
4971 structure: this means that the old CONF compatible routines can be
4972 retained (in particular wrt extensions) without having to duplicate the
4973 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
4976 *) Enhance the general user interface with mechanisms for inner control
4977 and with possibilities to have yes/no kind of prompts.
4980 *) Change all calls to low level digest routines in the library and
4981 applications to use EVP. Add missing calls to HMAC_cleanup() and
4982 don't assume HMAC_CTX can be copied using memcpy().
4983 [Verdon Walker <VWalker@novell.com>, Steve Henson]
4985 *) Add the possibility to control engines through control names but with
4986 arbitrary arguments instead of just a string.
4987 Change the key loaders to take a UI_METHOD instead of a callback
4988 function pointer. NOTE: this breaks binary compatibility with earlier
4989 versions of OpenSSL [engine].
4990 Adapt the nCipher code for these new conditions and add a card insertion
4994 *) Enhance the general user interface with mechanisms to better support
4995 dialog box interfaces, application-defined prompts, the possibility
4996 to use defaults (for example default passwords from somewhere else)
4997 and interrupts/cancellations.
5000 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
5001 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
5004 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
5005 tidy up some unnecessarily weird code in 'sk_new()').
5006 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
5008 *) Change the key loading routines for ENGINEs to use the same kind
5009 callback (pem_password_cb) as all other routines that need this
5013 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
5014 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
5015 than this minimum value is recommended.
5018 *) New random seeder for OpenVMS, using the system process statistics
5019 that are easily reachable.
5022 *) Windows apparently can't transparently handle global
5023 variables defined in DLLs. Initialisations such as:
5025 const ASN1_ITEM *it = &ASN1_INTEGER_it;
5027 wont compile. This is used by the any applications that need to
5028 declare their own ASN1 modules. This was fixed by adding the option
5029 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
5030 needed for static libraries under Win32.
5033 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
5034 setting of purpose and trust fields. New X509_STORE trust and
5035 purpose functions and tidy up setting in other SSL functions.
5038 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
5039 structure. These are inherited by X509_STORE_CTX when it is
5040 initialised. This allows various defaults to be set in the
5041 X509_STORE structure (such as flags for CRL checking and custom
5042 purpose or trust settings) for functions which only use X509_STORE_CTX
5043 internally such as S/MIME.
5045 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
5046 trust settings if they are not set in X509_STORE. This allows X509_STORE
5047 purposes and trust (in S/MIME for example) to override any set by default.
5049 Add command line options for CRL checking to smime, s_client and s_server
5053 *) Initial CRL based revocation checking. If the CRL checking flag(s)
5054 are set then the CRL is looked up in the X509_STORE structure and
5055 its validity and signature checked, then if the certificate is found
5056 in the CRL the verify fails with a revoked error.
5058 Various new CRL related callbacks added to X509_STORE_CTX structure.
5060 Command line options added to 'verify' application to support this.
5062 This needs some additional work, such as being able to handle multiple
5063 CRLs with different times, extension based lookup (rather than just
5064 by subject name) and ultimately more complete V2 CRL extension
5068 *) Add a general user interface API (crypto/ui/). This is designed
5069 to replace things like des_read_password and friends (backward
5070 compatibility functions using this new API are provided).
5071 The purpose is to remove prompting functions from the DES code
5072 section as well as provide for prompting through dialog boxes in
5073 a window system and the like.
5076 *) Add "ex_data" support to ENGINE so implementations can add state at a
5077 per-structure level rather than having to store it globally.
5080 *) Make it possible for ENGINE structures to be copied when retrieved by
5081 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
5082 This causes the "original" ENGINE structure to act like a template,
5083 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
5084 operational state can be localised to each ENGINE structure, despite the
5085 fact they all share the same "methods". New ENGINE structures returned in
5086 this case have no functional references and the return value is the single
5087 structural reference. This matches the single structural reference returned
5088 by ENGINE_by_id() normally, when it is incremented on the pre-existing
5092 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
5093 needs to match any other type at all we need to manually clear the
5097 *) Changes to the "openssl engine" utility to include;
5098 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
5099 about an ENGINE's available control commands.
5100 - executing control commands from command line arguments using the
5101 '-pre' and '-post' switches. '-post' is only used if '-t' is
5102 specified and the ENGINE is successfully initialised. The syntax for
5103 the individual commands are colon-separated, for example;
5104 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
5107 *) New dynamic control command support for ENGINEs. ENGINEs can now
5108 declare their own commands (numbers), names (strings), descriptions,
5109 and input types for run-time discovery by calling applications. A
5110 subset of these commands are implicitly classed as "executable"
5111 depending on their input type, and only these can be invoked through
5112 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
5113 can be based on user input, config files, etc). The distinction is
5114 that "executable" commands cannot return anything other than a boolean
5115 result and can only support numeric or string input, whereas some
5116 discoverable commands may only be for direct use through
5117 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
5118 pointers, or other custom uses. The "executable" commands are to
5119 support parameterisations of ENGINE behaviour that can be
5120 unambiguously defined by ENGINEs and used consistently across any
5121 OpenSSL-based application. Commands have been added to all the
5122 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
5123 control over shared-library paths without source code alterations.
5126 *) Changed all ENGINE implementations to dynamically allocate their
5127 ENGINEs rather than declaring them statically. Apart from this being
5128 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
5129 this also allows the implementations to compile without using the
5130 internal engine_int.h header.
5133 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
5134 'const' value. Any code that should be able to modify a RAND_METHOD
5135 should already have non-const pointers to it (ie. they should only
5136 modify their own ones).
5139 *) Made a variety of little tweaks to the ENGINE code.
5140 - "atalla" and "ubsec" string definitions were moved from header files
5141 to C code. "nuron" string definitions were placed in variables
5142 rather than hard-coded - allowing parameterisation of these values
5143 later on via ctrl() commands.
5144 - Removed unused "#if 0"'d code.
5145 - Fixed engine list iteration code so it uses ENGINE_free() to release
5146 structural references.
5147 - Constified the RAND_METHOD element of ENGINE structures.
5148 - Constified various get/set functions as appropriate and added
5149 missing functions (including a catch-all ENGINE_cpy that duplicates
5150 all ENGINE values onto a new ENGINE except reference counts/state).
5151 - Removed NULL parameter checks in get/set functions. Setting a method
5152 or function to NULL is a way of cancelling out a previously set
5153 value. Passing a NULL ENGINE parameter is just plain stupid anyway
5154 and doesn't justify the extra error symbols and code.
5155 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
5156 flags from engine_int.h to engine.h.
5157 - Changed prototypes for ENGINE handler functions (init(), finish(),
5158 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
5161 *) Implement binary inversion algorithm for BN_mod_inverse in addition
5162 to the algorithm using long division. The binary algorithm can be
5163 used only if the modulus is odd. On 32-bit systems, it is faster
5164 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
5165 roughly 5-15% for 256-bit moduli), so we use it only for moduli
5166 up to 450 bits. In 64-bit environments, the binary algorithm
5167 appears to be advantageous for much longer moduli; here we use it
5168 for moduli up to 2048 bits.
5171 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
5172 could not support the combine flag in choice fields.
5175 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
5176 extensions from a certificate request to the certificate.
5179 *) Allow multiple 'certopt' and 'nameopt' options to be separated
5180 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
5181 file: this allows the display of the certificate about to be
5182 signed to be customised, to allow certain fields to be included
5183 or excluded and extension details. The old system didn't display
5184 multicharacter strings properly, omitted fields not in the policy
5185 and couldn't display additional details such as extensions.
5188 *) Function EC_POINTs_mul for multiple scalar multiplication
5189 of an arbitrary number of elliptic curve points
5190 \sum scalars[i]*points[i],
5191 optionally including the generator defined for the EC_GROUP:
5192 scalar*generator + \sum scalars[i]*points[i].
5194 EC_POINT_mul is a simple wrapper function for the typical case
5195 that the point list has just one item (besides the optional
5199 *) First EC_METHODs for curves over GF(p):
5201 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
5202 operations and provides various method functions that can also
5203 operate with faster implementations of modular arithmetic.
5205 EC_GFp_mont_method() reuses most functions that are part of
5206 EC_GFp_simple_method, but uses Montgomery arithmetic.
5208 [Bodo Moeller; point addition and point doubling
5209 implementation directly derived from source code provided by
5210 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
5212 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
5213 crypto/ec/ec_lib.c):
5215 Curves are EC_GROUP objects (with an optional group generator)
5216 based on EC_METHODs that are built into the library.
5218 Points are EC_POINT objects based on EC_GROUP objects.
5220 Most of the framework would be able to handle curves over arbitrary
5221 finite fields, but as there are no obvious types for fields other
5222 than GF(p), some functions are limited to that for now.
5225 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
5226 that the file contains a complete HTTP response.
5229 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
5230 change the def and num file printf format specifier from "%-40sXXX"
5231 to "%-39s XXX". The latter will always guarantee a space after the
5232 field while the former will cause them to run together if the field
5233 is 40 of more characters long.
5236 *) Constify the cipher and digest 'method' functions and structures
5237 and modify related functions to take constant EVP_MD and EVP_CIPHER
5241 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
5242 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
5245 *) Modify EVP_Digest*() routines so they now return values. Although the
5246 internal software routines can never fail additional hardware versions
5250 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
5252 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
5253 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
5256 ERR_R_NESTED_ASN1_ERROR
5258 ERR_R_MISSING_ASN1_EOS
5259 were 4 .. 9, conflicting with
5260 ERR_LIB_RSA (= ERR_R_RSA_LIB)
5262 ERR_LIB_PEM (= ERR_R_PEM_LIB).
5263 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
5265 Add new error code 'ERR_R_INTERNAL_ERROR'.
5268 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
5272 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
5273 sets the subject name for a new request or supersedes the
5274 subject name in a given request. Formats that can be parsed are
5275 'CN=Some Name, OU=myOU, C=IT'
5277 'CN=Some Name/OU=myOU/C=IT'.
5279 Add options '-batch' and '-verbose' to 'openssl req'.
5280 [Massimiliano Pala <madwolf@hackmasters.net>]
5282 *) Introduce the possibility to access global variables through
5283 functions on platform were that's the best way to handle exporting
5284 global variables in shared libraries. To enable this functionality,
5285 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
5286 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
5287 is normally done by Configure or something similar).
5289 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
5290 in the source file (foo.c) like this:
5292 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
5293 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
5295 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
5296 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
5298 OPENSSL_DECLARE_GLOBAL(int,foo);
5299 #define foo OPENSSL_GLOBAL_REF(foo)
5300 OPENSSL_DECLARE_GLOBAL(double,bar);
5301 #define bar OPENSSL_GLOBAL_REF(bar)
5303 The #defines are very important, and therefore so is including the
5304 header file everywhere where the defined globals are used.
5306 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
5307 of ASN.1 items, but that structure is a bit different.
5309 The largest change is in util/mkdef.pl which has been enhanced with
5310 better and easier to understand logic to choose which symbols should
5311 go into the Windows .def files as well as a number of fixes and code
5312 cleanup (among others, algorithm keywords are now sorted
5313 lexicographically to avoid constant rewrites).
5316 *) In BN_div() keep a copy of the sign of 'num' before writing the
5317 result to 'rm' because if rm==num the value will be overwritten
5318 and produce the wrong result if 'num' is negative: this caused
5319 problems with BN_mod() and BN_nnmod().
5322 *) Function OCSP_request_verify(). This checks the signature on an
5323 OCSP request and verifies the signer certificate. The signer
5324 certificate is just checked for a generic purpose and OCSP request
5328 *) Add OCSP_check_validity() function to check the validity of OCSP
5329 responses. OCSP responses are prepared in real time and may only
5330 be a few seconds old. Simply checking that the current time lies
5331 between thisUpdate and nextUpdate max reject otherwise valid responses
5332 caused by either OCSP responder or client clock inaccuracy. Instead
5333 we allow thisUpdate and nextUpdate to fall within a certain period of
5334 the current time. The age of the response can also optionally be
5335 checked. Two new options -validity_period and -status_age added to
5339 *) If signature or public key algorithm is unrecognized print out its
5340 OID rather that just UNKNOWN.
5343 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
5344 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
5345 ID to be generated from the issuer certificate alone which can then be
5346 passed to OCSP_id_issuer_cmp().
5349 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
5350 ASN1 modules to export functions returning ASN1_ITEM pointers
5351 instead of the ASN1_ITEM structures themselves. This adds several
5352 new macros which allow the underlying ASN1 function/structure to
5353 be accessed transparently. As a result code should not use ASN1_ITEM
5354 references directly (such as &X509_it) but instead use the relevant
5355 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
5356 use of the new ASN1 code on platforms where exporting structures
5357 is problematical (for example in shared libraries) but exporting
5358 functions returning pointers to structures is not.
5361 *) Add support for overriding the generation of SSL/TLS session IDs.
5362 These callbacks can be registered either in an SSL_CTX or per SSL.
5363 The purpose of this is to allow applications to control, if they wish,
5364 the arbitrary values chosen for use as session IDs, particularly as it
5365 can be useful for session caching in multiple-server environments. A
5366 command-line switch for testing this (and any client code that wishes
5367 to use such a feature) has been added to "s_server".
5368 [Geoff Thorpe, Lutz Jaenicke]
5370 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
5371 of the form '#if defined(...) || defined(...) || ...' and
5372 '#if !defined(...) && !defined(...) && ...'. This also avoids
5373 the growing number of special cases it was previously handling.
5376 *) Make all configuration macros available for application by making
5377 sure they are available in opensslconf.h, by giving them names starting
5378 with "OPENSSL_" to avoid conflicts with other packages and by making
5379 sure e_os2.h will cover all platform-specific cases together with
5381 Additionally, it is now possible to define configuration/platform-
5382 specific names (called "system identities"). In the C code, these
5383 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
5384 macro with the name beginning with "OPENSSL_SYS_", which is determined
5385 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
5389 *) New option -set_serial to 'req' and 'x509' this allows the serial
5390 number to use to be specified on the command line. Previously self
5391 signed certificates were hard coded with serial number 0 and the
5392 CA options of 'x509' had to use a serial number in a file which was
5396 *) New options to 'ca' utility to support V2 CRL entry extensions.
5397 Currently CRL reason, invalidity date and hold instruction are
5398 supported. Add new CRL extensions to V3 code and some new objects.
5401 *) New function EVP_CIPHER_CTX_set_padding() this is used to
5402 disable standard block padding (aka PKCS#5 padding) in the EVP
5403 API, which was previously mandatory. This means that the data is
5404 not padded in any way and so the total length much be a multiple
5405 of the block size, otherwise an error occurs.
5408 *) Initial (incomplete) OCSP SSL support.
5411 *) New function OCSP_parse_url(). This splits up a URL into its host,
5412 port and path components: primarily to parse OCSP URLs. New -url
5413 option to ocsp utility.
5416 *) New nonce behavior. The return value of OCSP_check_nonce() now
5417 reflects the various checks performed. Applications can decide
5418 whether to tolerate certain situations such as an absent nonce
5419 in a response when one was present in a request: the ocsp application
5420 just prints out a warning. New function OCSP_add1_basic_nonce()
5421 this is to allow responders to include a nonce in a response even if
5422 the request is nonce-less.
5425 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
5426 skipped when using openssl x509 multiple times on a single input file,
5427 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
5430 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
5431 set string type: to handle setting ASN1_TIME structures. Fix ca
5432 utility to correctly initialize revocation date of CRLs.
5435 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
5436 the clients preferred ciphersuites and rather use its own preferences.
5437 Should help to work around M$ SGC (Server Gated Cryptography) bug in
5438 Internet Explorer by ensuring unchanged hash method during stepup.
5439 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
5442 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
5443 to aes and add a new 'exist' option to print out symbols that don't
5447 *) Additional options to ocsp utility to allow flags to be set and
5448 additional certificates supplied.
5451 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
5452 OCSP client a number of certificate to only verify the response
5456 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
5457 handle the new API. Currently only ECB, CBC modes supported. Add new
5460 Add TLS AES ciphersuites as described in RFC3268, "Advanced
5461 Encryption Standard (AES) Ciphersuites for Transport Layer
5462 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
5463 not enabled by default and were not part of the "ALL" ciphersuite
5464 alias because they were not yet official; they could be
5465 explicitly requested by specifying the "AESdraft" ciphersuite
5466 group alias. In the final release of OpenSSL 0.9.7, the group
5467 alias is called "AES" and is part of "ALL".)
5468 [Ben Laurie, Steve Henson, Bodo Moeller]
5470 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
5471 request to response.
5474 *) Functions for OCSP responders. OCSP_request_onereq_count(),
5475 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
5476 extract information from a certificate request. OCSP_response_create()
5477 creates a response and optionally adds a basic response structure.
5478 OCSP_basic_add1_status() adds a complete single response to a basic
5479 response and returns the OCSP_SINGLERESP structure just added (to allow
5480 extensions to be included for example). OCSP_basic_add1_cert() adds a
5481 certificate to a basic response and OCSP_basic_sign() signs a basic
5482 response with various flags. New helper functions ASN1_TIME_check()
5483 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
5484 (converts ASN1_TIME to GeneralizedTime).
5487 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
5488 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
5489 structure from a certificate. X509_pubkey_digest() digests the public_key
5490 contents: this is used in various key identifiers.
5493 *) Make sk_sort() tolerate a NULL argument.
5494 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
5496 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
5497 passed by the function are trusted implicitly. If any of them signed the
5498 response then it is assumed to be valid and is not verified.
5501 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
5502 to data. This was previously part of the PKCS7 ASN1 code. This
5503 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
5504 [Steve Henson, reported by Kenneth R. Robinette
5505 <support@securenetterm.com>]
5507 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
5508 routines: without these tracing memory leaks is very painful.
5509 Fix leaks in PKCS12 and PKCS7 routines.
5512 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
5513 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
5514 effectively meant GeneralizedTime would never be used. Now it
5515 is initialised to -1 but X509_time_adj() now has to check the value
5516 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
5517 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
5518 [Steve Henson, reported by Kenneth R. Robinette
5519 <support@securenetterm.com>]
5521 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
5522 result in a zero length in the ASN1_INTEGER structure which was
5523 not consistent with the structure when d2i_ASN1_INTEGER() was used
5524 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
5525 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
5526 where it did not print out a minus for negative ASN1_INTEGER.
5529 *) Add summary printout to ocsp utility. The various functions which
5530 convert status values to strings have been renamed to:
5531 OCSP_response_status_str(), OCSP_cert_status_str() and
5532 OCSP_crl_reason_str() and are no longer static. New options
5533 to verify nonce values and to disable verification. OCSP response
5534 printout format cleaned up.
5537 *) Add additional OCSP certificate checks. These are those specified
5538 in RFC2560. This consists of two separate checks: the CA of the
5539 certificate being checked must either be the OCSP signer certificate
5540 or the issuer of the OCSP signer certificate. In the latter case the
5541 OCSP signer certificate must contain the OCSP signing extended key
5542 usage. This check is performed by attempting to match the OCSP
5543 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
5544 in the OCSP_CERTID structures of the response.
5547 *) Initial OCSP certificate verification added to OCSP_basic_verify()
5548 and related routines. This uses the standard OpenSSL certificate
5549 verify routines to perform initial checks (just CA validity) and
5550 to obtain the certificate chain. Then additional checks will be
5551 performed on the chain. Currently the root CA is checked to see
5552 if it is explicitly trusted for OCSP signing. This is used to set
5553 a root CA as a global signing root: that is any certificate that
5554 chains to that CA is an acceptable OCSP signing certificate.
5557 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
5558 extensions from a separate configuration file.
5559 As when reading extensions from the main configuration file,
5560 the '-extensions ...' option may be used for specifying the
5562 [Massimiliano Pala <madwolf@comune.modena.it>]
5564 *) New OCSP utility. Allows OCSP requests to be generated or
5565 read. The request can be sent to a responder and the output
5566 parsed, outputed or printed in text form. Not complete yet:
5567 still needs to check the OCSP response validity.
5570 *) New subcommands for 'openssl ca':
5571 'openssl ca -status <serial>' prints the status of the cert with
5572 the given serial number (according to the index file).
5573 'openssl ca -updatedb' updates the expiry status of certificates
5575 [Massimiliano Pala <madwolf@comune.modena.it>]
5577 *) New '-newreq-nodes' command option to CA.pl. This is like
5578 '-newreq', but calls 'openssl req' with the '-nodes' option
5579 so that the resulting key is not encrypted.
5580 [Damien Miller <djm@mindrot.org>]
5582 *) New configuration for the GNU Hurd.
5583 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
5585 *) Initial code to implement OCSP basic response verify. This
5586 is currently incomplete. Currently just finds the signer's
5587 certificate and verifies the signature on the response.
5590 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
5591 value of OPENSSLDIR. This is available via the new '-d' option
5592 to 'openssl version', and is also included in 'openssl version -a'.
5595 *) Allowing defining memory allocation callbacks that will be given
5596 file name and line number information in additional arguments
5597 (a const char* and an int). The basic functionality remains, as
5598 well as the original possibility to just replace malloc(),
5599 realloc() and free() by functions that do not know about these
5600 additional arguments. To register and find out the current
5601 settings for extended allocation functions, the following
5602 functions are provided:
5604 CRYPTO_set_mem_ex_functions
5605 CRYPTO_set_locked_mem_ex_functions
5606 CRYPTO_get_mem_ex_functions
5607 CRYPTO_get_locked_mem_ex_functions
5609 These work the same way as CRYPTO_set_mem_functions and friends.
5610 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
5611 extended allocation function is enabled.
5612 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
5613 a conventional allocation function is enabled.
5614 [Richard Levitte, Bodo Moeller]
5616 *) Finish off removing the remaining LHASH function pointer casts.
5617 There should no longer be any prototype-casting required when using
5618 the LHASH abstraction, and any casts that remain are "bugs". See
5619 the callback types and macros at the head of lhash.h for details
5620 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
5623 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
5624 If /dev/[u]random devices are not available or do not return enough
5625 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
5627 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
5628 /etc/entropy will be queried once each in this sequence, quering stops
5629 when enough entropy was collected without querying more sockets.
5632 *) Change the Unix RAND_poll() variant to be able to poll several
5633 random devices, as specified by DEVRANDOM, until a sufficient amount
5634 of data has been collected. We spend at most 10 ms on each file
5635 (select timeout) and read in non-blocking mode. DEVRANDOM now
5636 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
5637 (previously it was just the string "/dev/urandom"), so on typical
5638 platforms the 10 ms delay will never occur.
5639 Also separate out the Unix variant to its own file, rand_unix.c.
5640 For VMS, there's a currently-empty rand_vms.c.
5643 *) Move OCSP client related routines to ocsp_cl.c. These
5644 provide utility functions which an application needing
5645 to issue a request to an OCSP responder and analyse the
5646 response will typically need: as opposed to those which an
5647 OCSP responder itself would need which will be added later.
5649 OCSP_request_sign() signs an OCSP request with an API similar
5650 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
5651 response. OCSP_response_get1_basic() extracts basic response
5652 from response. OCSP_resp_find_status(): finds and extracts status
5653 information from an OCSP_CERTID structure (which will be created
5654 when the request structure is built). These are built from lower
5655 level functions which work on OCSP_SINGLERESP structures but
5656 wont normally be used unless the application wishes to examine
5657 extensions in the OCSP response for example.
5659 Replace nonce routines with a pair of functions.
5660 OCSP_request_add1_nonce() adds a nonce value and optionally
5661 generates a random value. OCSP_check_nonce() checks the
5662 validity of the nonce in an OCSP response.
5665 *) Change function OCSP_request_add() to OCSP_request_add0_id().
5666 This doesn't copy the supplied OCSP_CERTID and avoids the
5667 need to free up the newly created id. Change return type
5668 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
5669 This can then be used to add extensions to the request.
5670 Deleted OCSP_request_new(), since most of its functionality
5671 is now in OCSP_REQUEST_new() (and the case insensitive name
5672 clash) apart from the ability to set the request name which
5673 will be added elsewhere.
5676 *) Update OCSP API. Remove obsolete extensions argument from
5677 various functions. Extensions are now handled using the new
5678 OCSP extension code. New simple OCSP HTTP function which
5679 can be used to send requests and parse the response.
5682 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
5683 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
5684 uses the special reorder version of SET OF to sort the attributes
5685 and reorder them to match the encoded order. This resolves a long
5686 standing problem: a verify on a PKCS7 structure just after signing
5687 it used to fail because the attribute order did not match the
5688 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
5689 it uses the received order. This is necessary to tolerate some broken
5690 software that does not order SET OF. This is handled by encoding
5691 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
5692 to produce the required SET OF.
5695 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
5696 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
5697 files to get correct declarations of the ASN.1 item variables.
5700 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
5701 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
5702 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
5703 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
5704 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
5705 ASN1_ITEM and no wrapper functions.
5708 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
5709 replace the old function pointer based I/O routines. Change most of
5710 the *_d2i_bio() and *_d2i_fp() functions to use these.
5713 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
5714 lines, recognice more "algorithms" that can be deselected, and make
5715 it complain about algorithm deselection that isn't recognised.
5718 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
5719 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
5720 to use new functions. Add NO_ASN1_OLD which can be set to remove
5721 some old style ASN1 functions: this can be used to determine if old
5722 code will still work when these eventually go away.
5725 *) New extension functions for OCSP structures, these follow the
5726 same conventions as certificates and CRLs.
5729 *) New function X509V3_add1_i2d(). This automatically encodes and
5730 adds an extension. Its behaviour can be customised with various
5731 flags to append, replace or delete. Various wrappers added for
5732 certifcates and CRLs.
5735 *) Fix to avoid calling the underlying ASN1 print routine when
5736 an extension cannot be parsed. Correct a typo in the
5737 OCSP_SERVICELOC extension. Tidy up print OCSP format.
5740 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
5741 entries for variables.
5744 *) Add functionality to apps/openssl.c for detecting locking
5745 problems: As the program is single-threaded, all we have
5746 to do is register a locking callback using an array for
5747 storing which locks are currently held by the program.
5750 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
5751 SSL_get_ex_data_X509_STORE_idx(), which is used in
5752 ssl_verify_cert_chain() and thus can be called at any time
5753 during TLS/SSL handshakes so that thread-safety is essential.
5754 Unfortunately, the ex_data design is not at all suited
5755 for multi-threaded use, so it probably should be abolished.
5758 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5759 [Broadcom, tweaked and integrated by Geoff Thorpe]
5761 *) Move common extension printing code to new function
5762 X509V3_print_extensions(). Reorganise OCSP print routines and
5763 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5766 *) New function X509_signature_print() to remove duplication in some
5770 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5771 set (this was treated exactly the same as SET OF previously). This
5772 is used to reorder the STACK representing the structure to match the
5773 encoding. This will be used to get round a problem where a PKCS7
5774 structure which was signed could not be verified because the STACK
5775 order did not reflect the encoded order.
5778 *) Reimplement the OCSP ASN1 module using the new code.
5781 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5782 for its ASN1 operations. The old style function pointers still exist
5783 for now but they will eventually go away.
5786 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
5787 completely replaces the old ASN1 functionality with a table driven
5788 encoder and decoder which interprets an ASN1_ITEM structure describing
5789 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
5790 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
5791 has also been converted to the new form.
5794 *) Change BN_mod_exp_recp so that negative moduli are tolerated
5795 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
5796 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
5797 for negative moduli.
5800 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
5801 of not touching the result's sign bit.
5804 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
5808 *) Changed the LHASH code to use prototypes for callbacks, and created
5809 macros to declare and implement thin (optionally static) functions
5810 that provide type-safety and avoid function pointer casting for the
5811 type-specific callbacks.
5814 *) Added Kerberos Cipher Suites to be used with TLS, as written in
5816 [Veers Staats <staatsvr@asc.hpc.mil>,
5817 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
5819 *) Reformat the FAQ so the different questions and answers can be divided
5820 in sections depending on the subject.
5823 *) Have the zlib compression code load ZLIB.DLL dynamically under
5827 *) New function BN_mod_sqrt for computing square roots modulo a prime
5828 (using the probabilistic Tonelli-Shanks algorithm unless
5829 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
5830 be handled deterministically).
5831 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5833 *) Make BN_mod_inverse faster by explicitly handling small quotients
5834 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
5835 512 bits], about 30% for larger ones [1024 or 2048 bits].)
5838 *) New function BN_kronecker.
5841 *) Fix BN_gcd so that it works on negative inputs; the result is
5842 positive unless both parameters are zero.
5843 Previously something reasonably close to an infinite loop was
5844 possible because numbers could be growing instead of shrinking
5845 in the implementation of Euclid's algorithm.
5848 *) Fix BN_is_word() and BN_is_one() macros to take into account the
5849 sign of the number in question.
5851 Fix BN_is_word(a,w) to work correctly for w == 0.
5853 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
5854 because its test if the absolute value of 'a' equals 'w'.
5855 Note that BN_abs_is_word does *not* handle w == 0 reliably;
5856 it exists mostly for use in the implementations of BN_is_zero(),
5857 BN_is_one(), and BN_is_word().
5860 *) New function BN_swap.
5863 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
5864 the exponentiation functions are more likely to produce reasonable
5865 results on negative inputs.
5868 *) Change BN_mod_mul so that the result is always non-negative.
5869 Previously, it could be negative if one of the factors was negative;
5870 I don't think anyone really wanted that behaviour.
5873 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
5874 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
5875 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
5876 and add new functions:
5885 BN_mod_lshift1_quick
5889 These functions always generate non-negative results.
5891 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
5892 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
5894 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
5895 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
5896 be reduced modulo m.
5897 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5900 The following entry accidentily appeared in the CHANGES file
5901 distributed with OpenSSL 0.9.7. The modifications described in
5902 it do *not* apply to OpenSSL 0.9.7.
5904 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5905 was actually never needed) and in BN_mul(). The removal in BN_mul()
5906 required a small change in bn_mul_part_recursive() and the addition
5907 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5908 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5909 bn_sub_words() and bn_add_words() except they take arrays with
5914 *) In 'openssl passwd', verify passwords read from the terminal
5915 unless the '-salt' option is used (which usually means that
5916 verification would just waste user's time since the resulting
5917 hash is going to be compared with some given password hash)
5918 or the new '-noverify' option is used.
5920 This is an incompatible change, but it does not affect
5921 non-interactive use of 'openssl passwd' (passwords on the command
5922 line, '-stdin' option, '-in ...' option) and thus should not
5926 *) Remove all references to RSAref, since there's no more need for it.
5929 *) Make DSO load along a path given through an environment variable
5930 (SHLIB_PATH) with shl_load().
5933 *) Constify the ENGINE code as a result of BIGNUM constification.
5934 Also constify the RSA code and most things related to it. In a
5935 few places, most notable in the depth of the ASN.1 code, ugly
5936 casts back to non-const were required (to be solved at a later
5940 *) Make it so the openssl application has all engines loaded by default.
5943 *) Constify the BIGNUM routines a little more.
5946 *) Add the following functions:
5948 ENGINE_load_cswift()
5950 ENGINE_load_atalla()
5952 ENGINE_load_builtin_engines()
5954 That way, an application can itself choose if external engines that
5955 are built-in in OpenSSL shall ever be used or not. The benefit is
5956 that applications won't have to be linked with libdl or other dso
5957 libraries unless it's really needed.
5959 Changed 'openssl engine' to load all engines on demand.
5960 Changed the engine header files to avoid the duplication of some
5961 declarations (they differed!).
5964 *) 'openssl engine' can now list capabilities.
5967 *) Better error reporting in 'openssl engine'.
5970 *) Never call load_dh_param(NULL) in s_server.
5973 *) Add engine application. It can currently list engines by name and
5974 identity, and test if they are actually available.
5977 *) Improve RPM specification file by forcing symbolic linking and making
5978 sure the installed documentation is also owned by root.root.
5979 [Damien Miller <djm@mindrot.org>]
5981 *) Give the OpenSSL applications more possibilities to make use of
5982 keys (public as well as private) handled by engines.
5985 *) Add OCSP code that comes from CertCo.
5988 *) Add VMS support for the Rijndael code.
5991 *) Added untested support for Nuron crypto accelerator.
5994 *) Add support for external cryptographic devices. This code was
5995 previously distributed separately as the "engine" branch.
5996 [Geoff Thorpe, Richard Levitte]
5998 *) Rework the filename-translation in the DSO code. It is now possible to
5999 have far greater control over how a "name" is turned into a filename
6000 depending on the operating environment and any oddities about the
6001 different shared library filenames on each system.
6004 *) Support threads on FreeBSD-elf in Configure.
6007 *) Fix for SHA1 assembly problem with MASM: it produces
6008 warnings about corrupt line number information when assembling
6009 with debugging information. This is caused by the overlapping
6011 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
6014 NCONF_get_number() has no error checking at all. As a replacement,
6015 NCONF_get_number_e() is defined (_e for "error checking") and is
6016 promoted strongly. The old NCONF_get_number is kept around for
6017 binary backward compatibility.
6018 Make it possible for methods to load from something other than a BIO,
6019 by providing a function pointer that is given a name instead of a BIO.
6020 For example, this could be used to load configuration data from an
6024 *) Fix for non blocking accept BIOs. Added new I/O special reason
6025 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
6026 with non blocking I/O was not possible because no retry code was
6027 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
6031 *) Added the beginnings of Rijndael support.
6034 *) Fix for bug in DirectoryString mask setting. Add support for
6035 X509_NAME_print_ex() in 'req' and X509_print_ex() function
6036 to allow certificate printing to more controllable, additional
6037 'certopt' option to 'x509' to allow new printing options to be
6041 *) Clean old EAY MD5 hack from e_os.h.
6044 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
6046 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6047 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6048 [Joe Orton, Steve Henson]
6050 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
6052 *) Fix additional bug revealed by the NISCC test suite:
6054 Stop bug triggering large recursion when presented with
6055 certain ASN.1 tags (CVE-2003-0851)
6058 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
6060 *) Fix various bugs revealed by running the NISCC test suite:
6062 Stop out of bounds reads in the ASN1 code when presented with
6063 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6065 If verify callback ignores invalid public key errors don't try to check
6066 certificate signature with the NULL public key.
6070 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6071 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6075 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6076 extra data after the compression methods not only for TLS 1.0
6077 but also for SSL 3.0 (as required by the specification).
6078 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6080 *) Change X509_certificate_type() to mark the key as exported/exportable
6081 when it's 512 *bits* long, not 512 bytes.
6084 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
6086 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6087 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6088 a protocol version number mismatch like a decryption error
6089 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6092 *) Turn on RSA blinding by default in the default implementation
6093 to avoid a timing attack. Applications that don't want it can call
6094 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6095 They would be ill-advised to do so in most cases.
6096 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6098 *) Change RSA blinding code so that it works when the PRNG is not
6099 seeded (in this case, the secret RSA exponent is abused as
6100 an unpredictable seed -- if it is not unpredictable, there
6101 is no point in blinding anyway). Make RSA blinding thread-safe
6102 by remembering the creator's thread ID in rsa->blinding and
6103 having all other threads use local one-time blinding factors
6104 (this requires more computation than sharing rsa->blinding, but
6105 avoids excessive locking; and if an RSA object is not shared
6106 between threads, blinding will still be very fast).
6109 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
6111 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6112 via timing by performing a MAC computation even if incorrrect
6113 block cipher padding has been found. This is a countermeasure
6114 against active attacks where the attacker has to distinguish
6115 between bad padding and a MAC verification error. (CVE-2003-0078)
6117 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6118 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6119 Martin Vuagnoux (EPFL, Ilion)]
6121 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
6123 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
6124 memory from it's contents. This is done with a counter that will
6125 place alternating values in each byte. This can be used to solve
6126 two issues: 1) the removal of calls to memset() by highly optimizing
6127 compilers, and 2) cleansing with other values than 0, since those can
6128 be read through on certain media, for example a swap space on disk.
6131 *) Bugfix: client side session caching did not work with external caching,
6132 because the session->cipher setting was not restored when reloading
6133 from the external cache. This problem was masked, when
6134 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
6135 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
6138 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
6139 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
6140 [Zeev Lieber <zeev-l@yahoo.com>]
6142 *) Undo an undocumented change introduced in 0.9.6e which caused
6143 repeated calls to OpenSSL_add_all_ciphers() and
6144 OpenSSL_add_all_digests() to be ignored, even after calling
6148 *) Change the default configuration reader to deal with last line not
6149 being properly terminated.
6152 *) Change X509_NAME_cmp() so it applies the special rules on handling
6153 DN values that are of type PrintableString, as well as RDNs of type
6154 emailAddress where the value has the type ia5String.
6155 [stefank@valicert.com via Richard Levitte]
6157 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
6158 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
6159 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
6160 the bitwise-OR of the two for use by the majority of applications
6161 wanting this behaviour, and update the docs. The documented
6162 behaviour and actual behaviour were inconsistent and had been
6163 changing anyway, so this is more a bug-fix than a behavioural
6165 [Geoff Thorpe, diagnosed by Nadav Har'El]
6167 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
6168 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
6171 *) Fix initialization code race conditions in
6172 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
6173 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
6174 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
6175 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
6176 ssl2_get_cipher_by_char(),
6177 ssl3_get_cipher_by_char().
6178 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
6180 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
6181 the cached sessions are flushed, as the remove_cb() might use ex_data
6182 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
6183 (see [openssl.org #212]).
6184 [Geoff Thorpe, Lutz Jaenicke]
6186 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
6187 length, instead of the encoding length to d2i_ASN1_OBJECT.
6190 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
6192 *) [In 0.9.6g-engine release:]
6193 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
6194 [Lynn Gazis <lgazis@rainbow.com>]
6196 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
6198 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
6199 and get fix the header length calculation.
6200 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
6201 Alon Kantor <alonk@checkpoint.com> (and others),
6204 *) Use proper error handling instead of 'assertions' in buffer
6205 overflow checks added in 0.9.6e. This prevents DoS (the
6206 assertions could call abort()).
6207 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
6209 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
6211 *) Add various sanity checks to asn1_get_length() to reject
6212 the ASN1 length bytes if they exceed sizeof(long), will appear
6213 negative or the content length exceeds the length of the
6215 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6217 *) Fix cipher selection routines: ciphers without encryption had no flags
6218 for the cipher strength set and where therefore not handled correctly
6219 by the selection routines (PR #130).
6222 *) Fix EVP_dsa_sha macro.
6226 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
6227 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
6228 that was added in OpenSSL 0.9.6d.
6230 As the countermeasure turned out to be incompatible with some
6231 broken SSL implementations, the new option is part of SSL_OP_ALL.
6232 SSL_OP_ALL is usually employed when compatibility with weird SSL
6233 implementations is desired (e.g. '-bugs' option to 's_client' and
6234 's_server'), so the new option is automatically set in many
6238 *) Changes in security patch:
6240 Changes marked "(CHATS)" were sponsored by the Defense Advanced
6241 Research Projects Agency (DARPA) and Air Force Research Laboratory,
6242 Air Force Materiel Command, USAF, under agreement number
6245 *) Add various sanity checks to asn1_get_length() to reject
6246 the ASN1 length bytes if they exceed sizeof(long), will appear
6247 negative or the content length exceeds the length of the
6248 supplied buffer. (CVE-2002-0659)
6249 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6251 *) Assertions for various potential buffer overflows, not known to
6253 [Ben Laurie (CHATS)]
6255 *) Various temporary buffers to hold ASCII versions of integers were
6256 too small for 64 bit platforms. (CVE-2002-0655)
6257 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
6259 *) Remote buffer overflow in SSL3 protocol - an attacker could
6260 supply an oversized session ID to a client. (CVE-2002-0656)
6261 [Ben Laurie (CHATS)]
6263 *) Remote buffer overflow in SSL2 protocol - an attacker could
6264 supply an oversized client master key. (CVE-2002-0656)
6265 [Ben Laurie (CHATS)]
6267 Changes between 0.9.6c and 0.9.6d [9 May 2002]
6269 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
6270 encoded as NULL) with id-dsa-with-sha1.
6271 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
6273 *) Check various X509_...() return values in apps/req.c.
6274 [Nils Larsch <nla@trustcenter.de>]
6276 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
6277 an end-of-file condition would erronously be flagged, when the CRLF
6278 was just at the end of a processed block. The bug was discovered when
6279 processing data through a buffering memory BIO handing the data to a
6280 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
6281 <ptsekov@syntrex.com> and Nedelcho Stanev.
6284 *) Implement a countermeasure against a vulnerability recently found
6285 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
6286 before application data chunks to avoid the use of known IVs
6287 with data potentially chosen by the attacker.
6290 *) Fix length checks in ssl3_get_client_hello().
6293 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
6294 to prevent ssl3_read_internal() from incorrectly assuming that
6295 ssl3_read_bytes() found application data while handshake
6296 processing was enabled when in fact s->s3->in_read_app_data was
6297 merely automatically cleared during the initial handshake.
6298 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
6300 *) Fix object definitions for Private and Enterprise: they were not
6301 recognized in their shortname (=lowercase) representation. Extend
6302 obj_dat.pl to issue an error when using undefined keywords instead
6303 of silently ignoring the problem (Svenning Sorensen
6304 <sss@sss.dnsalias.net>).
6307 *) Fix DH_generate_parameters() so that it works for 'non-standard'
6308 generators, i.e. generators other than 2 and 5. (Previously, the
6309 code did not properly initialise the 'add' and 'rem' values to
6310 BN_generate_prime().)
6312 In the new general case, we do not insist that 'generator' is
6313 actually a primitive root: This requirement is rather pointless;
6314 a generator of the order-q subgroup is just as good, if not
6318 *) Map new X509 verification errors to alerts. Discovered and submitted by
6319 Tom Wu <tom@arcot.com>.
6322 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
6323 returning non-zero before the data has been completely received
6324 when using non-blocking I/O.
6325 [Bodo Moeller; problem pointed out by John Hughes]
6327 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
6328 [Ben Laurie, Lutz Jaenicke]
6330 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
6331 Yoram Zahavi <YoramZ@gilian.com>).
6334 *) Add information about CygWin 1.3 and on, and preserve proper
6335 configuration for the versions before that.
6336 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6338 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
6339 check whether we deal with a copy of a session and do not delete from
6340 the cache in this case. Problem reported by "Izhar Shoshani Levi"
6341 <izhar@checkpoint.com>.
6344 *) Do not store session data into the internal session cache, if it
6345 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
6346 flag is set). Proposed by Aslam <aslam@funk.com>.
6349 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
6353 *) [In 0.9.6d-engine release:]
6354 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6355 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6357 *) Add the configuration target linux-s390x.
6358 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
6360 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
6361 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
6362 variable as an indication that a ClientHello message has been
6363 received. As the flag value will be lost between multiple
6364 invocations of ssl3_accept when using non-blocking I/O, the
6365 function may not be aware that a handshake has actually taken
6366 place, thus preventing a new session from being added to the
6369 To avoid this problem, we now set s->new_session to 2 instead of
6370 using a local variable.
6371 [Lutz Jaenicke, Bodo Moeller]
6373 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
6374 if the SSL_R_LENGTH_MISMATCH error is detected.
6375 [Geoff Thorpe, Bodo Moeller]
6377 *) New 'shared_ldflag' column in Configure platform table.
6380 *) Fix EVP_CIPHER_mode macro.
6381 ["Dan S. Camper" <dan@bti.net>]
6383 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
6384 type, we must throw them away by setting rr->length to 0.
6385 [D P Chang <dpc@qualys.com>]
6387 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
6389 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
6390 <Dominikus.Scherkl@biodata.com>. (The previous implementation
6391 worked incorrectly for those cases where range = 10..._2 and
6392 3*range is two bits longer than range.)
6395 *) Only add signing time to PKCS7 structures if it is not already
6399 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
6400 OBJ_ld_ce should be OBJ_id_ce.
6401 Also some ip-pda OIDs in crypto/objects/objects.txt were
6402 incorrect (cf. RFC 3039).
6403 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
6405 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
6406 returns early because it has nothing to do.
6407 [Andy Schneider <andy.schneider@bjss.co.uk>]
6409 *) [In 0.9.6c-engine release:]
6410 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
6411 [Andy Schneider <andy.schneider@bjss.co.uk>]
6413 *) [In 0.9.6c-engine release:]
6414 Add support for Cryptographic Appliance's keyserver technology.
6415 (Use engine 'keyclient')
6416 [Cryptographic Appliances and Geoff Thorpe]
6418 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
6419 is called via tools/c89.sh because arguments have to be
6420 rearranged (all '-L' options must appear before the first object
6422 [Richard Shapiro <rshapiro@abinitio.com>]
6424 *) [In 0.9.6c-engine release:]
6425 Add support for Broadcom crypto accelerator cards, backported
6427 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
6429 *) [In 0.9.6c-engine release:]
6430 Add support for SureWare crypto accelerator cards from
6431 Baltimore Technologies. (Use engine 'sureware')
6432 [Baltimore Technologies and Mark Cox]
6434 *) [In 0.9.6c-engine release:]
6435 Add support for crypto accelerator cards from Accelerated
6436 Encryption Processing, www.aep.ie. (Use engine 'aep')
6437 [AEP Inc. and Mark Cox]
6439 *) Add a configuration entry for gcc on UnixWare.
6440 [Gary Benson <gbenson@redhat.com>]
6442 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
6443 messages are stored in a single piece (fixed-length part and
6444 variable-length part combined) and fix various bugs found on the way.
6447 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
6448 instead. BIO_gethostbyname() does not know what timeouts are
6449 appropriate, so entries would stay in cache even when they have
6451 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
6453 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
6454 faced with a pathologically small ClientHello fragment that does
6455 not contain client_version: Instead of aborting with an error,
6456 simply choose the highest available protocol version (i.e.,
6457 TLS 1.0 unless it is disabled). In practice, ClientHello
6458 messages are never sent like this, but this change gives us
6459 strictly correct behaviour at least for TLS.
6462 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
6463 never resets s->method to s->ctx->method when called from within
6464 one of the SSL handshake functions.
6465 [Bodo Moeller; problem pointed out by Niko Baric]
6467 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
6468 (sent using the client's version number) if client_version is
6469 smaller than the protocol version in use. Also change
6470 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
6471 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
6472 the client will at least see that alert.
6475 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
6479 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
6480 client receives HelloRequest while in a handshake.
6481 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
6483 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
6484 should end in 'break', not 'goto end' which circuments various
6485 cleanups done in state SSL_ST_OK. But session related stuff
6486 must be disabled for SSL_ST_OK in the case that we just sent a
6489 Also avoid some overhead by not calling ssl_init_wbio_buffer()
6490 before just sending a HelloRequest.
6491 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
6493 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
6494 reveal whether illegal block cipher padding was found or a MAC
6495 verification error occured. (Neither SSLerr() codes nor alerts
6496 are directly visible to potential attackers, but the information
6497 may leak via logfiles.)
6499 Similar changes are not required for the SSL 2.0 implementation
6500 because the number of padding bytes is sent in clear for SSL 2.0,
6501 and the extra bytes are just ignored. However ssl/s2_pkt.c
6502 failed to verify that the purported number of padding bytes is in
6506 *) Add OpenUNIX-8 support including shared libraries
6507 (Boyd Lynn Gerber <gerberb@zenez.com>).
6510 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
6511 'wristwatch attack' using huge encoding parameters (cf.
6512 James H. Manger's CRYPTO 2001 paper). Note that the
6513 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
6514 encoding parameters and hence was not vulnerable.
6517 *) BN_sqr() bug fix.
6518 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
6520 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
6521 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
6522 followed by modular reduction.
6523 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
6525 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
6526 equivalent based on BN_pseudo_rand() instead of BN_rand().
6529 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
6530 This function was broken, as the check for a new client hello message
6531 to handle SGC did not allow these large messages.
6532 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
6535 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
6538 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
6539 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
6542 *) Rework the configuration and shared library support for Tru64 Unix.
6543 The configuration part makes use of modern compiler features and
6544 still retains old compiler behavior for those that run older versions
6545 of the OS. The shared library support part includes a variant that
6546 uses the RPATH feature, and is available through the special
6547 configuration target "alpha-cc-rpath", which will never be selected
6549 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
6551 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
6552 with the same message size as in ssl3_get_certificate_request().
6553 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
6554 messages might inadvertently be reject as too long.
6555 [Petr Lampa <lampa@fee.vutbr.cz>]
6557 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
6560 *) Modified SSL library such that the verify_callback that has been set
6561 specificly for an SSL object with SSL_set_verify() is actually being
6562 used. Before the change, a verify_callback set with this function was
6563 ignored and the verify_callback() set in the SSL_CTX at the time of
6564 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
6565 to allow the necessary settings.
6568 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
6569 explicitly to NULL, as at least on Solaris 8 this seems not always to be
6570 done automatically (in contradiction to the requirements of the C
6571 standard). This made problems when used from OpenSSH.
6574 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
6575 dh->length and always used
6577 BN_rand_range(priv_key, dh->p).
6579 BN_rand_range() is not necessary for Diffie-Hellman, and this
6580 specific range makes Diffie-Hellman unnecessarily inefficient if
6581 dh->length (recommended exponent length) is much smaller than the
6582 length of dh->p. We could use BN_rand_range() if the order of
6583 the subgroup was stored in the DH structure, but we only have
6588 BN_rand(priv_key, l, ...)
6590 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
6596 RSA_eay_public_encrypt
6597 RSA_eay_private_decrypt
6598 RSA_eay_private_encrypt (signing)
6599 RSA_eay_public_decrypt (signature verification)
6601 (default implementations for RSA_public_encrypt,
6602 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
6603 always reject numbers >= n.
6606 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
6607 to synchronize access to 'locking_thread'. This is necessary on
6608 systems where access to 'locking_thread' (an 'unsigned long'
6609 variable) is not atomic.
6612 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
6613 *before* setting the 'crypto_lock_rand' flag. The previous code had
6614 a race condition if 0 is a valid thread ID.
6615 [Travis Vitek <vitek@roguewave.com>]
6617 *) Add support for shared libraries under Irix.
6618 [Albert Chin-A-Young <china@thewrittenword.com>]
6620 *) Add configuration option to build on Linux on both big-endian and
6622 [Ralf Baechle <ralf@uni-koblenz.de>]
6624 *) Add the possibility to create shared libraries on HP-UX.
6627 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
6629 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
6630 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
6631 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
6632 PRNG state recovery was possible based on the output of
6633 one PRNG request appropriately sized to gain knowledge on
6634 'md' followed by enough consecutive 1-byte PRNG requests
6635 to traverse all of 'state'.
6637 1. When updating 'md_local' (the current thread's copy of 'md')
6638 during PRNG output generation, hash all of the previous
6639 'md_local' value, not just the half used for PRNG output.
6641 2. Make the number of bytes from 'state' included into the hash
6642 independent from the number of PRNG bytes requested.
6644 The first measure alone would be sufficient to avoid
6645 Markku-Juhani's attack. (Actually it had never occurred
6646 to me that the half of 'md_local' used for chaining was the
6647 half from which PRNG output bytes were taken -- I had always
6648 assumed that the secret half would be used.) The second
6649 measure makes sure that additional data from 'state' is never
6650 mixed into 'md_local' in small portions; this heuristically
6651 further strengthens the PRNG.
6654 *) Fix crypto/bn/asm/mips3.s.
6657 *) When only the key is given to "enc", the IV is undefined. Print out
6658 an error message in this case.
6661 *) Handle special case when X509_NAME is empty in X509 printing routines.
6664 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
6665 positive and less than q.
6668 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
6669 used: it isn't thread safe and the add_lock_callback should handle
6671 [Paul Rose <Paul.Rose@bridge.com>]
6673 *) Verify that incoming data obeys the block size in
6674 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
6678 [Ulf Möller, Bodo Möller]
6680 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
6681 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
6682 when fixing the server behaviour for backwards-compatible 'client
6683 hello' messages. (Note that the attack is impractical against
6684 SSL 3.0 and TLS 1.0 anyway because length and version checking
6685 means that the probability of guessing a valid ciphertext is
6686 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
6689 Before 0.9.5, the countermeasure (hide the error by generating a
6690 random 'decryption result') did not work properly because
6691 ERR_clear_error() was missing, meaning that SSL_get_error() would
6692 detect the supposedly ignored error.
6694 Both problems are now fixed.
6697 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
6698 (previously it was 1024).
6701 *) Fix for compatibility mode trust settings: ignore trust settings
6702 unless some valid trust or reject settings are present.
6705 *) Fix for blowfish EVP: its a variable length cipher.
6708 *) Fix various bugs related to DSA S/MIME verification. Handle missing
6709 parameters in DSA public key structures and return an error in the
6710 DSA routines if parameters are absent.
6713 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
6714 in the current directory if neither $RANDFILE nor $HOME was set.
6715 RAND_file_name() in 0.9.6a returned NULL in this case. This has
6716 caused some confusion to Windows users who haven't defined $HOME.
6717 Thus RAND_file_name() is changed again: e_os.h can define a
6718 DEFAULT_HOME, which will be used if $HOME is not set.
6719 For Windows, we use "C:"; on other platforms, we still require
6720 environment variables.
6722 *) Move 'if (!initialized) RAND_poll()' into regions protected by
6723 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
6724 having multiple threads call RAND_poll() concurrently.
6727 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
6728 combination of a flag and a thread ID variable.
6729 Otherwise while one thread is in ssleay_rand_bytes (which sets the
6730 flag), *other* threads can enter ssleay_add_bytes without obeying
6731 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
6732 that they do not hold after the first thread unsets add_do_not_lock).
6735 *) Change bctest again: '-x' expressions are not available in all
6739 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
6741 *) Fix a couple of memory leaks in PKCS7_dataDecode()
6742 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
6744 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
6745 the default extension for executables, if any. Also, make the perl
6746 scripts that use symlink() to test if it really exists and use "cp"
6747 if it doesn't. All this made OpenSSL compilable and installable in
6751 *) Fix for asn1_GetSequence() for indefinite length constructed data.
6752 If SEQUENCE is length is indefinite just set c->slen to the total
6753 amount of data available.
6754 [Steve Henson, reported by shige@FreeBSD.org]
6755 [This change does not apply to 0.9.7.]
6757 *) Change bctest to avoid here-documents inside command substitution
6758 (workaround for FreeBSD /bin/sh bug).
6759 For compatibility with Ultrix, avoid shell functions (introduced
6760 in the bctest version that searches along $PATH).
6763 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6764 with des_encrypt() defined on some operating systems, like Solaris
6768 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6769 On the Importance of Eliminating Errors in Cryptographic
6770 Computations, J. Cryptology 14 (2001) 2, 101-119,
6771 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6774 *) MIPS assembler BIGNUM division bug fix.
6777 *) Disabled incorrect Alpha assembler code.
6780 *) Fix PKCS#7 decode routines so they correctly update the length
6781 after reading an EOC for the EXPLICIT tag.
6783 [This change does not apply to 0.9.7.]
6785 *) Fix bug in PKCS#12 key generation routines. This was triggered
6786 if a 3DES key was generated with a 0 initial byte. Include
6787 PKCS12_BROKEN_KEYGEN compilation option to retain the old
6788 (but broken) behaviour.
6791 *) Enhance bctest to search for a working bc along $PATH and print
6793 [Tim Rice <tim@multitalents.net> via Richard Levitte]
6795 *) Fix memory leaks in err.c: free err_data string if necessary;
6796 don't write to the wrong index in ERR_set_error_data.
6799 *) Implement ssl23_peek (analogous to ssl23_read), which previously
6803 *) Replace rdtsc with _emit statements for VC++ version 5.
6804 [Jeremy Cooper <jeremy@baymoo.org>]
6806 *) Make it possible to reuse SSLv2 sessions.
6809 *) In copy_email() check for >= 0 as a return value for
6810 X509_NAME_get_index_by_NID() since 0 is a valid index.
6811 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
6813 *) Avoid coredump with unsupported or invalid public keys by checking if
6814 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
6815 PKCS7_verify() fails with non detached data.
6818 *) Don't use getenv in library functions when run as setuid/setgid.
6819 New function OPENSSL_issetugid().
6822 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
6823 due to incorrect handling of multi-threading:
6825 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
6827 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
6829 3. Count how many times MemCheck_off() has been called so that
6830 nested use can be treated correctly. This also avoids
6831 inband-signalling in the previous code (which relied on the
6832 assumption that thread ID 0 is impossible).
6835 *) Add "-rand" option also to s_client and s_server.
6838 *) Fix CPU detection on Irix 6.x.
6839 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
6840 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6842 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
6845 [This change does not apply to 0.9.7.]
6847 *) Use the cached encoding of an X509_NAME structure rather than
6848 copying it. This is apparently the reason for the libsafe "errors"
6849 but the code is actually correct.
6852 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
6853 Bleichenbacher's DSA attack.
6854 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
6855 to be set and top=0 forces the highest bit to be set; top=-1 is new
6856 and leaves the highest bit random.
6857 [Ulf Moeller, Bodo Moeller]
6859 *) In the NCONF_...-based implementations for CONF_... queries
6860 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
6861 a temporary CONF structure with the data component set to NULL
6862 (which gives segmentation faults in lh_retrieve).
6863 Instead, use NULL for the CONF pointer in CONF_get_string and
6864 CONF_get_number (which may use environment variables) and directly
6865 return NULL from CONF_get_section.
6868 *) Fix potential buffer overrun for EBCDIC.
6871 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
6872 keyUsage if basicConstraints absent for a CA.
6875 *) Make SMIME_write_PKCS7() write mail header values with a format that
6876 is more generally accepted (no spaces before the semicolon), since
6877 some programs can't parse those values properly otherwise. Also make
6878 sure BIO's that break lines after each write do not create invalid
6882 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
6883 macros previously used would not encode an empty SEQUENCE OF
6884 and break the signature.
6886 [This change does not apply to 0.9.7.]
6888 *) Zero the premaster secret after deriving the master secret in
6892 *) Add some EVP_add_digest_alias registrations (as found in
6893 OpenSSL_add_all_digests()) to SSL_library_init()
6894 aka OpenSSL_add_ssl_algorithms(). This provides improved
6895 compatibility with peers using X.509 certificates
6896 with unconventional AlgorithmIdentifier OIDs.
6899 *) Fix for Irix with NO_ASM.
6900 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6902 *) ./config script fixes.
6903 [Ulf Moeller, Richard Levitte]
6905 *) Fix 'openssl passwd -1'.
6908 *) Change PKCS12_key_gen_asc() so it can cope with non null
6909 terminated strings whose length is passed in the passlen
6910 parameter, for example from PEM callbacks. This was done
6911 by adding an extra length parameter to asc2uni().
6912 [Steve Henson, reported by <oddissey@samsung.co.kr>]
6914 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
6915 call failed, free the DSA structure.
6918 *) Fix to uni2asc() to cope with zero length Unicode strings.
6919 These are present in some PKCS#12 files.
6922 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
6923 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
6924 when writing a 32767 byte record.
6925 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
6927 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
6928 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
6930 (RSA objects have a reference count access to which is protected
6931 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
6932 so they are meant to be shared between threads.)
6933 [Bodo Moeller, Geoff Thorpe; original patch submitted by
6934 "Reddie, Steven" <Steven.Reddie@ca.com>]
6936 *) Fix a deadlock in CRYPTO_mem_leaks().
6939 *) Use better test patterns in bntest.
6942 *) rand_win.c fix for Borland C.
6945 *) BN_rshift bugfix for n == 0.
6948 *) Add a 'bctest' script that checks for some known 'bc' bugs
6949 so that 'make test' does not abort just because 'bc' is broken.
6952 *) Store verify_result within SSL_SESSION also for client side to
6953 avoid potential security hole. (Re-used sessions on the client side
6954 always resulted in verify_result==X509_V_OK, not using the original
6955 result of the server certificate verification.)
6958 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
6959 SSL3_RT_APPLICATION_DATA, return 0.
6960 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
6964 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
6965 releases, have been re-implemented by renaming the previous
6966 implementations of ssl2_read and ssl3_read to ssl2_read_internal
6967 and ssl3_read_internal, respectively, and adding 'peek' parameters
6968 to them. The new ssl[23]_{read,peek} functions are calls to
6969 ssl[23]_read_internal with the 'peek' flag set appropriately.
6970 A 'peek' parameter has also been added to ssl3_read_bytes, which
6971 does the actual work for ssl3_read_internal.
6974 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
6975 the method-specific "init()" handler. Also clean up ex_data after
6976 calling the method-specific "finish()" handler. Previously, this was
6977 happening the other way round.
6980 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
6981 The previous value, 12, was not always sufficient for BN_mod_exp().
6984 *) Make sure that shared libraries get the internal name engine with
6985 the full version number and not just 0. This should mark the
6986 shared libraries as not backward compatible. Of course, this should
6987 be changed again when we can guarantee backward binary compatibility.
6990 *) Fix typo in get_cert_by_subject() in by_dir.c
6991 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
6993 *) Rework the system to generate shared libraries:
6995 - Make note of the expected extension for the shared libraries and
6996 if there is a need for symbolic links from for example libcrypto.so.0
6997 to libcrypto.so.0.9.7. There is extended info in Configure for
7000 - Make as few rebuilds of the shared libraries as possible.
7002 - Still avoid linking the OpenSSL programs with the shared libraries.
7004 - When installing, install the shared libraries separately from the
7008 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
7010 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
7011 and not in SSL_clear because the latter is also used by the
7012 accept/connect functions; previously, the settings made by
7013 SSL_set_read_ahead would be lost during the handshake.
7014 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
7016 *) Correct util/mkdef.pl to be selective about disabled algorithms.
7017 Previously, it would create entries for disableed algorithms no
7021 *) Added several new manual pages for SSL_* function.
7024 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
7026 *) In ssl23_get_client_hello, generate an error message when faced
7027 with an initial SSL 3.0/TLS record that is too small to contain the
7028 first two bytes of the ClientHello message, i.e. client_version.
7029 (Note that this is a pathologic case that probably has never happened
7030 in real life.) The previous approach was to use the version number
7031 from the record header as a substitute; but our protocol choice
7032 should not depend on that one because it is not authenticated
7033 by the Finished messages.
7036 *) More robust randomness gathering functions for Windows.
7037 [Jeffrey Altman <jaltman@columbia.edu>]
7039 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
7040 not set then we don't setup the error code for issuer check errors
7041 to avoid possibly overwriting other errors which the callback does
7042 handle. If an application does set the flag then we assume it knows
7043 what it is doing and can handle the new informational codes
7047 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
7048 a general "ANY" type, as such it should be able to decode anything
7049 including tagged types. However it didn't check the class so it would
7050 wrongly interpret tagged types in the same way as their universal
7051 counterpart and unknown types were just rejected. Changed so that the
7052 tagged and unknown types are handled in the same way as a SEQUENCE:
7053 that is the encoding is stored intact. There is also a new type
7054 "V_ASN1_OTHER" which is used when the class is not universal, in this
7055 case we have no idea what the actual type is so we just lump them all
7059 *) On VMS, stdout may very well lead to a file that is written to
7060 in a record-oriented fashion. That means that every write() will
7061 write a separate record, which will be read separately by the
7062 programs trying to read from it. This can be very confusing.
7064 The solution is to put a BIO filter in the way that will buffer
7065 text until a linefeed is reached, and then write everything a
7066 line at a time, so every record written will be an actual line,
7067 not chunks of lines and not (usually doesn't happen, but I've
7068 seen it once) several lines in one record. BIO_f_linebuffer() is
7071 Currently, it's a VMS-only method, because that's where it has
7072 been tested well enough.
7075 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
7076 it can return incorrect results.
7077 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
7078 but it was in 0.9.6-beta[12].)
7081 *) Disable the check for content being present when verifying detached
7082 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
7083 include zero length content when signing messages.
7086 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
7087 BIO_ctrl (for BIO pairs).
7090 *) Add DSO method for VMS.
7093 *) Bug fix: Montgomery multiplication could produce results with the
7097 *) Add RPM specification openssl.spec and modify it to build three
7098 packages. The default package contains applications, application
7099 documentation and run-time libraries. The devel package contains
7100 include files, static libraries and function documentation. The
7101 doc package contains the contents of the doc directory. The original
7102 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
7105 *) Add a large number of documentation files for many SSL routines.
7106 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7108 *) Add a configuration entry for Sony News 4.
7109 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
7111 *) Don't set the two most significant bits to one when generating a
7112 random number < q in the DSA library.
7115 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
7116 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
7117 the underlying transport is blocking) if a handshake took place.
7118 (The default behaviour is needed by applications such as s_client
7119 and s_server that use select() to determine when to use SSL_read;
7120 but for applications that know in advance when to expect data, it
7121 just makes things more complicated.)
7124 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
7128 *) Add a few more EBCDIC conditionals that make `req' and `x509'
7129 work better on such systems.
7130 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7132 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
7133 Update PKCS12_parse() so it copies the friendlyName and the
7134 keyid to the certificates aux info.
7137 *) Fix bug in PKCS7_verify() which caused an infinite loop
7138 if there was more than one signature.
7139 [Sven Uszpelkat <su@celocom.de>]
7141 *) Major change in util/mkdef.pl to include extra information
7142 about each symbol, as well as presentig variables as well
7143 as functions. This change means that there's n more need
7144 to rebuild the .num files when some algorithms are excluded.
7147 *) Allow the verify time to be set by an application,
7148 rather than always using the current time.
7151 *) Phase 2 verify code reorganisation. The certificate
7152 verify code now looks up an issuer certificate by a
7153 number of criteria: subject name, authority key id
7154 and key usage. It also verifies self signed certificates
7155 by the same criteria. The main comparison function is
7156 X509_check_issued() which performs these checks.
7158 Lot of changes were necessary in order to support this
7159 without completely rewriting the lookup code.
7161 Authority and subject key identifier are now cached.
7163 The LHASH 'certs' is X509_STORE has now been replaced
7164 by a STACK_OF(X509_OBJECT). This is mainly because an
7165 LHASH can't store or retrieve multiple objects with
7166 the same hash value.
7168 As a result various functions (which were all internal
7169 use only) have changed to handle the new X509_STORE
7170 structure. This will break anything that messed round
7171 with X509_STORE internally.
7173 The functions X509_STORE_add_cert() now checks for an
7174 exact match, rather than just subject name.
7176 The X509_STORE API doesn't directly support the retrieval
7177 of multiple certificates matching a given criteria, however
7178 this can be worked round by performing a lookup first
7179 (which will fill the cache with candidate certificates)
7180 and then examining the cache for matches. This is probably
7181 the best we can do without throwing out X509_LOOKUP
7182 entirely (maybe later...).
7184 The X509_VERIFY_CTX structure has been enhanced considerably.
7186 All certificate lookup operations now go via a get_issuer()
7187 callback. Although this currently uses an X509_STORE it
7188 can be replaced by custom lookups. This is a simple way
7189 to bypass the X509_STORE hackery necessary to make this
7190 work and makes it possible to use more efficient techniques
7191 in future. A very simple version which uses a simple
7192 STACK for its trusted certificate store is also provided
7193 using X509_STORE_CTX_trusted_stack().
7195 The verify_cb() and verify() callbacks now have equivalents
7196 in the X509_STORE_CTX structure.
7198 X509_STORE_CTX also has a 'flags' field which can be used
7199 to customise the verify behaviour.
7202 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
7203 excludes S/MIME capabilities.
7206 *) When a certificate request is read in keep a copy of the
7207 original encoding of the signed data and use it when outputing
7208 again. Signatures then use the original encoding rather than
7209 a decoded, encoded version which may cause problems if the
7210 request is improperly encoded.
7213 *) For consistency with other BIO_puts implementations, call
7214 buffer_write(b, ...) directly in buffer_puts instead of calling
7217 In BIO_puts, increment b->num_write as in BIO_write.
7218 [Peter.Sylvester@EdelWeb.fr]
7220 *) Fix BN_mul_word for the case where the word is 0. (We have to use
7221 BN_zero, we may not return a BIGNUM with an array consisting of
7225 *) Avoid calling abort() from within the library when problems are
7226 detected, except if preprocessor symbols have been defined
7227 (such as REF_CHECK, BN_DEBUG etc.).
7230 *) New openssl application 'rsautl'. This utility can be
7231 used for low level RSA operations. DER public key
7232 BIO/fp routines also added.
7235 *) New Configure entry and patches for compiling on QNX 4.
7236 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
7238 *) A demo state-machine implementation was sponsored by
7239 Nuron (http://www.nuron.com/) and is now available in
7240 demos/state_machine.
7243 *) New options added to the 'dgst' utility for signature
7244 generation and verification.
7247 *) Unrecognized PKCS#7 content types are now handled via a
7248 catch all ASN1_TYPE structure. This allows unsupported
7249 types to be stored as a "blob" and an application can
7250 encode and decode it manually.
7253 *) Fix various signed/unsigned issues to make a_strex.c
7255 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
7257 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
7258 length if passed a buffer. ASN1_INTEGER_to_BN failed
7259 if passed a NULL BN and its argument was negative.
7260 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
7262 *) Modification to PKCS#7 encoding routines to output definite
7263 length encoding. Since currently the whole structures are in
7264 memory there's not real point in using indefinite length
7265 constructed encoding. However if OpenSSL is compiled with
7266 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
7269 *) Added BIO_vprintf() and BIO_vsnprintf().
7272 *) Added more prefixes to parse for in the the strings written
7273 through a logging bio, to cover all the levels that are available
7274 through syslog. The prefixes are now:
7276 PANIC, EMERG, EMR => LOG_EMERG
7277 ALERT, ALR => LOG_ALERT
7278 CRIT, CRI => LOG_CRIT
7279 ERROR, ERR => LOG_ERR
7280 WARNING, WARN, WAR => LOG_WARNING
7281 NOTICE, NOTE, NOT => LOG_NOTICE
7282 INFO, INF => LOG_INFO
7283 DEBUG, DBG => LOG_DEBUG
7285 and as before, if none of those prefixes are present at the
7286 beginning of the string, LOG_ERR is chosen.
7288 On Win32, the LOG_* levels are mapped according to this:
7290 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
7291 LOG_WARNING => EVENTLOG_WARNING_TYPE
7292 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
7296 *) Made it possible to reconfigure with just the configuration
7297 argument "reconf" or "reconfigure". The command line arguments
7298 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
7299 and are retrieved from there when reconfiguring.
7303 [Assar Westerlund <assar@sics.se>, Richard Levitte]
7305 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
7308 *) The obj_dat.pl script was messing up the sorting of object
7309 names. The reason was that it compared the quoted version
7310 of strings as a result "OCSP" > "OCSP Signing" because
7311 " > SPACE. Changed script to store unquoted versions of
7312 names and add quotes on output. It was also omitting some
7313 names from the lookup table if they were given a default
7314 value (that is if SN is missing it is given the same
7315 value as LN and vice versa), these are now added on the
7316 grounds that if an object has a name we should be able to
7317 look it up. Finally added warning output when duplicate
7318 short or long names are found.
7321 *) Changes needed for Tandem NSK.
7322 [Scott Uroff <scott@xypro.com>]
7324 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
7325 RSA_padding_check_SSLv23(), special padding was never detected
7326 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
7327 version rollback attacks was not effective.
7329 In s23_clnt.c, don't use special rollback-attack detection padding
7330 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
7331 client; similarly, in s23_srvr.c, don't do the rollback check if
7332 SSL 2.0 is the only protocol enabled in the server.
7335 *) Make it possible to get hexdumps of unprintable data with 'openssl
7336 asn1parse'. By implication, the functions ASN1_parse_dump() and
7337 BIO_dump_indent() are added.
7340 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
7341 these print out strings and name structures based on various
7342 flags including RFC2253 support and proper handling of
7343 multibyte characters. Added options to the 'x509' utility
7344 to allow the various flags to be set.
7347 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
7348 Also change the functions X509_cmp_current_time() and
7349 X509_gmtime_adj() work with an ASN1_TIME structure,
7350 this will enable certificates using GeneralizedTime in validity
7351 dates to be checked.
7354 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
7355 negative public key encodings) on by default,
7356 NO_NEG_PUBKEY_BUG can be set to disable it.
7359 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
7360 content octets. An i2c_ASN1_OBJECT is unnecessary because
7361 the encoding can be trivially obtained from the structure.
7364 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
7365 not read locks (CRYPTO_r_[un]lock).
7368 *) A first attempt at creating official support for shared
7369 libraries through configuration. I've kept it so the
7370 default is static libraries only, and the OpenSSL programs
7371 are always statically linked for now, but there are
7372 preparations for dynamic linking in place.
7373 This has been tested on Linux and Tru64.
7376 *) Randomness polling function for Win9x, as described in:
7377 Peter Gutmann, Software Generation of Practically Strong
7381 *) Fix so PRNG is seeded in req if using an already existing
7385 *) New options to smime application. -inform and -outform
7386 allow alternative formats for the S/MIME message including
7387 PEM and DER. The -content option allows the content to be
7388 specified separately. This should allow things like Netscape
7389 form signing output easier to verify.
7392 *) Fix the ASN1 encoding of tags using the 'long form'.
7395 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
7396 STRING types. These convert content octets to and from the
7397 underlying type. The actual tag and length octets are
7398 already assumed to have been read in and checked. These
7399 are needed because all other string types have virtually
7400 identical handling apart from the tag. By having versions
7401 of the ASN1 functions that just operate on content octets
7402 IMPLICIT tagging can be handled properly. It also allows
7403 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
7404 and ASN1_INTEGER are identical apart from the tag.
7407 *) Change the handling of OID objects as follows:
7409 - New object identifiers are inserted in objects.txt, following
7410 the syntax given in objects.README.
7411 - objects.pl is used to process obj_mac.num and create a new
7413 - obj_dat.pl is used to create a new obj_dat.h, using the data in
7416 This is currently kind of a hack, and the perl code in objects.pl
7417 isn't very elegant, but it works as I intended. The simplest way
7418 to check that it worked correctly is to look in obj_dat.h and
7419 check the array nid_objs and make sure the objects haven't moved
7420 around (this is important!). Additions are OK, as well as
7421 consistent name changes.
7424 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
7427 *) Addition of the command line parameter '-rand file' to 'openssl req'.
7428 The given file adds to whatever has already been seeded into the
7429 random pool through the RANDFILE configuration file option or
7430 environment variable, or the default random state file.
7433 *) mkstack.pl now sorts each macro group into lexical order.
7434 Previously the output order depended on the order the files
7435 appeared in the directory, resulting in needless rewriting
7439 *) Patches to make OpenSSL compile under Win32 again. Mostly
7440 work arounds for the VC++ problem that it treats func() as
7441 func(void). Also stripped out the parts of mkdef.pl that
7442 added extra typesafe functions: these no longer exist.
7445 *) Reorganisation of the stack code. The macros are now all
7446 collected in safestack.h . Each macro is defined in terms of
7447 a "stack macro" of the form SKM_<name>(type, a, b). The
7448 DEBUG_SAFESTACK is now handled in terms of function casts,
7449 this has the advantage of retaining type safety without the
7450 use of additional functions. If DEBUG_SAFESTACK is not defined
7451 then the non typesafe macros are used instead. Also modified the
7452 mkstack.pl script to handle the new form. Needs testing to see
7453 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
7454 the default if no major problems. Similar behaviour for ASN1_SET_OF
7455 and PKCS12_STACK_OF.
7458 *) When some versions of IIS use the 'NET' form of private key the
7459 key derivation algorithm is different. Normally MD5(password) is
7460 used as a 128 bit RC4 key. In the modified case
7461 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
7462 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
7463 as the old Netscape_RSA functions except they have an additional
7464 'sgckey' parameter which uses the modified algorithm. Also added
7465 an -sgckey command line option to the rsa utility. Thanks to
7466 Adrian Peck <bertie@ncipher.com> for posting details of the modified
7467 algorithm to openssl-dev.
7470 *) The evp_local.h macros were using 'c.##kname' which resulted in
7471 invalid expansion on some systems (SCO 5.0.5 for example).
7472 Corrected to 'c.kname'.
7473 [Phillip Porch <root@theporch.com>]
7475 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
7476 a STACK of email addresses from a certificate or request, these look
7477 in the subject name and the subject alternative name extensions and
7478 omit any duplicate addresses.
7481 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
7482 This makes DSA verification about 2 % faster.
7485 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
7486 (meaning that now 2^5 values will be precomputed, which is only 4 KB
7487 plus overhead for 1024 bit moduli).
7488 This makes exponentiations about 0.5 % faster for 1024 bit
7489 exponents (as measured by "openssl speed rsa2048").
7492 *) Rename memory handling macros to avoid conflicts with other
7494 Malloc => OPENSSL_malloc
7495 Malloc_locked => OPENSSL_malloc_locked
7496 Realloc => OPENSSL_realloc
7497 Free => OPENSSL_free
7500 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
7501 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
7504 *) CygWin32 support.
7505 [John Jarvie <jjarvie@newsguy.com>]
7507 *) The type-safe stack code has been rejigged. It is now only compiled
7508 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
7509 by default all type-specific stack functions are "#define"d back to
7510 standard stack functions. This results in more streamlined output
7511 but retains the type-safety checking possibilities of the original
7515 *) The STACK code has been cleaned up, and certain type declarations
7516 that didn't make a lot of sense have been brought in line. This has
7517 also involved a cleanup of sorts in safestack.h to more correctly
7518 map type-safe stack functions onto their plain stack counterparts.
7519 This work has also resulted in a variety of "const"ifications of
7520 lots of the code, especially "_cmp" operations which should normally
7521 be prototyped with "const" parameters anyway.
7524 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
7525 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
7526 (The PRNG state consists of two parts, the large pool 'state' and 'md',
7527 where all of 'md' is used each time the PRNG is used, but 'state'
7528 is used only indexed by a cyclic counter. As entropy may not be
7529 well distributed from the beginning, 'md' is important as a
7530 chaining variable. However, the output function chains only half
7531 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
7532 all of 'md', and seeding with STATE_SIZE dummy bytes will result
7533 in all of 'state' being rewritten, with the new values depending
7534 on virtually all of 'md'. This overcomes the 80 bit limitation.)
7537 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
7538 the handshake is continued after ssl_verify_cert_chain();
7539 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
7540 can lead to 'unexplainable' connection aborts later.
7541 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
7543 *) Major EVP API cipher revision.
7544 Add hooks for extra EVP features. This allows various cipher
7545 parameters to be set in the EVP interface. Support added for variable
7546 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
7547 setting of RC2 and RC5 parameters.
7549 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
7552 Remove lots of duplicated code from the EVP library. For example *every*
7553 cipher init() function handles the 'iv' in the same way according to the
7554 cipher mode. They also all do nothing if the 'key' parameter is NULL and
7555 for CFB and OFB modes they zero ctx->num.
7557 New functionality allows removal of S/MIME code RC2 hack.
7559 Most of the routines have the same form and so can be declared in terms
7562 By shifting this to the top level EVP_CipherInit() it can be removed from
7563 all individual ciphers. If the cipher wants to handle IVs or keys
7564 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
7567 Change lots of functions like EVP_EncryptUpdate() to now return a
7568 value: although software versions of the algorithms cannot fail
7569 any installed hardware versions can.
7572 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
7573 this option is set, tolerate broken clients that send the negotiated
7574 protocol version number instead of the requested protocol version
7578 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
7579 i.e. non-zero for export ciphersuites, zero otherwise.
7580 Previous versions had this flag inverted, inconsistent with
7581 rsa_tmp_cb (..._TMP_RSA_CB).
7582 [Bodo Moeller; problem reported by Amit Chopra]
7584 *) Add missing DSA library text string. Work around for some IIS
7585 key files with invalid SEQUENCE encoding.
7588 *) Add a document (doc/standards.txt) that list all kinds of standards
7589 and so on that are implemented in OpenSSL.
7592 *) Enhance c_rehash script. Old version would mishandle certificates
7593 with the same subject name hash and wouldn't handle CRLs at all.
7594 Added -fingerprint option to crl utility, to support new c_rehash
7598 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
7601 *) Fix for SSL server purpose checking. Server checking was
7602 rejecting certificates which had extended key usage present
7603 but no ssl client purpose.
7604 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
7606 *) Make PKCS#12 code work with no password. The PKCS#12 spec
7607 is a little unclear about how a blank password is handled.
7608 Since the password in encoded as a BMPString with terminating
7609 double NULL a zero length password would end up as just the
7610 double NULL. However no password at all is different and is
7611 handled differently in the PKCS#12 key generation code. NS
7612 treats a blank password as zero length. MSIE treats it as no
7613 password on export: but it will try both on import. We now do
7614 the same: PKCS12_parse() tries zero length and no password if
7615 the password is set to "" or NULL (NULL is now a valid password:
7616 it wasn't before) as does the pkcs12 application.
7619 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
7620 perror when PEM_read_bio_X509_REQ fails, the error message must
7621 be obtained from the error queue.
7624 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
7625 it in ERR_remove_state if appropriate, and change ERR_get_state
7626 accordingly to avoid race conditions (this is necessary because
7627 thread_hash is no longer constant once set).
7630 *) Bugfix for linux-elf makefile.one.
7633 *) RSA_get_default_method() will now cause a default
7634 RSA_METHOD to be chosen if one doesn't exist already.
7635 Previously this was only set during a call to RSA_new()
7636 or RSA_new_method(NULL) meaning it was possible for
7637 RSA_get_default_method() to return NULL.
7640 *) Added native name translation to the existing DSO code
7641 that will convert (if the flag to do so is set) filenames
7642 that are sufficiently small and have no path information
7643 into a canonical native form. Eg. "blah" converted to
7644 "libblah.so" or "blah.dll" etc.
7647 *) New function ERR_error_string_n(e, buf, len) which is like
7648 ERR_error_string(e, buf), but writes at most 'len' bytes
7649 including the 0 terminator. For ERR_error_string_n, 'buf'
7651 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
7653 *) CONF library reworked to become more general. A new CONF
7654 configuration file reader "class" is implemented as well as a
7655 new functions (NCONF_*, for "New CONF") to handle it. The now
7656 old CONF_* functions are still there, but are reimplemented to
7657 work in terms of the new functions. Also, a set of functions
7658 to handle the internal storage of the configuration data is
7659 provided to make it easier to write new configuration file
7660 reader "classes" (I can definitely see something reading a
7661 configuration file in XML format, for example), called _CONF_*,
7662 or "the configuration storage API"...
7664 The new configuration file reading functions are:
7666 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
7667 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
7669 NCONF_default, NCONF_WIN32
7671 NCONF_dump_fp, NCONF_dump_bio
7673 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
7674 NCONF_new creates a new CONF object. This works in the same way
7675 as other interfaces in OpenSSL, like the BIO interface.
7676 NCONF_dump_* dump the internal storage of the configuration file,
7677 which is useful for debugging. All other functions take the same
7678 arguments as the old CONF_* functions wth the exception of the
7679 first that must be a `CONF *' instead of a `LHASH *'.
7681 To make it easer to use the new classes with the old CONF_* functions,
7682 the function CONF_set_default_method is provided.
7685 *) Add '-tls1' option to 'openssl ciphers', which was already
7686 mentioned in the documentation but had not been implemented.
7687 (This option is not yet really useful because even the additional
7688 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
7691 *) Initial DSO code added into libcrypto for letting OpenSSL (and
7692 OpenSSL-based applications) load shared libraries and bind to
7693 them in a portable way.
7694 [Geoff Thorpe, with contributions from Richard Levitte]
7696 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
7698 *) Make sure _lrotl and _lrotr are only used with MSVC.
7700 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
7701 (the default implementation of RAND_status).
7703 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
7704 to '-clrext' (= clear extensions), as intended and documented.
7705 [Bodo Moeller; inconsistency pointed out by Michael Attili
7708 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
7709 was larger than the MD block size.
7710 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
7712 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
7713 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
7714 using the passed key: if the passed key was a private key the result
7715 of X509_print(), for example, would be to print out all the private key
7719 *) des_quad_cksum() byte order bug fix.
7720 [Ulf Möller, using the problem description in krb4-0.9.7, where
7721 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
7723 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
7725 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
7727 *) For easily testing in shell scripts whether some command
7728 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
7729 returns with exit code 0 iff no command of the given name is available.
7730 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
7731 the output goes to stdout and nothing is printed to stderr.
7732 Additional arguments are always ignored.
7734 Since for each cipher there is a command of the same name,
7735 the 'no-cipher' compilation switches can be tested this way.
7737 ('openssl no-XXX' is not able to detect pseudo-commands such
7738 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
7741 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
7744 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
7745 is set; it will be thrown away anyway because each handshake creates
7747 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
7748 to parameters -- in previous versions (since OpenSSL 0.9.3) the
7749 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
7750 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7753 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7754 'Q' and 'R' lose their special meanings (quit/renegotiate).
7755 This is part of what -quiet does; unlike -quiet, -ign_eof
7756 does not suppress any output.
7759 *) Add compatibility options to the purpose and trust code. The
7760 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7761 accepts a certificate or CA, this was the previous behaviour,
7762 with all the associated security issues.
7764 X509_TRUST_COMPAT is the old trust behaviour: only and
7765 automatically trust self signed roots in certificate store. A
7766 new trust setting X509_TRUST_DEFAULT is used to specify that
7767 a purpose has no associated trust setting and it should instead
7768 use the value in the default purpose.
7771 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7772 and fix a memory leak.
7775 *) In util/mkerr.pl (which implements 'make errors'), preserve
7776 reason strings from the previous version of the .c file, as
7777 the default to have only downcase letters (and digits) in
7778 automatically generated reasons codes is not always appropriate.
7781 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7782 using strerror. Previously, ERR_reason_error_string() returned
7783 library names as reason strings for SYSerr; but SYSerr is a special
7784 case where small numbers are errno values, not library numbers.
7787 *) Add '-dsaparam' option to 'openssl dhparam' application. This
7788 converts DSA parameters into DH parameters. (When creating parameters,
7789 DSA_generate_parameters is used.)
7792 *) Include 'length' (recommended exponent length) in C code generated
7793 by 'openssl dhparam -C'.
7796 *) The second argument to set_label in perlasm was already being used
7797 so couldn't be used as a "file scope" flag. Moved to third argument
7801 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
7802 instead of RAND_bytes for encryption IVs and salts.
7805 *) Include RAND_status() into RAND_METHOD instead of implementing
7806 it only for md_rand.c Otherwise replacing the PRNG by calling
7807 RAND_set_rand_method would be impossible.
7810 *) Don't let DSA_generate_key() enter an infinite loop if the random
7811 number generation fails.
7814 *) New 'rand' application for creating pseudo-random output.
7817 *) Added configuration support for Linux/IA64
7818 [Rolf Haberrecker <rolf@suse.de>]
7820 *) Assembler module support for Mingw32.
7823 *) Shared library support for HPUX (in shlib/).
7824 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
7826 *) Shared library support for Solaris gcc.
7827 [Lutz Behnke <behnke@trustcenter.de>]
7829 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
7831 *) PKCS7_encrypt() was adding text MIME headers twice because they
7832 were added manually and by SMIME_crlf_copy().
7835 *) In bntest.c don't call BN_rand with zero bits argument.
7836 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
7838 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
7839 case was implemented. This caused BN_div_recp() to fail occasionally.
7842 *) Add an optional second argument to the set_label() in the perl
7843 assembly language builder. If this argument exists and is set
7844 to 1 it signals that the assembler should use a symbol whose
7845 scope is the entire file, not just the current function. This
7846 is needed with MASM which uses the format label:: for this scope.
7847 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
7849 *) Change the ASN1 types so they are typedefs by default. Before
7850 almost all types were #define'd to ASN1_STRING which was causing
7851 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
7855 *) Change names of new functions to the new get1/get0 naming
7856 convention: After 'get1', the caller owns a reference count
7857 and has to call ..._free; 'get0' returns a pointer to some
7858 data structure without incrementing reference counters.
7859 (Some of the existing 'get' functions increment a reference
7860 counter, some don't.)
7861 Similarly, 'set1' and 'add1' functions increase reference
7862 counters or duplicate objects.
7865 *) Allow for the possibility of temp RSA key generation failure:
7866 the code used to assume it always worked and crashed on failure.
7869 *) Fix potential buffer overrun problem in BIO_printf().
7870 [Ulf Möller, using public domain code by Patrick Powell; problem
7871 pointed out by David Sacerdote <das33@cornell.edu>]
7873 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
7874 RAND_egd() and RAND_status(). In the command line application,
7875 the EGD socket can be specified like a seed file using RANDFILE
7879 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
7880 Some CAs (e.g. Verisign) distribute certificates in this form.
7883 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
7884 list to exclude them. This means that no special compilation option
7885 is needed to use anonymous DH: it just needs to be included in the
7889 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
7890 EVP_MD_type. The old functionality is available in a new macro called
7891 EVP_MD_md(). Change code that uses it and update docs.
7894 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
7895 where the 'void *' argument is replaced by a function pointer argument.
7896 Previously 'void *' was abused to point to functions, which works on
7897 many platforms, but is not correct. As these functions are usually
7898 called by macros defined in OpenSSL header files, most source code
7899 should work without changes.
7902 *) <openssl/opensslconf.h> (which is created by Configure) now contains
7903 sections with information on -D... compiler switches used for
7904 compiling the library so that applications can see them. To enable
7905 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
7906 must be defined. E.g.,
7907 #define OPENSSL_ALGORITHM_DEFINES
7908 #include <openssl/opensslconf.h>
7909 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
7910 [Richard Levitte, Ulf and Bodo Möller]
7912 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
7916 *) Change the 'other' type in certificate aux info to a STACK_OF
7917 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
7918 the required ASN1 format: arbitrary types determined by an OID.
7921 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
7922 argument to 'req'. This is not because the function is newer or
7923 better than others it just uses the work 'NEW' in the certificate
7924 request header lines. Some software needs this.
7927 *) Reorganise password command line arguments: now passwords can be
7928 obtained from various sources. Delete the PEM_cb function and make
7929 it the default behaviour: i.e. if the callback is NULL and the
7930 usrdata argument is not NULL interpret it as a null terminated pass
7931 phrase. If usrdata and the callback are NULL then the pass phrase
7932 is prompted for as usual.
7935 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
7936 the support is automatically enabled. The resulting binaries will
7937 autodetect the card and use it if present.
7938 [Ben Laurie and Compaq Inc.]
7940 *) Work around for Netscape hang bug. This sends certificate request
7941 and server done in one record. Since this is perfectly legal in the
7942 SSL/TLS protocol it isn't a "bug" option and is on by default. See
7943 the bugs/SSLv3 entry for more info.
7946 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
7949 *) Add -rand argument to smime and pkcs12 applications and read/write
7953 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
7956 *) Add command line password options to the remaining applications.
7959 *) Bug fix for BN_div_recp() for numerators with an even number of
7963 *) More tests in bntest.c, and changed test_bn output.
7966 *) ./config recognizes MacOS X now.
7969 *) Bug fix for BN_div() when the first words of num and divsor are
7970 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
7973 *) Add support for various broken PKCS#8 formats, and command line
7974 options to produce them.
7977 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
7978 get temporary BIGNUMs from a BN_CTX.
7981 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
7985 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
7986 include a #define from the old name to the new. The original intent
7987 was that statically linked binaries could for example just call
7988 SSLeay_add_all_ciphers() to just add ciphers to the table and not
7989 link with digests. This never worked becayse SSLeay_add_all_digests()
7990 and SSLeay_add_all_ciphers() were in the same source file so calling
7991 one would link with the other. They are now in separate source files.
7994 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
7997 *) Use a less unusual form of the Miller-Rabin primality test (it used
7998 a binary algorithm for exponentiation integrated into the Miller-Rabin
7999 loop, our standard modexp algorithms are faster).
8002 *) Support for the EBCDIC character set completed.
8003 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
8005 *) Source code cleanups: use const where appropriate, eliminate casts,
8006 use void * instead of char * in lhash.
8009 *) Bugfix: ssl3_send_server_key_exchange was not restartable
8010 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
8011 this the server could overwrite ephemeral keys that the client
8015 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
8016 using 50 iterations of the Rabin-Miller test.
8018 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
8019 iterations of the Rabin-Miller test as required by the appendix
8020 to FIPS PUB 186[-1]) instead of DSA_is_prime.
8021 As BN_is_prime_fasttest includes trial division, DSA parameter
8022 generation becomes much faster.
8024 This implies a change for the callback functions in DSA_is_prime
8025 and DSA_generate_parameters: The callback function is called once
8026 for each positive witness in the Rabin-Miller test, not just
8027 occasionally in the inner loop; and the parameters to the
8028 callback function now provide an iteration count for the outer
8029 loop rather than for the current invocation of the inner loop.
8030 DSA_generate_parameters additionally can call the callback
8031 function with an 'iteration count' of -1, meaning that a
8032 candidate has passed the trial division test (when q is generated
8033 from an application-provided seed, trial division is skipped).
8036 *) New function BN_is_prime_fasttest that optionally does trial
8037 division before starting the Rabin-Miller test and has
8038 an additional BN_CTX * argument (whereas BN_is_prime always
8039 has to allocate at least one BN_CTX).
8040 'callback(1, -1, cb_arg)' is called when a number has passed the
8041 trial division stage.
8044 *) Fix for bug in CRL encoding. The validity dates weren't being handled
8048 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
8051 *) New function BN_pseudo_rand().
8054 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
8055 bignum version of BN_from_montgomery() with the working code from
8056 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
8060 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
8061 made it impossible to use the same SSL_SESSION data structure in
8062 SSL2 clients in multiple threads.
8065 *) The return value of RAND_load_file() no longer counts bytes obtained
8066 by stat(). RAND_load_file(..., -1) is new and uses the complete file
8067 to seed the PRNG (previously an explicit byte count was required).
8068 [Ulf Möller, Bodo Möller]
8070 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
8071 used (char *) instead of (void *) and had casts all over the place.
8074 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
8077 *) Retain source code compatibility for BN_prime_checks macro:
8078 BN_is_prime(..., BN_prime_checks, ...) now uses
8079 BN_prime_checks_for_size to determine the appropriate number of
8080 Rabin-Miller iterations.
8083 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
8084 DH_CHECK_P_NOT_SAFE_PRIME.
8085 (Check if this is true? OpenPGP calls them "strong".)
8088 *) Merge the functionality of "dh" and "gendh" programs into a new program
8089 "dhparam". The old programs are retained for now but will handle DH keys
8090 (instead of parameters) in future.
8093 *) Make the ciphers, s_server and s_client programs check the return values
8094 when a new cipher list is set.
8097 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
8098 ciphers. Before when the 56bit ciphers were enabled the sorting was
8101 The syntax for the cipher sorting has been extended to support sorting by
8102 cipher-strength (using the strength_bits hard coded in the tables).
8103 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
8105 Fix a bug in the cipher-command parser: when supplying a cipher command
8106 string with an "undefined" symbol (neither command nor alphanumeric
8107 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
8108 an error is flagged.
8110 Due to the strength-sorting extension, the code of the
8111 ssl_create_cipher_list() function was completely rearranged. I hope that
8112 the readability was also increased :-)
8113 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8115 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
8116 for the first serial number and places 2 in the serial number file. This
8117 avoids problems when the root CA is created with serial number zero and
8118 the first user certificate has the same issuer name and serial number
8122 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
8123 the new code. Add documentation for this stuff.
8126 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
8127 X509_*() to X509at_*() on the grounds that they don't handle X509
8128 structures and behave in an analagous way to the X509v3 functions:
8129 they shouldn't be called directly but wrapper functions should be used
8132 So we also now have some wrapper functions that call the X509at functions
8133 when passed certificate requests. (TO DO: similar things can be done with
8134 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
8135 things. Some of these need some d2i or i2d and print functionality
8136 because they handle more complex structures.)
8139 *) Add missing #ifndefs that caused missing symbols when building libssl
8140 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
8141 NO_RSA in ssl/s2*.c.
8142 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
8144 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
8145 has a return value which indicates the quality of the random data
8146 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
8147 error queue. New function RAND_pseudo_bytes() generates output that is
8148 guaranteed to be unique but not unpredictable. RAND_add is like
8149 RAND_seed, but takes an extra argument for an entropy estimate
8150 (RAND_seed always assumes full entropy).
8153 *) Do more iterations of Rabin-Miller probable prime test (specifically,
8154 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
8155 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
8156 in crypto/bn/bn_prime.c for the complete table). This guarantees a
8157 false-positive rate of at most 2^-80 for random input.
8160 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
8163 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
8164 in the 0.9.5 release), this returns the chain
8165 from an X509_CTX structure with a dup of the stack and all
8166 the X509 reference counts upped: so the stack will exist
8167 after X509_CTX_cleanup() has been called. Modify pkcs12.c
8170 Also make SSL_SESSION_print() print out the verify return
8174 *) Add manpage for the pkcs12 command. Also change the default
8175 behaviour so MAC iteration counts are used unless the new
8176 -nomaciter option is used. This improves file security and
8177 only older versions of MSIE (4.0 for example) need it.
8180 *) Honor the no-xxx Configure options when creating .DEF files.
8183 *) Add PKCS#10 attributes to field table: challengePassword,
8184 unstructuredName and unstructuredAddress. These are taken from
8185 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
8186 international characters are used.
8188 More changes to X509_ATTRIBUTE code: allow the setting of types
8189 based on strings. Remove the 'loc' parameter when adding
8190 attributes because these will be a SET OF encoding which is sorted
8194 *) Initial changes to the 'req' utility to allow request generation
8195 automation. This will allow an application to just generate a template
8196 file containing all the field values and have req construct the
8199 Initial support for X509_ATTRIBUTE handling. Stacks of these are
8200 used all over the place including certificate requests and PKCS#7
8201 structures. They are currently handled manually where necessary with
8202 some primitive wrappers for PKCS#7. The new functions behave in a
8203 manner analogous to the X509 extension functions: they allow
8204 attributes to be looked up by NID and added.
8206 Later something similar to the X509V3 code would be desirable to
8207 automatically handle the encoding, decoding and printing of the
8208 more complex types. The string types like challengePassword can
8209 be handled by the string table functions.
8211 Also modified the multi byte string table handling. Now there is
8212 a 'global mask' which masks out certain types. The table itself
8213 can use the flag STABLE_NO_MASK to ignore the mask setting: this
8214 is useful when for example there is only one permissible type
8215 (as in countryName) and using the mask might result in no valid
8219 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
8220 SSL_get_peer_finished to allow applications to obtain the latest
8221 Finished messages sent to the peer or expected from the peer,
8222 respectively. (SSL_get_peer_finished is usually the Finished message
8223 actually received from the peer, otherwise the protocol will be aborted.)
8225 As the Finished message are message digests of the complete handshake
8226 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
8227 be used for external authentication procedures when the authentication
8228 provided by SSL/TLS is not desired or is not enough.
8231 *) Enhanced support for Alpha Linux is added. Now ./config checks if
8232 the host supports BWX extension and if Compaq C is present on the
8233 $PATH. Just exploiting of the BWX extension results in 20-30%
8234 performance kick for some algorithms, e.g. DES and RC4 to mention
8235 a couple. Compaq C in turn generates ~20% faster code for MD5 and
8239 *) Add support for MS "fast SGC". This is arguably a violation of the
8240 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
8241 weak crypto and after checking the certificate is SGC a second one
8242 with strong crypto. MS SGC stops the first handshake after receiving
8243 the server certificate message and sends a second client hello. Since
8244 a server will typically do all the time consuming operations before
8245 expecting any further messages from the client (server key exchange
8246 is the most expensive) there is little difference between the two.
8248 To get OpenSSL to support MS SGC we have to permit a second client
8249 hello message after we have sent server done. In addition we have to
8250 reset the MAC if we do get this second client hello.
8253 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
8254 if a DER encoded private key is RSA or DSA traditional format. Changed
8255 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
8256 format DER encoded private key. Newer code should use PKCS#8 format which
8257 has the key type encoded in the ASN1 structure. Added DER private key
8258 support to pkcs8 application.
8261 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
8262 ciphersuites has been selected (as required by the SSL 3/TLS 1
8263 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
8264 is set, we interpret this as a request to violate the specification
8265 (the worst that can happen is a handshake failure, and 'correct'
8266 behaviour would result in a handshake failure anyway).
8269 *) In SSL_CTX_add_session, take into account that there might be multiple
8270 SSL_SESSION structures with the same session ID (e.g. when two threads
8271 concurrently obtain them from an external cache).
8272 The internal cache can handle only one SSL_SESSION with a given ID,
8273 so if there's a conflict, we now throw out the old one to achieve
8277 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
8278 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
8279 some routines that use cipher OIDs: some ciphers do not have OIDs
8280 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
8284 *) Simplify the trust setting structure and code. Now we just have
8285 two sequences of OIDs for trusted and rejected settings. These will
8286 typically have values the same as the extended key usage extension
8287 and any application specific purposes.
8289 The trust checking code now has a default behaviour: it will just
8290 check for an object with the same NID as the passed id. Functions can
8291 be provided to override either the default behaviour or the behaviour
8292 for a given id. SSL client, server and email already have functions
8293 in place for compatibility: they check the NID and also return "trusted"
8294 if the certificate is self signed.
8297 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
8298 traditional format into an EVP_PKEY structure.
8301 *) Add a password callback function PEM_cb() which either prompts for
8302 a password if usr_data is NULL or otherwise assumes it is a null
8303 terminated password. Allow passwords to be passed on command line
8304 environment or config files in a few more utilities.
8307 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
8308 keys. Add some short names for PKCS#8 PBE algorithms and allow them
8309 to be specified on the command line for the pkcs8 and pkcs12 utilities.
8310 Update documentation.
8313 *) Support for ASN1 "NULL" type. This could be handled before by using
8314 ASN1_TYPE but there wasn't any function that would try to read a NULL
8315 and produce an error if it couldn't. For compatibility we also have
8316 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
8317 don't allocate anything because they don't need to.
8320 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
8322 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
8324 *) Rebuild of the memory allocation routines used by OpenSSL code and
8325 possibly others as well. The purpose is to make an interface that
8326 provide hooks so anyone can build a separate set of allocation and
8327 deallocation routines to be used by OpenSSL, for example memory
8328 pool implementations, or something else, which was previously hard
8329 since Malloc(), Realloc() and Free() were defined as macros having
8330 the values malloc, realloc and free, respectively (except for Win32
8331 compilations). The same is provided for memory debugging code.
8332 OpenSSL already comes with functionality to find memory leaks, but
8333 this gives people a chance to debug other memory problems.
8335 With these changes, a new set of functions and macros have appeared:
8337 CRYPTO_set_mem_debug_functions() [F]
8338 CRYPTO_get_mem_debug_functions() [F]
8339 CRYPTO_dbg_set_options() [F]
8340 CRYPTO_dbg_get_options() [F]
8341 CRYPTO_malloc_debug_init() [M]
8343 The memory debug functions are NULL by default, unless the library
8344 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
8345 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
8346 gives the standard debugging functions that come with OpenSSL) or
8347 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
8348 provided by the library user) must be used. When the standard
8349 debugging functions are used, CRYPTO_dbg_set_options can be used to
8350 request additional information:
8351 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
8352 the CRYPTO_MDEBUG_xxx macro when compiling the library.
8354 Also, things like CRYPTO_set_mem_functions will always give the
8355 expected result (the new set of functions is used for allocation
8356 and deallocation) at all times, regardless of platform and compiler
8359 To finish it up, some functions that were never use in any other
8360 way than through macros have a new API and new semantic:
8363 CRYPTO_dbg_realloc()
8366 All macros of value have retained their old syntax.
8367 [Richard Levitte and Bodo Moeller]
8369 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
8370 ordering of SMIMECapabilities wasn't in "strength order" and there
8371 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
8375 *) Some ASN1 types with illegal zero length encoding (INTEGER,
8376 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
8377 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
8379 *) Merge in my S/MIME library for OpenSSL. This provides a simple
8380 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
8381 functionality to handle multipart/signed properly) and a utility
8382 called 'smime' to call all this stuff. This is based on code I
8383 originally wrote for Celo who have kindly allowed it to be
8384 included in OpenSSL.
8387 *) Add variants des_set_key_checked and des_set_key_unchecked of
8388 des_set_key (aka des_key_sched). Global variable des_check_key
8389 decides which of these is called by des_set_key; this way
8390 des_check_key behaves as it always did, but applications and
8391 the library itself, which was buggy for des_check_key == 1,
8392 have a cleaner way to pick the version they need.
8395 *) New function PKCS12_newpass() which changes the password of a
8399 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
8400 dynamic mix. In both cases the ids can be used as an index into the
8401 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
8402 functions so they accept a list of the field values and the
8403 application doesn't need to directly manipulate the X509_TRUST
8407 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
8411 *) Modify the way the V3 extension code looks up extensions. This now
8412 works in a similar way to the object code: we have some "standard"
8413 extensions in a static table which is searched with OBJ_bsearch()
8414 and the application can add dynamic ones if needed. The file
8415 crypto/x509v3/ext_dat.h now has the info: this file needs to be
8416 updated whenever a new extension is added to the core code and kept
8417 in ext_nid order. There is a simple program 'tabtest.c' which checks
8418 this. New extensions are not added too often so this file can readily
8419 be maintained manually.
8421 There are two big advantages in doing things this way. The extensions
8422 can be looked up immediately and no longer need to be "added" using
8423 X509V3_add_standard_extensions(): this function now does nothing.
8424 [Side note: I get *lots* of email saying the extension code doesn't
8425 work because people forget to call this function]
8426 Also no dynamic allocation is done unless new extensions are added:
8427 so if we don't add custom extensions there is no need to call
8428 X509V3_EXT_cleanup().
8431 *) Modify enc utility's salting as follows: make salting the default. Add a
8432 magic header, so unsalted files fail gracefully instead of just decrypting
8433 to garbage. This is because not salting is a big security hole, so people
8434 should be discouraged from doing it.
8437 *) Fixes and enhancements to the 'x509' utility. It allowed a message
8438 digest to be passed on the command line but it only used this
8439 parameter when signing a certificate. Modified so all relevant
8440 operations are affected by the digest parameter including the
8441 -fingerprint and -x509toreq options. Also -x509toreq choked if a
8442 DSA key was used because it didn't fix the digest.
8445 *) Initial certificate chain verify code. Currently tests the untrusted
8446 certificates for consistency with the verify purpose (which is set
8447 when the X509_STORE_CTX structure is set up) and checks the pathlength.
8449 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
8450 this is because it will reject chains with invalid extensions whereas
8451 every previous version of OpenSSL and SSLeay made no checks at all.
8453 Trust code: checks the root CA for the relevant trust settings. Trust
8454 settings have an initial value consistent with the verify purpose: e.g.
8455 if the verify purpose is for SSL client use it expects the CA to be
8456 trusted for SSL client use. However the default value can be changed to
8457 permit custom trust settings: one example of this would be to only trust
8458 certificates from a specific "secure" set of CAs.
8460 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
8461 which should be used for version portability: especially since the
8462 verify structure is likely to change more often now.
8464 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
8465 to set them. If not set then assume SSL clients will verify SSL servers
8468 Two new options to the verify program: -untrusted allows a set of
8469 untrusted certificates to be passed in and -purpose which sets the
8470 intended purpose of the certificate. If a purpose is set then the
8471 new chain verify code is used to check extension consistency.
8474 *) Support for the authority information access extension.
8477 *) Modify RSA and DSA PEM read routines to transparently handle
8478 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
8479 public keys in a format compatible with certificate
8480 SubjectPublicKeyInfo structures. Unfortunately there were already
8481 functions called *_PublicKey_* which used various odd formats so
8482 these are retained for compatibility: however the DSA variants were
8483 never in a public release so they have been deleted. Changed dsa/rsa
8484 utilities to handle the new format: note no releases ever handled public
8485 keys so we should be OK.
8487 The primary motivation for this change is to avoid the same fiasco
8488 that dogs private keys: there are several incompatible private key
8489 formats some of which are standard and some OpenSSL specific and
8490 require various evil hacks to allow partial transparent handling and
8491 even then it doesn't work with DER formats. Given the option anything
8492 other than PKCS#8 should be dumped: but the other formats have to
8493 stay in the name of compatibility.
8495 With public keys and the benefit of hindsight one standard format
8496 is used which works with EVP_PKEY, RSA or DSA structures: though
8497 it clearly returns an error if you try to read the wrong kind of key.
8499 Added a -pubkey option to the 'x509' utility to output the public key.
8500 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
8501 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
8502 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
8503 that do the same as the EVP_PKEY_assign_*() except they up the
8504 reference count of the added key (they don't "swallow" the
8508 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
8509 CRLs would fail if the file contained no certificates or no CRLs:
8510 added a new function to read in both types and return the number
8511 read: this means that if none are read it will be an error. The
8512 DER versions of the certificate and CRL reader would always fail
8513 because it isn't possible to mix certificates and CRLs in DER format
8514 without choking one or the other routine. Changed this to just read
8515 a certificate: this is the best we can do. Also modified the code
8516 in apps/verify.c to take notice of return codes: it was previously
8517 attempting to read in certificates from NULL pointers and ignoring
8518 any errors: this is one reason why the cert and CRL reader seemed
8519 to work. It doesn't check return codes from the default certificate
8520 routines: these may well fail if the certificates aren't installed.
8523 *) Code to support otherName option in GeneralName.
8526 *) First update to verify code. Change the verify utility
8527 so it warns if it is passed a self signed certificate:
8528 for consistency with the normal behaviour. X509_verify
8529 has been modified to it will now verify a self signed
8530 certificate if *exactly* the same certificate appears
8531 in the store: it was previously impossible to trust a
8532 single self signed certificate. This means that:
8533 openssl verify ss.pem
8534 now gives a warning about a self signed certificate but
8535 openssl verify -CAfile ss.pem ss.pem
8539 *) For servers, store verify_result in SSL_SESSION data structure
8540 (and add it to external session representation).
8541 This is needed when client certificate verifications fails,
8542 but an application-provided verification callback (set by
8543 SSL_CTX_set_cert_verify_callback) allows accepting the session
8544 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
8545 but returns 1): When the session is reused, we have to set
8546 ssl->verify_result to the appropriate error code to avoid
8548 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
8550 *) Fix a bug in the new PKCS#7 code: it didn't consider the
8551 case in PKCS7_dataInit() where the signed PKCS7 structure
8552 didn't contain any existing data because it was being created.
8553 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
8555 *) Add a salt to the key derivation routines in enc.c. This
8556 forms the first 8 bytes of the encrypted file. Also add a
8557 -S option to allow a salt to be input on the command line.
8560 *) New function X509_cmp(). Oddly enough there wasn't a function
8561 to compare two certificates. We do this by working out the SHA1
8562 hash and comparing that. X509_cmp() will be needed by the trust
8566 *) SSL_get1_session() is like SSL_get_session(), but increments
8567 the reference count in the SSL_SESSION returned.
8568 [Geoff Thorpe <geoff@eu.c2.net>]
8570 *) Fix for 'req': it was adding a null to request attributes.
8571 Also change the X509_LOOKUP and X509_INFO code to handle
8572 certificate auxiliary information.
8575 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
8579 *) Add the possibility to add extra information to the memory leak
8580 detecting output, to form tracebacks, showing from where each
8581 allocation was originated: CRYPTO_push_info("constant string") adds
8582 the string plus current file name and line number to a per-thread
8583 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
8584 is like calling CYRPTO_pop_info() until the stack is empty.
8585 Also updated memory leak detection code to be multi-thread-safe.
8588 *) Add options -text and -noout to pkcs7 utility and delete the
8589 encryption options which never did anything. Update docs.
8592 *) Add options to some of the utilities to allow the pass phrase
8593 to be included on either the command line (not recommended on
8594 OSes like Unix) or read from the environment. Update the
8595 manpages and fix a few bugs.
8598 *) Add a few manpages for some of the openssl commands.
8601 *) Fix the -revoke option in ca. It was freeing up memory twice,
8602 leaking and not finding already revoked certificates.
8605 *) Extensive changes to support certificate auxiliary information.
8606 This involves the use of X509_CERT_AUX structure and X509_AUX
8607 functions. An X509_AUX function such as PEM_read_X509_AUX()
8608 can still read in a certificate file in the usual way but it
8609 will also read in any additional "auxiliary information". By
8610 doing things this way a fair degree of compatibility can be
8611 retained: existing certificates can have this information added
8612 using the new 'x509' options.
8614 Current auxiliary information includes an "alias" and some trust
8615 settings. The trust settings will ultimately be used in enhanced
8616 certificate chain verification routines: currently a certificate
8617 can only be trusted if it is self signed and then it is trusted
8621 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
8622 The problem was that one of the replacement routines had not been working
8623 since SSLeay releases. For now the offending routine has been replaced
8624 with non-optimised assembler. Even so, this now gives around 95%
8625 performance improvement for 1024 bit RSA signs.
8628 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
8629 handling. Most clients have the effective key size in bits equal to
8630 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
8631 A few however don't do this and instead use the size of the decrypted key
8632 to determine the RC2 key length and the AlgorithmIdentifier to determine
8633 the effective key length. In this case the effective key length can still
8634 be 40 bits but the key length can be 168 bits for example. This is fixed
8635 by manually forcing an RC2 key into the EVP_PKEY structure because the
8636 EVP code can't currently handle unusual RC2 key sizes: it always assumes
8637 the key length and effective key length are equal.
8640 *) Add a bunch of functions that should simplify the creation of
8641 X509_NAME structures. Now you should be able to do:
8642 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
8643 and have it automatically work out the correct field type and fill in
8644 the structures. The more adventurous can try:
8645 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
8646 and it will (hopefully) work out the correct multibyte encoding.
8649 *) Change the 'req' utility to use the new field handling and multibyte
8650 copy routines. Before the DN field creation was handled in an ad hoc
8651 way in req, ca, and x509 which was rather broken and didn't support
8652 BMPStrings or UTF8Strings. Since some software doesn't implement
8653 BMPStrings or UTF8Strings yet, they can be enabled using the config file
8654 using the dirstring_type option. See the new comment in the default
8655 openssl.cnf for more info.
8658 *) Make crypto/rand/md_rand.c more robust:
8659 - Assure unique random numbers after fork().
8660 - Make sure that concurrent threads access the global counter and
8661 md serializably so that we never lose entropy in them
8662 or use exactly the same state in multiple threads.
8663 Access to the large state is not always serializable because
8664 the additional locking could be a performance killer, and
8665 md should be large enough anyway.
8668 *) New file apps/app_rand.c with commonly needed functionality
8669 for handling the random seed file.
8671 Use the random seed file in some applications that previously did not:
8673 dsaparam -genkey (which also ignored its '-rand' option),
8676 x509 (when signing).
8677 Except on systems with /dev/urandom, it is crucial to have a random
8678 seed file at least for key creation, DSA signing, and for DH exchanges;
8679 for RSA signatures we could do without one.
8681 gendh and gendsa (unlike genrsa) used to read only the first byte
8682 of each file listed in the '-rand' option. The function as previously
8683 found in genrsa is now in app_rand.c and is used by all programs
8684 that support '-rand'.
8687 *) In RAND_write_file, use mode 0600 for creating files;
8688 don't just chmod when it may be too late.
8691 *) Report an error from X509_STORE_load_locations
8692 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
8695 *) New function ASN1_mbstring_copy() this copies a string in either
8696 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
8697 into an ASN1_STRING type. A mask of permissible types is passed
8698 and it chooses the "minimal" type to use or an error if not type
8702 *) Add function equivalents to the various macros in asn1.h. The old
8703 macros are retained with an M_ prefix. Code inside the library can
8704 use the M_ macros. External code (including the openssl utility)
8705 should *NOT* in order to be "shared library friendly".
8708 *) Add various functions that can check a certificate's extensions
8709 to see if it usable for various purposes such as SSL client,
8710 server or S/MIME and CAs of these types. This is currently
8711 VERY EXPERIMENTAL but will ultimately be used for certificate chain
8712 verification. Also added a -purpose flag to x509 utility to
8713 print out all the purposes.
8716 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
8720 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
8721 for, obtain and decode and extension and obtain its critical flag.
8722 This allows all the necessary extension code to be handled in a
8723 single function call.
8726 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
8727 platforms. See crypto/rc4/rc4_enc.c for further details.
8730 *) New -noout option to asn1parse. This causes no output to be produced
8731 its main use is when combined with -strparse and -out to extract data
8732 from a file (which may not be in ASN.1 format).
8735 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
8736 when producing the local key id.
8737 [Richard Levitte <levitte@stacken.kth.se>]
8739 *) New option -dhparam in s_server. This allows a DH parameter file to be
8740 stated explicitly. If it is not stated then it tries the first server
8741 certificate file. The previous behaviour hard coded the filename
8745 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
8746 a public key to be input or output. For example:
8747 openssl rsa -in key.pem -pubout -out pubkey.pem
8748 Also added necessary DSA public key functions to handle this.
8751 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
8752 in the message. This was handled by allowing
8753 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8754 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8756 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8757 to the end of the strings whereas this didn't. This would cause problems
8758 if strings read with d2i_ASN1_bytes() were later modified.
8759 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8761 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8762 data and it contains EOF it will end up returning an error. This is
8763 caused by input 46 bytes long. The cause is due to the way base64
8764 BIOs find the start of base64 encoded data. They do this by trying a
8765 trial decode on each line until they find one that works. When they
8766 do a flag is set and it starts again knowing it can pass all the
8767 data directly through the decoder. Unfortunately it doesn't reset
8768 the context it uses. This means that if EOF is reached an attempt
8769 is made to pass two EOFs through the context and this causes the
8770 resulting error. This can also cause other problems as well. As is
8771 usual with these problems it takes *ages* to find and the fix is
8772 trivial: move one line.
8773 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8775 *) Ugly workaround to get s_client and s_server working under Windows. The
8776 old code wouldn't work because it needed to select() on sockets and the
8777 tty (for keypresses and to see if data could be written). Win32 only
8778 supports select() on sockets so we select() with a 1s timeout on the
8779 sockets and then see if any characters are waiting to be read, if none
8780 are present then we retry, we also assume we can always write data to
8781 the tty. This isn't nice because the code then blocks until we've
8782 received a complete line of data and it is effectively polling the
8783 keyboard at 1s intervals: however it's quite a bit better than not
8784 working at all :-) A dedicated Windows application might handle this
8785 with an event loop for example.
8788 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
8789 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
8790 will be called when RSA_sign() and RSA_verify() are used. This is useful
8791 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
8792 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
8793 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
8794 This necessitated the support of an extra signature type NID_md5_sha1
8795 for SSL signatures and modifications to the SSL library to use it instead
8796 of calling RSA_public_decrypt() and RSA_private_encrypt().
8799 *) Add new -verify -CAfile and -CApath options to the crl program, these
8800 will lookup a CRL issuers certificate and verify the signature in a
8801 similar way to the verify program. Tidy up the crl program so it
8802 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
8803 less strict. It will now permit CRL extensions even if it is not
8804 a V2 CRL: this will allow it to tolerate some broken CRLs.
8807 *) Initialize all non-automatic variables each time one of the openssl
8808 sub-programs is started (this is necessary as they may be started
8809 multiple times from the "OpenSSL>" prompt).
8810 [Lennart Bang, Bodo Moeller]
8812 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
8813 removing all other RSA functionality (this is what NO_RSA does). This
8814 is so (for example) those in the US can disable those operations covered
8815 by the RSA patent while allowing storage and parsing of RSA keys and RSA
8819 *) Non-copying interface to BIO pairs.
8820 (still largely untested)
8823 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
8824 ASCII string. This was handled independently in various places before.
8827 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
8828 UTF8 strings a character at a time.
8831 *) Use client_version from client hello to select the protocol
8832 (s23_srvr.c) and for RSA client key exchange verification
8833 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
8836 *) Add various utility functions to handle SPKACs, these were previously
8837 handled by poking round in the structure internals. Added new function
8838 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
8839 print, verify and generate SPKACs. Based on an original idea from
8840 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
8843 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
8846 *) Allow the config file extension section to be overwritten on the
8847 command line. Based on an original idea from Massimiliano Pala
8848 <madwolf@comune.modena.it>. The new option is called -extensions
8849 and can be applied to ca, req and x509. Also -reqexts to override
8850 the request extensions in req and -crlexts to override the crl extensions
8854 *) Add new feature to the SPKAC handling in ca. Now you can include
8855 the same field multiple times by preceding it by "XXXX." for example:
8858 this is the same syntax as used in the req config file.
8861 *) Allow certificate extensions to be added to certificate requests. These
8862 are specified in a 'req_extensions' option of the req section of the
8863 config file. They can be printed out with the -text option to req but
8864 are otherwise ignored at present.
8867 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
8868 data read consists of only the final block it would not decrypted because
8869 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
8870 A misplaced 'break' also meant the decrypted final block might not be
8871 copied until the next read.
8874 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
8875 a few extra parameters to the DH structure: these will be useful if
8876 for example we want the value of 'q' or implement X9.42 DH.
8879 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
8880 provides hooks that allow the default DSA functions or functions on a
8881 "per key" basis to be replaced. This allows hardware acceleration and
8882 hardware key storage to be handled without major modification to the
8883 library. Also added low level modexp hooks and CRYPTO_EX structure and
8884 associated functions.
8887 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
8888 as "read only": it can't be written to and the buffer it points to will
8889 not be freed. Reading from a read only BIO is much more efficient than
8890 a normal memory BIO. This was added because there are several times when
8891 an area of memory needs to be read from a BIO. The previous method was
8892 to create a memory BIO and write the data to it, this results in two
8893 copies of the data and an O(n^2) reading algorithm. There is a new
8894 function BIO_new_mem_buf() which creates a read only memory BIO from
8895 an area of memory. Also modified the PKCS#7 routines to use read only
8899 *) Bugfix: ssl23_get_client_hello did not work properly when called in
8900 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
8901 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
8902 but a retry condition occured while trying to read the rest.
8905 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
8906 NID_pkcs7_encrypted by default: this was wrong since this should almost
8907 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
8908 the encrypted data type: this is a more sensible place to put it and it
8909 allows the PKCS#12 code to be tidied up that duplicated this
8913 *) Changed obj_dat.pl script so it takes its input and output files on
8914 the command line. This should avoid shell escape redirection problems
8918 *) Initial support for certificate extension requests, these are included
8919 in things like Xenroll certificate requests. Included functions to allow
8920 extensions to be obtained and added.
8923 *) -crlf option to s_client and s_server for sending newlines as
8924 CRLF (as required by many protocols).
8927 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
8929 *) Install libRSAglue.a when OpenSSL is built with RSAref.
8930 [Ralf S. Engelschall]
8932 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
8933 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
8935 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
8939 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
8940 DH parameters/keys (q is lost during that conversion, but the resulting
8941 DH parameters contain its length).
8943 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
8944 much faster than DH_generate_parameters (which creates parameters
8945 where p = 2*q + 1), and also the smaller q makes DH computations
8946 much more efficient (160-bit exponentiation instead of 1024-bit
8947 exponentiation); so this provides a convenient way to support DHE
8948 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
8949 utter importance to use
8950 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8952 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8953 when such DH parameters are used, because otherwise small subgroup
8954 attacks may become possible!
8957 *) Avoid memory leak in i2d_DHparams.
8960 *) Allow the -k option to be used more than once in the enc program:
8961 this allows the same encrypted message to be read by multiple recipients.
8964 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
8965 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
8966 it will always use the numerical form of the OID, even if it has a short
8970 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
8971 method only got called if p,q,dmp1,dmq1,iqmp components were present,
8972 otherwise bn_mod_exp was called. In the case of hardware keys for example
8973 no private key components need be present and it might store extra data
8974 in the RSA structure, which cannot be accessed from bn_mod_exp.
8975 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
8976 private key operations.
8979 *) Added support for SPARC Linux.
8982 *) pem_password_cb function type incompatibly changed from
8983 typedef int pem_password_cb(char *buf, int size, int rwflag);
8985 ....(char *buf, int size, int rwflag, void *userdata);
8986 so that applications can pass data to their callbacks:
8987 The PEM[_ASN1]_{read,write}... functions and macros now take an
8988 additional void * argument, which is just handed through whenever
8989 the password callback is called.
8990 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
8992 New function SSL_CTX_set_default_passwd_cb_userdata.
8994 Compatibility note: As many C implementations push function arguments
8995 onto the stack in reverse order, the new library version is likely to
8996 interoperate with programs that have been compiled with the old
8997 pem_password_cb definition (PEM_whatever takes some data that
8998 happens to be on the stack as its last argument, and the callback
8999 just ignores this garbage); but there is no guarantee whatsoever that
9002 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
9003 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
9004 problems not only on Windows, but also on some Unix platforms.
9005 To avoid problematic command lines, these definitions are now in an
9006 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
9007 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
9010 *) MIPS III/IV assembler module is reimplemented.
9013 *) More DES library cleanups: remove references to srand/rand and
9014 delete an unused file.
9017 *) Add support for the the free Netwide assembler (NASM) under Win32,
9018 since not many people have MASM (ml) and it can be hard to obtain.
9019 This is currently experimental but it seems to work OK and pass all
9020 the tests. Check out INSTALL.W32 for info.
9023 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
9024 without temporary keys kept an extra copy of the server key,
9025 and connections with temporary keys did not free everything in case
9029 *) New function RSA_check_key and new openssl rsa option -check
9030 for verifying the consistency of RSA keys.
9031 [Ulf Moeller, Bodo Moeller]
9033 *) Various changes to make Win32 compile work:
9034 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
9035 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
9036 comparison" warnings.
9037 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
9040 *) Add a debugging option to PKCS#5 v2 key generation function: when
9041 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
9042 derived keys are printed to stderr.
9045 *) Copy the flags in ASN1_STRING_dup().
9046 [Roman E. Pavlov <pre@mo.msk.ru>]
9048 *) The x509 application mishandled signing requests containing DSA
9049 keys when the signing key was also DSA and the parameters didn't match.
9051 It was supposed to omit the parameters when they matched the signing key:
9052 the verifying software was then supposed to automatically use the CA's
9053 parameters if they were absent from the end user certificate.
9055 Omitting parameters is no longer recommended. The test was also
9056 the wrong way round! This was probably due to unusual behaviour in
9057 EVP_cmp_parameters() which returns 1 if the parameters match.
9058 This meant that parameters were omitted when they *didn't* match and
9059 the certificate was useless. Certificates signed with 'ca' didn't have
9061 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
9063 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
9064 The interface is as follows:
9065 Applications can use
9066 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
9067 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
9068 "off" is now the default.
9069 The library internally uses
9070 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
9071 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
9072 to disable memory-checking temporarily.
9074 Some inconsistent states that previously were possible (and were
9075 even the default) are now avoided.
9077 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
9078 with each memory chunk allocated; this is occasionally more helpful
9079 than just having a counter.
9081 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
9083 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
9087 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
9088 which largely parallels "options", but is for changing API behaviour,
9089 whereas "options" are about protocol behaviour.
9090 Initial "mode" flags are:
9092 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
9093 a single record has been written.
9094 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
9095 retries use the same buffer location.
9096 (But all of the contents must be
9100 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
9103 *) Fix problems with no-hmac etc.
9104 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
9106 *) New functions RSA_get_default_method(), RSA_set_method() and
9107 RSA_get_method(). These allows replacement of RSA_METHODs without having
9108 to mess around with the internals of an RSA structure.
9111 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
9112 Also really enable memory leak checks in openssl.c and in some
9114 [Chad C. Mulligan, Bodo Moeller]
9116 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
9117 up the length of negative integers. This has now been simplified to just
9118 store the length when it is first determined and use it later, rather
9119 than trying to keep track of where data is copied and updating it to
9121 [Steve Henson, reported by Brien Wheeler
9122 <bwheeler@authentica-security.com>]
9124 *) Add a new function PKCS7_signatureVerify. This allows the verification
9125 of a PKCS#7 signature but with the signing certificate passed to the
9126 function itself. This contrasts with PKCS7_dataVerify which assumes the
9127 certificate is present in the PKCS#7 structure. This isn't always the
9128 case: certificates can be omitted from a PKCS#7 structure and be
9129 distributed by "out of band" means (such as a certificate database).
9132 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
9133 function prototypes in pem.h, also change util/mkdef.pl to add the
9134 necessary function names.
9137 *) mk1mf.pl (used by Windows builds) did not properly read the
9138 options set by Configure in the top level Makefile, and Configure
9139 was not even able to write more than one option correctly.
9140 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
9143 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
9144 file to be loaded from a BIO or FILE pointer. The BIO version will
9145 for example allow memory BIOs to contain config info.
9148 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
9149 Whoever hopes to achieve shared-library compatibility across versions
9150 must use this, not the compile-time macro.
9151 (Exercise 0.9.4: Which is the minimum library version required by
9153 Note: All this applies only to multi-threaded programs, others don't
9157 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
9158 through a BIO pair triggered the default case, i.e.
9159 SSLerr(...,SSL_R_UNKNOWN_STATE).
9162 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
9163 can use the SSL library even if none of the specific BIOs is
9167 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
9168 for the encoded length.
9169 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
9171 *) Add initial documentation of the X509V3 functions.
9174 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
9175 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
9176 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
9177 secure PKCS#8 private key format with a high iteration count.
9180 *) Fix determination of Perl interpreter: A perl or perl5
9181 _directory_ in $PATH was also accepted as the interpreter.
9182 [Ralf S. Engelschall]
9184 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
9185 wrong with it but it was very old and did things like calling
9186 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
9190 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
9191 to use the new extension code.
9194 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
9195 with macros. This should make it easier to change their form, add extra
9196 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
9200 *) Add to configuration table a new entry that can specify an alternative
9201 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
9202 according to Mark Crispin <MRC@Panda.COM>.
9206 *) DES CBC did not update the IV. Weird.
9209 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
9210 Changing the behaviour of the former might break existing programs --
9211 where IV updating is needed, des_ncbc_encrypt can be used.
9214 *) When bntest is run from "make test" it drives bc to check its
9215 calculations, as well as internally checking them. If an internal check
9216 fails, it needs to cause bc to give a non-zero result or make test carries
9217 on without noticing the failure. Fixed.
9220 *) DES library cleanups.
9223 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
9224 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
9225 ciphers. NOTE: although the key derivation function has been verified
9226 against some published test vectors it has not been extensively tested
9227 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
9231 *) Instead of "mkdir -p", which is not fully portable, use new
9232 Perl script "util/mkdir-p.pl".
9235 *) Rewrite the way password based encryption (PBE) is handled. It used to
9236 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
9237 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
9238 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
9239 the 'parameter' field of the AlgorithmIdentifier is passed to the
9240 underlying key generation function so it must do its own ASN1 parsing.
9241 This has also changed the EVP_PBE_CipherInit() function which now has a
9242 'parameter' argument instead of literal salt and iteration count values
9243 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
9246 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
9247 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
9248 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
9249 KEY" because this clashed with PKCS#8 unencrypted string. Since this
9250 value was just used as a "magic string" and not used directly its
9251 value doesn't matter.
9254 *) Introduce some semblance of const correctness to BN. Shame C doesn't
9258 *) "linux-sparc64" configuration (ultrapenguin).
9259 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
9260 "linux-sparc" configuration.
9261 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
9263 *) config now generates no-xxx options for missing ciphers.
9266 *) Support the EBCDIC character set (work in progress).
9267 File ebcdic.c not yet included because it has a different license.
9268 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9270 *) Support BS2000/OSD-POSIX.
9271 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9273 *) Make callbacks for key generation use void * instead of char *.
9276 *) Make S/MIME samples compile (not yet tested).
9279 *) Additional typesafe stacks.
9282 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
9286 Changes between 0.9.3 and 0.9.3a [29 May 1999]
9288 *) New configuration variant "sco5-gcc".
9290 *) Updated some demos.
9291 [Sean O Riordain, Wade Scholine]
9293 *) Add missing BIO_free at exit of pkcs12 application.
9296 *) Fix memory leak in conf.c.
9299 *) Updates for Win32 to assembler version of MD5.
9302 *) Set #! path to perl in apps/der_chop to where we found it
9303 instead of using a fixed path.
9306 *) SHA library changes for irix64-mips4-cc.
9309 *) Improvements for VMS support.
9313 Changes between 0.9.2b and 0.9.3 [24 May 1999]
9315 *) Bignum library bug fix. IRIX 6 passes "make test" now!
9316 This also avoids the problems with SC4.2 and unpatched SC5.
9317 [Andy Polyakov <appro@fy.chalmers.se>]
9319 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
9320 These are required because of the typesafe stack would otherwise break
9321 existing code. If old code used a structure member which used to be STACK
9322 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
9323 sk_num or sk_value it would produce an error because the num, data members
9324 are not present in STACK_OF. Now it just produces a warning. sk_set
9325 replaces the old method of assigning a value to sk_value
9326 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
9327 that does this will no longer work (and should use sk_set instead) but
9328 this could be regarded as a "questionable" behaviour anyway.
9331 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
9332 correctly handle encrypted S/MIME data.
9335 *) Change type of various DES function arguments from des_cblock
9336 (which means, in function argument declarations, pointer to char)
9337 to des_cblock * (meaning pointer to array with 8 char elements),
9338 which allows the compiler to do more typechecking; it was like
9339 that back in SSLeay, but with lots of ugly casts.
9341 Introduce new type const_des_cblock.
9344 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
9345 problems: find RecipientInfo structure that matches recipient certificate
9346 and initialise the ASN1 structures properly based on passed cipher.
9349 *) Belatedly make the BN tests actually check the results.
9352 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
9353 to and from BNs: it was completely broken. New compilation option
9354 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
9355 key elements as negative integers.
9358 *) Reorganize and speed up MD5.
9359 [Andy Polyakov <appro@fy.chalmers.se>]
9362 [Richard Levitte <richard@levitte.org>]
9364 *) New option -out to asn1parse to allow the parsed structure to be
9365 output to a file. This is most useful when combined with the -strparse
9366 option to examine the output of things like OCTET STRINGS.
9369 *) Make SSL library a little more fool-proof by not requiring any longer
9370 that SSL_set_{accept,connect}_state be called before
9371 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
9372 in many applications because usually everything *appeared* to work as
9373 intended anyway -- now it really works as intended).
9376 *) Move openssl.cnf out of lib/.
9379 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
9380 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
9381 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
9382 [Ralf S. Engelschall]
9384 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
9385 handle PKCS#7 enveloped data properly.
9386 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
9388 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
9389 copying pointers. The cert_st handling is changed by this in
9390 various ways (and thus what used to be known as ctx->default_cert
9391 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
9392 any longer when s->cert does not give us what we need).
9393 ssl_cert_instantiate becomes obsolete by this change.
9394 As soon as we've got the new code right (possibly it already is?),
9395 we have solved a couple of bugs of the earlier code where s->cert
9396 was used as if it could not have been shared with other SSL structures.
9398 Note that using the SSL API in certain dirty ways now will result
9399 in different behaviour than observed with earlier library versions:
9400 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
9401 does not influence s as it used to.
9403 In order to clean up things more thoroughly, inside SSL_SESSION
9404 we don't use CERT any longer, but a new structure SESS_CERT
9405 that holds per-session data (if available); currently, this is
9406 the peer's certificate chain and, for clients, the server's certificate
9407 and temporary key. CERT holds only those values that can have
9408 meaningful defaults in an SSL_CTX.
9411 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
9412 from the internal representation. Various PKCS#7 fixes: remove some
9413 evil casts and set the enc_dig_alg field properly based on the signing
9417 *) Allow PKCS#12 password to be set from the command line or the
9418 environment. Let 'ca' get its config file name from the environment
9419 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
9423 *) Allow certificate policies extension to use an IA5STRING for the
9424 organization field. This is contrary to the PKIX definition but
9425 VeriSign uses it and IE5 only recognises this form. Document 'x509'
9429 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
9430 without disallowing inline assembler and the like for non-pedantic builds.
9433 *) Support Borland C++ builder.
9434 [Janez Jere <jj@void.si>, modified by Ulf Möller]
9439 *) SHA-1 cleanups and performance enhancements.
9440 [Andy Polyakov <appro@fy.chalmers.se>]
9442 *) Sparc v8plus assembler for the bignum library.
9443 [Andy Polyakov <appro@fy.chalmers.se>]
9445 *) Accept any -xxx and +xxx compiler options in Configure.
9448 *) Update HPUX configuration.
9451 *) Add missing sk_<type>_unshift() function to safestack.h
9452 [Ralf S. Engelschall]
9454 *) New function SSL_CTX_use_certificate_chain_file that sets the
9455 "extra_cert"s in addition to the certificate. (This makes sense
9456 only for "PEM" format files, as chains as a whole are not
9460 *) Support verify_depth from the SSL API.
9461 x509_vfy.c had what can be considered an off-by-one-error:
9462 Its depth (which was not part of the external interface)
9463 was actually counting the number of certificates in a chain;
9464 now it really counts the depth.
9467 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
9468 instead of X509err, which often resulted in confusing error
9469 messages since the error codes are not globally unique
9470 (e.g. an alleged error in ssl3_accept when a certificate
9471 didn't match the private key).
9473 *) New function SSL_CTX_set_session_id_context that allows to set a default
9474 value (so that you don't need SSL_set_session_id_context for each
9475 connection using the SSL_CTX).
9478 *) OAEP decoding bug fix.
9481 *) Support INSTALL_PREFIX for package builders, as proposed by
9485 *) New Configure options "threads" and "no-threads". For systems
9486 where the proper compiler options are known (currently Solaris
9487 and Linux), "threads" is the default.
9490 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
9493 *) Install various scripts to $(OPENSSLDIR)/misc, not to
9494 $(INSTALLTOP)/bin -- they shouldn't clutter directories
9495 such as /usr/local/bin.
9498 *) "make linux-shared" to build shared libraries.
9499 [Niels Poppe <niels@netbox.org>]
9501 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
9504 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
9505 extension adding in x509 utility.
9508 *) Remove NOPROTO sections and error code comments.
9511 *) Partial rewrite of the DEF file generator to now parse the ANSI
9515 *) New Configure options --prefix=DIR and --openssldir=DIR.
9518 *) Complete rewrite of the error code script(s). It is all now handled
9519 by one script at the top level which handles error code gathering,
9520 header rewriting and C source file generation. It should be much better
9521 than the old method: it now uses a modified version of Ulf's parser to
9522 read the ANSI prototypes in all header files (thus the old K&R definitions
9523 aren't needed for error creation any more) and do a better job of
9524 translating function codes into names. The old 'ASN1 error code imbedded
9525 in a comment' is no longer necessary and it doesn't use .err files which
9526 have now been deleted. Also the error code call doesn't have to appear all
9527 on one line (which resulted in some large lines...).
9530 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
9533 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
9534 0 (which usually indicates a closed connection), but continue reading.
9537 *) Fix some race conditions.
9540 *) Add support for CRL distribution points extension. Add Certificate
9541 Policies and CRL distribution points documentation.
9544 *) Move the autogenerated header file parts to crypto/opensslconf.h.
9547 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
9548 8 of keying material. Merlin has also confirmed interop with this fix
9549 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
9550 [Merlin Hughes <merlin@baltimore.ie>]
9552 *) Fix lots of warnings.
9553 [Richard Levitte <levitte@stacken.kth.se>]
9555 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
9556 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
9557 [Richard Levitte <levitte@stacken.kth.se>]
9559 *) Fix problems with sizeof(long) == 8.
9560 [Andy Polyakov <appro@fy.chalmers.se>]
9562 *) Change functions to ANSI C.
9565 *) Fix typos in error codes.
9566 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
9568 *) Remove defunct assembler files from Configure.
9571 *) SPARC v8 assembler BIGNUM implementation.
9572 [Andy Polyakov <appro@fy.chalmers.se>]
9574 *) Support for Certificate Policies extension: both print and set.
9575 Various additions to support the r2i method this uses.
9578 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
9579 return a const string when you are expecting an allocated buffer.
9582 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
9583 types DirectoryString and DisplayText.
9586 *) Add code to allow r2i extensions to access the configuration database,
9587 add an LHASH database driver and add several ctx helper functions.
9590 *) Fix an evil bug in bn_expand2() which caused various BN functions to
9591 fail when they extended the size of a BIGNUM.
9594 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
9595 support typesafe stack.
9598 *) Fix typo in SSL_[gs]et_options().
9599 [Nils Frostberg <nils@medcom.se>]
9601 *) Delete various functions and files that belonged to the (now obsolete)
9602 old X509V3 handling code.
9605 *) New Configure option "rsaref".
9608 *) Don't auto-generate pem.h.
9611 *) Introduce type-safe ASN.1 SETs.
9614 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
9615 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
9617 *) Introduce type-safe STACKs. This will almost certainly break lots of code
9618 that links with OpenSSL (well at least cause lots of warnings), but fear
9619 not: the conversion is trivial, and it eliminates loads of evil casts. A
9620 few STACKed things have been converted already. Feel free to convert more.
9621 In the fullness of time, I'll do away with the STACK type altogether.
9624 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
9625 specified in <certfile> by updating the entry in the index.txt file.
9626 This way one no longer has to edit the index.txt file manually for
9627 revoking a certificate. The -revoke option does the gory details now.
9628 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
9630 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
9631 `-text' option at all and this way the `-noout -text' combination was
9632 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
9633 [Ralf S. Engelschall]
9635 *) Make sure a corresponding plain text error message exists for the
9636 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
9637 verify callback function determined that a certificate was revoked.
9638 [Ralf S. Engelschall]
9640 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
9641 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
9642 all available cipers including rc5, which was forgotten until now.
9643 In order to let the testing shell script know which algorithms
9644 are available, a new (up to now undocumented) command
9645 "openssl list-cipher-commands" is used.
9648 *) Bugfix: s_client occasionally would sleep in select() when
9649 it should have checked SSL_pending() first.
9652 *) New functions DSA_do_sign and DSA_do_verify to provide access to
9653 the raw DSA values prior to ASN.1 encoding.
9656 *) Tweaks to Configure
9657 [Niels Poppe <niels@netbox.org>]
9659 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
9663 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
9666 *) New config option to avoid instructions that are illegal on the 80386.
9667 The default code is faster, but requires at least a 486.
9670 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
9671 SSL2_SERVER_VERSION (not used at all) macros, which are now the
9672 same as SSL2_VERSION anyway.
9675 *) New "-showcerts" option for s_client.
9678 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
9679 application. Various cleanups and fixes.
9682 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
9683 modify error routines to work internally. Add error codes and PBE init
9684 to library startup routines.
9687 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
9688 packing functions to asn1 and evp. Changed function names and error
9689 codes along the way.
9692 *) PKCS12 integration: and so it begins... First of several patches to
9693 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
9694 objects to objects.h
9697 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
9698 and display support for Thawte strong extranet extension.
9701 *) Add LinuxPPC support.
9702 [Jeff Dubrule <igor@pobox.org>]
9704 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
9705 bn_div_words in alpha.s.
9706 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
9708 *) Make sure the RSA OAEP test is skipped under -DRSAref because
9709 OAEP isn't supported when OpenSSL is built with RSAref.
9710 [Ulf Moeller <ulf@fitug.de>]
9712 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
9713 so they no longer are missing under -DNOPROTO.
9714 [Soren S. Jorvang <soren@t.dk>]
9717 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
9719 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
9720 doesn't work when the session is reused. Coming soon!
9723 *) Fix a security hole, that allows sessions to be reused in the wrong
9724 context thus bypassing client cert protection! All software that uses
9725 client certs and session caches in multiple contexts NEEDS PATCHING to
9726 allow session reuse! A fuller solution is in the works.
9727 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
9729 *) Some more source tree cleanups (removed obsolete files
9730 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
9731 permission on "config" script to be executable) and a fix for the INSTALL
9733 [Ulf Moeller <ulf@fitug.de>]
9735 *) Remove some legacy and erroneous uses of malloc, free instead of
9737 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
9739 *) Make rsa_oaep_test return non-zero on error.
9740 [Ulf Moeller <ulf@fitug.de>]
9742 *) Add support for native Solaris shared libraries. Configure
9743 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
9744 if someone would make that last step automatic.
9745 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
9747 *) ctx_size was not built with the right compiler during "make links". Fixed.
9750 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
9751 except NULL ciphers". This means the default cipher list will no longer
9752 enable NULL ciphers. They need to be specifically enabled e.g. with
9753 the string "DEFAULT:eNULL".
9756 *) Fix to RSA private encryption routines: if p < q then it would
9757 occasionally produce an invalid result. This will only happen with
9758 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9761 *) Be less restrictive and allow also `perl util/perlpath.pl
9762 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9763 because this way one can also use an interpreter named `perl5' (which is
9764 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9765 installed as `perl').
9766 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9768 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9769 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9771 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9772 advapi32.lib to Win32 build and change the pem test comparision
9773 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9774 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9775 and crypto/des/ede_cbcm_enc.c.
9778 *) DES quad checksum was broken on big-endian architectures. Fixed.
9781 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9782 Win32 test batch file so it (might) work again. The Win32 test batch file
9783 is horrible: I feel ill....
9786 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
9787 in e_os.h. Audit of header files to check ANSI and non ANSI
9788 sections: 10 functions were absent from non ANSI section and not exported
9789 from Windows DLLs. Fixed up libeay.num for new functions.
9792 *) Make `openssl version' output lines consistent.
9793 [Ralf S. Engelschall]
9795 *) Fix Win32 symbol export lists for BIO functions: Added
9796 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
9797 to ms/libeay{16,32}.def.
9798 [Ralf S. Engelschall]
9800 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
9801 fine under Unix and passes some trivial tests I've now added. But the
9802 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
9803 added to make sure no one expects that this stuff really works in the
9804 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
9805 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
9807 [Ralf S. Engelschall]
9809 *) Fix the generation of two part addresses in perl.
9810 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
9812 *) Add config entry for Linux on MIPS.
9813 [John Tobey <jtobey@channel1.com>]
9815 *) Make links whenever Configure is run, unless we are on Windoze.
9818 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
9819 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
9823 *) Add a useful kludge to allow package maintainers to specify compiler and
9824 other platforms details on the command line without having to patch the
9825 Configure script everytime: One now can use ``perl Configure
9826 <id>:<details>'', i.e. platform ids are allowed to have details appended
9827 to them (seperated by colons). This is treated as there would be a static
9828 pre-configured entry in Configure's %table under key <id> with value
9829 <details> and ``perl Configure <id>'' is called. So, when you want to
9830 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
9831 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
9832 now, which overrides the FreeBSD-elf entry on-the-fly.
9833 [Ralf S. Engelschall]
9835 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
9838 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
9839 on the `perl Configure ...' command line. This way one can compile
9840 OpenSSL libraries with Position Independent Code (PIC) which is needed
9841 for linking it into DSOs.
9842 [Ralf S. Engelschall]
9844 *) Remarkably, export ciphers were totally broken and no-one had noticed!
9848 *) Cleaned up the LICENSE document: The official contact for any license
9849 questions now is the OpenSSL core team under openssl-core@openssl.org.
9850 And add a paragraph about the dual-license situation to make sure people
9851 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
9852 to the OpenSSL toolkit.
9853 [Ralf S. Engelschall]
9855 *) General source tree makefile cleanups: Made `making xxx in yyy...'
9856 display consistent in the source tree and replaced `/bin/rm' by `rm'.
9857 Additonally cleaned up the `make links' target: Remove unnecessary
9858 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
9859 to speed processing and no longer clutter the display with confusing
9860 stuff. Instead only the actually done links are displayed.
9861 [Ralf S. Engelschall]
9863 *) Permit null encryption ciphersuites, used for authentication only. It used
9864 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
9865 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
9869 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
9870 signed attributes when verifying signatures (this would break them),
9871 the detached data encoding was wrong and public keys obtained using
9872 X509_get_pubkey() weren't freed.
9875 *) Add text documentation for the BUFFER functions. Also added a work around
9876 to a Win95 console bug. This was triggered by the password read stuff: the
9877 last character typed gets carried over to the next fread(). If you were
9878 generating a new cert request using 'req' for example then the last
9879 character of the passphrase would be CR which would then enter the first
9883 *) Added the new `Includes OpenSSL Cryptography Software' button as
9884 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
9885 button and can be used by applications based on OpenSSL to show the
9886 relationship to the OpenSSL project.
9887 [Ralf S. Engelschall]
9889 *) Remove confusing variables in function signatures in files
9890 ssl/ssl_lib.c and ssl/ssl.h.
9891 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9893 *) Don't install bss_file.c under PREFIX/include/
9894 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9896 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
9897 functions that return function pointers and has support for NT specific
9898 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
9899 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
9900 unsigned to signed types: this was killing the Win32 compile.
9903 *) Add new certificate file to stack functions,
9904 SSL_add_dir_cert_subjects_to_stack() and
9905 SSL_add_file_cert_subjects_to_stack(). These largely supplant
9906 SSL_load_client_CA_file(), and can be used to add multiple certs easily
9907 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
9908 This means that Apache-SSL and similar packages don't have to mess around
9909 to add as many CAs as they want to the preferred list.
9912 *) Experiment with doxygen documentation. Currently only partially applied to
9914 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
9915 openssl.doxy as the configuration file.
9918 *) Get rid of remaining C++-style comments which strict C compilers hate.
9919 [Ralf S. Engelschall, pointed out by Carlos Amengual]
9921 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
9922 compiled in by default: it has problems with large keys.
9925 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
9926 DH private keys and/or callback functions which directly correspond to
9927 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
9928 is needed for applications which have to configure certificates on a
9929 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
9931 For the RSA certificate situation is makes no difference, but
9932 for the DSA certificate situation this fixes the "no shared cipher"
9933 problem where the OpenSSL cipher selection procedure failed because the
9934 temporary keys were not overtaken from the context and the API provided
9935 no way to reconfigure them.
9936 The new functions now let applications reconfigure the stuff and they
9937 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
9938 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
9939 non-public-API function ssl_cert_instantiate() is used as a helper
9940 function and also to reduce code redundancy inside ssl_rsa.c.
9941 [Ralf S. Engelschall]
9943 *) Move s_server -dcert and -dkey options out of the undocumented feature
9944 area because they are useful for the DSA situation and should be
9945 recognized by the users.
9946 [Ralf S. Engelschall]
9948 *) Fix the cipher decision scheme for export ciphers: the export bits are
9949 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
9950 SSL_EXP_MASK. So, the original variable has to be used instead of the
9951 already masked variable.
9952 [Richard Levitte <levitte@stacken.kth.se>]
9954 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
9955 [Richard Levitte <levitte@stacken.kth.se>]
9957 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
9958 from `int' to `unsigned int' because it's a length and initialized by
9959 EVP_DigestFinal() which expects an `unsigned int *'.
9960 [Richard Levitte <levitte@stacken.kth.se>]
9962 *) Don't hard-code path to Perl interpreter on shebang line of Configure
9963 script. Instead use the usual Shell->Perl transition trick.
9964 [Ralf S. Engelschall]
9966 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
9967 (in addition to RSA certificates) to match the behaviour of `openssl dsa
9968 -noout -modulus' as it's already the case for `openssl rsa -noout
9969 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
9970 currently the public key is printed (a decision which was already done by
9971 `openssl dsa -modulus' in the past) which serves a similar purpose.
9972 Additionally the NO_RSA no longer completely removes the whole -modulus
9973 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
9975 [Ralf S. Engelschall]
9977 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
9978 BIO. See the source (crypto/evp/bio_ok.c) for more info.
9979 [Arne Ansper <arne@ats.cyber.ee>]
9981 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
9982 to be added. Now both 'req' and 'ca' can use new objects defined in the
9986 *) Add cool BIO that does syslog (or event log on NT).
9987 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
9989 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
9990 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
9991 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
9992 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
9995 *) Add preliminary config info for new extension code.
9998 *) Make RSA_NO_PADDING really use no padding.
9999 [Ulf Moeller <ulf@fitug.de>]
10001 *) Generate errors when private/public key check is done.
10004 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
10005 for some CRL extensions and new objects added.
10008 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
10009 key usage extension and fuller support for authority key id.
10012 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
10013 padding method for RSA, which is recommended for new applications in PKCS
10014 #1 v2.0 (RFC 2437, October 1998).
10015 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
10016 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
10017 against Bleichbacher's attack on RSA.
10018 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
10021 *) Updates to the new SSL compression code
10022 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10024 *) Fix so that the version number in the master secret, when passed
10025 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
10026 (because the server will not accept higher), that the version number
10027 is 0x03,0x01, not 0x03,0x00
10028 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10030 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
10031 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
10032 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
10035 *) Support for RAW extensions where an arbitrary extension can be
10036 created by including its DER encoding. See apps/openssl.cnf for
10040 *) Make sure latest Perl versions don't interpret some generated C array
10041 code as Perl array code in the crypto/err/err_genc.pl script.
10042 [Lars Weber <3weber@informatik.uni-hamburg.de>]
10044 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
10045 not many people have the assembler. Various Win32 compilation fixes and
10046 update to the INSTALL.W32 file with (hopefully) more accurate Win32
10047 build instructions.
10050 *) Modify configure script 'Configure' to automatically create crypto/date.h
10051 file under Win32 and also build pem.h from pem.org. New script
10052 util/mkfiles.pl to create the MINFO file on environments that can't do a
10053 'make files': perl util/mkfiles.pl >MINFO should work.
10056 *) Major rework of DES function declarations, in the pursuit of correctness
10057 and purity. As a result, many evil casts evaporated, and some weirdness,
10058 too. You may find this causes warnings in your code. Zapping your evil
10059 casts will probably fix them. Mostly.
10062 *) Fix for a typo in asn1.h. Bug fix to object creation script
10063 obj_dat.pl. It considered a zero in an object definition to mean
10064 "end of object": none of the objects in objects.h have any zeros
10065 so it wasn't spotted.
10066 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
10068 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
10069 Masking (CBCM). In the absence of test vectors, the best I have been able
10070 to do is check that the decrypt undoes the encrypt, so far. Send me test
10071 vectors if you have them.
10074 *) Correct calculation of key length for export ciphers (too much space was
10075 allocated for null ciphers). This has not been tested!
10078 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
10079 message is now correct (it understands "crypto" and "ssl" on its
10080 command line). There is also now an "update" option. This will update
10081 the util/ssleay.num and util/libeay.num files with any new functions.
10083 perl util/mkdef.pl crypto ssl update
10084 it will update them.
10087 *) Overhauled the Perl interface (perl/*):
10088 - ported BN stuff to OpenSSL's different BN library
10089 - made the perl/ source tree CVS-aware
10090 - renamed the package from SSLeay to OpenSSL (the files still contain
10091 their history because I've copied them in the repository)
10092 - removed obsolete files (the test scripts will be replaced
10093 by better Test::Harness variants in the future)
10094 [Ralf S. Engelschall]
10096 *) First cut for a very conservative source tree cleanup:
10097 1. merge various obsolete readme texts into doc/ssleay.txt
10098 where we collect the old documents and readme texts.
10099 2. remove the first part of files where I'm already sure that we no
10100 longer need them because of three reasons: either they are just temporary
10101 files which were left by Eric or they are preserved original files where
10102 I've verified that the diff is also available in the CVS via "cvs diff
10103 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
10104 the crypto/md/ stuff).
10105 [Ralf S. Engelschall]
10107 *) More extension code. Incomplete support for subject and issuer alt
10108 name, issuer and authority key id. Change the i2v function parameters
10109 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
10110 what that's for :-) Fix to ASN1 macro which messed up
10111 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
10114 *) Preliminary support for ENUMERATED type. This is largely copied from the
10118 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
10119 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10121 *) Make sure `make rehash' target really finds the `openssl' program.
10122 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10124 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
10125 like to hear about it if this slows down other processors.
10128 *) Add CygWin32 platform information to Configure script.
10129 [Alan Batie <batie@aahz.jf.intel.com>]
10131 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
10132 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
10134 *) New program nseq to manipulate netscape certificate sequences
10137 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
10141 *) Fixes to BN code. Previously the default was to define BN_RECURSION
10142 but the BN code had some problems that would cause failures when
10143 doing certificate verification and some other functions.
10144 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10146 *) Add ASN1 and PEM code to support netscape certificate sequences.
10149 *) Add ASN1 and PEM code to support netscape certificate sequences.
10152 *) Add several PKIX and private extended key usage OIDs.
10155 *) Modify the 'ca' program to handle the new extension code. Modify
10156 openssl.cnf for new extension format, add comments.
10159 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
10160 and add a sample to openssl.cnf so req -x509 now adds appropriate
10164 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
10165 error code, add initial support to X509_print() and x509 application.
10168 *) Takes a deep breath and start addding X509 V3 extension support code. Add
10169 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
10170 stuff is currently isolated and isn't even compiled yet.
10173 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
10174 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
10175 Removed the versions check from X509 routines when loading extensions:
10176 this allows certain broken certificates that don't set the version
10177 properly to be processed.
10180 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
10181 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
10182 can still be regenerated with "make depend".
10185 *) Spelling mistake in C version of CAST-128.
10186 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
10188 *) Changes to the error generation code. The perl script err-code.pl
10189 now reads in the old error codes and retains the old numbers, only
10190 adding new ones if necessary. It also only changes the .err files if new
10191 codes are added. The makefiles have been modified to only insert errors
10192 when needed (to avoid needlessly modifying header files). This is done
10193 by only inserting errors if the .err file is newer than the auto generated
10194 C file. To rebuild all the error codes from scratch (the old behaviour)
10195 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
10196 or delete all the .err files.
10199 *) CAST-128 was incorrectly implemented for short keys. The C version has
10200 been fixed, but is untested. The assembler versions are also fixed, but
10201 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
10202 to regenerate it if needed.
10203 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
10204 Hagino <itojun@kame.net>]
10206 *) File was opened incorrectly in randfile.c.
10207 [Ulf Möller <ulf@fitug.de>]
10209 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
10210 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
10211 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
10212 al: it's just almost always a UTCTime. Note this patch adds new error
10213 codes so do a "make errors" if there are problems.
10216 *) Correct Linux 1 recognition in config.
10217 [Ulf Möller <ulf@fitug.de>]
10219 *) Remove pointless MD5 hash when using DSA keys in ca.
10220 [Anonymous <nobody@replay.com>]
10222 *) Generate an error if given an empty string as a cert directory. Also
10223 generate an error if handed NULL (previously returned 0 to indicate an
10224 error, but didn't set one).
10225 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
10227 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
10230 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
10231 parameters. This was causing a warning which killed off the Win32 compile.
10234 *) Remove C++ style comments from crypto/bn/bn_local.h.
10235 [Neil Costigan <neil.costigan@celocom.com>]
10237 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
10238 based on a text string, looking up short and long names and finally
10239 "dot" format. The "dot" format stuff didn't work. Added new function
10240 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
10241 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
10242 OID is not part of the table.
10245 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
10246 X509_LOOKUP_by_alias().
10249 *) Sort openssl functions by name.
10252 *) Get the gendsa program working (hopefully) and add it to app list. Remove
10253 encryption from sample DSA keys (in case anyone is interested the password
10257 *) Make _all_ *_free functions accept a NULL pointer.
10258 [Frans Heymans <fheymans@isaserver.be>]
10260 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
10262 [Anonymous <nobody@replay.com>]
10264 *) s_server should send the CAfile as acceptable CAs, not its own cert.
10265 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10267 *) Don't blow it for numeric -newkey arguments to apps/req.
10268 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10270 *) Temp key "for export" tests were wrong in s3_srvr.c.
10271 [Anonymous <nobody@replay.com>]
10273 *) Add prototype for temp key callback functions
10274 SSL_CTX_set_tmp_{rsa,dh}_callback().
10277 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
10278 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
10281 *) X509_name_add_entry() freed the wrong thing after an error.
10282 [Arne Ansper <arne@ats.cyber.ee>]
10284 *) rsa_eay.c would attempt to free a NULL context.
10285 [Arne Ansper <arne@ats.cyber.ee>]
10287 *) BIO_s_socket() had a broken should_retry() on Windoze.
10288 [Arne Ansper <arne@ats.cyber.ee>]
10290 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
10291 [Arne Ansper <arne@ats.cyber.ee>]
10293 *) Make sure the already existing X509_STORE->depth variable is initialized
10294 in X509_STORE_new(), but document the fact that this variable is still
10295 unused in the certificate verification process.
10296 [Ralf S. Engelschall]
10298 *) Fix the various library and apps files to free up pkeys obtained from
10299 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
10302 *) Fix reference counting in X509_PUBKEY_get(). This makes
10303 demos/maurice/example2.c work, amongst others, probably.
10304 [Steve Henson and Ben Laurie]
10306 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
10307 `openssl' and second, the shortcut symlinks for the `openssl <command>'
10308 are no longer created. This way we have a single and consistent command
10309 line interface `openssl <command>', similar to `cvs <command>'.
10310 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
10312 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
10313 BIT STRING wrapper always have zero unused bits.
10316 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
10319 *) Make the top-level INSTALL documentation easier to understand.
10322 *) Makefiles updated to exit if an error occurs in a sub-directory
10323 make (including if user presses ^C) [Paul Sutton]
10325 *) Make Montgomery context stuff explicit in RSA data structure.
10328 *) Fix build order of pem and err to allow for generated pem.h.
10331 *) Fix renumbering bug in X509_NAME_delete_entry().
10334 *) Enhanced the err-ins.pl script so it makes the error library number
10335 global and can add a library name. This is needed for external ASN1 and
10336 other error libraries.
10339 *) Fixed sk_insert which never worked properly.
10342 *) Fix ASN1 macros so they can handle indefinite length construted
10343 EXPLICIT tags. Some non standard certificates use these: they can now
10347 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
10348 into a single doc/ssleay.txt bundle. This way the information is still
10349 preserved but no longer messes up this directory. Now it's new room for
10350 the new set of documenation files.
10351 [Ralf S. Engelschall]
10353 *) SETs were incorrectly DER encoded. This was a major pain, because they
10354 shared code with SEQUENCEs, which aren't coded the same. This means that
10355 almost everything to do with SETs or SEQUENCEs has either changed name or
10356 number of arguments.
10357 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
10359 *) Fix test data to work with the above.
10362 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
10363 was already fixed by Eric for 0.9.1 it seems.
10364 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
10366 *) Autodetect FreeBSD3.
10369 *) Fix various bugs in Configure. This affects the following platforms:
10373 unixware-2.0-pentium
10377 *) Eliminate generated files from CVS. Reorder tests to regenerate files
10378 before they are needed.
10381 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
10385 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
10387 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
10388 changed SSLeay to OpenSSL in version strings.
10389 [Ralf S. Engelschall]
10391 *) Some fixups to the top-level documents.
10394 *) Fixed the nasty bug where rsaref.h was not found under compile-time
10395 because the symlink to include/ was missing.
10396 [Ralf S. Engelschall]
10398 *) Incorporated the popular no-RSA/DSA-only patches
10399 which allow to compile a RSA-free SSLeay.
10400 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
10402 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
10403 when "ssleay" is still not found.
10404 [Ralf S. Engelschall]
10406 *) Added more platforms to Configure: Cray T3E, HPUX 11,
10407 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
10409 *) Updated the README file.
10410 [Ralf S. Engelschall]
10412 *) Added various .cvsignore files in the CVS repository subdirs
10413 to make a "cvs update" really silent.
10414 [Ralf S. Engelschall]
10416 *) Recompiled the error-definition header files and added
10417 missing symbols to the Win32 linker tables.
10418 [Ralf S. Engelschall]
10420 *) Cleaned up the top-level documents;
10421 o new files: CHANGES and LICENSE
10422 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
10423 o merged COPYRIGHT into LICENSE
10424 o removed obsolete TODO file
10425 o renamed MICROSOFT to INSTALL.W32
10426 [Ralf S. Engelschall]
10428 *) Removed dummy files from the 0.9.1b source tree:
10429 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
10430 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
10431 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
10432 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
10433 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
10434 [Ralf S. Engelschall]
10436 *) Added various platform portability fixes.
10439 *) The Genesis of the OpenSSL rpject:
10440 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
10441 Young and Tim J. Hudson created while they were working for C2Net until
10443 [The OpenSSL Project]
10446 Changes between 0.9.0b and 0.9.1b [not released]
10448 *) Updated a few CA certificates under certs/
10451 *) Changed some BIGNUM api stuff.
10454 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
10455 DGUX x86, Linux Alpha, etc.
10458 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
10459 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
10463 *) Add -strparse option to asn1pars program which parses nested
10465 [Dr Stephen Henson <shenson@bigfoot.com>]
10467 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
10470 *) DSA fix for "ca" program.
10473 *) Added "-genkey" option to "dsaparam" program.
10476 *) Added RIPE MD160 (rmd160) message digest.
10479 *) Added -a (all) option to "ssleay version" command.
10482 *) Added PLATFORM define which is the id given to Configure.
10485 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
10488 *) Extended the ASN.1 parser routines.
10491 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
10494 *) Added a BN_CTX to the BN library.
10497 *) Fixed the weak key values in DES library
10500 *) Changed API in EVP library for cipher aliases.
10503 *) Added support for RC2/64bit cipher.
10506 *) Converted the lhash library to the crypto/mem.c functions.
10509 *) Added more recognized ASN.1 object ids.
10512 *) Added more RSA padding checks for SSL/TLS.
10515 *) Added BIO proxy/filter functionality.
10518 *) Added extra_certs to SSL_CTX which can be used
10519 send extra CA certificates to the client in the CA cert chain sending
10520 process. It can be configured with SSL_CTX_add_extra_chain_cert().
10523 *) Now Fortezza is denied in the authentication phase because
10524 this is key exchange mechanism is not supported by SSLeay at all.
10527 *) Additional PKCS1 checks.
10530 *) Support the string "TLSv1" for all TLS v1 ciphers.
10533 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
10534 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
10537 *) Fixed a few memory leaks.
10540 *) Fixed various code and comment typos.
10543 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
10544 bytes sent in the client random.
10545 [Edward Bishop <ebishop@spyglass.com>]