5 Changes between 1.0.1m and 1.0.1n [xx XXX xxxx]
9 Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
11 *) Segmentation fault in ASN1_TYPE_cmp fix
13 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
14 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
15 certificate signature algorithm consistency this can be used to crash any
16 certificate verification operation and exploited in a DoS attack. Any
17 application which performs certificate verification is vulnerable including
18 OpenSSL clients and servers which enable client authentication.
22 *) ASN.1 structure reuse memory corruption fix
24 Reusing a structure in ASN.1 parsing may allow an attacker to cause
25 memory corruption via an invalid write. Such reuse is and has been
26 strongly discouraged and is believed to be rare.
28 Applications that parse structures containing CHOICE or ANY DEFINED BY
29 components may be affected. Certificate parsing (d2i_X509 and related
30 functions) are however not affected. OpenSSL clients and servers are
35 *) PKCS7 NULL pointer dereferences fix
37 The PKCS#7 parsing code does not handle missing outer ContentInfo
38 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
39 missing content and trigger a NULL pointer dereference on parsing.
41 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
42 otherwise parse PKCS#7 structures from untrusted sources are
43 affected. OpenSSL clients and servers are not affected.
45 This issue was reported to OpenSSL by Michal Zalewski (Google).
49 *) DoS via reachable assert in SSLv2 servers fix
51 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
52 servers that both support SSLv2 and enable export cipher suites by sending
53 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
55 This issue was discovered by Sean Burford (Google) and Emilia Käsper
56 (OpenSSL development team).
60 *) Use After Free following d2i_ECPrivatekey error fix
62 A malformed EC private key file consumed via the d2i_ECPrivateKey function
63 could cause a use after free condition. This, in turn, could cause a double
64 free in several private key parsing functions (such as d2i_PrivateKey
65 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
66 for applications that receive EC private keys from untrusted
67 sources. This scenario is considered rare.
69 This issue was discovered by the BoringSSL project and fixed in their
74 *) X509_to_X509_REQ NULL pointer deref fix
76 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
77 the certificate key is invalid. This function is rarely used in practice.
79 This issue was discovered by Brian Carpenter.
83 *) Removed the export ciphers from the DEFAULT ciphers
86 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
88 *) Build fixes for the Windows and OpenVMS platforms
89 [Matt Caswell and Richard Levitte]
91 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
93 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
94 message can cause a segmentation fault in OpenSSL due to a NULL pointer
95 dereference. This could lead to a Denial Of Service attack. Thanks to
96 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
100 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
101 dtls1_buffer_record function under certain conditions. In particular this
102 could occur if an attacker sent repeated DTLS records with the same
103 sequence number but for the next epoch. The memory leak could be exploited
104 by an attacker in a Denial of Service attack through memory exhaustion.
105 Thanks to Chris Mueller for reporting this issue.
109 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
110 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
111 method would be set to NULL which could later result in a NULL pointer
112 dereference. Thanks to Frank Schmirler for reporting this issue.
116 *) Abort handshake if server key exchange message is omitted for ephemeral
119 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
120 reporting this issue.
124 *) Remove non-export ephemeral RSA code on client and server. This code
125 violated the TLS standard by allowing the use of temporary RSA keys in
126 non-export ciphersuites and could be used by a server to effectively
127 downgrade the RSA key length used to a value smaller than the server
128 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
129 INRIA or reporting this issue.
133 *) Fixed issue where DH client certificates are accepted without verification.
134 An OpenSSL server will accept a DH certificate for client authentication
135 without the certificate verify message. This effectively allows a client to
136 authenticate without the use of a private key. This only affects servers
137 which trust a client certificate authority which issues certificates
138 containing DH keys: these are extremely rare and hardly ever encountered.
139 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
144 *) Ensure that the session ID context of an SSL is updated when its
145 SSL_CTX is updated via SSL_set_SSL_CTX.
147 The session ID context is typically set from the parent SSL_CTX,
148 and can vary with the CTX.
151 *) Fix various certificate fingerprint issues.
153 By using non-DER or invalid encodings outside the signed portion of a
154 certificate the fingerprint can be changed without breaking the signature.
155 Although no details of the signed portion of the certificate can be changed
156 this can cause problems with some applications: e.g. those using the
157 certificate fingerprint for blacklists.
159 1. Reject signatures with non zero unused bits.
161 If the BIT STRING containing the signature has non zero unused bits reject
162 the signature. All current signature algorithms require zero unused bits.
164 2. Check certificate algorithm consistency.
166 Check the AlgorithmIdentifier inside TBS matches the one in the
167 certificate signature. NB: this will result in signature failure
168 errors for some broken certificates.
170 Thanks to Konrad Kraszewski from Google for reporting this issue.
172 3. Check DSA/ECDSA signatures use DER.
174 Reencode DSA/ECDSA signatures and compare with the original received
175 signature. Return an error if there is a mismatch.
177 This will reject various cases including garbage after signature
178 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
179 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
180 (negative or with leading zeroes).
182 Further analysis was conducted and fixes were developed by Stephen Henson
183 of the OpenSSL core team.
188 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
189 results on some platforms, including x86_64. This bug occurs at random
190 with a very low probability, and is not known to be exploitable in any
191 way, though its exact impact is difficult to determine. Thanks to Pieter
192 Wuille (Blockstream) who reported this issue and also suggested an initial
193 fix. Further analysis was conducted by the OpenSSL development team and
194 Adam Langley of Google. The final fix was developed by Andy Polyakov of
195 the OpenSSL core team.
199 *) Do not resume sessions on the server if the negotiated protocol
200 version does not match the session's version. Resuming with a different
201 version, while not strictly forbidden by the RFC, is of questionable
202 sanity and breaks all known clients.
203 [David Benjamin, Emilia Käsper]
205 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
206 early CCS messages during renegotiation. (Note that because
207 renegotiation is encrypted, this early CCS was not exploitable.)
210 *) Tighten client-side session ticket handling during renegotiation:
211 ensure that the client only accepts a session ticket if the server sends
212 the extension anew in the ServerHello. Previously, a TLS client would
213 reuse the old extension state and thus accept a session ticket if one was
214 announced in the initial ServerHello.
216 Similarly, ensure that the client requires a session ticket if one
217 was advertised in the ServerHello. Previously, a TLS client would
218 ignore a missing NewSessionTicket message.
221 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
225 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
226 sends a carefully crafted handshake message, to cause OpenSSL to fail
227 to free up to 64k of memory causing a memory leak. This could be
228 exploited in a Denial Of Service attack. This issue affects OpenSSL
229 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
230 whether SRTP is used or configured. Implementations of OpenSSL that
231 have been compiled with OPENSSL_NO_SRTP defined are not affected.
233 The fix was developed by the OpenSSL team.
237 *) Session Ticket Memory Leak.
239 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
240 integrity of that ticket is first verified. In the event of a session
241 ticket integrity check failing, OpenSSL will fail to free memory
242 causing a memory leak. By sending a large number of invalid session
243 tickets an attacker could exploit this issue in a Denial Of Service
248 *) Build option no-ssl3 is incomplete.
250 When OpenSSL is configured with "no-ssl3" as a build option, servers
251 could accept and complete a SSL 3.0 handshake, and clients could be
252 configured to send them.
254 [Akamai and the OpenSSL team]
256 *) Add support for TLS_FALLBACK_SCSV.
257 Client applications doing fallback retries should call
258 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
260 [Adam Langley, Bodo Moeller]
262 *) Add additional DigestInfo checks.
264 Reencode DigestInto in DER and check against the original when
265 verifying RSA signature: this will reject any improperly encoded
266 DigestInfo structures.
268 Note: this is a precautionary measure and no attacks are currently known.
272 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
274 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
275 SRP code can be overrun an internal buffer. Add sanity check that
276 g, A, B < N to SRP code.
278 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
279 Group for discovering this issue.
283 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
284 TLS 1.0 instead of higher protocol versions when the ClientHello message
285 is badly fragmented. This allows a man-in-the-middle attacker to force a
286 downgrade to TLS 1.0 even if both the server and the client support a
287 higher protocol version, by modifying the client's TLS records.
289 Thanks to David Benjamin and Adam Langley (Google) for discovering and
290 researching this issue.
294 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
295 to a denial of service attack. A malicious server can crash the client
296 with a null pointer dereference (read) by specifying an anonymous (EC)DH
297 ciphersuite and sending carefully crafted handshake messages.
299 Thanks to Felix Gröbert (Google) for discovering and researching this
304 *) By sending carefully crafted DTLS packets an attacker could cause openssl
305 to leak memory. This can be exploited through a Denial of Service attack.
306 Thanks to Adam Langley for discovering and researching this issue.
310 *) An attacker can force openssl to consume large amounts of memory whilst
311 processing DTLS handshake messages. This can be exploited through a
312 Denial of Service attack.
313 Thanks to Adam Langley for discovering and researching this issue.
317 *) An attacker can force an error condition which causes openssl to crash
318 whilst processing DTLS packets due to memory being freed twice. This
319 can be exploited through a Denial of Service attack.
320 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
325 *) If a multithreaded client connects to a malicious server using a resumed
326 session and the server sends an ec point format extension it could write
327 up to 255 bytes to freed memory.
329 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
334 *) A malicious server can crash an OpenSSL client with a null pointer
335 dereference (read) by specifying an SRP ciphersuite even though it was not
336 properly negotiated with the client. This can be exploited through a
337 Denial of Service attack.
339 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
340 discovering and researching this issue.
344 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
345 X509_name_oneline, X509_name_print_ex et al. to leak some information
346 from the stack. Applications may be affected if they echo pretty printing
347 output to the attacker.
349 Thanks to Ivan Fratric (Google) for discovering this issue.
351 [Emilia Käsper, and Steve Henson]
353 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
354 for corner cases. (Certain input points at infinity could lead to
355 bogus results, with non-infinity inputs mapped to infinity too.)
358 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
360 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
361 handshake can force the use of weak keying material in OpenSSL
362 SSL/TLS clients and servers.
364 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
365 researching this issue. (CVE-2014-0224)
366 [KIKUCHI Masashi, Steve Henson]
368 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
369 OpenSSL DTLS client the code can be made to recurse eventually crashing
372 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
374 [Imre Rad, Steve Henson]
376 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
377 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
378 client or server. This is potentially exploitable to run arbitrary
379 code on a vulnerable client or server.
381 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
382 [Jüri Aedla, Steve Henson]
384 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
385 are subject to a denial of service attack.
387 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
388 this issue. (CVE-2014-3470)
389 [Felix Gröbert, Ivan Fratric, Steve Henson]
391 *) Harmonize version and its documentation. -f flag is used to display
393 [mancha <mancha1@zoho.com>]
395 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
397 [mancha <mancha1@zoho.com>]
399 *) Fix some double frees. These are not thought to be exploitable.
400 [mancha <mancha1@zoho.com>]
402 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
404 *) A missing bounds check in the handling of the TLS heartbeat extension
405 can be used to reveal up to 64k of memory to a connected client or
408 Thanks for Neel Mehta of Google Security for discovering this bug and to
409 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
410 preparing the fix (CVE-2014-0160)
411 [Adam Langley, Bodo Moeller]
413 *) Fix for the attack described in the paper "Recovering OpenSSL
414 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
415 by Yuval Yarom and Naomi Benger. Details can be obtained from:
416 http://eprint.iacr.org/2014/140
418 Thanks to Yuval Yarom and Naomi Benger for discovering this
419 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
420 [Yuval Yarom and Naomi Benger]
422 *) TLS pad extension: draft-agl-tls-padding-03
424 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
425 TLS client Hello record length value would otherwise be > 255 and
426 less that 512 pad with a dummy extension containing zeroes so it
427 is at least 512 bytes long.
429 [Adam Langley, Steve Henson]
431 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
433 *) Fix for TLS record tampering bug. A carefully crafted invalid
434 handshake could crash OpenSSL with a NULL pointer exception.
435 Thanks to Anton Johansson for reporting this issues.
438 *) Keep original DTLS digest and encryption contexts in retransmission
439 structures so we can use the previous session parameters if they need
440 to be resent. (CVE-2013-6450)
443 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
444 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
445 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
446 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
447 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
448 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
449 [Rob Stradling, Adam Langley]
451 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
453 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
454 supporting platforms or when small records were transferred.
455 [Andy Polyakov, Steve Henson]
457 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
459 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
461 This addresses the flaw in CBC record processing discovered by
462 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
463 at: http://www.isg.rhul.ac.uk/tls/
465 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
466 Security Group at Royal Holloway, University of London
467 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
468 Emilia Käsper for the initial patch.
470 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
472 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
473 ciphersuites which can be exploited in a denial of service attack.
474 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
475 and detecting this bug and to Wolfgang Ettlinger
476 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
480 *) Return an error when checking OCSP signatures when key is NULL.
481 This fixes a DoS attack. (CVE-2013-0166)
484 *) Make openssl verify return errors.
485 [Chris Palmer <palmer@google.com> and Ben Laurie]
487 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
488 the right response is stapled. Also change SSL_get_certificate()
489 so it returns the certificate actually sent.
490 See http://rt.openssl.org/Ticket/Display.html?id=2836.
491 [Rob Stradling <rob.stradling@comodo.com>]
493 *) Fix possible deadlock when decoding public keys.
496 *) Don't use TLS 1.0 record version number in initial client hello
500 Changes between 1.0.1b and 1.0.1c [10 May 2012]
502 *) Sanity check record length before skipping explicit IV in TLS
503 1.2, 1.1 and DTLS to fix DoS attack.
505 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
506 fuzzing as a service testing platform.
510 *) Initialise tkeylen properly when encrypting CMS messages.
511 Thanks to Solar Designer of Openwall for reporting this issue.
514 *) In FIPS mode don't try to use composite ciphers as they are not
518 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
520 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
521 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
522 mean any application compiled against OpenSSL 1.0.0 headers setting
523 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
524 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
525 0x10000000L Any application which was previously compiled against
526 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
527 will need to be recompiled as a result. Letting be results in
528 inability to disable specifically TLS 1.1 and in client context,
529 in unlike event, limit maximum offered version to TLS 1.0 [see below].
532 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
533 disable just protocol X, but all protocols above X *if* there are
534 protocols *below* X still enabled. In more practical terms it means
535 that if application wants to disable TLS1.0 in favor of TLS1.1 and
536 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
537 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
541 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
543 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
544 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
545 in CRYPTO_realloc_clean.
547 Thanks to Tavis Ormandy, Google Security Team, for discovering this
548 issue and to Adam Langley <agl@chromium.org> for fixing it.
550 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
552 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
555 *) Workarounds for some broken servers that "hang" if a client hello
556 record length exceeds 255 bytes.
558 1. Do not use record version number > TLS 1.0 in initial client
559 hello: some (but not all) hanging servers will now work.
560 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
561 the number of ciphers sent in the client hello. This should be
562 set to an even number, such as 50, for example by passing:
563 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
564 Most broken servers should now work.
565 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
566 TLS 1.2 client support entirely.
569 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
572 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
574 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
575 STRING form instead of a DigestInfo.
578 *) The format used for MDC2 RSA signatures is inconsistent between EVP
579 and the RSA_sign/RSA_verify functions. This was made more apparent when
580 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
581 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
582 the correct format in RSA_verify so both forms transparently work.
585 *) Some servers which support TLS 1.0 can choke if we initially indicate
586 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
587 encrypted premaster secret. As a workaround use the maximum pemitted
588 client version in client hello, this should keep such servers happy
589 and still work with previous versions of OpenSSL.
592 *) Add support for TLS/DTLS heartbeats.
593 [Robin Seggelmann <seggelmann@fh-muenster.de>]
595 *) Add support for SCTP.
596 [Robin Seggelmann <seggelmann@fh-muenster.de>]
598 *) Improved PRNG seeding for VOS.
599 [Paul Green <Paul.Green@stratus.com>]
601 *) Extensive assembler packs updates, most notably:
603 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
604 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
605 - x86_64: bit-sliced AES implementation;
606 - ARM: NEON support, contemporary platforms optimizations;
607 - s390x: z196 support;
608 - *: GHASH and GF(2^m) multiplication implementations;
612 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
613 (removal of unnecessary code)
614 [Peter Sylvester <peter.sylvester@edelweb.fr>]
616 *) Add TLS key material exporter from RFC 5705.
619 *) Add DTLS-SRTP negotiation from RFC 5764.
622 *) Add Next Protocol Negotiation,
623 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
624 disabled with a no-npn flag to config or Configure. Code donated
626 [Adam Langley <agl@google.com> and Ben Laurie]
628 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
629 NIST-P256, NIST-P521, with constant-time single point multiplication on
630 typical inputs. Compiler support for the nonstandard type __uint128_t is
631 required to use this (present in gcc 4.4 and later, for 64-bit builds).
632 Code made available under Apache License version 2.0.
634 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
635 line to include this in your build of OpenSSL, and run "make depend" (or
636 "make update"). This enables the following EC_METHODs:
638 EC_GFp_nistp224_method()
639 EC_GFp_nistp256_method()
640 EC_GFp_nistp521_method()
642 EC_GROUP_new_by_curve_name() will automatically use these (while
643 EC_GROUP_new_curve_GFp() currently prefers the more flexible
645 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
647 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
648 all platforms. Move ssize_t definition from e_os.h to the public
649 header file e_os2.h as it now appears in public header file cms.h
652 *) New -sigopt option to the ca, req and x509 utilities. Additional
653 signature parameters can be passed using this option and in
657 *) Add RSA PSS signing function. This will generate and set the
658 appropriate AlgorithmIdentifiers for PSS based on those in the
659 corresponding EVP_MD_CTX structure. No application support yet.
662 *) Support for companion algorithm specific ASN1 signing routines.
663 New function ASN1_item_sign_ctx() signs a pre-initialised
664 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
665 the appropriate parameters.
668 *) Add new algorithm specific ASN1 verification initialisation function
669 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
670 handling will be the same no matter what EVP_PKEY_METHOD is used.
671 Add a PSS handler to support verification of PSS signatures: checked
672 against a number of sample certificates.
675 *) Add signature printing for PSS. Add PSS OIDs.
676 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
678 *) Add algorithm specific signature printing. An individual ASN1 method
679 can now print out signatures instead of the standard hex dump.
681 More complex signatures (e.g. PSS) can print out more meaningful
682 information. Include DSA version that prints out the signature
686 *) Password based recipient info support for CMS library: implementing
690 *) Split password based encryption into PBES2 and PBKDF2 functions. This
691 neatly separates the code into cipher and PBE sections and is required
692 for some algorithms that split PBES2 into separate pieces (such as
696 *) Session-handling fixes:
697 - Fix handling of connections that are resuming with a session ID,
698 but also support Session Tickets.
699 - Fix a bug that suppressed issuing of a new ticket if the client
700 presented a ticket with an expired session.
701 - Try to set the ticket lifetime hint to something reasonable.
702 - Make tickets shorter by excluding irrelevant information.
703 - On the client side, don't ignore renewed tickets.
704 [Adam Langley, Bodo Moeller (Google)]
706 *) Fix PSK session representation.
709 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
711 This work was sponsored by Intel.
714 *) Add GCM support to TLS library. Some custom code is needed to split
715 the IV between the fixed (from PRF) and explicit (from TLS record)
716 portions. This adds all GCM ciphersuites supported by RFC5288 and
717 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
718 add a special AESGCM string for GCM only.
721 *) Expand range of ctrls for AES GCM. Permit setting invocation
722 field on decrypt and retrieval of invocation field only on encrypt.
725 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
726 As required by RFC5289 these ciphersuites cannot be used if for
727 versions of TLS earlier than 1.2.
730 *) For FIPS capable OpenSSL interpret a NULL default public key method
731 as unset and return the appopriate default but do *not* set the default.
732 This means we can return the appopriate method in applications that
733 swicth between FIPS and non-FIPS modes.
736 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
737 ENGINE is used then we cannot handle that in the FIPS module so we
738 keep original code iff non-FIPS operations are allowed.
741 *) Add -attime option to openssl utilities.
742 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
744 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
747 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
748 FIPS EC methods unconditionally for now.
751 *) New build option no-ec2m to disable characteristic 2 code.
754 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
755 all cases can be covered as some introduce binary incompatibilities.
758 *) Redirect RSA operations to FIPS module including keygen,
759 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
762 *) Add similar low level API blocking to ciphers.
765 *) Low level digest APIs are not approved in FIPS mode: any attempt
766 to use these will cause a fatal error. Applications that *really* want
767 to use them can use the private_* version instead.
770 *) Redirect cipher operations to FIPS module for FIPS builds.
773 *) Redirect digest operations to FIPS module for FIPS builds.
776 *) Update build system to add "fips" flag which will link in fipscanister.o
777 for static and shared library builds embedding a signature if needed.
780 *) Output TLS supported curves in preference order instead of numerical
781 order. This is currently hardcoded for the highest order curves first.
782 This should be configurable so applications can judge speed vs strength.
785 *) Add TLS v1.2 server support for client authentication.
788 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
792 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
793 FIPS modules versions.
796 *) Add TLS v1.2 client side support for client authentication. Keep cache
797 of handshake records longer as we don't know the hash algorithm to use
798 until after the certificate request message is received.
801 *) Initial TLS v1.2 client support. Add a default signature algorithms
802 extension including all the algorithms we support. Parse new signature
803 format in client key exchange. Relax some ECC signing restrictions for
804 TLS v1.2 as indicated in RFC5246.
807 *) Add server support for TLS v1.2 signature algorithms extension. Switch
808 to new signature format when needed using client digest preference.
809 All server ciphersuites should now work correctly in TLS v1.2. No client
810 support yet and no support for client certificates.
813 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
814 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
815 ciphersuites. At present only RSA key exchange ciphersuites work with
816 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
817 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
818 and version checking.
821 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
822 with this defined it will not be affected by any changes to ssl internal
823 structures. Add several utility functions to allow openssl application
824 to work with OPENSSL_NO_SSL_INTERN defined.
828 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
830 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
833 *) Permit abbreviated handshakes when renegotiating using the function
834 SSL_renegotiate_abbreviated().
835 [Robin Seggelmann <seggelmann@fh-muenster.de>]
837 *) Add call to ENGINE_register_all_complete() to
838 ENGINE_load_builtin_engines(), so some implementations get used
839 automatically instead of needing explicit application support.
842 *) Add support for TLS key exporter as described in RFC5705.
843 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
845 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
846 a few changes are required:
848 Add SSL_OP_NO_TLSv1_1 flag.
850 Update version checking logic to handle version 1.1.
851 Add explicit IV handling (ported from DTLS code).
852 Add command line options to s_client/s_server.
855 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
857 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
858 in CMS and PKCS7 code. When RSA decryption fails use a random key for
859 content decryption and always return the same error. Note: this attack
860 needs on average 2^20 messages so it only affects automated senders. The
861 old behaviour can be reenabled in the CMS code by setting the
862 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
863 an MMA defence is not necessary.
864 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
865 this issue. (CVE-2012-0884)
868 *) Fix CVE-2011-4619: make sure we really are receiving a
869 client hello before rejecting multiple SGC restarts. Thanks to
870 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
873 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
875 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
876 Thanks to Antonio Martin, Enterprise Secure Access Research and
877 Development, Cisco Systems, Inc. for discovering this bug and
878 preparing a fix. (CVE-2012-0050)
881 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
883 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
884 of the Vaudenay padding oracle attack on CBC mode encryption
885 which enables an efficient plaintext recovery attack against
886 the OpenSSL implementation of DTLS. Their attack exploits timing
887 differences arising during decryption processing. A research
888 paper describing this attack can be found at:
889 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
890 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
891 Security Group at Royal Holloway, University of London
892 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
893 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
894 for preparing the fix. (CVE-2011-4108)
895 [Robin Seggelmann, Michael Tuexen]
897 *) Clear bytes used for block padding of SSL 3.0 records.
899 [Adam Langley (Google)]
901 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
902 Kadianakis <desnacked@gmail.com> for discovering this issue and
903 Adam Langley for preparing the fix. (CVE-2011-4619)
904 [Adam Langley (Google)]
906 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
907 [Andrey Kulikov <amdeich@gmail.com>]
909 *) Prevent malformed RFC3779 data triggering an assertion failure.
910 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
911 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
912 [Rob Austein <sra@hactrn.net>]
914 *) Improved PRNG seeding for VOS.
915 [Paul Green <Paul.Green@stratus.com>]
917 *) Fix ssl_ciph.c set-up race.
918 [Adam Langley (Google)]
920 *) Fix spurious failures in ecdsatest.c.
921 [Emilia Käsper (Google)]
923 *) Fix the BIO_f_buffer() implementation (which was mixing different
924 interpretations of the '..._len' fields).
925 [Adam Langley (Google)]
927 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
928 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
929 threads won't reuse the same blinding coefficients.
931 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
932 lock to call BN_BLINDING_invert_ex, and avoids one use of
933 BN_BLINDING_update for each BN_BLINDING structure (previously,
934 the last update always remained unused).
935 [Emilia Käsper (Google)]
937 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
938 [Bob Buckholz (Google)]
940 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
942 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
943 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
944 [Kaspar Brand <ossl@velox.ch>]
946 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
947 for multi-threaded use of ECDH. (CVE-2011-3210)
948 [Adam Langley (Google)]
950 *) Fix x509_name_ex_d2i memory leak on bad inputs.
953 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
954 signature public key algorithm by using OID xref utilities instead.
955 Before this you could only use some ECC ciphersuites with SHA1 only.
958 *) Add protection against ECDSA timing attacks as mentioned in the paper
959 by Billy Bob Brumley and Nicola Tuveri, see:
961 http://eprint.iacr.org/2011/232.pdf
963 [Billy Bob Brumley and Nicola Tuveri]
965 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
967 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
968 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
970 *) Fix bug in string printing code: if *any* escaping is enabled we must
971 escape the escape character (backslash) or the resulting string is
975 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
977 *) Disable code workaround for ancient and obsolete Netscape browsers
978 and servers: an attacker can use it in a ciphersuite downgrade attack.
979 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
982 *) Fixed J-PAKE implementation error, originally discovered by
983 Sebastien Martini, further info and confirmation from Stefan
984 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
987 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
989 *) Fix extension code to avoid race conditions which can result in a buffer
990 overrun vulnerability: resumed sessions must not be modified as they can
991 be shared by multiple threads. CVE-2010-3864
994 *) Fix WIN32 build system to correctly link an ENGINE directory into
998 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
1000 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
1002 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
1004 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
1006 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
1007 context. The operation can be customised via the ctrl mechanism in
1008 case ENGINEs want to include additional functionality.
1011 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
1014 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
1015 output hashes compatible with older versions of OpenSSL.
1016 [Willy Weisz <weisz@vcpc.univie.ac.at>]
1018 *) Fix compression algorithm handling: if resuming a session use the
1019 compression algorithm of the resumed session instead of determining
1020 it from client hello again. Don't allow server to change algorithm.
1023 *) Add load_crls() function to apps tidying load_certs() too. Add option
1024 to verify utility to allow additional CRLs to be included.
1027 *) Update OCSP request code to permit adding custom headers to the request:
1028 some responders need this.
1031 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
1033 [Julia Lawall <julia@diku.dk>]
1035 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
1036 needlessly dereferenced structures, used obsolete functions and
1037 didn't handle all updated verify codes correctly.
1040 *) Disable MD2 in the default configuration.
1043 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
1044 indicate the initial BIO being pushed or popped. This makes it possible
1045 to determine whether the BIO is the one explicitly called or as a result
1046 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
1047 it handles reference counts correctly and doesn't zero out the I/O bio
1048 when it is not being explicitly popped. WARNING: applications which
1049 included workarounds for the old buggy behaviour will need to be modified
1050 or they could free up already freed BIOs.
1053 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
1054 renaming to all platforms (within the 0.9.8 branch, this was
1055 done conditionally on Netware platforms to avoid a name clash).
1056 [Guenter <lists@gknw.net>]
1058 *) Add ECDHE and PSK support to DTLS.
1059 [Michael Tuexen <tuexen@fh-muenster.de>]
1061 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
1065 *) Add "missing" function EVP_MD_flags() (without this the only way to
1066 retrieve a digest flags is by accessing the structure directly. Update
1067 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
1068 or cipher is registered as in the "from" argument. Print out all
1069 registered digests in the dgst usage message instead of manually
1070 attempting to work them out.
1073 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
1074 this allows the use of compression and extensions. Change default cipher
1075 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
1076 by default unless an application cipher string requests it.
1079 *) Alter match criteria in PKCS12_parse(). It used to try to use local
1080 key ids to find matching certificates and keys but some PKCS#12 files
1081 don't follow the (somewhat unwritten) rules and this strategy fails.
1082 Now just gather all certificates together and the first private key
1083 then look for the first certificate that matches the key.
1086 *) Support use of registered digest and cipher names for dgst and cipher
1087 commands instead of having to add each one as a special case. So now
1094 openssl dgst -sha256 foo
1096 and this works for ENGINE based algorithms too.
1100 *) Update Gost ENGINE to support parameter files.
1101 [Victor B. Wagner <vitus@cryptocom.ru>]
1103 *) Support GeneralizedTime in ca utility.
1104 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
1106 *) Enhance the hash format used for certificate directory links. The new
1107 form uses the canonical encoding (meaning equivalent names will work
1108 even if they aren't identical) and uses SHA1 instead of MD5. This form
1109 is incompatible with the older format and as a result c_rehash should
1110 be used to rebuild symbolic links.
1113 *) Make PKCS#8 the default write format for private keys, replacing the
1114 traditional format. This form is standardised, more secure and doesn't
1115 include an implicit MD5 dependency.
1118 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
1119 committed to OpenSSL should pass this lot as a minimum.
1122 *) Add session ticket override functionality for use by EAP-FAST.
1123 [Jouni Malinen <j@w1.fi>]
1125 *) Modify HMAC functions to return a value. Since these can be implemented
1126 in an ENGINE errors can occur.
1129 *) Type-checked OBJ_bsearch_ex.
1132 *) Type-checked OBJ_bsearch. Also some constification necessitated
1133 by type-checking. Still to come: TXT_DB, bsearch(?),
1134 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
1138 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
1139 seconds to a tm structure directly, instead of going through OS
1140 specific date routines. This avoids any issues with OS routines such
1141 as the year 2038 bug. New *_adj() functions for ASN1 time structures
1142 and X509_time_adj_ex() to cover the extended range. The existing
1143 X509_time_adj() is still usable and will no longer have any date issues.
1146 *) Delta CRL support. New use deltas option which will attempt to locate
1147 and search any appropriate delta CRLs available.
1149 This work was sponsored by Google.
1152 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
1153 code and add additional score elements. Validate alternate CRL paths
1154 as part of the CRL checking and indicate a new error "CRL path validation
1155 error" in this case. Applications wanting additional details can use
1156 the verify callback and check the new "parent" field. If this is not
1157 NULL CRL path validation is taking place. Existing applications wont
1158 see this because it requires extended CRL support which is off by
1161 This work was sponsored by Google.
1164 *) Support for freshest CRL extension.
1166 This work was sponsored by Google.
1169 *) Initial indirect CRL support. Currently only supported in the CRLs
1170 passed directly and not via lookup. Process certificate issuer
1171 CRL entry extension and lookup CRL entries by bother issuer name
1172 and serial number. Check and process CRL issuer entry in IDP extension.
1174 This work was sponsored by Google.
1177 *) Add support for distinct certificate and CRL paths. The CRL issuer
1178 certificate is validated separately in this case. Only enabled if
1179 an extended CRL support flag is set: this flag will enable additional
1180 CRL functionality in future.
1182 This work was sponsored by Google.
1185 *) Add support for policy mappings extension.
1187 This work was sponsored by Google.
1190 *) Fixes to pathlength constraint, self issued certificate handling,
1191 policy processing to align with RFC3280 and PKITS tests.
1193 This work was sponsored by Google.
1196 *) Support for name constraints certificate extension. DN, email, DNS
1197 and URI types are currently supported.
1199 This work was sponsored by Google.
1202 *) To cater for systems that provide a pointer-based thread ID rather
1203 than numeric, deprecate the current numeric thread ID mechanism and
1204 replace it with a structure and associated callback type. This
1205 mechanism allows a numeric "hash" to be extracted from a thread ID in
1206 either case, and on platforms where pointers are larger than 'long',
1207 mixing is done to help ensure the numeric 'hash' is usable even if it
1208 can't be guaranteed unique. The default mechanism is to use "&errno"
1209 as a pointer-based thread ID to distinguish between threads.
1211 Applications that want to provide their own thread IDs should now use
1212 CRYPTO_THREADID_set_callback() to register a callback that will call
1213 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
1215 Note that ERR_remove_state() is now deprecated, because it is tied
1216 to the assumption that thread IDs are numeric. ERR_remove_state(0)
1217 to free the current thread's error state should be replaced by
1218 ERR_remove_thread_state(NULL).
1220 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
1221 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
1222 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
1223 application was previously providing a numeric thread callback that
1224 was inappropriate for distinguishing threads, then uniqueness might
1225 have been obtained with &errno that happened immediately in the
1226 intermediate development versions of OpenSSL; this is no longer the
1227 case, the numeric thread callback will now override the automatic use
1229 [Geoff Thorpe, with help from Bodo Moeller]
1231 *) Initial support for different CRL issuing certificates. This covers a
1232 simple case where the self issued certificates in the chain exist and
1233 the real CRL issuer is higher in the existing chain.
1235 This work was sponsored by Google.
1238 *) Removed effectively defunct crypto/store from the build.
1241 *) Revamp of STACK to provide stronger type-checking. Still to come:
1242 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
1243 ASN1_STRING, CONF_VALUE.
1246 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
1247 RAM on SSL connections. This option can save about 34k per idle SSL.
1250 *) Revamp of LHASH to provide stronger type-checking. Still to come:
1251 STACK, TXT_DB, bsearch, qsort.
1254 *) Initial support for Cryptographic Message Syntax (aka CMS) based
1255 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
1256 support for data, signedData, compressedData, digestedData and
1257 encryptedData, envelopedData types included. Scripts to check against
1258 RFC4134 examples draft and interop and consistency checks of many
1259 content types and variants.
1262 *) Add options to enc utility to support use of zlib compression BIO.
1265 *) Extend mk1mf to support importing of options and assembly language
1266 files from Configure script, currently only included in VC-WIN32.
1267 The assembly language rules can now optionally generate the source
1268 files from the associated perl scripts.
1271 *) Implement remaining functionality needed to support GOST ciphersuites.
1272 Interop testing has been performed using CryptoPro implementations.
1273 [Victor B. Wagner <vitus@cryptocom.ru>]
1275 *) s390x assembler pack.
1278 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
1282 *) Implement Opaque PRF Input TLS extension as specified in
1283 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
1284 official specification yet and no extension type assignment by
1285 IANA exists, this extension (for now) will have to be explicitly
1286 enabled when building OpenSSL by providing the extension number
1287 to use. For example, specify an option
1289 -DTLSEXT_TYPE_opaque_prf_input=0x9527
1291 to the "config" or "Configure" script to enable the extension,
1292 assuming extension number 0x9527 (which is a completely arbitrary
1293 and unofficial assignment based on the MD5 hash of the Internet
1294 Draft). Note that by doing so, you potentially lose
1295 interoperability with other TLS implementations since these might
1296 be using the same extension number for other purposes.
1298 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
1299 opaque PRF input value to use in the handshake. This will create
1300 an interal copy of the length-'len' string at 'src', and will
1301 return non-zero for success.
1303 To get more control and flexibility, provide a callback function
1306 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
1307 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
1311 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
1314 Callback function 'cb' will be called in handshakes, and is
1315 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
1316 Argument 'arg' is for application purposes (the value as given to
1317 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
1318 be provided to the callback function). The callback function
1319 has to return non-zero to report success: usually 1 to use opaque
1320 PRF input just if possible, or 2 to enforce use of the opaque PRF
1321 input. In the latter case, the library will abort the handshake
1322 if opaque PRF input is not successfully negotiated.
1324 Arguments 'peerinput' and 'len' given to the callback function
1325 will always be NULL and 0 in the case of a client. A server will
1326 see the client's opaque PRF input through these variables if
1327 available (NULL and 0 otherwise). Note that if the server
1328 provides an opaque PRF input, the length must be the same as the
1329 length of the client's opaque PRF input.
1331 Note that the callback function will only be called when creating
1332 a new session (session resumption can resume whatever was
1333 previously negotiated), and will not be called in SSL 2.0
1334 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
1335 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
1336 for applications that need to enforce opaque PRF input.
1340 *) Update ssl code to support digests other than SHA1+MD5 for handshake
1343 [Victor B. Wagner <vitus@cryptocom.ru>]
1345 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1346 RFC4507bis. The encrypted ticket format is an encrypted encoded
1347 SSL_SESSION structure, that way new session features are automatically
1350 If a client application caches session in an SSL_SESSION structure
1351 support is transparent because tickets are now stored in the encoded
1354 The SSL_CTX structure automatically generates keys for ticket
1355 protection in servers so again support should be possible
1356 with no application modification.
1358 If a client or server wishes to disable RFC4507 support then the option
1359 SSL_OP_NO_TICKET can be set.
1361 Add a TLS extension debugging callback to allow the contents of any client
1362 or server extensions to be examined.
1364 This work was sponsored by Google.
1367 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
1368 OpenSSL should now compile cleanly on gcc 4.2
1369 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
1371 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
1372 support including streaming MAC support: this is required for GOST
1373 ciphersuite support.
1374 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
1376 *) Add option -stream to use PKCS#7 streaming in smime utility. New
1377 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
1378 to output in BER and PEM format.
1381 *) Experimental support for use of HMAC via EVP_PKEY interface. This
1382 allows HMAC to be handled via the EVP_DigestSign*() interface. The
1383 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
1384 ENGINE support for HMAC keys which are unextractable. New -mac and
1385 -macopt options to dgst utility.
1388 *) New option -sigopt to dgst utility. Update dgst to use
1389 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
1390 alternative signing paramaters such as X9.31 or PSS in the dgst
1394 *) Change ssl_cipher_apply_rule(), the internal function that does
1395 the work each time a ciphersuite string requests enabling
1396 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
1397 removing ("!foo+bar") a class of ciphersuites: Now it maintains
1398 the order of disabled ciphersuites such that those ciphersuites
1399 that most recently went from enabled to disabled not only stay
1400 in order with respect to each other, but also have higher priority
1401 than other disabled ciphersuites the next time ciphersuites are
1404 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
1405 the same ciphersuites as with "HIGH" alone, but in a specific
1406 order where the PSK ciphersuites come first (since they are the
1407 most recently disabled ciphersuites when "HIGH" is parsed).
1409 Also, change ssl_create_cipher_list() (using this new
1410 funcionality) such that between otherwise identical
1411 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
1415 *) Change ssl_create_cipher_list() so that it automatically
1416 arranges the ciphersuites in reasonable order before starting
1417 to process the rule string. Thus, the definition for "DEFAULT"
1418 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
1419 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
1420 This makes it much easier to arrive at a reasonable default order
1421 in applications for which anonymous ciphers are OK (meaning
1422 that you can't actually use DEFAULT).
1423 [Bodo Moeller; suggested by Victor Duchovni]
1425 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
1426 processing) into multiple integers instead of setting
1427 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
1428 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
1429 (These masks as well as the individual bit definitions are hidden
1430 away into the non-exported interface ssl/ssl_locl.h, so this
1431 change to the definition of the SSL_CIPHER structure shouldn't
1432 affect applications.) This give us more bits for each of these
1433 categories, so there is no longer a need to coagulate AES128 and
1434 AES256 into a single algorithm bit, and to coagulate Camellia128
1435 and Camellia256 into a single algorithm bit, which has led to all
1438 Thus, among other things, the kludge introduced in 0.9.7m and
1439 0.9.8e for masking out AES256 independently of AES128 or masking
1440 out Camellia256 independently of AES256 is not needed here in 0.9.9.
1442 With the change, we also introduce new ciphersuite aliases that
1443 so far were missing: "AES128", "AES256", "CAMELLIA128", and
1447 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
1448 Use the leftmost N bytes of the signature input if the input is
1449 larger than the prime q (with N being the size in bytes of q).
1452 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
1453 it yet and it is largely untested.
1456 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
1459 *) Initial incomplete changes to avoid need for function casts in OpenSSL
1460 some compilers (gcc 4.2 and later) reject their use. Safestack is
1461 reimplemented. Update ASN1 to avoid use of legacy functions.
1464 *) Win32/64 targets are linked with Winsock2.
1467 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
1468 to external functions. This can be used to increase CRL handling
1469 efficiency especially when CRLs are very large by (for example) storing
1470 the CRL revoked certificates in a database.
1473 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
1474 new CRLs added to a directory can be used. New command line option
1475 -verify_return_error to s_client and s_server. This causes real errors
1476 to be returned by the verify callback instead of carrying on no matter
1477 what. This reflects the way a "real world" verify callback would behave.
1480 *) GOST engine, supporting several GOST algorithms and public key formats.
1481 Kindly donated by Cryptocom.
1484 *) Partial support for Issuing Distribution Point CRL extension. CRLs
1485 partitioned by DP are handled but no indirect CRL or reason partitioning
1486 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
1487 selected via a scoring technique which handles IDP and AKID in CRLs.
1490 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
1491 will ultimately be used for all verify operations: this will remove the
1492 X509_STORE dependency on certificate verification and allow alternative
1493 lookup methods. X509_STORE based implementations of these two callbacks.
1496 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
1497 Modify get_crl() to find a valid (unexpired) CRL if possible.
1500 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
1501 this would be called X509_CRL_cmp() but that name is already used by
1502 a function that just compares CRL issuer names. Cache several CRL
1503 extensions in X509_CRL structure and cache CRLDP in X509.
1506 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
1507 this maps equivalent X509_NAME structures into a consistent structure.
1508 Name comparison can then be performed rapidly using memcmp().
1511 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
1515 *) Allow digests to supply their own micalg string for S/MIME type using
1516 the ctrl EVP_MD_CTRL_MICALG.
1519 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
1520 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
1521 ctrl. It can then customise the structure before and/or after signing
1525 *) New function OBJ_add_sigid() to allow application defined signature OIDs
1526 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
1527 to free up any added signature OIDs.
1530 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
1531 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
1532 digest and cipher tables. New options added to openssl utility:
1533 list-message-digest-algorithms and list-cipher-algorithms.
1536 *) Change the array representation of binary polynomials: the list
1537 of degrees of non-zero coefficients is now terminated with -1.
1538 Previously it was terminated with 0, which was also part of the
1539 value; thus, the array representation was not applicable to
1540 polynomials where t^0 has coefficient zero. This change makes
1541 the array representation useful in a more general context.
1544 *) Various modifications and fixes to SSL/TLS cipher string
1545 handling. For ECC, the code now distinguishes between fixed ECDH
1546 with RSA certificates on the one hand and with ECDSA certificates
1547 on the other hand, since these are separate ciphersuites. The
1548 unused code for Fortezza ciphersuites has been removed.
1550 For consistency with EDH, ephemeral ECDH is now called "EECDH"
1551 (not "ECDHE"). For consistency with the code for DH
1552 certificates, use of ECDH certificates is now considered ECDH
1553 authentication, not RSA or ECDSA authentication (the latter is
1554 merely the CA's signing algorithm and not actively used in the
1557 The temporary ciphersuite alias "ECCdraft" is no longer
1558 available, and ECC ciphersuites are no longer excluded from "ALL"
1559 and "DEFAULT". The following aliases now exist for RFC 4492
1560 ciphersuites, most of these by analogy with the DH case:
1562 kECDHr - ECDH cert, signed with RSA
1563 kECDHe - ECDH cert, signed with ECDSA
1564 kECDH - ECDH cert (signed with either RSA or ECDSA)
1565 kEECDH - ephemeral ECDH
1566 ECDH - ECDH cert or ephemeral ECDH
1572 AECDH - anonymous ECDH
1573 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
1577 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
1578 Use correct micalg parameters depending on digest(s) in signed message.
1581 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
1582 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
1585 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
1586 an engine to register a method. Add ENGINE lookups for methods and
1587 functional reference processing.
1590 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
1591 EVP_{Sign,Verify}* which allow an application to customise the signature
1595 *) New -resign option to smime utility. This adds one or more signers
1596 to an existing PKCS#7 signedData structure. Also -md option to use an
1597 alternative message digest algorithm for signing.
1600 *) Tidy up PKCS#7 routines and add new functions to make it easier to
1601 create PKCS7 structures containing multiple signers. Update smime
1602 application to support multiple signers.
1605 *) New -macalg option to pkcs12 utility to allow setting of an alternative
1609 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
1610 Reorganize PBE internals to lookup from a static table using NIDs,
1611 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
1612 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
1613 PRF which will be automatically used with PBES2.
1616 *) Replace the algorithm specific calls to generate keys in "req" with the
1620 *) Update PKCS#7 enveloped data routines to use new API. This is now
1621 supported by any public key method supporting the encrypt operation. A
1622 ctrl is added to allow the public key algorithm to examine or modify
1623 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
1627 *) Add a ctrl to asn1 method to allow a public key algorithm to express
1628 a default digest type to use. In most cases this will be SHA1 but some
1629 algorithms (such as GOST) need to specify an alternative digest. The
1630 return value indicates how strong the prefernce is 1 means optional and
1631 2 is mandatory (that is it is the only supported type). Modify
1632 ASN1_item_sign() to accept a NULL digest argument to indicate it should
1633 use the default md. Update openssl utilities to use the default digest
1634 type for signing if it is not explicitly indicated.
1637 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
1638 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
1639 signing method from the key type. This effectively removes the link
1640 between digests and public key types.
1643 *) Add an OID cross reference table and utility functions. Its purpose is to
1644 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
1645 rsaEncryption. This will allow some of the algorithm specific hackery
1646 needed to use the correct OID to be removed.
1649 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
1650 structures for PKCS7_sign(). They are now set up by the relevant public
1654 *) Add provisional EC pkey method with support for ECDSA and ECDH.
1657 *) Add support for key derivation (agreement) in the API, DH method and
1661 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
1662 public and private key formats. As a side effect these add additional
1663 command line functionality not previously available: DSA signatures can be
1664 generated and verified using pkeyutl and DH key support and generation in
1669 [Oliver Tappe <zooey@hirschkaefer.de>]
1671 *) New make target "install_html_docs" installs HTML renditions of the
1673 [Oliver Tappe <zooey@hirschkaefer.de>]
1675 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
1676 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
1677 support key and parameter generation and add initial key generation
1678 functionality for RSA.
1681 *) Add functions for main EVP_PKEY_method operations. The undocumented
1682 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
1683 EVP_PKEY_{encrypt,decrypt}_old.
1686 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
1687 key API, doesn't do much yet.
1690 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
1691 public key algorithms. New option to openssl utility:
1692 "list-public-key-algorithms" to print out info.
1695 *) Implement the Supported Elliptic Curves Extension for
1696 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1699 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
1700 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
1703 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
1704 utilities such as rsa, dsa, dsaparam etc except they process any key
1708 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1709 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1710 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
1714 *) Initial support for pluggable public key ASN1.
1715 De-spaghettify the public key ASN1 handling. Move public and private
1716 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
1717 algorithm specific handling to a single module within the relevant
1718 algorithm directory. Add functions to allow (near) opaque processing
1719 of public and private key structures.
1722 *) Implement the Supported Point Formats Extension for
1723 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1726 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
1727 for the psk identity [hint] and the psk callback functions to the
1728 SSL_SESSION, SSL and SSL_CTX structure.
1731 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
1735 SSL_CTX_use_psk_identity_hint
1736 SSL_get_psk_identity_hint
1737 SSL_get_psk_identity
1738 SSL_use_psk_identity_hint
1740 [Mika Kousa and Pasi Eronen of Nokia Corporation]
1742 *) Add RFC 3161 compliant time stamp request creation, response generation
1743 and response verification functionality.
1744 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
1746 *) Add initial support for TLS extensions, specifically for the server_name
1747 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1748 have new members for a host name. The SSL data structure has an
1749 additional member SSL_CTX *initial_ctx so that new sessions can be
1750 stored in that context to allow for session resumption, even after the
1751 SSL has been switched to a new SSL_CTX in reaction to a client's
1752 server_name extension.
1754 New functions (subject to change):
1756 SSL_get_servername()
1757 SSL_get_servername_type()
1760 New CTRL codes and macros (subject to change):
1762 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1763 - SSL_CTX_set_tlsext_servername_callback()
1764 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1765 - SSL_CTX_set_tlsext_servername_arg()
1766 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1768 openssl s_client has a new '-servername ...' option.
1770 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1771 '-key2 ...', '-servername_fatal' (subject to change). This allows
1772 testing the HostName extension for a specific single host name ('-cert'
1773 and '-key' remain fallbacks for handshakes without HostName
1774 negotiation). If the unrecogninzed_name alert has to be sent, this by
1775 default is a warning; it becomes fatal with the '-servername_fatal'
1778 [Peter Sylvester, Remy Allais, Christophe Renou]
1780 *) Whirlpool hash implementation is added.
1783 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1784 bn(64,32). Because of instruction set limitations it doesn't have
1785 any negative impact on performance. This was done mostly in order
1786 to make it possible to share assembler modules, such as bn_mul_mont
1787 implementations, between 32- and 64-bit builds without hassle.
1790 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1791 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1795 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1796 dedicated Montgomery multiplication procedure, is introduced.
1797 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1798 "64-bit" performance on certain 32-bit targets.
1801 *) New option SSL_OP_NO_COMP to disable use of compression selectively
1802 in SSL structures. New SSL ctrl to set maximum send fragment size.
1803 Save memory by seeting the I/O buffer sizes dynamically instead of
1804 using the maximum available value.
1807 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
1808 in addition to the text details.
1811 *) Very, very preliminary EXPERIMENTAL support for printing of general
1812 ASN1 structures. This currently produces rather ugly output and doesn't
1813 handle several customised structures at all.
1816 *) Integrated support for PVK file format and some related formats such
1817 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
1818 these in the 'rsa' and 'dsa' utilities.
1821 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
1824 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
1825 place for the (very old) "NETSCAPE" format certificates which are now
1826 handled using new ASN1 code equivalents.
1829 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
1830 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
1831 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
1834 *) Modify CRL distribution points extension code to print out previously
1835 unsupported fields. Enhance extension setting code to allow setting of
1839 *) Add print and set support for Issuing Distribution Point CRL extension.
1842 *) Change 'Configure' script to enable Camellia by default.
1845 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
1847 *) When rejecting SSL/TLS records due to an incorrect version number, never
1848 update s->server with a new major version number. As of
1849 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
1850 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
1851 the previous behavior could result in a read attempt at NULL when
1852 receiving specific incorrect SSL/TLS records once record payload
1853 protection is active. (CVE-2010-0740)
1854 [Bodo Moeller, Adam Langley <agl@chromium.org>]
1856 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
1857 could be crashed if the relevant tables were not present (e.g. chrooted).
1858 [Tomas Hoger <thoger@redhat.com>]
1860 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
1862 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
1863 [Martin Olsson, Neel Mehta]
1865 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
1866 accommodate for stack sorting, always a write lock!).
1869 *) On some versions of WIN32 Heap32Next is very slow. This can cause
1870 excessive delays in the RAND_poll(): over a minute. As a workaround
1871 include a time check in the inner Heap32Next loop too.
1874 *) The code that handled flushing of data in SSL/TLS originally used the
1875 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
1876 the problem outlined in PR#1949. The fix suggested there however can
1877 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
1878 of Apache). So instead simplify the code to flush unconditionally.
1879 This should be fine since flushing with no data to flush is a no op.
1882 *) Handle TLS versions 2.0 and later properly and correctly use the
1883 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
1884 off ancient servers have a habit of sticking around for a while...
1887 *) Modify compression code so it frees up structures without using the
1888 ex_data callbacks. This works around a problem where some applications
1889 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
1890 restarting) then use compression (e.g. SSL with compression) later.
1891 This results in significant per-connection memory leaks and
1892 has caused some security issues including CVE-2008-1678 and
1896 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
1897 change when encrypting or decrypting.
1900 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
1901 connect and renegotiate with servers which do not support RI.
1902 Until RI is more widely deployed this option is enabled by default.
1905 *) Add "missing" ssl ctrls to clear options and mode.
1908 *) If client attempts to renegotiate and doesn't support RI respond with
1909 a no_renegotiation alert as required by RFC5746. Some renegotiating
1910 TLS clients will continue a connection gracefully when they receive
1911 the alert. Unfortunately OpenSSL mishandled this alert and would hang
1912 waiting for a server hello which it will never receive. Now we treat a
1913 received no_renegotiation alert as a fatal error. This is because
1914 applications requesting a renegotiation might well expect it to succeed
1915 and would have no code in place to handle the server denying it so the
1916 only safe thing to do is to terminate the connection.
1919 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
1920 peer supports secure renegotiation and 0 otherwise. Print out peer
1921 renegotiation support in s_client/s_server.
1924 *) Replace the highly broken and deprecated SPKAC certification method with
1925 the updated NID creation version. This should correctly handle UTF8.
1928 *) Implement RFC5746. Re-enable renegotiation but require the extension
1929 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
1930 turns out to be a bad idea. It has been replaced by
1931 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
1932 SSL_CTX_set_options(). This is really not recommended unless you
1933 know what you are doing.
1934 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
1936 *) Fixes to stateless session resumption handling. Use initial_ctx when
1937 issuing and attempting to decrypt tickets in case it has changed during
1938 servername handling. Use a non-zero length session ID when attempting
1939 stateless session resumption: this makes it possible to determine if
1940 a resumption has occurred immediately after receiving server hello
1941 (several places in OpenSSL subtly assume this) instead of later in
1945 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
1946 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
1947 fixes for a few places where the return code is not checked
1949 [Julia Lawall <julia@diku.dk>]
1951 *) Add --strict-warnings option to Configure script to include devteam
1952 warnings in other configurations.
1955 *) Add support for --libdir option and LIBDIR variable in makefiles. This
1956 makes it possible to install openssl libraries in locations which
1957 have names other than "lib", for example "/usr/lib64" which some
1959 [Steve Henson, based on patch from Jeremy Utley]
1961 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
1962 X690 8.9.12 and can produce some misleading textual output of OIDs.
1963 [Steve Henson, reported by Dan Kaminsky]
1965 *) Delete MD2 from algorithm tables. This follows the recommendation in
1966 several standards that it is not used in new applications due to
1967 several cryptographic weaknesses. For binary compatibility reasons
1968 the MD2 API is still compiled in by default.
1971 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
1975 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
1976 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
1978 [Guenter <lists@gknw.net>]
1980 *) Fix the server certificate chain building code to use X509_verify_cert(),
1981 it used to have an ad-hoc builder which was unable to cope with anything
1982 other than a simple chain.
1983 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
1985 *) Don't check self signed certificate signatures in X509_verify_cert()
1986 by default (a flag can override this): it just wastes time without
1987 adding any security. As a useful side effect self signed root CAs
1988 with non-FIPS digests are now usable in FIPS mode.
1991 *) In dtls1_process_out_of_seq_message() the check if the current message
1992 is already buffered was missing. For every new message was memory
1993 allocated, allowing an attacker to perform an denial of service attack
1994 with sending out of seq handshake messages until there is no memory
1995 left. Additionally every future messege was buffered, even if the
1996 sequence number made no sense and would be part of another handshake.
1997 So only messages with sequence numbers less than 10 in advance will be
1998 buffered. (CVE-2009-1378)
1999 [Robin Seggelmann, discovered by Daniel Mentz]
2001 *) Records are buffered if they arrive with a future epoch to be
2002 processed after finishing the corresponding handshake. There is
2003 currently no limitation to this buffer allowing an attacker to perform
2004 a DOS attack with sending records with future epochs until there is no
2005 memory left. This patch adds the pqueue_size() function to detemine
2006 the size of a buffer and limits the record buffer to 100 entries.
2008 [Robin Seggelmann, discovered by Daniel Mentz]
2010 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
2011 parent structure is freed. (CVE-2009-1379)
2014 *) Handle non-blocking I/O properly in SSL_shutdown() call.
2015 [Darryl Miles <darryl-mailinglists@netbauds.net>]
2018 [Ilya O. <vrghost@gmail.com>]
2020 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
2022 *) Disable renegotiation completely - this fixes a severe security
2023 problem (CVE-2009-3555) at the cost of breaking all
2024 renegotiation. Renegotiation can be re-enabled by setting
2025 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
2026 run-time. This is really not recommended unless you know what
2030 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
2032 *) Don't set val to NULL when freeing up structures, it is freed up by
2033 underlying code. If sizeof(void *) > sizeof(long) this can result in
2034 zeroing past the valid field. (CVE-2009-0789)
2035 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
2037 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
2038 checked correctly. This would allow some invalid signed attributes to
2039 appear to verify correctly. (CVE-2009-0591)
2040 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2042 *) Reject UniversalString and BMPString types with invalid lengths. This
2043 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
2044 a legal length. (CVE-2009-0590)
2047 *) Set S/MIME signing as the default purpose rather than setting it
2048 unconditionally. This allows applications to override it at the store
2052 *) Permit restricted recursion of ASN1 strings. This is needed in practice
2053 to handle some structures.
2056 *) Improve efficiency of mem_gets: don't search whole buffer each time
2058 [Jeremy Shapiro <jnshapir@us.ibm.com>]
2060 *) New -hex option for openssl rand.
2063 *) Print out UTF8String and NumericString when parsing ASN1.
2066 *) Support NumericString type for name components.
2069 *) Allow CC in the environment to override the automatically chosen
2070 compiler. Note that nothing is done to ensure flags work with the
2074 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
2076 *) Properly check EVP_VerifyFinal() and similar return values
2078 [Ben Laurie, Bodo Moeller, Google Security Team]
2080 *) Enable TLS extensions by default.
2083 *) Allow the CHIL engine to be loaded, whether the application is
2084 multithreaded or not. (This does not release the developer from the
2085 obligation to set up the dynamic locking callbacks.)
2086 [Sander Temme <sander@temme.net>]
2088 *) Use correct exit code if there is an error in dgst command.
2089 [Steve Henson; problem pointed out by Roland Dirlewanger]
2091 *) Tweak Configure so that you need to say "experimental-jpake" to enable
2092 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
2095 *) Add experimental JPAKE support, including demo authentication in
2096 s_client and s_server.
2099 *) Set the comparison function in v3_addr_canonize().
2100 [Rob Austein <sra@hactrn.net>]
2102 *) Add support for XMPP STARTTLS in s_client.
2103 [Philip Paeps <philip@freebsd.org>]
2105 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
2106 to ensure that even with this option, only ciphersuites in the
2107 server's preference list will be accepted. (Note that the option
2108 applies only when resuming a session, so the earlier behavior was
2109 just about the algorithm choice for symmetric cryptography.)
2112 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
2114 *) Fix NULL pointer dereference if a DTLS server received
2115 ChangeCipherSpec as first record (CVE-2009-1386).
2118 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
2119 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
2122 *) The fix in 0.9.8c that supposedly got rid of unsafe
2123 double-checked locking was incomplete for RSA blinding,
2124 addressing just one layer of what turns out to have been
2125 doubly unsafe triple-checked locking.
2127 So now fix this for real by retiring the MONT_HELPER macro
2128 in crypto/rsa/rsa_eay.c.
2130 [Bodo Moeller; problem pointed out by Marius Schilder]
2132 *) Various precautionary measures:
2134 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
2136 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
2137 (NB: This would require knowledge of the secret session ticket key
2138 to exploit, in which case you'd be SOL either way.)
2140 - Change bn_nist.c so that it will properly handle input BIGNUMs
2141 outside the expected range.
2143 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
2146 [Neel Mehta, Bodo Moeller]
2148 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
2149 the load fails. Useful for distros.
2150 [Ben Laurie and the FreeBSD team]
2152 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
2155 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
2158 *) Expand ENGINE to support engine supplied SSL client certificate functions.
2160 This work was sponsored by Logica.
2163 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
2164 keystores. Support for SSL/TLS client authentication too.
2165 Not compiled unless enable-capieng specified to Configure.
2167 This work was sponsored by Logica.
2170 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
2171 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
2172 attribute creation routines such as certifcate requests and PKCS#12
2176 Changes between 0.9.8g and 0.9.8h [28 May 2008]
2178 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
2179 handshake which could lead to a cilent crash as found using the
2180 Codenomicon TLS test suite (CVE-2008-1672)
2181 [Steve Henson, Mark Cox]
2183 *) Fix double free in TLS server name extensions which could lead to
2184 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
2187 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
2189 Clear the error queue to ensure that error entries left from
2190 older function calls do not interfere with the correct operation.
2191 [Lutz Jaenicke, Erik de Castro Lopo]
2193 *) Remove root CA certificates of commercial CAs:
2195 The OpenSSL project does not recommend any specific CA and does not
2196 have any policy with respect to including or excluding any CA.
2197 Therefore it does not make any sense to ship an arbitrary selection
2198 of root CA certificates with the OpenSSL software.
2201 *) RSA OAEP patches to fix two separate invalid memory reads.
2202 The first one involves inputs when 'lzero' is greater than
2203 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
2204 before the beginning of from). The second one involves inputs where
2205 the 'db' section contains nothing but zeroes (there is a one-byte
2206 invalid read after the end of 'db').
2207 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2209 *) Partial backport from 0.9.9-dev:
2211 Introduce bn_mul_mont (dedicated Montgomery multiplication
2212 procedure) as a candidate for BIGNUM assembler implementation.
2213 While 0.9.9-dev uses assembler for various architectures, only
2214 x86_64 is available by default here in the 0.9.8 branch, and
2215 32-bit x86 is available through a compile-time setting.
2217 To try the 32-bit x86 assembler implementation, use Configure
2218 option "enable-montasm" (which exists only for this backport).
2220 As "enable-montasm" for 32-bit x86 disclaims code stability
2221 anyway, in this constellation we activate additional code
2222 backported from 0.9.9-dev for further performance improvements,
2223 namely BN_from_montgomery_word. (To enable this otherwise,
2224 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
2226 [Andy Polyakov (backport partially by Bodo Moeller)]
2228 *) Add TLS session ticket callback. This allows an application to set
2229 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
2230 values. This is useful for key rollover for example where several key
2231 sets may exist with different names.
2234 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
2235 This was broken until now in 0.9.8 releases, such that the only way
2236 a registered ENGINE could be used (assuming it initialises
2237 successfully on the host) was to explicitly set it as the default
2238 for the relevant algorithms. This is in contradiction with 0.9.7
2239 behaviour and the documentation. With this fix, when an ENGINE is
2240 registered into a given algorithm's table of implementations, the
2241 'uptodate' flag is reset so that auto-discovery will be used next
2242 time a new context for that algorithm attempts to select an
2244 [Ian Lister (tweaked by Geoff Thorpe)]
2246 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
2247 implemention in the following ways:
2249 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
2252 Lack of BER streaming support means one pass streaming processing is
2253 only supported if data is detached: setting the streaming flag is
2254 ignored for embedded content.
2256 CMS support is disabled by default and must be explicitly enabled
2257 with the enable-cms configuration option.
2260 *) Update the GMP engine glue to do direct copies between BIGNUM and
2261 mpz_t when openssl and GMP use the same limb size. Otherwise the
2262 existing "conversion via a text string export" trick is still used.
2263 [Paul Sheer <paulsheer@gmail.com>]
2265 *) Zlib compression BIO. This is a filter BIO which compressed and
2266 uncompresses any data passed through it.
2269 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
2270 RFC3394 compatible AES key wrapping.
2273 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
2274 sets string data without copying. X509_ALGOR_set0() and
2275 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
2276 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
2277 from an X509_ATTRIBUTE structure optionally checking it occurs only
2278 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
2282 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
2283 to get the expected BN_FLG_CONSTTIME behavior.
2284 [Bodo Moeller (Google)]
2288 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
2289 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
2290 - added some more tests to do_tests.pl
2291 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
2292 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
2293 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
2294 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
2295 - various changes to netware.pl to enable gcc-cross builds on Win32
2297 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
2298 - various changes to fix missing prototype warnings
2299 - fixed x86nasm.pl to create correct asm files for NASM COFF output
2300 - added AES, WHIRLPOOL and CPUID assembler code to build files
2301 - added missing AES assembler make rules to mk1mf.pl
2302 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
2303 [Guenter Knauf <eflash@gmx.net>]
2305 *) Implement certificate status request TLS extension defined in RFC3546.
2306 A client can set the appropriate parameters and receive the encoded
2307 OCSP response via a callback. A server can query the supplied parameters
2308 and set the encoded OCSP response in the callback. Add simplified examples
2309 to s_client and s_server.
2312 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
2314 *) Fix various bugs:
2315 + Binary incompatibility of ssl_ctx_st structure
2316 + DTLS interoperation with non-compliant servers
2317 + Don't call get_session_cb() without proposed session
2318 + Fix ia64 assembler code
2319 [Andy Polyakov, Steve Henson]
2321 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
2323 *) DTLS Handshake overhaul. There were longstanding issues with
2324 OpenSSL DTLS implementation, which were making it impossible for
2325 RFC 4347 compliant client to communicate with OpenSSL server.
2326 Unfortunately just fixing these incompatibilities would "cut off"
2327 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
2328 server keeps tolerating non RFC compliant syntax. The opposite is
2329 not true, 0.9.8f client can not communicate with earlier server.
2330 This update even addresses CVE-2007-4995.
2333 *) Changes to avoid need for function casts in OpenSSL: some compilers
2334 (gcc 4.2 and later) reject their use.
2335 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
2338 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2339 RFC4507bis. The encrypted ticket format is an encrypted encoded
2340 SSL_SESSION structure, that way new session features are automatically
2343 If a client application caches session in an SSL_SESSION structure
2344 support is transparent because tickets are now stored in the encoded
2347 The SSL_CTX structure automatically generates keys for ticket
2348 protection in servers so again support should be possible
2349 with no application modification.
2351 If a client or server wishes to disable RFC4507 support then the option
2352 SSL_OP_NO_TICKET can be set.
2354 Add a TLS extension debugging callback to allow the contents of any client
2355 or server extensions to be examined.
2357 This work was sponsored by Google.
2360 *) Add initial support for TLS extensions, specifically for the server_name
2361 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2362 have new members for a host name. The SSL data structure has an
2363 additional member SSL_CTX *initial_ctx so that new sessions can be
2364 stored in that context to allow for session resumption, even after the
2365 SSL has been switched to a new SSL_CTX in reaction to a client's
2366 server_name extension.
2368 New functions (subject to change):
2370 SSL_get_servername()
2371 SSL_get_servername_type()
2374 New CTRL codes and macros (subject to change):
2376 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2377 - SSL_CTX_set_tlsext_servername_callback()
2378 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2379 - SSL_CTX_set_tlsext_servername_arg()
2380 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2382 openssl s_client has a new '-servername ...' option.
2384 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2385 '-key2 ...', '-servername_fatal' (subject to change). This allows
2386 testing the HostName extension for a specific single host name ('-cert'
2387 and '-key' remain fallbacks for handshakes without HostName
2388 negotiation). If the unrecogninzed_name alert has to be sent, this by
2389 default is a warning; it becomes fatal with the '-servername_fatal'
2392 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
2394 *) Add AES and SSE2 assembly language support to VC++ build.
2397 *) Mitigate attack on final subtraction in Montgomery reduction.
2400 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
2401 (which previously caused an internal error).
2404 *) Squeeze another 10% out of IGE mode when in != out.
2407 *) AES IGE mode speedup.
2408 [Dean Gaudet (Google)]
2410 *) Add the Korean symmetric 128-bit cipher SEED (see
2411 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
2412 add SEED ciphersuites from RFC 4162:
2414 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
2415 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
2416 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
2417 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
2419 To minimize changes between patchlevels in the OpenSSL 0.9.8
2420 series, SEED remains excluded from compilation unless OpenSSL
2421 is configured with 'enable-seed'.
2422 [KISA, Bodo Moeller]
2424 *) Mitigate branch prediction attacks, which can be practical if a
2425 single processor is shared, allowing a spy process to extract
2426 information. For detailed background information, see
2427 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
2428 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
2429 and Necessary Software Countermeasures"). The core of the change
2430 are new versions BN_div_no_branch() and
2431 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
2432 respectively, which are slower, but avoid the security-relevant
2433 conditional branches. These are automatically called by BN_div()
2434 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
2435 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
2436 remove a conditional branch.
2438 BN_FLG_CONSTTIME is the new name for the previous
2439 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
2440 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
2441 in the exponent causes BN_mod_exp_mont() to use the alternative
2442 implementation in BN_mod_exp_mont_consttime().) The old name
2443 remains as a deprecated alias.
2445 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
2446 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
2447 constant-time implementations for more than just exponentiation.
2448 Here too the old name is kept as a deprecated alias.
2450 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2451 the BN_BLINDING structure gets an independent copy of the
2452 modulus. This means that the previous "BIGNUM *m" argument to
2453 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2454 essentially becomes "const BIGNUM *m", although we can't actually
2455 change this in the header file before 0.9.9. It allows
2456 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2457 enable BN_FLG_CONSTTIME.
2459 [Matthew D Wood (Intel Corp)]
2461 *) In the SSL/TLS server implementation, be strict about session ID
2462 context matching (which matters if an application uses a single
2463 external cache for different purposes). Previously,
2464 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2465 set. This did ensure strict client verification, but meant that,
2466 with applications using a single external cache for quite
2467 different requirements, clients could circumvent ciphersuite
2468 restrictions for a given session ID context by starting a session
2469 in a different context.
2472 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2473 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2474 authentication-only ciphersuites.
2477 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
2478 not complete and could lead to a possible single byte overflow
2479 (CVE-2007-5135) [Ben Laurie]
2481 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
2483 *) Since AES128 and AES256 (and similarly Camellia128 and
2484 Camellia256) share a single mask bit in the logic of
2485 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2486 kludge to work properly if AES128 is available and AES256 isn't
2487 (or if Camellia128 is available and Camellia256 isn't).
2490 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
2491 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
2492 When a point or a seed is encoded in a BIT STRING, we need to
2493 prevent the removal of trailing zero bits to get the proper DER
2494 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
2495 of a NamedBitList, for which trailing 0 bits need to be removed.)
2498 *) Have SSL/TLS server implementation tolerate "mismatched" record
2499 protocol version while receiving ClientHello even if the
2500 ClientHello is fragmented. (The server can't insist on the
2501 particular protocol version it has chosen before the ServerHello
2502 message has informed the client about his choice.)
2505 *) Add RFC 3779 support.
2506 [Rob Austein for ARIN, Ben Laurie]
2508 *) Load error codes if they are not already present instead of using a
2509 static variable. This allows them to be cleanly unloaded and reloaded.
2510 Improve header file function name parsing.
2513 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
2514 or CAPABILITY handshake as required by RFCs.
2517 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
2519 *) Introduce limits to prevent malicious keys being able to
2520 cause a denial of service. (CVE-2006-2940)
2521 [Steve Henson, Bodo Moeller]
2523 *) Fix ASN.1 parsing of certain invalid structures that can result
2524 in a denial of service. (CVE-2006-2937) [Steve Henson]
2526 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2527 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2529 *) Fix SSL client code which could crash if connecting to a
2530 malicious SSLv2 server. (CVE-2006-4343)
2531 [Tavis Ormandy and Will Drewry, Google Security Team]
2533 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
2534 match only those. Before that, "AES256-SHA" would be interpreted
2535 as a pattern and match "AES128-SHA" too (since AES128-SHA got
2536 the same strength classification in 0.9.7h) as we currently only
2537 have a single AES bit in the ciphersuite description bitmap.
2538 That change, however, also applied to ciphersuite strings such as
2539 "RC4-MD5" that intentionally matched multiple ciphersuites --
2540 namely, SSL 2.0 ciphersuites in addition to the more common ones
2541 from SSL 3.0/TLS 1.0.
2543 So we change the selection algorithm again: Naming an explicit
2544 ciphersuite selects this one ciphersuite, and any other similar
2545 ciphersuite (same bitmap) from *other* protocol versions.
2546 Thus, "RC4-MD5" again will properly select both the SSL 2.0
2547 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
2549 Since SSL 2.0 does not have any ciphersuites for which the
2550 128/256 bit distinction would be relevant, this works for now.
2551 The proper fix will be to use different bits for AES128 and
2552 AES256, which would have avoided the problems from the beginning;
2553 however, bits are scarce, so we can only do this in a new release
2554 (not just a patchlevel) when we can change the SSL_CIPHER
2555 definition to split the single 'unsigned long mask' bitmap into
2556 multiple values to extend the available space.
2560 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2562 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2563 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2565 *) Add AES IGE and biIGE modes.
2568 *) Change the Unix randomness entropy gathering to use poll() when
2569 possible instead of select(), since the latter has some
2570 undesirable limitations.
2571 [Darryl Miles via Richard Levitte and Bodo Moeller]
2573 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2574 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2575 cannot be implicitly activated as part of, e.g., the "AES" alias.
2576 However, please upgrade to OpenSSL 0.9.9[-dev] for
2577 non-experimental use of the ECC ciphersuites to get TLS extension
2578 support, which is required for curve and point format negotiation
2579 to avoid potential handshake problems.
2582 *) Disable rogue ciphersuites:
2584 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2585 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2586 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2588 The latter two were purportedly from
2589 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2592 Also deactivate the remaining ciphersuites from
2593 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2594 unofficial, and the ID has long expired.
2597 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2598 dual-core machines) and other potential thread-safety issues.
2601 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
2602 versions), which is now available for royalty-free use
2603 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
2604 Also, add Camellia TLS ciphersuites from RFC 4132.
2606 To minimize changes between patchlevels in the OpenSSL 0.9.8
2607 series, Camellia remains excluded from compilation unless OpenSSL
2608 is configured with 'enable-camellia'.
2611 *) Disable the padding bug check when compression is in use. The padding
2612 bug check assumes the first packet is of even length, this is not
2613 necessarily true if compresssion is enabled and can result in false
2614 positives causing handshake failure. The actual bug test is ancient
2615 code so it is hoped that implementations will either have fixed it by
2616 now or any which still have the bug do not support compression.
2619 Changes between 0.9.8a and 0.9.8b [04 May 2006]
2621 *) When applying a cipher rule check to see if string match is an explicit
2622 cipher suite and only match that one cipher suite if it is.
2625 *) Link in manifests for VC++ if needed.
2626 [Austin Ziegler <halostatue@gmail.com>]
2628 *) Update support for ECC-based TLS ciphersuites according to
2629 draft-ietf-tls-ecc-12.txt with proposed changes (but without
2630 TLS extensions, which are supported starting with the 0.9.9
2631 branch, not in the OpenSSL 0.9.8 branch).
2634 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
2635 opaque EVP_CIPHER_CTX handling.
2638 *) Fixes and enhancements to zlib compression code. We now only use
2639 "zlib1.dll" and use the default __cdecl calling convention on Win32
2640 to conform with the standards mentioned here:
2641 http://www.zlib.net/DLL_FAQ.txt
2642 Static zlib linking now works on Windows and the new --with-zlib-include
2643 --with-zlib-lib options to Configure can be used to supply the location
2644 of the headers and library. Gracefully handle case where zlib library
2648 *) Several fixes and enhancements to the OID generation code. The old code
2649 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
2650 handle numbers larger than ULONG_MAX, truncated printing and had a
2651 non standard OBJ_obj2txt() behaviour.
2654 *) Add support for building of engines under engine/ as shared libraries
2655 under VC++ build system.
2658 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
2659 Hopefully, we will not see any false combination of paths any more.
2662 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
2664 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2665 (part of SSL_OP_ALL). This option used to disable the
2666 countermeasure against man-in-the-middle protocol-version
2667 rollback in the SSL 2.0 server implementation, which is a bad
2668 idea. (CVE-2005-2969)
2670 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2671 for Information Security, National Institute of Advanced Industrial
2672 Science and Technology [AIST], Japan)]
2674 *) Add two function to clear and return the verify parameter flags.
2677 *) Keep cipherlists sorted in the source instead of sorting them at
2678 runtime, thus removing the need for a lock.
2681 *) Avoid some small subgroup attacks in Diffie-Hellman.
2682 [Nick Mathewson and Ben Laurie]
2684 *) Add functions for well-known primes.
2687 *) Extended Windows CE support.
2688 [Satoshi Nakamura and Andy Polyakov]
2690 *) Initialize SSL_METHOD structures at compile time instead of during
2691 runtime, thus removing the need for a lock.
2694 *) Make PKCS7_decrypt() work even if no certificate is supplied by
2695 attempting to decrypt each encrypted key in turn. Add support to
2699 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
2701 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
2704 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
2707 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
2708 key into the same file any more.
2711 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
2714 *) Add -utf8 command line and config file option to 'ca'.
2715 [Stefan <stf@udoma.org]
2717 *) Removed the macro des_crypt(), as it seems to conflict with some
2718 libraries. Use DES_crypt().
2721 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
2722 involves renaming the source and generated shared-libs for
2723 both. The engines will accept the corrected or legacy ids
2724 ('ncipher' and '4758_cca' respectively) when binding. NB,
2725 this only applies when building 'shared'.
2726 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
2728 *) Add attribute functions to EVP_PKEY structure. Modify
2729 PKCS12_create() to recognize a CSP name attribute and
2730 use it. Make -CSP option work again in pkcs12 utility.
2733 *) Add new functionality to the bn blinding code:
2734 - automatic re-creation of the BN_BLINDING parameters after
2735 a fixed number of uses (currently 32)
2736 - add new function for parameter creation
2737 - introduce flags to control the update behaviour of the
2738 BN_BLINDING parameters
2739 - hide BN_BLINDING structure
2740 Add a second BN_BLINDING slot to the RSA structure to improve
2741 performance when a single RSA object is shared among several
2745 *) Add support for DTLS.
2746 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
2748 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
2749 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
2752 *) Remove buggy and incompletet DH cert support from
2753 ssl/ssl_rsa.c and ssl/s3_both.c
2756 *) Use SHA-1 instead of MD5 as the default digest algorithm for
2757 the apps/openssl applications.
2760 *) Compile clean with "-Wall -Wmissing-prototypes
2761 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
2762 DEBUG_SAFESTACK must also be set.
2765 *) Change ./Configure so that certain algorithms can be disabled by default.
2766 The new counterpiece to "no-xxx" is "enable-xxx".
2768 The patented RC5 and MDC2 algorithms will now be disabled unless
2769 "enable-rc5" and "enable-mdc2", respectively, are specified.
2771 (IDEA remains enabled despite being patented. This is because IDEA
2772 is frequently required for interoperability, and there is no license
2773 fee for non-commercial use. As before, "no-idea" can be used to
2774 avoid this algorithm.)
2778 *) Add processing of proxy certificates (see RFC 3820). This work was
2779 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
2780 EGEE (Enabling Grids for E-science in Europe).
2783 *) RC4 performance overhaul on modern architectures/implementations, such
2784 as Intel P4, IA-64 and AMD64.
2787 *) New utility extract-section.pl. This can be used specify an alternative
2788 section number in a pod file instead of having to treat each file as
2789 a separate case in Makefile. This can be done by adding two lines to the
2792 =for comment openssl_section:XXX
2794 The blank line is mandatory.
2798 *) New arguments -certform, -keyform and -pass for s_client and s_server
2799 to allow alternative format key and certificate files and passphrase
2803 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
2804 update associated structures and add various utility functions.
2806 Add new policy related verify parameters, include policy checking in
2807 standard verify code. Enhance 'smime' application with extra parameters
2808 to support policy checking and print out.
2811 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
2812 Nehemiah processors. These extensions support AES encryption in hardware
2813 as well as RNG (though RNG support is currently disabled).
2814 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
2816 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
2819 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
2820 [Andy Polyakov and a number of other people]
2822 *) Improved PowerPC platform support. Most notably BIGNUM assembler
2823 implementation contributed by IBM.
2824 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
2826 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
2827 exponent rather than 'unsigned long'. There is a corresponding change to
2828 the new 'rsa_keygen' element of the RSA_METHOD structure.
2829 [Jelte Jansen, Geoff Thorpe]
2831 *) Functionality for creating the initial serial number file is now
2832 moved from CA.pl to the 'ca' utility with a new option -create_serial.
2834 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
2835 number file to 1, which is bound to cause problems. To avoid
2836 the problems while respecting compatibility between different 0.9.7
2837 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
2838 CA.pl for serial number initialization. With the new release 0.9.8,
2839 we can fix the problem directly in the 'ca' utility.)
2842 *) Reduced header interdepencies by declaring more opaque objects in
2843 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
2844 give fewer recursive includes, which could break lazy source code - so
2845 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
2846 developers should define this symbol when building and using openssl to
2847 ensure they track the recommended behaviour, interfaces, [etc], but
2848 backwards-compatible behaviour prevails when this isn't defined.
2851 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
2854 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
2855 This will generate a random key of the appropriate length based on the
2856 cipher context. The EVP_CIPHER can provide its own random key generation
2857 routine to support keys of a specific form. This is used in the des and
2858 3des routines to generate a key of the correct parity. Update S/MIME
2859 code to use new functions and hence generate correct parity DES keys.
2860 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
2861 valid (weak or incorrect parity).
2864 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
2865 as looking them up. This is useful when the verified structure may contain
2866 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
2867 present unless the new PKCS7_NO_CRL flag is asserted.
2870 *) Extend ASN1 oid configuration module. It now additionally accepts the
2873 shortName = some long name, 1.2.3.4
2876 *) Reimplemented the BN_CTX implementation. There is now no more static
2877 limitation on the number of variables it can handle nor the depth of the
2878 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
2879 information can now expand as required, and rather than having a single
2880 static array of bignums, BN_CTX now uses a linked-list of such arrays
2881 allowing it to expand on demand whilst maintaining the usefulness of
2882 BN_CTX's "bundling".
2885 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
2886 to allow all RSA operations to function using a single BN_CTX.
2889 *) Preliminary support for certificate policy evaluation and checking. This
2890 is initially intended to pass the tests outlined in "Conformance Testing
2891 of Relying Party Client Certificate Path Processing Logic" v1.07.
2894 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
2895 remained unused and not that useful. A variety of other little bignum
2896 tweaks and fixes have also been made continuing on from the audit (see
2900 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
2901 associated ASN1, EVP and SSL functions and old ASN1 macros.
2904 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
2905 and this should never fail. So the return value from the use of
2906 BN_set_word() (which can fail due to needless expansion) is now deprecated;
2907 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
2910 *) BN_CTX_get() should return zero-valued bignums, providing the same
2911 initialised value as BN_new().
2912 [Geoff Thorpe, suggested by Ulf Möller]
2914 *) Support for inhibitAnyPolicy certificate extension.
2917 *) An audit of the BIGNUM code is underway, for which debugging code is
2918 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
2919 is considered valid when processing BIGNUMs, and causes execution to
2920 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
2921 further steps are taken to deliberately pollute unused data in BIGNUM
2922 structures to try and expose faulty code further on. For now, openssl will
2923 (in its default mode of operation) continue to tolerate the inconsistent
2924 forms that it has tolerated in the past, but authors and packagers should
2925 consider trying openssl and their own applications when compiled with
2926 these debugging symbols defined. It will help highlight potential bugs in
2927 their own code, and will improve the test coverage for OpenSSL itself. At
2928 some point, these tighter rules will become openssl's default to improve
2929 maintainability, though the assert()s and other overheads will remain only
2930 in debugging configurations. See bn.h for more details.
2931 [Geoff Thorpe, Nils Larsch, Ulf Möller]
2933 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
2934 that can only be obtained through BN_CTX_new() (which implicitly
2935 initialises it). The presence of this function only made it possible
2936 to overwrite an existing structure (and cause memory leaks).
2939 *) Because of the callback-based approach for implementing LHASH as a
2940 template type, lh_insert() adds opaque objects to hash-tables and
2941 lh_doall() or lh_doall_arg() are typically used with a destructor callback
2942 to clean up those corresponding objects before destroying the hash table
2943 (and losing the object pointers). So some over-zealous constifications in
2944 LHASH have been relaxed so that lh_insert() does not take (nor store) the
2945 objects as "const" and the lh_doall[_arg] callback wrappers are not
2946 prototyped to have "const" restrictions on the object pointers they are
2947 given (and so aren't required to cast them away any more).
2950 *) The tmdiff.h API was so ugly and minimal that our own timing utility
2951 (speed) prefers to use its own implementation. The two implementations
2952 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
2953 its object type properly exposed (MS_TM) instead of casting to/from "char
2954 *". This may still change yet if someone realises MS_TM and "ms_time_***"
2955 aren't necessarily the greatest nomenclatures - but this is what was used
2956 internally to the implementation so I've used that for now.
2959 *) Ensure that deprecated functions do not get compiled when
2960 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
2961 the self-tests were still using deprecated key-generation functions so
2962 these have been updated also.
2965 *) Reorganise PKCS#7 code to separate the digest location functionality
2966 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
2967 New function PKCS7_set_digest() to set the digest type for PKCS#7
2968 digestedData type. Add additional code to correctly generate the
2969 digestedData type and add support for this type in PKCS7 initialization
2973 *) New function PKCS7_set0_type_other() this initializes a PKCS7
2974 structure of type "other".
2977 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
2978 sure the loop does correctly stop and breaking ("division by zero")
2979 modulus operations are not performed. The (pre-generated) prime
2980 table crypto/bn/bn_prime.h was already correct, but it could not be
2981 re-generated on some platforms because of the "division by zero"
2982 situation in the script.
2983 [Ralf S. Engelschall]
2985 *) Update support for ECC-based TLS ciphersuites according to
2986 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
2987 SHA-1 now is only used for "small" curves (where the
2988 representation of a field element takes up to 24 bytes); for
2989 larger curves, the field element resulting from ECDH is directly
2990 used as premaster secret.
2991 [Douglas Stebila (Sun Microsystems Laboratories)]
2993 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
2994 curve secp160r1 to the tests.
2995 [Douglas Stebila (Sun Microsystems Laboratories)]
2997 *) Add the possibility to load symbols globally with DSO.
2998 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
3000 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
3001 control of the error stack.
3004 *) Add support for STORE in ENGINE.
3007 *) Add the STORE type. The intention is to provide a common interface
3008 to certificate and key stores, be they simple file-based stores, or
3009 HSM-type store, or LDAP stores, or...
3010 NOTE: The code is currently UNTESTED and isn't really used anywhere.
3013 *) Add a generic structure called OPENSSL_ITEM. This can be used to
3014 pass a list of arguments to any function as well as provide a way
3015 for a function to pass data back to the caller.
3018 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
3019 works like BUF_strdup() but can be used to duplicate a portion of
3020 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
3024 *) Add the function sk_find_ex() which works like sk_find(), but will
3025 return an index to an element even if an exact match couldn't be
3026 found. The index is guaranteed to point at the element where the
3027 searched-for key would be inserted to preserve sorting order.
3030 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
3031 takes an extra flags argument for optional functionality. Currently,
3032 the following flags are defined:
3034 OBJ_BSEARCH_VALUE_ON_NOMATCH
3035 This one gets OBJ_bsearch_ex() to return a pointer to the first
3036 element where the comparing function returns a negative or zero
3039 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
3040 This one gets OBJ_bsearch_ex() to return a pointer to the first
3041 element where the comparing function returns zero. This is useful
3042 if there are more than one element where the comparing function
3046 *) Make it possible to create self-signed certificates with 'openssl ca'
3047 in such a way that the self-signed certificate becomes part of the
3048 CA database and uses the same mechanisms for serial number generation
3049 as all other certificate signing. The new flag '-selfsign' enables
3050 this functionality. Adapt CA.sh and CA.pl.in.
3053 *) Add functionality to check the public key of a certificate request
3054 against a given private. This is useful to check that a certificate
3055 request can be signed by that key (self-signing).
3058 *) Make it possible to have multiple active certificates with the same
3059 subject in the CA index file. This is done only if the keyword
3060 'unique_subject' is set to 'no' in the main CA section (default
3061 if 'CA_default') of the configuration file. The value is saved
3062 with the database itself in a separate index attribute file,
3063 named like the index file with '.attr' appended to the name.
3066 *) Generate muti valued AVAs using '+' notation in config files for
3070 *) Support for nameConstraints certificate extension.
3073 *) Support for policyConstraints certificate extension.
3076 *) Support for policyMappings certificate extension.
3079 *) Make sure the default DSA_METHOD implementation only uses its
3080 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
3081 and change its own handlers to be NULL so as to remove unnecessary
3082 indirection. This lets alternative implementations fallback to the
3083 default implementation more easily.
3086 *) Support for directoryName in GeneralName related extensions
3090 *) Make it possible to link applications using Makefile.shared.
3091 Make that possible even when linking against static libraries!
3094 *) Support for single pass processing for S/MIME signing. This now
3095 means that S/MIME signing can be done from a pipe, in addition
3096 cleartext signing (multipart/signed type) is effectively streaming
3097 and the signed data does not need to be all held in memory.
3099 This is done with a new flag PKCS7_STREAM. When this flag is set
3100 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
3101 is done after the data is output (and digests calculated) in
3102 SMIME_write_PKCS7().
3105 *) Add full support for -rpath/-R, both in shared libraries and
3106 applications, at least on the platforms where it's known how
3110 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
3111 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
3112 will now compute a table of multiples of the generator that
3113 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
3114 faster (notably in the case of a single point multiplication,
3115 scalar * generator).
3116 [Nils Larsch, Bodo Moeller]
3118 *) IPv6 support for certificate extensions. The various extensions
3119 which use the IP:a.b.c.d can now take IPv6 addresses using the
3120 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
3124 *) Added an ENGINE that implements RSA by performing private key
3125 exponentiations with the GMP library. The conversions to and from
3126 GMP's mpz_t format aren't optimised nor are any montgomery forms
3127 cached, and on x86 it appears OpenSSL's own performance has caught up.
3128 However there are likely to be other architectures where GMP could
3129 provide a boost. This ENGINE is not built in by default, but it can be
3130 specified at Configure time and should be accompanied by the necessary
3131 linker additions, eg;
3132 ./config -DOPENSSL_USE_GMP -lgmp
3135 *) "openssl engine" will not display ENGINE/DSO load failure errors when
3136 testing availability of engines with "-t" - the old behaviour is
3137 produced by increasing the feature's verbosity with "-tt".
3140 *) ECDSA routines: under certain error conditions uninitialized BN objects
3141 could be freed. Solution: make sure initialization is performed early
3142 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
3146 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
3147 and DH_METHOD (eg. by ENGINE implementations) to override the normal
3148 software implementations. For DSA and DH, parameter generation can
3149 also be overriden by providing the appropriate method callbacks.
3152 *) Change the "progress" mechanism used in key-generation and
3153 primality testing to functions that take a new BN_GENCB pointer in
3154 place of callback/argument pairs. The new API functions have "_ex"
3155 postfixes and the older functions are reimplemented as wrappers for
3156 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
3157 declarations of the old functions to help (graceful) attempts to
3158 migrate to the new functions. Also, the new key-generation API
3159 functions operate on a caller-supplied key-structure and return
3160 success/failure rather than returning a key or NULL - this is to
3161 help make "keygen" another member function of RSA_METHOD etc.
3163 Example for using the new callback interface:
3165 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
3169 BN_GENCB_set(&my_cb, my_callback, my_arg);
3171 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
3172 /* For the meaning of a, b in calls to my_callback(), see the
3173 * documentation of the function that calls the callback.
3174 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
3175 * my_callback should return 1 if it wants BN_is_prime_ex()
3176 * to continue, or 0 to stop.
3181 *) Change the ZLIB compression method to be stateful, and make it
3182 available to TLS with the number defined in
3183 draft-ietf-tls-compression-04.txt.
3186 *) Add the ASN.1 structures and functions for CertificatePair, which
3187 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
3189 CertificatePair ::= SEQUENCE {
3190 forward [0] Certificate OPTIONAL,
3191 reverse [1] Certificate OPTIONAL,
3192 -- at least one of the pair shall be present -- }
3194 Also implement the PEM functions to read and write certificate
3195 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
3197 This needed to be defined, mostly for the sake of the LDAP
3198 attribute crossCertificatePair, but may prove useful elsewhere as
3202 *) Make it possible to inhibit symlinking of shared libraries in
3203 Makefile.shared, for Cygwin's sake.
3206 *) Extend the BIGNUM API by creating a function
3207 void BN_set_negative(BIGNUM *a, int neg);
3208 and a macro that behave like
3209 int BN_is_negative(const BIGNUM *a);
3211 to avoid the need to access 'a->neg' directly in applications.
3214 *) Implement fast modular reduction for pseudo-Mersenne primes
3215 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
3216 EC_GROUP_new_curve_GFp() will now automatically use this
3218 [Nils Larsch <nla@trustcenter.de>]
3220 *) Add new lock type (CRYPTO_LOCK_BN).
3223 *) Change the ENGINE framework to automatically load engines
3224 dynamically from specific directories unless they could be
3225 found to already be built in or loaded. Move all the
3226 current engines except for the cryptodev one to a new
3228 The engines in engines/ are built as shared libraries if
3229 the "shared" options was given to ./Configure or ./config.
3230 Otherwise, they are inserted in libcrypto.a.
3231 /usr/local/ssl/engines is the default directory for dynamic
3232 engines, but that can be overriden at configure time through
3233 the usual use of --prefix and/or --openssldir, and at run
3234 time with the environment variable OPENSSL_ENGINES.
3235 [Geoff Thorpe and Richard Levitte]
3237 *) Add Makefile.shared, a helper makefile to build shared
3238 libraries. Addapt Makefile.org.
3241 *) Add version info to Win32 DLLs.
3242 [Peter 'Luna' Runestig" <peter@runestig.com>]
3244 *) Add new 'medium level' PKCS#12 API. Certificates and keys
3245 can be added using this API to created arbitrary PKCS#12
3246 files while avoiding the low level API.
3248 New options to PKCS12_create(), key or cert can be NULL and
3249 will then be omitted from the output file. The encryption
3250 algorithm NIDs can be set to -1 for no encryption, the mac
3251 iteration count can be set to 0 to omit the mac.
3253 Enhance pkcs12 utility by making the -nokeys and -nocerts
3254 options work when creating a PKCS#12 file. New option -nomac
3255 to omit the mac, NONE can be set for an encryption algorithm.
3256 New code is modified to use the enhanced PKCS12_create()
3257 instead of the low level API.
3260 *) Extend ASN1 encoder to support indefinite length constructed
3261 encoding. This can output sequences tags and octet strings in
3262 this form. Modify pk7_asn1.c to support indefinite length
3263 encoding. This is experimental and needs additional code to
3264 be useful, such as an ASN1 bio and some enhanced streaming
3267 Extend template encode functionality so that tagging is passed
3268 down to the template encoder.
3271 *) Let 'openssl req' fail if an argument to '-newkey' is not
3272 recognized instead of using RSA as a default.
3275 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
3276 As these are not official, they are not included in "ALL";
3277 the "ECCdraft" ciphersuite group alias can be used to select them.
3278 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
3280 *) Add ECDH engine support.
3281 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
3283 *) Add ECDH in new directory crypto/ecdh/.
3284 [Douglas Stebila (Sun Microsystems Laboratories)]
3286 *) Let BN_rand_range() abort with an error after 100 iterations
3287 without success (which indicates a broken PRNG).
3290 *) Change BN_mod_sqrt() so that it verifies that the input value
3291 is really the square of the return value. (Previously,
3292 BN_mod_sqrt would show GIGO behaviour.)
3295 *) Add named elliptic curves over binary fields from X9.62, SECG,
3296 and WAP/WTLS; add OIDs that were still missing.
3298 [Sheueling Chang Shantz and Douglas Stebila
3299 (Sun Microsystems Laboratories)]
3301 *) Extend the EC library for elliptic curves over binary fields
3302 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
3305 EC_GF2m_simple_method
3309 EC_GROUP_new_curve_GF2m
3310 EC_GROUP_set_curve_GF2m
3311 EC_GROUP_get_curve_GF2m
3312 EC_POINT_set_affine_coordinates_GF2m
3313 EC_POINT_get_affine_coordinates_GF2m
3314 EC_POINT_set_compressed_coordinates_GF2m
3316 Point compression for binary fields is disabled by default for
3317 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
3320 As binary polynomials are represented as BIGNUMs, various members
3321 of the EC_GROUP and EC_POINT data structures can be shared
3322 between the implementations for prime fields and binary fields;
3323 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
3324 are essentially identical to their ..._GFp counterparts.
3325 (For simplicity, the '..._GFp' prefix has been dropped from
3326 various internal method names.)
3328 An internal 'field_div' method (similar to 'field_mul' and
3329 'field_sqr') has been added; this is used only for binary fields.
3331 [Sheueling Chang Shantz and Douglas Stebila
3332 (Sun Microsystems Laboratories)]
3334 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
3335 through methods ('mul', 'precompute_mult').
3337 The generic implementations (now internally called 'ec_wNAF_mul'
3338 and 'ec_wNAF_precomputed_mult') remain the default if these
3339 methods are undefined.
3341 [Sheueling Chang Shantz and Douglas Stebila
3342 (Sun Microsystems Laboratories)]
3344 *) New function EC_GROUP_get_degree, which is defined through
3345 EC_METHOD. For curves over prime fields, this returns the bit
3346 length of the modulus.
3348 [Sheueling Chang Shantz and Douglas Stebila
3349 (Sun Microsystems Laboratories)]
3351 *) New functions EC_GROUP_dup, EC_POINT_dup.
3352 (These simply call ..._new and ..._copy).
3354 [Sheueling Chang Shantz and Douglas Stebila
3355 (Sun Microsystems Laboratories)]
3357 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
3358 Polynomials are represented as BIGNUMs (where the sign bit is not
3359 used) in the following functions [macros]:
3362 BN_GF2m_sub [= BN_GF2m_add]
3363 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
3364 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
3365 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
3367 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
3368 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
3369 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
3370 BN_GF2m_cmp [= BN_ucmp]
3372 (Note that only the 'mod' functions are actually for fields GF(2^m).
3373 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
3375 For some functions, an the irreducible polynomial defining a
3376 field can be given as an 'unsigned int[]' with strictly
3377 decreasing elements giving the indices of those bits that are set;
3378 i.e., p[] represents the polynomial
3379 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
3381 p[0] > p[1] > ... > p[k] = 0.
3382 This applies to the following functions:
3387 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
3388 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
3390 BN_GF2m_mod_sqrt_arr
3391 BN_GF2m_mod_solve_quad_arr
3395 Conversion can be performed by the following functions:
3400 bntest.c has additional tests for binary polynomial arithmetic.
3402 Two implementations for BN_GF2m_mod_div() are available.
3403 The default algorithm simply uses BN_GF2m_mod_inv() and
3404 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
3405 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
3406 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
3408 [Sheueling Chang Shantz and Douglas Stebila
3409 (Sun Microsystems Laboratories)]
3411 *) Add new error code 'ERR_R_DISABLED' that can be used when some
3412 functionality is disabled at compile-time.
3413 [Douglas Stebila <douglas.stebila@sun.com>]
3415 *) Change default behaviour of 'openssl asn1parse' so that more
3416 information is visible when viewing, e.g., a certificate:
3418 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
3419 mode the content of non-printable OCTET STRINGs is output in a
3420 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
3421 avoid the appearance of a printable string.
3422 [Nils Larsch <nla@trustcenter.de>]
3424 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
3426 EC_GROUP_set_asn1_flag()
3427 EC_GROUP_get_asn1_flag()
3428 EC_GROUP_set_point_conversion_form()
3429 EC_GROUP_get_point_conversion_form()
3430 These control ASN1 encoding details:
3431 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
3432 has been set to OPENSSL_EC_NAMED_CURVE.
3433 - Points are encoded in uncompressed form by default; options for
3434 asn1_for are as for point2oct, namely
3435 POINT_CONVERSION_COMPRESSED
3436 POINT_CONVERSION_UNCOMPRESSED
3437 POINT_CONVERSION_HYBRID
3439 Also add 'seed' and 'seed_len' members to EC_GROUP with access
3442 EC_GROUP_get0_seed()
3443 EC_GROUP_get_seed_len()
3444 This is used only for ASN1 purposes (so far).
3445 [Nils Larsch <nla@trustcenter.de>]
3447 *) Add 'field_type' member to EC_METHOD, which holds the NID
3448 of the appropriate field type OID. The new function
3449 EC_METHOD_get_field_type() returns this value.
3450 [Nils Larsch <nla@trustcenter.de>]
3455 EC_POINT_point2hex()
3456 EC_POINT_hex2point()
3457 providing useful interfaces to EC_POINT_point2oct() and
3458 EC_POINT_oct2point().
3459 [Nils Larsch <nla@trustcenter.de>]
3461 *) Change internals of the EC library so that the functions
3462 EC_GROUP_set_generator()
3463 EC_GROUP_get_generator()
3464 EC_GROUP_get_order()
3465 EC_GROUP_get_cofactor()
3466 are implemented directly in crypto/ec/ec_lib.c and not dispatched
3467 to methods, which would lead to unnecessary code duplication when
3468 adding different types of curves.
3469 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
3471 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
3472 arithmetic, and such that modified wNAFs are generated
3473 (which avoid length expansion in many cases).
3476 *) Add a function EC_GROUP_check_discriminant() (defined via
3477 EC_METHOD) that verifies that the curve discriminant is non-zero.
3479 Add a function EC_GROUP_check() that makes some sanity tests
3480 on a EC_GROUP, its generator and order. This includes
3481 EC_GROUP_check_discriminant().
3482 [Nils Larsch <nla@trustcenter.de>]
3484 *) Add ECDSA in new directory crypto/ecdsa/.
3486 Add applications 'openssl ecparam' and 'openssl ecdsa'
3487 (these are based on 'openssl dsaparam' and 'openssl dsa').
3489 ECDSA support is also included in various other files across the
3490 library. Most notably,
3491 - 'openssl req' now has a '-newkey ecdsa:file' option;
3492 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
3493 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
3494 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
3495 them suitable for ECDSA where domain parameters must be
3496 extracted before the specific public key;
3497 - ECDSA engine support has been added.
3498 [Nils Larsch <nla@trustcenter.de>]
3500 *) Include some named elliptic curves, and add OIDs from X9.62,
3501 SECG, and WAP/WTLS. Each curve can be obtained from the new
3503 EC_GROUP_new_by_curve_name(),
3504 and the list of available named curves can be obtained with
3505 EC_get_builtin_curves().
3506 Also add a 'curve_name' member to EC_GROUP objects, which can be
3508 EC_GROUP_set_curve_name()
3509 EC_GROUP_get_curve_name()
3510 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
3512 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3513 was actually never needed) and in BN_mul(). The removal in BN_mul()
3514 required a small change in bn_mul_part_recursive() and the addition
3515 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3516 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3517 bn_sub_words() and bn_add_words() except they take arrays with
3521 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
3523 *) Cleanse PEM buffers before freeing them since they may contain
3525 [Benjamin Bennett <ben@psc.edu>]
3527 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3528 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3529 authentication-only ciphersuites.
3532 *) Since AES128 and AES256 share a single mask bit in the logic of
3533 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3534 kludge to work properly if AES128 is available and AES256 isn't.
3537 *) Expand security boundary to match 1.1.1 module.
3540 *) Remove redundant features: hash file source, editing of test vectors
3541 modify fipsld to use external fips_premain.c signature.
3544 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
3545 run algorithm test programs.
3548 *) Make algorithm test programs more tolerant of whitespace.
3551 *) Have SSL/TLS server implementation tolerate "mismatched" record
3552 protocol version while receiving ClientHello even if the
3553 ClientHello is fragmented. (The server can't insist on the
3554 particular protocol version it has chosen before the ServerHello
3555 message has informed the client about his choice.)
3558 *) Load error codes if they are not already present instead of using a
3559 static variable. This allows them to be cleanly unloaded and reloaded.
3562 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3564 *) Introduce limits to prevent malicious keys being able to
3565 cause a denial of service. (CVE-2006-2940)
3566 [Steve Henson, Bodo Moeller]
3568 *) Fix ASN.1 parsing of certain invalid structures that can result
3569 in a denial of service. (CVE-2006-2937) [Steve Henson]
3571 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3572 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3574 *) Fix SSL client code which could crash if connecting to a
3575 malicious SSLv2 server. (CVE-2006-4343)
3576 [Tavis Ormandy and Will Drewry, Google Security Team]
3578 *) Change ciphersuite string processing so that an explicit
3579 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3580 will no longer include "AES128-SHA"), and any other similar
3581 ciphersuite (same bitmap) from *other* protocol versions (so that
3582 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3583 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3584 changes from 0.9.8b and 0.9.8d.
3587 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
3589 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3590 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3592 *) Change the Unix randomness entropy gathering to use poll() when
3593 possible instead of select(), since the latter has some
3594 undesirable limitations.
3595 [Darryl Miles via Richard Levitte and Bodo Moeller]
3597 *) Disable rogue ciphersuites:
3599 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3600 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3601 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3603 The latter two were purportedly from
3604 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3607 Also deactive the remaining ciphersuites from
3608 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3609 unofficial, and the ID has long expired.
3612 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3613 dual-core machines) and other potential thread-safety issues.
3616 Changes between 0.9.7i and 0.9.7j [04 May 2006]
3618 *) Adapt fipsld and the build system to link against the validated FIPS
3619 module in FIPS mode.
3622 *) Fixes for VC++ 2005 build under Windows.
3625 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
3626 from a Windows bash shell such as MSYS. It is autodetected from the
3627 "config" script when run from a VC++ environment. Modify standard VC++
3628 build to use fipscanister.o from the GNU make build.
3631 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
3633 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
3634 The value now differs depending on if you build for FIPS or not.
3635 BEWARE! A program linked with a shared FIPSed libcrypto can't be
3636 safely run with a non-FIPSed libcrypto, as it may crash because of
3637 the difference induced by this change.
3640 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
3642 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3643 (part of SSL_OP_ALL). This option used to disable the
3644 countermeasure against man-in-the-middle protocol-version
3645 rollback in the SSL 2.0 server implementation, which is a bad
3646 idea. (CVE-2005-2969)
3648 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3649 for Information Security, National Institute of Advanced Industrial
3650 Science and Technology [AIST], Japan)]
3652 *) Minimal support for X9.31 signatures and PSS padding modes. This is
3653 mainly for FIPS compliance and not fully integrated at this stage.
3656 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
3657 the exponentiation using a fixed-length exponent. (Otherwise,
3658 the information leaked through timing could expose the secret key
3659 after many signatures; cf. Bleichenbacher's attack on DSA with
3663 *) Make a new fixed-window mod_exp implementation the default for
3664 RSA, DSA, and DH private-key operations so that the sequence of
3665 squares and multiplies and the memory access pattern are
3666 independent of the particular secret key. This will mitigate
3667 cache-timing and potential related attacks.
3669 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
3670 and this is automatically used by BN_mod_exp_mont() if the new flag
3671 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
3672 will use this BN flag for private exponents unless the flag
3673 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
3674 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
3676 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
3678 *) Change the client implementation for SSLv23_method() and
3679 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
3680 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
3681 (Previously, the SSL 2.0 backwards compatible Client Hello
3682 message format would be used even with SSL_OP_NO_SSLv2.)
3685 *) Add support for smime-type MIME parameter in S/MIME messages which some
3689 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
3690 a threadsafe manner. Modify rsa code to use new function and add calls
3691 to dsa and dh code (which had race conditions before).
3694 *) Include the fixed error library code in the C error file definitions
3695 instead of fixing them up at runtime. This keeps the error code
3696 structures constant.
3699 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
3701 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
3704 *) Fixes for newer kerberos headers. NB: the casts are needed because
3705 the 'length' field is signed on one version and unsigned on another
3706 with no (?) obvious way to tell the difference, without these VC++
3707 complains. Also the "definition" of FAR (blank) is no longer included
3708 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
3709 some needed definitions.
3712 *) Undo Cygwin change.
3715 *) Added support for proxy certificates according to RFC 3820.
3716 Because they may be a security thread to unaware applications,
3717 they must be explicitely allowed in run-time. See
3718 docs/HOWTO/proxy_certificates.txt for further information.
3721 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
3723 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
3724 server and client random values. Previously
3725 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
3726 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
3728 This change has negligible security impact because:
3730 1. Server and client random values still have 24 bytes of pseudo random
3733 2. Server and client random values are sent in the clear in the initial
3736 3. The master secret is derived using the premaster secret (48 bytes in
3737 size for static RSA ciphersuites) as well as client server and random
3740 The OpenSSL team would like to thank the UK NISCC for bringing this issue
3743 [Stephen Henson, reported by UK NISCC]
3745 *) Use Windows randomness collection on Cygwin.
3748 *) Fix hang in EGD/PRNGD query when communication socket is closed
3749 prematurely by EGD/PRNGD.
3750 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
3752 *) Prompt for pass phrases when appropriate for PKCS12 input format.
3755 *) Back-port of selected performance improvements from development
3756 branch, as well as improved support for PowerPC platforms.
3759 *) Add lots of checks for memory allocation failure, error codes to indicate
3760 failure and freeing up memory if a failure occurs.
3761 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
3763 *) Add new -passin argument to dgst.
3766 *) Perform some character comparisons of different types in X509_NAME_cmp:
3767 this is needed for some certificates that reencode DNs into UTF8Strings
3768 (in violation of RFC3280) and can't or wont issue name rollover
3772 *) Make an explicit check during certificate validation to see that
3773 the CA setting in each certificate on the chain is correct. As a
3774 side effect always do the following basic checks on extensions,
3775 not just when there's an associated purpose to the check:
3777 - if there is an unhandled critical extension (unless the user
3778 has chosen to ignore this fault)
3779 - if the path length has been exceeded (if one is set at all)
3780 - that certain extensions fit the associated purpose (if one has
3784 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
3786 *) Avoid a race condition when CRLs are checked in a multi threaded
3787 environment. This would happen due to the reordering of the revoked
3788 entries during signature checking and serial number lookup. Now the
3789 encoding is cached and the serial number sort performed under a lock.
3790 Add new STACK function sk_is_sorted().
3793 *) Add Delta CRL to the extension code.
3796 *) Various fixes to s3_pkt.c so alerts are sent properly.
3797 [David Holmes <d.holmes@f5.com>]
3799 *) Reduce the chances of duplicate issuer name and serial numbers (in
3800 violation of RFC3280) using the OpenSSL certificate creation utilities.
3801 This is done by creating a random 64 bit value for the initial serial
3802 number when a serial number file is created or when a self signed
3803 certificate is created using 'openssl req -x509'. The initial serial
3804 number file is created using 'openssl x509 -next_serial' in CA.pl
3805 rather than being initialized to 1.
3808 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
3810 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3811 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3812 [Joe Orton, Steve Henson]
3814 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
3816 [Joe Orton, Steve Henson]
3818 *) Make it possible to have multiple active certificates with the same
3819 subject in the CA index file. This is done only if the keyword
3820 'unique_subject' is set to 'no' in the main CA section (default
3821 if 'CA_default') of the configuration file. The value is saved
3822 with the database itself in a separate index attribute file,
3823 named like the index file with '.attr' appended to the name.
3826 *) X509 verify fixes. Disable broken certificate workarounds when
3827 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
3828 keyUsage extension present. Don't accept CRLs with unhandled critical
3829 extensions: since verify currently doesn't process CRL extensions this
3830 rejects a CRL with *any* critical extensions. Add new verify error codes
3834 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
3835 A clarification of RFC2560 will require the use of OCTET STRINGs and
3836 some implementations cannot handle the current raw format. Since OpenSSL
3837 copies and compares OCSP nonces as opaque blobs without any attempt at
3838 parsing them this should not create any compatibility issues.
3841 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
3842 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
3843 this HMAC (and other) operations are several times slower than OpenSSL
3847 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
3848 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3850 *) Use the correct content when signing type "other".
3853 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
3855 *) Fix various bugs revealed by running the NISCC test suite:
3857 Stop out of bounds reads in the ASN1 code when presented with
3858 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3860 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
3862 If verify callback ignores invalid public key errors don't try to check
3863 certificate signature with the NULL public key.
3867 *) New -ignore_err option in ocsp application to stop the server
3868 exiting on the first error in a request.
3871 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3872 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3876 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3877 extra data after the compression methods not only for TLS 1.0
3878 but also for SSL 3.0 (as required by the specification).
3879 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3881 *) Change X509_certificate_type() to mark the key as exported/exportable
3882 when it's 512 *bits* long, not 512 bytes.
3885 *) Change AES_cbc_encrypt() so it outputs exact multiple of
3886 blocks during encryption.
3889 *) Various fixes to base64 BIO and non blocking I/O. On write
3890 flushes were not handled properly if the BIO retried. On read
3891 data was not being buffered properly and had various logic bugs.
3892 This also affects blocking I/O when the data being decoded is a
3896 *) Various S/MIME bugfixes and compatibility changes:
3897 output correct application/pkcs7 MIME type if
3898 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
3899 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
3900 of files as .eml work). Correctly handle very long lines in MIME
3904 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
3906 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3907 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3908 a protocol version number mismatch like a decryption error
3909 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3912 *) Turn on RSA blinding by default in the default implementation
3913 to avoid a timing attack. Applications that don't want it can call
3914 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3915 They would be ill-advised to do so in most cases.
3916 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3918 *) Change RSA blinding code so that it works when the PRNG is not
3919 seeded (in this case, the secret RSA exponent is abused as
3920 an unpredictable seed -- if it is not unpredictable, there
3921 is no point in blinding anyway). Make RSA blinding thread-safe
3922 by remembering the creator's thread ID in rsa->blinding and
3923 having all other threads use local one-time blinding factors
3924 (this requires more computation than sharing rsa->blinding, but
3925 avoids excessive locking; and if an RSA object is not shared
3926 between threads, blinding will still be very fast).
3929 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
3930 ENGINE as defaults for all supported algorithms irrespective of
3931 the 'flags' parameter. 'flags' is now honoured, so applications
3932 should make sure they are passing it correctly.
3935 *) Target "mingw" now allows native Windows code to be generated in
3936 the Cygwin environment as well as with the MinGW compiler.
3939 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
3941 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3942 via timing by performing a MAC computation even if incorrrect
3943 block cipher padding has been found. This is a countermeasure
3944 against active attacks where the attacker has to distinguish
3945 between bad padding and a MAC verification error. (CVE-2003-0078)
3947 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3948 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3949 Martin Vuagnoux (EPFL, Ilion)]
3951 *) Make the no-err option work as intended. The intention with no-err
3952 is not to have the whole error stack handling routines removed from
3953 libcrypto, it's only intended to remove all the function name and
3954 reason texts, thereby removing some of the footprint that may not
3955 be interesting if those errors aren't displayed anyway.
3957 NOTE: it's still possible for any application or module to have it's
3958 own set of error texts inserted. The routines are there, just not
3959 used by default when no-err is given.
3962 *) Add support for FreeBSD on IA64.
3963 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
3965 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
3966 Kerberos function mit_des_cbc_cksum(). Before this change,
3967 the value returned by DES_cbc_cksum() was like the one from
3968 mit_des_cbc_cksum(), except the bytes were swapped.
3969 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
3971 *) Allow an application to disable the automatic SSL chain building.
3972 Before this a rather primitive chain build was always performed in
3973 ssl3_output_cert_chain(): an application had no way to send the
3974 correct chain if the automatic operation produced an incorrect result.
3976 Now the chain builder is disabled if either:
3978 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
3980 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
3982 The reasoning behind this is that an application would not want the
3983 auto chain building to take place if extra chain certificates are
3984 present and it might also want a means of sending no additional
3985 certificates (for example the chain has two certificates and the
3989 *) Add the possibility to build without the ENGINE framework.
3990 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3992 *) Under Win32 gmtime() can return NULL: check return value in
3993 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
3996 *) DSA routines: under certain error conditions uninitialized BN objects
3997 could be freed. Solution: make sure initialization is performed early
3998 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
3999 Nils Larsch <nla@trustcenter.de> via PR#459)
4002 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
4003 checked on reconnect on the client side, therefore session resumption
4004 could still fail with a "ssl session id is different" error. This
4005 behaviour is masked when SSL_OP_ALL is used due to
4006 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
4007 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4008 followup to PR #377.
4011 *) IA-32 assembler support enhancements: unified ELF targets, support
4012 for SCO/Caldera platforms, fix for Cygwin shared build.
4015 *) Add support for FreeBSD on sparc64. As a consequence, support for
4016 FreeBSD on non-x86 processors is separate from x86 processors on
4017 the config script, much like the NetBSD support.
4018 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
4020 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
4022 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
4025 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
4026 code (06) was taken as the first octet of the session ID and the last
4027 octet was ignored consequently. As a result SSLv2 client side session
4028 caching could not have worked due to the session ID mismatch between
4030 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4034 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
4035 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
4039 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
4040 seems that in spite of existing for more than a year, many application
4041 author have done nothing to provide the necessary callbacks, which
4042 means that this particular engine will not work properly anywhere.
4043 This is a very unfortunate situation which forces us, in the name
4044 of usability, to give the hw_ncipher.c a static lock, which is part
4046 NOTE: This is for the 0.9.7 series ONLY. This hack will never
4047 appear in 0.9.8 or later. We EXPECT application authors to have
4048 dealt properly with this when 0.9.8 is released (unless we actually
4049 make such changes in the libcrypto locking code that changes will
4050 have to be made anyway).
4053 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
4054 octets have been read, EOF or an error occurs. Without this change
4055 some truncated ASN1 structures will not produce an error.
4058 *) Disable Heimdal support, since it hasn't been fully implemented.
4059 Still give the possibility to force the use of Heimdal, but with
4060 warnings and a request that patches get sent to openssl-dev.
4063 *) Add the VC-CE target, introduce the WINCE sysname, and add
4064 INSTALL.WCE and appropriate conditionals to make it build.
4065 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4067 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
4068 cygssl-x.y.z.dll, where x, y and z are the major, minor and
4069 edit numbers of the version.
4070 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4072 *) Introduce safe string copy and catenation functions
4073 (BUF_strlcpy() and BUF_strlcat()).
4074 [Ben Laurie (CHATS) and Richard Levitte]
4076 *) Avoid using fixed-size buffers for one-line DNs.
4077 [Ben Laurie (CHATS)]
4079 *) Add BUF_MEM_grow_clean() to avoid information leakage when
4080 resizing buffers containing secrets, and use where appropriate.
4081 [Ben Laurie (CHATS)]
4083 *) Avoid using fixed size buffers for configuration file location.
4084 [Ben Laurie (CHATS)]
4086 *) Avoid filename truncation for various CA files.
4087 [Ben Laurie (CHATS)]
4089 *) Use sizeof in preference to magic numbers.
4090 [Ben Laurie (CHATS)]
4092 *) Avoid filename truncation in cert requests.
4093 [Ben Laurie (CHATS)]
4095 *) Add assertions to check for (supposedly impossible) buffer
4097 [Ben Laurie (CHATS)]
4099 *) Don't cache truncated DNS entries in the local cache (this could
4100 potentially lead to a spoofing attack).
4101 [Ben Laurie (CHATS)]
4103 *) Fix various buffers to be large enough for hex/decimal
4104 representations in a platform independent manner.
4105 [Ben Laurie (CHATS)]
4107 *) Add CRYPTO_realloc_clean() to avoid information leakage when
4108 resizing buffers containing secrets, and use where appropriate.
4109 [Ben Laurie (CHATS)]
4111 *) Add BIO_indent() to avoid much slightly worrying code to do
4113 [Ben Laurie (CHATS)]
4115 *) Convert sprintf()/BIO_puts() to BIO_printf().
4116 [Ben Laurie (CHATS)]
4118 *) buffer_gets() could terminate with the buffer only half
4120 [Ben Laurie (CHATS)]
4122 *) Add assertions to prevent user-supplied crypto functions from
4123 overflowing internal buffers by having large block sizes, etc.
4124 [Ben Laurie (CHATS)]
4126 *) New OPENSSL_assert() macro (similar to assert(), but enabled
4128 [Ben Laurie (CHATS)]
4130 *) Eliminate unused copy of key in RC4.
4131 [Ben Laurie (CHATS)]
4133 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
4134 [Ben Laurie (CHATS)]
4136 *) Fix off-by-one error in EGD path.
4137 [Ben Laurie (CHATS)]
4139 *) If RANDFILE path is too long, ignore instead of truncating.
4140 [Ben Laurie (CHATS)]
4142 *) Eliminate unused and incorrectly sized X.509 structure
4144 [Ben Laurie (CHATS)]
4146 *) Eliminate unused and dangerous function knumber().
4147 [Ben Laurie (CHATS)]
4149 *) Eliminate unused and dangerous structure, KSSL_ERR.
4150 [Ben Laurie (CHATS)]
4152 *) Protect against overlong session ID context length in an encoded
4153 session object. Since these are local, this does not appear to be
4155 [Ben Laurie (CHATS)]
4157 *) Change from security patch (see 0.9.6e below) that did not affect
4158 the 0.9.6 release series:
4160 Remote buffer overflow in SSL3 protocol - an attacker could
4161 supply an oversized master key in Kerberos-enabled versions.
4163 [Ben Laurie (CHATS)]
4165 *) Change the SSL kerb5 codes to match RFC 2712.
4168 *) Make -nameopt work fully for req and add -reqopt switch.
4169 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
4171 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
4172 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
4174 *) Make sure tests can be performed even if the corresponding algorithms
4175 have been removed entirely. This was also the last step to make
4176 OpenSSL compilable with DJGPP under all reasonable conditions.
4177 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
4179 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
4180 to allow version independent disabling of normally unselected ciphers,
4181 which may be activated as a side-effect of selecting a single cipher.
4183 (E.g., cipher list string "RSA" enables ciphersuites that are left
4184 out of "ALL" because they do not provide symmetric encryption.
4185 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
4186 [Lutz Jaenicke, Bodo Moeller]
4188 *) Add appropriate support for separate platform-dependent build
4189 directories. The recommended way to make a platform-dependent
4190 build directory is the following (tested on Linux), maybe with
4193 # Place yourself outside of the OpenSSL source tree. In
4194 # this example, the environment variable OPENSSL_SOURCE
4195 # is assumed to contain the absolute OpenSSL source directory.
4196 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
4197 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
4198 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
4199 mkdir -p `dirname $F`
4200 ln -s $OPENSSL_SOURCE/$F $F
4203 To be absolutely sure not to disturb the source tree, a "make clean"
4204 is a good thing. If it isn't successfull, don't worry about it,
4205 it probably means the source directory is very clean.
4208 *) Make sure any ENGINE control commands make local copies of string
4209 pointers passed to them whenever necessary. Otherwise it is possible
4210 the caller may have overwritten (or deallocated) the original string
4211 data when a later ENGINE operation tries to use the stored values.
4212 [Götz Babin-Ebell <babinebell@trustcenter.de>]
4214 *) Improve diagnostics in file reading and command-line digests.
4215 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
4217 *) Add AES modes CFB and OFB to the object database. Correct an
4218 error in AES-CFB decryption.
4221 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
4222 allows existing EVP_CIPHER_CTX structures to be reused after
4223 calling EVP_*Final(). This behaviour is used by encryption
4224 BIOs and some applications. This has the side effect that
4225 applications must explicitly clean up cipher contexts with
4226 EVP_CIPHER_CTX_cleanup() or they will leak memory.
4229 *) Check the values of dna and dnb in bn_mul_recursive before calling
4230 bn_mul_comba (a non zero value means the a or b arrays do not contain
4231 n2 elements) and fallback to bn_mul_normal if either is not zero.
4234 *) Fix escaping of non-ASCII characters when using the -subj option
4235 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
4238 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
4239 form for "surname", serialNumber has no short form.
4240 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
4241 therefore remove "mail" short name for "internet 7".
4242 The OID for unique identifiers in X509 certificates is
4243 x500UniqueIdentifier, not uniqueIdentifier.
4244 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
4247 *) Add an "init" command to the ENGINE config module and auto initialize
4248 ENGINEs. Without any "init" command the ENGINE will be initialized
4249 after all ctrl commands have been executed on it. If init=1 the
4250 ENGINE is initailized at that point (ctrls before that point are run
4251 on the uninitialized ENGINE and after on the initialized one). If
4252 init=0 then the ENGINE will not be iniatialized at all.
4255 *) Fix the 'app_verify_callback' interface so that the user-defined
4256 argument is actually passed to the callback: In the
4257 SSL_CTX_set_cert_verify_callback() prototype, the callback
4258 declaration has been changed from
4261 int (*cb)(X509_STORE_CTX *,void *);
4262 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
4263 i=s->ctx->app_verify_callback(&ctx)
4264 has been changed into
4265 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
4267 To update applications using SSL_CTX_set_cert_verify_callback(),
4268 a dummy argument can be added to their callback functions.
4269 [D. K. Smetters <smetters@parc.xerox.com>]
4271 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
4272 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
4274 *) Add and OPENSSL_LOAD_CONF define which will cause
4275 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
4276 This allows older applications to transparently support certain
4277 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
4278 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
4279 load the config file and OPENSSL_add_all_algorithms_conf() which will
4280 always load it have also been added.
4283 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
4284 Adjust NIDs and EVP layer.
4285 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4287 *) Config modules support in openssl utility.
4289 Most commands now load modules from the config file,
4290 though in a few (such as version) this isn't done
4291 because it couldn't be used for anything.
4293 In the case of ca and req the config file used is
4294 the same as the utility itself: that is the -config
4295 command line option can be used to specify an
4299 *) Move default behaviour from OPENSSL_config(). If appname is NULL
4300 use "openssl_conf" if filename is NULL use default openssl config file.
4303 *) Add an argument to OPENSSL_config() to allow the use of an alternative
4304 config section name. Add a new flag to tolerate a missing config file
4305 and move code to CONF_modules_load_file().
4308 *) Support for crypto accelerator cards from Accelerated Encryption
4309 Processing, www.aep.ie. (Use engine 'aep')
4310 The support was copied from 0.9.6c [engine] and adapted/corrected
4311 to work with the new engine framework.
4312 [AEP Inc. and Richard Levitte]
4314 *) Support for SureWare crypto accelerator cards from Baltimore
4315 Technologies. (Use engine 'sureware')
4316 The support was copied from 0.9.6c [engine] and adapted
4317 to work with the new engine framework.
4320 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
4321 make the newer ENGINE framework commands for the CHIL engine work.
4322 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
4324 *) Make it possible to produce shared libraries on ReliantUNIX.
4325 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
4327 *) Add the configuration target debug-linux-ppro.
4328 Make 'openssl rsa' use the general key loading routines
4329 implemented in apps.c, and make those routines able to
4330 handle the key format FORMAT_NETSCAPE and the variant
4332 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4334 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4335 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4337 *) Add -keyform to rsautl, and document -engine.
4338 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
4340 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
4341 BIO_R_NO_SUCH_FILE error code rather than the generic
4342 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
4345 *) Add new functions
4347 ERR_peek_last_error_line
4348 ERR_peek_last_error_line_data.
4349 These are similar to
4352 ERR_peek_error_line_data,
4353 but report on the latest error recorded rather than the first one
4354 still in the error queue.
4355 [Ben Laurie, Bodo Moeller]
4357 *) default_algorithms option in ENGINE config module. This allows things
4359 default_algorithms = ALL
4360 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
4363 *) Prelminary ENGINE config module.
4366 *) New experimental application configuration code.
4369 *) Change the AES code to follow the same name structure as all other
4370 symmetric ciphers, and behave the same way. Move everything to
4371 the directory crypto/aes, thereby obsoleting crypto/rijndael.
4372 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4374 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
4375 [Ben Laurie and Theo de Raadt]
4377 *) Add option to output public keys in req command.
4378 [Massimiliano Pala madwolf@openca.org]
4380 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
4381 (up to about 10% better than before for P-192 and P-224).
4384 *) New functions/macros
4386 SSL_CTX_set_msg_callback(ctx, cb)
4387 SSL_CTX_set_msg_callback_arg(ctx, arg)
4388 SSL_set_msg_callback(ssl, cb)
4389 SSL_set_msg_callback_arg(ssl, arg)
4391 to request calling a callback function
4393 void cb(int write_p, int version, int content_type,
4394 const void *buf, size_t len, SSL *ssl, void *arg)
4396 whenever a protocol message has been completely received
4397 (write_p == 0) or sent (write_p == 1). Here 'version' is the
4398 protocol version according to which the SSL library interprets
4399 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
4400 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
4401 the content type as defined in the SSL 3.0/TLS 1.0 protocol
4402 specification (change_cipher_spec(20), alert(21), handshake(22)).
4403 'buf' and 'len' point to the actual message, 'ssl' to the
4404 SSL object, and 'arg' is the application-defined value set by
4405 SSL[_CTX]_set_msg_callback_arg().
4407 'openssl s_client' and 'openssl s_server' have new '-msg' options
4408 to enable a callback that displays all protocol messages.
4411 *) Change the shared library support so shared libraries are built as
4412 soon as the corresponding static library is finished, and thereby get
4413 openssl and the test programs linked against the shared library.
4414 This still only happens when the keyword "shard" has been given to
4415 the configuration scripts.
4417 NOTE: shared library support is still an experimental thing, and
4418 backward binary compatibility is still not guaranteed.
4419 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
4421 *) Add support for Subject Information Access extension.
4422 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4424 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
4425 additional bytes when new memory had to be allocated, not just
4426 when reusing an existing buffer.
4429 *) New command line and configuration option 'utf8' for the req command.
4430 This allows field values to be specified as UTF8 strings.
4433 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
4434 runs for the former and machine-readable output for the latter.
4437 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
4438 of the e-mail address in the DN (i.e., it will go into a certificate
4439 extension only). The new configuration file option 'email_in_dn = no'
4440 has the same effect.
4441 [Massimiliano Pala madwolf@openca.org]
4443 *) Change all functions with names starting with des_ to be starting
4444 with DES_ instead. Add wrappers that are compatible with libdes,
4445 but are named _ossl_old_des_*. Finally, add macros that map the
4446 des_* symbols to the corresponding _ossl_old_des_* if libdes
4447 compatibility is desired. If OpenSSL 0.9.6c compatibility is
4448 desired, the des_* symbols will be mapped to DES_*, with one
4451 Since we provide two compatibility mappings, the user needs to
4452 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
4453 compatibility is desired. The default (i.e., when that macro
4454 isn't defined) is OpenSSL 0.9.6c compatibility.
4456 There are also macros that enable and disable the support of old
4457 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
4458 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
4459 are defined, the default will apply: to support the old des routines.
4461 In either case, one must include openssl/des.h to get the correct
4462 definitions. Do not try to just include openssl/des_old.h, that
4465 NOTE: This is a major break of an old API into a new one. Software
4466 authors are encouraged to switch to the DES_ style functions. Some
4467 time in the future, des_old.h and the libdes compatibility functions
4468 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
4469 default), and then completely removed.
4472 *) Test for certificates which contain unsupported critical extensions.
4473 If such a certificate is found during a verify operation it is
4474 rejected by default: this behaviour can be overridden by either
4475 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
4476 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
4477 X509_supported_extension() has also been added which returns 1 if a
4478 particular extension is supported.
4481 *) Modify the behaviour of EVP cipher functions in similar way to digests
4482 to retain compatibility with existing code.
4485 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
4486 compatibility with existing code. In particular the 'ctx' parameter does
4487 not have to be to be initialized before the call to EVP_DigestInit() and
4488 it is tidied up after a call to EVP_DigestFinal(). New function
4489 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
4490 EVP_MD_CTX_copy() changed to not require the destination to be
4491 initialized valid and new function EVP_MD_CTX_copy_ex() added which
4492 requires the destination to be valid.
4494 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
4495 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
4498 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
4499 so that complete 'Handshake' protocol structures are kept in memory
4500 instead of overwriting 'msg_type' and 'length' with 'body' data.
4503 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
4504 [Massimo Santin via Richard Levitte]
4506 *) Major restructuring to the underlying ENGINE code. This includes
4507 reduction of linker bloat, separation of pure "ENGINE" manipulation
4508 (initialisation, etc) from functionality dealing with implementations
4509 of specific crypto iterfaces. This change also introduces integrated
4510 support for symmetric ciphers and digest implementations - so ENGINEs
4511 can now accelerate these by providing EVP_CIPHER and EVP_MD
4512 implementations of their own. This is detailed in crypto/engine/README
4513 as it couldn't be adequately described here. However, there are a few
4514 API changes worth noting - some RSA, DSA, DH, and RAND functions that
4515 were changed in the original introduction of ENGINE code have now
4516 reverted back - the hooking from this code to ENGINE is now a good
4517 deal more passive and at run-time, operations deal directly with
4518 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
4519 dereferencing through an ENGINE pointer any more. Also, the ENGINE
4520 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
4521 they were not being used by the framework as there is no concept of a
4522 BIGNUM_METHOD and they could not be generalised to the new
4523 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
4524 ENGINE_cpy() has been removed as it cannot be consistently defined in
4528 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
4531 *) Change mkdef.pl to sort symbols that get the same entry number,
4532 and make sure the automatically generated functions ERR_load_*
4533 become part of libeay.num as well.
4536 *) New function SSL_renegotiate_pending(). This returns true once
4537 renegotiation has been requested (either SSL_renegotiate() call
4538 or HelloRequest/ClientHello receveived from the peer) and becomes
4539 false once a handshake has been completed.
4540 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
4541 sends a HelloRequest, but does not ensure that a handshake takes
4542 place. SSL_renegotiate_pending() is useful for checking if the
4543 client has followed the request.)
4546 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
4547 By default, clients may request session resumption even during
4548 renegotiation (if session ID contexts permit); with this option,
4549 session resumption is possible only in the first handshake.
4551 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4552 more bits available for options that should not be part of
4553 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4556 *) Add some demos for certificate and certificate request creation.
4559 *) Make maximum certificate chain size accepted from the peer application
4560 settable (SSL*_get/set_max_cert_list()), as proposed by
4561 "Douglas E. Engert" <deengert@anl.gov>.
4564 *) Add support for shared libraries for Unixware-7
4565 (Boyd Lynn Gerber <gerberb@zenez.com>).
4568 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4569 be done prior to destruction. Use this to unload error strings from
4570 ENGINEs that load their own error strings. NB: This adds two new API
4571 functions to "get" and "set" this destroy handler in an ENGINE.
4574 *) Alter all existing ENGINE implementations (except "openssl" and
4575 "openbsd") to dynamically instantiate their own error strings. This
4576 makes them more flexible to be built both as statically-linked ENGINEs
4577 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4578 Also, add stub code to each that makes building them as self-contained
4579 shared-libraries easier (see README.ENGINE).
4582 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4583 implementations into applications that are completely implemented in
4584 self-contained shared-libraries. The "dynamic" ENGINE exposes control
4585 commands that can be used to configure what shared-library to load and
4586 to control aspects of the way it is handled. Also, made an update to
4587 the README.ENGINE file that brings its information up-to-date and
4588 provides some information and instructions on the "dynamic" ENGINE
4589 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
4592 *) Make it possible to unload ranges of ERR strings with a new
4593 "ERR_unload_strings" function.
4596 *) Add a copy() function to EVP_MD.
4599 *) Make EVP_MD routines take a context pointer instead of just the
4600 md_data void pointer.
4603 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
4604 that the digest can only process a single chunk of data
4605 (typically because it is provided by a piece of
4606 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
4607 is only going to provide a single chunk of data, and hence the
4608 framework needn't accumulate the data for oneshot drivers.
4611 *) As with "ERR", make it possible to replace the underlying "ex_data"
4612 functions. This change also alters the storage and management of global
4613 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
4614 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
4615 index counters. The API functions that use this state have been changed
4616 to take a "class_index" rather than pointers to the class's local STACK
4617 and counter, and there is now an API function to dynamically create new
4618 classes. This centralisation allows us to (a) plug a lot of the
4619 thread-safety problems that existed, and (b) makes it possible to clean
4620 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
4621 such data would previously have always leaked in application code and
4622 workarounds were in place to make the memory debugging turn a blind eye
4623 to it. Application code that doesn't use this new function will still
4624 leak as before, but their memory debugging output will announce it now
4625 rather than letting it slide.
4627 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
4628 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
4629 has a return value to indicate success or failure.
4632 *) Make it possible to replace the underlying "ERR" functions such that the
4633 global state (2 LHASH tables and 2 locks) is only used by the "default"
4634 implementation. This change also adds two functions to "get" and "set"
4635 the implementation prior to it being automatically set the first time
4636 any other ERR function takes place. Ie. an application can call "get",
4637 pass the return value to a module it has just loaded, and that module
4638 can call its own "set" function using that value. This means the
4639 module's "ERR" operations will use (and modify) the error state in the
4640 application and not in its own statically linked copy of OpenSSL code.
4643 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
4644 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
4645 the operation, and provides a more encapsulated way for external code
4646 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
4647 to use these functions rather than manually incrementing the counts.
4649 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
4652 *) Add EVP test program.
4655 *) Add symmetric cipher support to ENGINE. Expect the API to change!
4658 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
4659 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
4660 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
4661 These allow a CRL to be built without having to access X509_CRL fields
4662 directly. Modify 'ca' application to use new functions.
4665 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
4666 bug workarounds. Rollback attack detection is a security feature.
4667 The problem will only arise on OpenSSL servers when TLSv1 is not
4668 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
4669 Software authors not wanting to support TLSv1 will have special reasons
4670 for their choice and can explicitly enable this option.
4671 [Bodo Moeller, Lutz Jaenicke]
4673 *) Rationalise EVP so it can be extended: don't include a union of
4674 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
4675 (similar to those existing for EVP_CIPHER_CTX).
4680 EVP_MD_CTX_init(&md); /* new function call */
4681 EVP_DigestInit(&md, EVP_sha1());
4682 EVP_DigestUpdate(&md, in, len);
4683 EVP_DigestFinal(&md, out, NULL);
4684 EVP_MD_CTX_cleanup(&md); /* new function call */
4688 *) Make DES key schedule conform to the usual scheme, as well as
4689 correcting its structure. This means that calls to DES functions
4690 now have to pass a pointer to a des_key_schedule instead of a
4691 plain des_key_schedule (which was actually always a pointer
4694 des_key_schedule ks;
4696 des_set_key_checked(..., &ks);
4697 des_ncbc_encrypt(..., &ks, ...);
4699 (Note that a later change renames 'des_...' into 'DES_...'.)
4702 *) Initial reduction of linker bloat: the use of some functions, such as
4703 PEM causes large amounts of unused functions to be linked in due to
4704 poor organisation. For example pem_all.c contains every PEM function
4705 which has a knock on effect of linking in large amounts of (unused)
4706 ASN1 code. Grouping together similar functions and splitting unrelated
4707 functions prevents this.
4710 *) Cleanup of EVP macros.
4713 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
4714 correct _ecb suffix.
4717 *) Add initial OCSP responder support to ocsp application. The
4718 revocation information is handled using the text based index
4719 use by the ca application. The responder can either handle
4720 requests generated internally, supplied in files (for example
4721 via a CGI script) or using an internal minimal server.
4724 *) Add configuration choices to get zlib compression for TLS.
4727 *) Changes to Kerberos SSL for RFC 2712 compliance:
4728 1. Implemented real KerberosWrapper, instead of just using
4729 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
4730 2. Implemented optional authenticator field of KerberosWrapper.
4732 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
4733 and authenticator structs; see crypto/krb5/.
4735 Generalized Kerberos calls to support multiple Kerberos libraries.
4736 [Vern Staats <staatsvr@asc.hpc.mil>,
4737 Jeffrey Altman <jaltman@columbia.edu>
4738 via Richard Levitte]
4740 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
4741 already does with RSA. testdsa.h now has 'priv_key/pub_key'
4742 values for each of the key sizes rather than having just
4743 parameters (and 'speed' generating keys each time).
4746 *) Speed up EVP routines.
4749 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
4750 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
4751 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
4752 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
4754 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
4755 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
4756 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
4759 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
4761 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
4764 *) Added the OS2-EMX target.
4765 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
4767 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
4768 to support NCONF routines in extension code. New function CONF_set_nconf()
4769 to allow functions which take an NCONF to also handle the old LHASH
4770 structure: this means that the old CONF compatible routines can be
4771 retained (in particular wrt extensions) without having to duplicate the
4772 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
4775 *) Enhance the general user interface with mechanisms for inner control
4776 and with possibilities to have yes/no kind of prompts.
4779 *) Change all calls to low level digest routines in the library and
4780 applications to use EVP. Add missing calls to HMAC_cleanup() and
4781 don't assume HMAC_CTX can be copied using memcpy().
4782 [Verdon Walker <VWalker@novell.com>, Steve Henson]
4784 *) Add the possibility to control engines through control names but with
4785 arbitrary arguments instead of just a string.
4786 Change the key loaders to take a UI_METHOD instead of a callback
4787 function pointer. NOTE: this breaks binary compatibility with earlier
4788 versions of OpenSSL [engine].
4789 Adapt the nCipher code for these new conditions and add a card insertion
4793 *) Enhance the general user interface with mechanisms to better support
4794 dialog box interfaces, application-defined prompts, the possibility
4795 to use defaults (for example default passwords from somewhere else)
4796 and interrupts/cancellations.
4799 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
4800 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
4803 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
4804 tidy up some unnecessarily weird code in 'sk_new()').
4805 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
4807 *) Change the key loading routines for ENGINEs to use the same kind
4808 callback (pem_password_cb) as all other routines that need this
4812 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
4813 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
4814 than this minimum value is recommended.
4817 *) New random seeder for OpenVMS, using the system process statistics
4818 that are easily reachable.
4821 *) Windows apparently can't transparently handle global
4822 variables defined in DLLs. Initialisations such as:
4824 const ASN1_ITEM *it = &ASN1_INTEGER_it;
4826 wont compile. This is used by the any applications that need to
4827 declare their own ASN1 modules. This was fixed by adding the option
4828 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
4829 needed for static libraries under Win32.
4832 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
4833 setting of purpose and trust fields. New X509_STORE trust and
4834 purpose functions and tidy up setting in other SSL functions.
4837 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
4838 structure. These are inherited by X509_STORE_CTX when it is
4839 initialised. This allows various defaults to be set in the
4840 X509_STORE structure (such as flags for CRL checking and custom
4841 purpose or trust settings) for functions which only use X509_STORE_CTX
4842 internally such as S/MIME.
4844 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
4845 trust settings if they are not set in X509_STORE. This allows X509_STORE
4846 purposes and trust (in S/MIME for example) to override any set by default.
4848 Add command line options for CRL checking to smime, s_client and s_server
4852 *) Initial CRL based revocation checking. If the CRL checking flag(s)
4853 are set then the CRL is looked up in the X509_STORE structure and
4854 its validity and signature checked, then if the certificate is found
4855 in the CRL the verify fails with a revoked error.
4857 Various new CRL related callbacks added to X509_STORE_CTX structure.
4859 Command line options added to 'verify' application to support this.
4861 This needs some additional work, such as being able to handle multiple
4862 CRLs with different times, extension based lookup (rather than just
4863 by subject name) and ultimately more complete V2 CRL extension
4867 *) Add a general user interface API (crypto/ui/). This is designed
4868 to replace things like des_read_password and friends (backward
4869 compatibility functions using this new API are provided).
4870 The purpose is to remove prompting functions from the DES code
4871 section as well as provide for prompting through dialog boxes in
4872 a window system and the like.
4875 *) Add "ex_data" support to ENGINE so implementations can add state at a
4876 per-structure level rather than having to store it globally.
4879 *) Make it possible for ENGINE structures to be copied when retrieved by
4880 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
4881 This causes the "original" ENGINE structure to act like a template,
4882 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
4883 operational state can be localised to each ENGINE structure, despite the
4884 fact they all share the same "methods". New ENGINE structures returned in
4885 this case have no functional references and the return value is the single
4886 structural reference. This matches the single structural reference returned
4887 by ENGINE_by_id() normally, when it is incremented on the pre-existing
4891 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
4892 needs to match any other type at all we need to manually clear the
4896 *) Changes to the "openssl engine" utility to include;
4897 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
4898 about an ENGINE's available control commands.
4899 - executing control commands from command line arguments using the
4900 '-pre' and '-post' switches. '-post' is only used if '-t' is
4901 specified and the ENGINE is successfully initialised. The syntax for
4902 the individual commands are colon-separated, for example;
4903 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
4906 *) New dynamic control command support for ENGINEs. ENGINEs can now
4907 declare their own commands (numbers), names (strings), descriptions,
4908 and input types for run-time discovery by calling applications. A
4909 subset of these commands are implicitly classed as "executable"
4910 depending on their input type, and only these can be invoked through
4911 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
4912 can be based on user input, config files, etc). The distinction is
4913 that "executable" commands cannot return anything other than a boolean
4914 result and can only support numeric or string input, whereas some
4915 discoverable commands may only be for direct use through
4916 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
4917 pointers, or other custom uses. The "executable" commands are to
4918 support parameterisations of ENGINE behaviour that can be
4919 unambiguously defined by ENGINEs and used consistently across any
4920 OpenSSL-based application. Commands have been added to all the
4921 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
4922 control over shared-library paths without source code alterations.
4925 *) Changed all ENGINE implementations to dynamically allocate their
4926 ENGINEs rather than declaring them statically. Apart from this being
4927 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
4928 this also allows the implementations to compile without using the
4929 internal engine_int.h header.
4932 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
4933 'const' value. Any code that should be able to modify a RAND_METHOD
4934 should already have non-const pointers to it (ie. they should only
4935 modify their own ones).
4938 *) Made a variety of little tweaks to the ENGINE code.
4939 - "atalla" and "ubsec" string definitions were moved from header files
4940 to C code. "nuron" string definitions were placed in variables
4941 rather than hard-coded - allowing parameterisation of these values
4942 later on via ctrl() commands.
4943 - Removed unused "#if 0"'d code.
4944 - Fixed engine list iteration code so it uses ENGINE_free() to release
4945 structural references.
4946 - Constified the RAND_METHOD element of ENGINE structures.
4947 - Constified various get/set functions as appropriate and added
4948 missing functions (including a catch-all ENGINE_cpy that duplicates
4949 all ENGINE values onto a new ENGINE except reference counts/state).
4950 - Removed NULL parameter checks in get/set functions. Setting a method
4951 or function to NULL is a way of cancelling out a previously set
4952 value. Passing a NULL ENGINE parameter is just plain stupid anyway
4953 and doesn't justify the extra error symbols and code.
4954 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
4955 flags from engine_int.h to engine.h.
4956 - Changed prototypes for ENGINE handler functions (init(), finish(),
4957 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
4960 *) Implement binary inversion algorithm for BN_mod_inverse in addition
4961 to the algorithm using long division. The binary algorithm can be
4962 used only if the modulus is odd. On 32-bit systems, it is faster
4963 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
4964 roughly 5-15% for 256-bit moduli), so we use it only for moduli
4965 up to 450 bits. In 64-bit environments, the binary algorithm
4966 appears to be advantageous for much longer moduli; here we use it
4967 for moduli up to 2048 bits.
4970 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
4971 could not support the combine flag in choice fields.
4974 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
4975 extensions from a certificate request to the certificate.
4978 *) Allow multiple 'certopt' and 'nameopt' options to be separated
4979 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
4980 file: this allows the display of the certificate about to be
4981 signed to be customised, to allow certain fields to be included
4982 or excluded and extension details. The old system didn't display
4983 multicharacter strings properly, omitted fields not in the policy
4984 and couldn't display additional details such as extensions.
4987 *) Function EC_POINTs_mul for multiple scalar multiplication
4988 of an arbitrary number of elliptic curve points
4989 \sum scalars[i]*points[i],
4990 optionally including the generator defined for the EC_GROUP:
4991 scalar*generator + \sum scalars[i]*points[i].
4993 EC_POINT_mul is a simple wrapper function for the typical case
4994 that the point list has just one item (besides the optional
4998 *) First EC_METHODs for curves over GF(p):
5000 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
5001 operations and provides various method functions that can also
5002 operate with faster implementations of modular arithmetic.
5004 EC_GFp_mont_method() reuses most functions that are part of
5005 EC_GFp_simple_method, but uses Montgomery arithmetic.
5007 [Bodo Moeller; point addition and point doubling
5008 implementation directly derived from source code provided by
5009 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
5011 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
5012 crypto/ec/ec_lib.c):
5014 Curves are EC_GROUP objects (with an optional group generator)
5015 based on EC_METHODs that are built into the library.
5017 Points are EC_POINT objects based on EC_GROUP objects.
5019 Most of the framework would be able to handle curves over arbitrary
5020 finite fields, but as there are no obvious types for fields other
5021 than GF(p), some functions are limited to that for now.
5024 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
5025 that the file contains a complete HTTP response.
5028 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
5029 change the def and num file printf format specifier from "%-40sXXX"
5030 to "%-39s XXX". The latter will always guarantee a space after the
5031 field while the former will cause them to run together if the field
5032 is 40 of more characters long.
5035 *) Constify the cipher and digest 'method' functions and structures
5036 and modify related functions to take constant EVP_MD and EVP_CIPHER
5040 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
5041 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
5044 *) Modify EVP_Digest*() routines so they now return values. Although the
5045 internal software routines can never fail additional hardware versions
5049 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
5051 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
5052 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
5055 ERR_R_NESTED_ASN1_ERROR
5057 ERR_R_MISSING_ASN1_EOS
5058 were 4 .. 9, conflicting with
5059 ERR_LIB_RSA (= ERR_R_RSA_LIB)
5061 ERR_LIB_PEM (= ERR_R_PEM_LIB).
5062 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
5064 Add new error code 'ERR_R_INTERNAL_ERROR'.
5067 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
5071 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
5072 sets the subject name for a new request or supersedes the
5073 subject name in a given request. Formats that can be parsed are
5074 'CN=Some Name, OU=myOU, C=IT'
5076 'CN=Some Name/OU=myOU/C=IT'.
5078 Add options '-batch' and '-verbose' to 'openssl req'.
5079 [Massimiliano Pala <madwolf@hackmasters.net>]
5081 *) Introduce the possibility to access global variables through
5082 functions on platform were that's the best way to handle exporting
5083 global variables in shared libraries. To enable this functionality,
5084 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
5085 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
5086 is normally done by Configure or something similar).
5088 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
5089 in the source file (foo.c) like this:
5091 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
5092 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
5094 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
5095 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
5097 OPENSSL_DECLARE_GLOBAL(int,foo);
5098 #define foo OPENSSL_GLOBAL_REF(foo)
5099 OPENSSL_DECLARE_GLOBAL(double,bar);
5100 #define bar OPENSSL_GLOBAL_REF(bar)
5102 The #defines are very important, and therefore so is including the
5103 header file everywhere where the defined globals are used.
5105 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
5106 of ASN.1 items, but that structure is a bit different.
5108 The largest change is in util/mkdef.pl which has been enhanced with
5109 better and easier to understand logic to choose which symbols should
5110 go into the Windows .def files as well as a number of fixes and code
5111 cleanup (among others, algorithm keywords are now sorted
5112 lexicographically to avoid constant rewrites).
5115 *) In BN_div() keep a copy of the sign of 'num' before writing the
5116 result to 'rm' because if rm==num the value will be overwritten
5117 and produce the wrong result if 'num' is negative: this caused
5118 problems with BN_mod() and BN_nnmod().
5121 *) Function OCSP_request_verify(). This checks the signature on an
5122 OCSP request and verifies the signer certificate. The signer
5123 certificate is just checked for a generic purpose and OCSP request
5127 *) Add OCSP_check_validity() function to check the validity of OCSP
5128 responses. OCSP responses are prepared in real time and may only
5129 be a few seconds old. Simply checking that the current time lies
5130 between thisUpdate and nextUpdate max reject otherwise valid responses
5131 caused by either OCSP responder or client clock inaccuracy. Instead
5132 we allow thisUpdate and nextUpdate to fall within a certain period of
5133 the current time. The age of the response can also optionally be
5134 checked. Two new options -validity_period and -status_age added to
5138 *) If signature or public key algorithm is unrecognized print out its
5139 OID rather that just UNKNOWN.
5142 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
5143 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
5144 ID to be generated from the issuer certificate alone which can then be
5145 passed to OCSP_id_issuer_cmp().
5148 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
5149 ASN1 modules to export functions returning ASN1_ITEM pointers
5150 instead of the ASN1_ITEM structures themselves. This adds several
5151 new macros which allow the underlying ASN1 function/structure to
5152 be accessed transparently. As a result code should not use ASN1_ITEM
5153 references directly (such as &X509_it) but instead use the relevant
5154 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
5155 use of the new ASN1 code on platforms where exporting structures
5156 is problematical (for example in shared libraries) but exporting
5157 functions returning pointers to structures is not.
5160 *) Add support for overriding the generation of SSL/TLS session IDs.
5161 These callbacks can be registered either in an SSL_CTX or per SSL.
5162 The purpose of this is to allow applications to control, if they wish,
5163 the arbitrary values chosen for use as session IDs, particularly as it
5164 can be useful for session caching in multiple-server environments. A
5165 command-line switch for testing this (and any client code that wishes
5166 to use such a feature) has been added to "s_server".
5167 [Geoff Thorpe, Lutz Jaenicke]
5169 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
5170 of the form '#if defined(...) || defined(...) || ...' and
5171 '#if !defined(...) && !defined(...) && ...'. This also avoids
5172 the growing number of special cases it was previously handling.
5175 *) Make all configuration macros available for application by making
5176 sure they are available in opensslconf.h, by giving them names starting
5177 with "OPENSSL_" to avoid conflicts with other packages and by making
5178 sure e_os2.h will cover all platform-specific cases together with
5180 Additionally, it is now possible to define configuration/platform-
5181 specific names (called "system identities"). In the C code, these
5182 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
5183 macro with the name beginning with "OPENSSL_SYS_", which is determined
5184 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
5188 *) New option -set_serial to 'req' and 'x509' this allows the serial
5189 number to use to be specified on the command line. Previously self
5190 signed certificates were hard coded with serial number 0 and the
5191 CA options of 'x509' had to use a serial number in a file which was
5195 *) New options to 'ca' utility to support V2 CRL entry extensions.
5196 Currently CRL reason, invalidity date and hold instruction are
5197 supported. Add new CRL extensions to V3 code and some new objects.
5200 *) New function EVP_CIPHER_CTX_set_padding() this is used to
5201 disable standard block padding (aka PKCS#5 padding) in the EVP
5202 API, which was previously mandatory. This means that the data is
5203 not padded in any way and so the total length much be a multiple
5204 of the block size, otherwise an error occurs.
5207 *) Initial (incomplete) OCSP SSL support.
5210 *) New function OCSP_parse_url(). This splits up a URL into its host,
5211 port and path components: primarily to parse OCSP URLs. New -url
5212 option to ocsp utility.
5215 *) New nonce behavior. The return value of OCSP_check_nonce() now
5216 reflects the various checks performed. Applications can decide
5217 whether to tolerate certain situations such as an absent nonce
5218 in a response when one was present in a request: the ocsp application
5219 just prints out a warning. New function OCSP_add1_basic_nonce()
5220 this is to allow responders to include a nonce in a response even if
5221 the request is nonce-less.
5224 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
5225 skipped when using openssl x509 multiple times on a single input file,
5226 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
5229 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
5230 set string type: to handle setting ASN1_TIME structures. Fix ca
5231 utility to correctly initialize revocation date of CRLs.
5234 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
5235 the clients preferred ciphersuites and rather use its own preferences.
5236 Should help to work around M$ SGC (Server Gated Cryptography) bug in
5237 Internet Explorer by ensuring unchanged hash method during stepup.
5238 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
5241 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
5242 to aes and add a new 'exist' option to print out symbols that don't
5246 *) Additional options to ocsp utility to allow flags to be set and
5247 additional certificates supplied.
5250 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
5251 OCSP client a number of certificate to only verify the response
5255 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
5256 handle the new API. Currently only ECB, CBC modes supported. Add new
5259 Add TLS AES ciphersuites as described in RFC3268, "Advanced
5260 Encryption Standard (AES) Ciphersuites for Transport Layer
5261 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
5262 not enabled by default and were not part of the "ALL" ciphersuite
5263 alias because they were not yet official; they could be
5264 explicitly requested by specifying the "AESdraft" ciphersuite
5265 group alias. In the final release of OpenSSL 0.9.7, the group
5266 alias is called "AES" and is part of "ALL".)
5267 [Ben Laurie, Steve Henson, Bodo Moeller]
5269 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
5270 request to response.
5273 *) Functions for OCSP responders. OCSP_request_onereq_count(),
5274 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
5275 extract information from a certificate request. OCSP_response_create()
5276 creates a response and optionally adds a basic response structure.
5277 OCSP_basic_add1_status() adds a complete single response to a basic
5278 response and returns the OCSP_SINGLERESP structure just added (to allow
5279 extensions to be included for example). OCSP_basic_add1_cert() adds a
5280 certificate to a basic response and OCSP_basic_sign() signs a basic
5281 response with various flags. New helper functions ASN1_TIME_check()
5282 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
5283 (converts ASN1_TIME to GeneralizedTime).
5286 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
5287 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
5288 structure from a certificate. X509_pubkey_digest() digests the public_key
5289 contents: this is used in various key identifiers.
5292 *) Make sk_sort() tolerate a NULL argument.
5293 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
5295 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
5296 passed by the function are trusted implicitly. If any of them signed the
5297 response then it is assumed to be valid and is not verified.
5300 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
5301 to data. This was previously part of the PKCS7 ASN1 code. This
5302 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
5303 [Steve Henson, reported by Kenneth R. Robinette
5304 <support@securenetterm.com>]
5306 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
5307 routines: without these tracing memory leaks is very painful.
5308 Fix leaks in PKCS12 and PKCS7 routines.
5311 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
5312 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
5313 effectively meant GeneralizedTime would never be used. Now it
5314 is initialised to -1 but X509_time_adj() now has to check the value
5315 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
5316 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
5317 [Steve Henson, reported by Kenneth R. Robinette
5318 <support@securenetterm.com>]
5320 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
5321 result in a zero length in the ASN1_INTEGER structure which was
5322 not consistent with the structure when d2i_ASN1_INTEGER() was used
5323 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
5324 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
5325 where it did not print out a minus for negative ASN1_INTEGER.
5328 *) Add summary printout to ocsp utility. The various functions which
5329 convert status values to strings have been renamed to:
5330 OCSP_response_status_str(), OCSP_cert_status_str() and
5331 OCSP_crl_reason_str() and are no longer static. New options
5332 to verify nonce values and to disable verification. OCSP response
5333 printout format cleaned up.
5336 *) Add additional OCSP certificate checks. These are those specified
5337 in RFC2560. This consists of two separate checks: the CA of the
5338 certificate being checked must either be the OCSP signer certificate
5339 or the issuer of the OCSP signer certificate. In the latter case the
5340 OCSP signer certificate must contain the OCSP signing extended key
5341 usage. This check is performed by attempting to match the OCSP
5342 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
5343 in the OCSP_CERTID structures of the response.
5346 *) Initial OCSP certificate verification added to OCSP_basic_verify()
5347 and related routines. This uses the standard OpenSSL certificate
5348 verify routines to perform initial checks (just CA validity) and
5349 to obtain the certificate chain. Then additional checks will be
5350 performed on the chain. Currently the root CA is checked to see
5351 if it is explicitly trusted for OCSP signing. This is used to set
5352 a root CA as a global signing root: that is any certificate that
5353 chains to that CA is an acceptable OCSP signing certificate.
5356 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
5357 extensions from a separate configuration file.
5358 As when reading extensions from the main configuration file,
5359 the '-extensions ...' option may be used for specifying the
5361 [Massimiliano Pala <madwolf@comune.modena.it>]
5363 *) New OCSP utility. Allows OCSP requests to be generated or
5364 read. The request can be sent to a responder and the output
5365 parsed, outputed or printed in text form. Not complete yet:
5366 still needs to check the OCSP response validity.
5369 *) New subcommands for 'openssl ca':
5370 'openssl ca -status <serial>' prints the status of the cert with
5371 the given serial number (according to the index file).
5372 'openssl ca -updatedb' updates the expiry status of certificates
5374 [Massimiliano Pala <madwolf@comune.modena.it>]
5376 *) New '-newreq-nodes' command option to CA.pl. This is like
5377 '-newreq', but calls 'openssl req' with the '-nodes' option
5378 so that the resulting key is not encrypted.
5379 [Damien Miller <djm@mindrot.org>]
5381 *) New configuration for the GNU Hurd.
5382 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
5384 *) Initial code to implement OCSP basic response verify. This
5385 is currently incomplete. Currently just finds the signer's
5386 certificate and verifies the signature on the response.
5389 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
5390 value of OPENSSLDIR. This is available via the new '-d' option
5391 to 'openssl version', and is also included in 'openssl version -a'.
5394 *) Allowing defining memory allocation callbacks that will be given
5395 file name and line number information in additional arguments
5396 (a const char* and an int). The basic functionality remains, as
5397 well as the original possibility to just replace malloc(),
5398 realloc() and free() by functions that do not know about these
5399 additional arguments. To register and find out the current
5400 settings for extended allocation functions, the following
5401 functions are provided:
5403 CRYPTO_set_mem_ex_functions
5404 CRYPTO_set_locked_mem_ex_functions
5405 CRYPTO_get_mem_ex_functions
5406 CRYPTO_get_locked_mem_ex_functions
5408 These work the same way as CRYPTO_set_mem_functions and friends.
5409 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
5410 extended allocation function is enabled.
5411 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
5412 a conventional allocation function is enabled.
5413 [Richard Levitte, Bodo Moeller]
5415 *) Finish off removing the remaining LHASH function pointer casts.
5416 There should no longer be any prototype-casting required when using
5417 the LHASH abstraction, and any casts that remain are "bugs". See
5418 the callback types and macros at the head of lhash.h for details
5419 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
5422 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
5423 If /dev/[u]random devices are not available or do not return enough
5424 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
5426 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
5427 /etc/entropy will be queried once each in this sequence, quering stops
5428 when enough entropy was collected without querying more sockets.
5431 *) Change the Unix RAND_poll() variant to be able to poll several
5432 random devices, as specified by DEVRANDOM, until a sufficient amount
5433 of data has been collected. We spend at most 10 ms on each file
5434 (select timeout) and read in non-blocking mode. DEVRANDOM now
5435 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
5436 (previously it was just the string "/dev/urandom"), so on typical
5437 platforms the 10 ms delay will never occur.
5438 Also separate out the Unix variant to its own file, rand_unix.c.
5439 For VMS, there's a currently-empty rand_vms.c.
5442 *) Move OCSP client related routines to ocsp_cl.c. These
5443 provide utility functions which an application needing
5444 to issue a request to an OCSP responder and analyse the
5445 response will typically need: as opposed to those which an
5446 OCSP responder itself would need which will be added later.
5448 OCSP_request_sign() signs an OCSP request with an API similar
5449 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
5450 response. OCSP_response_get1_basic() extracts basic response
5451 from response. OCSP_resp_find_status(): finds and extracts status
5452 information from an OCSP_CERTID structure (which will be created
5453 when the request structure is built). These are built from lower
5454 level functions which work on OCSP_SINGLERESP structures but
5455 wont normally be used unless the application wishes to examine
5456 extensions in the OCSP response for example.
5458 Replace nonce routines with a pair of functions.
5459 OCSP_request_add1_nonce() adds a nonce value and optionally
5460 generates a random value. OCSP_check_nonce() checks the
5461 validity of the nonce in an OCSP response.
5464 *) Change function OCSP_request_add() to OCSP_request_add0_id().
5465 This doesn't copy the supplied OCSP_CERTID and avoids the
5466 need to free up the newly created id. Change return type
5467 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
5468 This can then be used to add extensions to the request.
5469 Deleted OCSP_request_new(), since most of its functionality
5470 is now in OCSP_REQUEST_new() (and the case insensitive name
5471 clash) apart from the ability to set the request name which
5472 will be added elsewhere.
5475 *) Update OCSP API. Remove obsolete extensions argument from
5476 various functions. Extensions are now handled using the new
5477 OCSP extension code. New simple OCSP HTTP function which
5478 can be used to send requests and parse the response.
5481 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
5482 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
5483 uses the special reorder version of SET OF to sort the attributes
5484 and reorder them to match the encoded order. This resolves a long
5485 standing problem: a verify on a PKCS7 structure just after signing
5486 it used to fail because the attribute order did not match the
5487 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
5488 it uses the received order. This is necessary to tolerate some broken
5489 software that does not order SET OF. This is handled by encoding
5490 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
5491 to produce the required SET OF.
5494 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
5495 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
5496 files to get correct declarations of the ASN.1 item variables.
5499 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
5500 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
5501 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
5502 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
5503 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
5504 ASN1_ITEM and no wrapper functions.
5507 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
5508 replace the old function pointer based I/O routines. Change most of
5509 the *_d2i_bio() and *_d2i_fp() functions to use these.
5512 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
5513 lines, recognice more "algorithms" that can be deselected, and make
5514 it complain about algorithm deselection that isn't recognised.
5517 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
5518 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
5519 to use new functions. Add NO_ASN1_OLD which can be set to remove
5520 some old style ASN1 functions: this can be used to determine if old
5521 code will still work when these eventually go away.
5524 *) New extension functions for OCSP structures, these follow the
5525 same conventions as certificates and CRLs.
5528 *) New function X509V3_add1_i2d(). This automatically encodes and
5529 adds an extension. Its behaviour can be customised with various
5530 flags to append, replace or delete. Various wrappers added for
5531 certifcates and CRLs.
5534 *) Fix to avoid calling the underlying ASN1 print routine when
5535 an extension cannot be parsed. Correct a typo in the
5536 OCSP_SERVICELOC extension. Tidy up print OCSP format.
5539 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
5540 entries for variables.
5543 *) Add functionality to apps/openssl.c for detecting locking
5544 problems: As the program is single-threaded, all we have
5545 to do is register a locking callback using an array for
5546 storing which locks are currently held by the program.
5549 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
5550 SSL_get_ex_data_X509_STORE_idx(), which is used in
5551 ssl_verify_cert_chain() and thus can be called at any time
5552 during TLS/SSL handshakes so that thread-safety is essential.
5553 Unfortunately, the ex_data design is not at all suited
5554 for multi-threaded use, so it probably should be abolished.
5557 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5558 [Broadcom, tweaked and integrated by Geoff Thorpe]
5560 *) Move common extension printing code to new function
5561 X509V3_print_extensions(). Reorganise OCSP print routines and
5562 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5565 *) New function X509_signature_print() to remove duplication in some
5569 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5570 set (this was treated exactly the same as SET OF previously). This
5571 is used to reorder the STACK representing the structure to match the
5572 encoding. This will be used to get round a problem where a PKCS7
5573 structure which was signed could not be verified because the STACK
5574 order did not reflect the encoded order.
5577 *) Reimplement the OCSP ASN1 module using the new code.
5580 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5581 for its ASN1 operations. The old style function pointers still exist
5582 for now but they will eventually go away.
5585 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
5586 completely replaces the old ASN1 functionality with a table driven
5587 encoder and decoder which interprets an ASN1_ITEM structure describing
5588 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
5589 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
5590 has also been converted to the new form.
5593 *) Change BN_mod_exp_recp so that negative moduli are tolerated
5594 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
5595 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
5596 for negative moduli.
5599 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
5600 of not touching the result's sign bit.
5603 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
5607 *) Changed the LHASH code to use prototypes for callbacks, and created
5608 macros to declare and implement thin (optionally static) functions
5609 that provide type-safety and avoid function pointer casting for the
5610 type-specific callbacks.
5613 *) Added Kerberos Cipher Suites to be used with TLS, as written in
5615 [Veers Staats <staatsvr@asc.hpc.mil>,
5616 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
5618 *) Reformat the FAQ so the different questions and answers can be divided
5619 in sections depending on the subject.
5622 *) Have the zlib compression code load ZLIB.DLL dynamically under
5626 *) New function BN_mod_sqrt for computing square roots modulo a prime
5627 (using the probabilistic Tonelli-Shanks algorithm unless
5628 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
5629 be handled deterministically).
5630 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5632 *) Make BN_mod_inverse faster by explicitly handling small quotients
5633 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
5634 512 bits], about 30% for larger ones [1024 or 2048 bits].)
5637 *) New function BN_kronecker.
5640 *) Fix BN_gcd so that it works on negative inputs; the result is
5641 positive unless both parameters are zero.
5642 Previously something reasonably close to an infinite loop was
5643 possible because numbers could be growing instead of shrinking
5644 in the implementation of Euclid's algorithm.
5647 *) Fix BN_is_word() and BN_is_one() macros to take into account the
5648 sign of the number in question.
5650 Fix BN_is_word(a,w) to work correctly for w == 0.
5652 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
5653 because its test if the absolute value of 'a' equals 'w'.
5654 Note that BN_abs_is_word does *not* handle w == 0 reliably;
5655 it exists mostly for use in the implementations of BN_is_zero(),
5656 BN_is_one(), and BN_is_word().
5659 *) New function BN_swap.
5662 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
5663 the exponentiation functions are more likely to produce reasonable
5664 results on negative inputs.
5667 *) Change BN_mod_mul so that the result is always non-negative.
5668 Previously, it could be negative if one of the factors was negative;
5669 I don't think anyone really wanted that behaviour.
5672 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
5673 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
5674 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
5675 and add new functions:
5684 BN_mod_lshift1_quick
5688 These functions always generate non-negative results.
5690 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
5691 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
5693 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
5694 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
5695 be reduced modulo m.
5696 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5699 The following entry accidentily appeared in the CHANGES file
5700 distributed with OpenSSL 0.9.7. The modifications described in
5701 it do *not* apply to OpenSSL 0.9.7.
5703 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5704 was actually never needed) and in BN_mul(). The removal in BN_mul()
5705 required a small change in bn_mul_part_recursive() and the addition
5706 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5707 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5708 bn_sub_words() and bn_add_words() except they take arrays with
5713 *) In 'openssl passwd', verify passwords read from the terminal
5714 unless the '-salt' option is used (which usually means that
5715 verification would just waste user's time since the resulting
5716 hash is going to be compared with some given password hash)
5717 or the new '-noverify' option is used.
5719 This is an incompatible change, but it does not affect
5720 non-interactive use of 'openssl passwd' (passwords on the command
5721 line, '-stdin' option, '-in ...' option) and thus should not
5725 *) Remove all references to RSAref, since there's no more need for it.
5728 *) Make DSO load along a path given through an environment variable
5729 (SHLIB_PATH) with shl_load().
5732 *) Constify the ENGINE code as a result of BIGNUM constification.
5733 Also constify the RSA code and most things related to it. In a
5734 few places, most notable in the depth of the ASN.1 code, ugly
5735 casts back to non-const were required (to be solved at a later
5739 *) Make it so the openssl application has all engines loaded by default.
5742 *) Constify the BIGNUM routines a little more.
5745 *) Add the following functions:
5747 ENGINE_load_cswift()
5749 ENGINE_load_atalla()
5751 ENGINE_load_builtin_engines()
5753 That way, an application can itself choose if external engines that
5754 are built-in in OpenSSL shall ever be used or not. The benefit is
5755 that applications won't have to be linked with libdl or other dso
5756 libraries unless it's really needed.
5758 Changed 'openssl engine' to load all engines on demand.
5759 Changed the engine header files to avoid the duplication of some
5760 declarations (they differed!).
5763 *) 'openssl engine' can now list capabilities.
5766 *) Better error reporting in 'openssl engine'.
5769 *) Never call load_dh_param(NULL) in s_server.
5772 *) Add engine application. It can currently list engines by name and
5773 identity, and test if they are actually available.
5776 *) Improve RPM specification file by forcing symbolic linking and making
5777 sure the installed documentation is also owned by root.root.
5778 [Damien Miller <djm@mindrot.org>]
5780 *) Give the OpenSSL applications more possibilities to make use of
5781 keys (public as well as private) handled by engines.
5784 *) Add OCSP code that comes from CertCo.
5787 *) Add VMS support for the Rijndael code.
5790 *) Added untested support for Nuron crypto accelerator.
5793 *) Add support for external cryptographic devices. This code was
5794 previously distributed separately as the "engine" branch.
5795 [Geoff Thorpe, Richard Levitte]
5797 *) Rework the filename-translation in the DSO code. It is now possible to
5798 have far greater control over how a "name" is turned into a filename
5799 depending on the operating environment and any oddities about the
5800 different shared library filenames on each system.
5803 *) Support threads on FreeBSD-elf in Configure.
5806 *) Fix for SHA1 assembly problem with MASM: it produces
5807 warnings about corrupt line number information when assembling
5808 with debugging information. This is caused by the overlapping
5810 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
5813 NCONF_get_number() has no error checking at all. As a replacement,
5814 NCONF_get_number_e() is defined (_e for "error checking") and is
5815 promoted strongly. The old NCONF_get_number is kept around for
5816 binary backward compatibility.
5817 Make it possible for methods to load from something other than a BIO,
5818 by providing a function pointer that is given a name instead of a BIO.
5819 For example, this could be used to load configuration data from an
5823 *) Fix for non blocking accept BIOs. Added new I/O special reason
5824 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
5825 with non blocking I/O was not possible because no retry code was
5826 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
5830 *) Added the beginnings of Rijndael support.
5833 *) Fix for bug in DirectoryString mask setting. Add support for
5834 X509_NAME_print_ex() in 'req' and X509_print_ex() function
5835 to allow certificate printing to more controllable, additional
5836 'certopt' option to 'x509' to allow new printing options to be
5840 *) Clean old EAY MD5 hack from e_os.h.
5843 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
5845 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5846 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5847 [Joe Orton, Steve Henson]
5849 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
5851 *) Fix additional bug revealed by the NISCC test suite:
5853 Stop bug triggering large recursion when presented with
5854 certain ASN.1 tags (CVE-2003-0851)
5857 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
5859 *) Fix various bugs revealed by running the NISCC test suite:
5861 Stop out of bounds reads in the ASN1 code when presented with
5862 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5864 If verify callback ignores invalid public key errors don't try to check
5865 certificate signature with the NULL public key.
5869 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5870 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5874 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5875 extra data after the compression methods not only for TLS 1.0
5876 but also for SSL 3.0 (as required by the specification).
5877 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5879 *) Change X509_certificate_type() to mark the key as exported/exportable
5880 when it's 512 *bits* long, not 512 bytes.
5883 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
5885 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5886 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5887 a protocol version number mismatch like a decryption error
5888 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5891 *) Turn on RSA blinding by default in the default implementation
5892 to avoid a timing attack. Applications that don't want it can call
5893 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5894 They would be ill-advised to do so in most cases.
5895 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5897 *) Change RSA blinding code so that it works when the PRNG is not
5898 seeded (in this case, the secret RSA exponent is abused as
5899 an unpredictable seed -- if it is not unpredictable, there
5900 is no point in blinding anyway). Make RSA blinding thread-safe
5901 by remembering the creator's thread ID in rsa->blinding and
5902 having all other threads use local one-time blinding factors
5903 (this requires more computation than sharing rsa->blinding, but
5904 avoids excessive locking; and if an RSA object is not shared
5905 between threads, blinding will still be very fast).
5908 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
5910 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5911 via timing by performing a MAC computation even if incorrrect
5912 block cipher padding has been found. This is a countermeasure
5913 against active attacks where the attacker has to distinguish
5914 between bad padding and a MAC verification error. (CVE-2003-0078)
5916 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5917 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5918 Martin Vuagnoux (EPFL, Ilion)]
5920 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
5922 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
5923 memory from it's contents. This is done with a counter that will
5924 place alternating values in each byte. This can be used to solve
5925 two issues: 1) the removal of calls to memset() by highly optimizing
5926 compilers, and 2) cleansing with other values than 0, since those can
5927 be read through on certain media, for example a swap space on disk.
5930 *) Bugfix: client side session caching did not work with external caching,
5931 because the session->cipher setting was not restored when reloading
5932 from the external cache. This problem was masked, when
5933 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
5934 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
5937 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
5938 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
5939 [Zeev Lieber <zeev-l@yahoo.com>]
5941 *) Undo an undocumented change introduced in 0.9.6e which caused
5942 repeated calls to OpenSSL_add_all_ciphers() and
5943 OpenSSL_add_all_digests() to be ignored, even after calling
5947 *) Change the default configuration reader to deal with last line not
5948 being properly terminated.
5951 *) Change X509_NAME_cmp() so it applies the special rules on handling
5952 DN values that are of type PrintableString, as well as RDNs of type
5953 emailAddress where the value has the type ia5String.
5954 [stefank@valicert.com via Richard Levitte]
5956 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
5957 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
5958 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
5959 the bitwise-OR of the two for use by the majority of applications
5960 wanting this behaviour, and update the docs. The documented
5961 behaviour and actual behaviour were inconsistent and had been
5962 changing anyway, so this is more a bug-fix than a behavioural
5964 [Geoff Thorpe, diagnosed by Nadav Har'El]
5966 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
5967 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
5970 *) Fix initialization code race conditions in
5971 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
5972 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
5973 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
5974 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
5975 ssl2_get_cipher_by_char(),
5976 ssl3_get_cipher_by_char().
5977 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
5979 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
5980 the cached sessions are flushed, as the remove_cb() might use ex_data
5981 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
5982 (see [openssl.org #212]).
5983 [Geoff Thorpe, Lutz Jaenicke]
5985 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
5986 length, instead of the encoding length to d2i_ASN1_OBJECT.
5989 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
5991 *) [In 0.9.6g-engine release:]
5992 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
5993 [Lynn Gazis <lgazis@rainbow.com>]
5995 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
5997 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
5998 and get fix the header length calculation.
5999 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
6000 Alon Kantor <alonk@checkpoint.com> (and others),
6003 *) Use proper error handling instead of 'assertions' in buffer
6004 overflow checks added in 0.9.6e. This prevents DoS (the
6005 assertions could call abort()).
6006 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
6008 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
6010 *) Add various sanity checks to asn1_get_length() to reject
6011 the ASN1 length bytes if they exceed sizeof(long), will appear
6012 negative or the content length exceeds the length of the
6014 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6016 *) Fix cipher selection routines: ciphers without encryption had no flags
6017 for the cipher strength set and where therefore not handled correctly
6018 by the selection routines (PR #130).
6021 *) Fix EVP_dsa_sha macro.
6025 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
6026 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
6027 that was added in OpenSSL 0.9.6d.
6029 As the countermeasure turned out to be incompatible with some
6030 broken SSL implementations, the new option is part of SSL_OP_ALL.
6031 SSL_OP_ALL is usually employed when compatibility with weird SSL
6032 implementations is desired (e.g. '-bugs' option to 's_client' and
6033 's_server'), so the new option is automatically set in many
6037 *) Changes in security patch:
6039 Changes marked "(CHATS)" were sponsored by the Defense Advanced
6040 Research Projects Agency (DARPA) and Air Force Research Laboratory,
6041 Air Force Materiel Command, USAF, under agreement number
6044 *) Add various sanity checks to asn1_get_length() to reject
6045 the ASN1 length bytes if they exceed sizeof(long), will appear
6046 negative or the content length exceeds the length of the
6047 supplied buffer. (CVE-2002-0659)
6048 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6050 *) Assertions for various potential buffer overflows, not known to
6052 [Ben Laurie (CHATS)]
6054 *) Various temporary buffers to hold ASCII versions of integers were
6055 too small for 64 bit platforms. (CVE-2002-0655)
6056 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
6058 *) Remote buffer overflow in SSL3 protocol - an attacker could
6059 supply an oversized session ID to a client. (CVE-2002-0656)
6060 [Ben Laurie (CHATS)]
6062 *) Remote buffer overflow in SSL2 protocol - an attacker could
6063 supply an oversized client master key. (CVE-2002-0656)
6064 [Ben Laurie (CHATS)]
6066 Changes between 0.9.6c and 0.9.6d [9 May 2002]
6068 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
6069 encoded as NULL) with id-dsa-with-sha1.
6070 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
6072 *) Check various X509_...() return values in apps/req.c.
6073 [Nils Larsch <nla@trustcenter.de>]
6075 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
6076 an end-of-file condition would erronously be flagged, when the CRLF
6077 was just at the end of a processed block. The bug was discovered when
6078 processing data through a buffering memory BIO handing the data to a
6079 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
6080 <ptsekov@syntrex.com> and Nedelcho Stanev.
6083 *) Implement a countermeasure against a vulnerability recently found
6084 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
6085 before application data chunks to avoid the use of known IVs
6086 with data potentially chosen by the attacker.
6089 *) Fix length checks in ssl3_get_client_hello().
6092 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
6093 to prevent ssl3_read_internal() from incorrectly assuming that
6094 ssl3_read_bytes() found application data while handshake
6095 processing was enabled when in fact s->s3->in_read_app_data was
6096 merely automatically cleared during the initial handshake.
6097 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
6099 *) Fix object definitions for Private and Enterprise: they were not
6100 recognized in their shortname (=lowercase) representation. Extend
6101 obj_dat.pl to issue an error when using undefined keywords instead
6102 of silently ignoring the problem (Svenning Sorensen
6103 <sss@sss.dnsalias.net>).
6106 *) Fix DH_generate_parameters() so that it works for 'non-standard'
6107 generators, i.e. generators other than 2 and 5. (Previously, the
6108 code did not properly initialise the 'add' and 'rem' values to
6109 BN_generate_prime().)
6111 In the new general case, we do not insist that 'generator' is
6112 actually a primitive root: This requirement is rather pointless;
6113 a generator of the order-q subgroup is just as good, if not
6117 *) Map new X509 verification errors to alerts. Discovered and submitted by
6118 Tom Wu <tom@arcot.com>.
6121 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
6122 returning non-zero before the data has been completely received
6123 when using non-blocking I/O.
6124 [Bodo Moeller; problem pointed out by John Hughes]
6126 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
6127 [Ben Laurie, Lutz Jaenicke]
6129 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
6130 Yoram Zahavi <YoramZ@gilian.com>).
6133 *) Add information about CygWin 1.3 and on, and preserve proper
6134 configuration for the versions before that.
6135 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6137 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
6138 check whether we deal with a copy of a session and do not delete from
6139 the cache in this case. Problem reported by "Izhar Shoshani Levi"
6140 <izhar@checkpoint.com>.
6143 *) Do not store session data into the internal session cache, if it
6144 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
6145 flag is set). Proposed by Aslam <aslam@funk.com>.
6148 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
6152 *) [In 0.9.6d-engine release:]
6153 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6154 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6156 *) Add the configuration target linux-s390x.
6157 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
6159 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
6160 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
6161 variable as an indication that a ClientHello message has been
6162 received. As the flag value will be lost between multiple
6163 invocations of ssl3_accept when using non-blocking I/O, the
6164 function may not be aware that a handshake has actually taken
6165 place, thus preventing a new session from being added to the
6168 To avoid this problem, we now set s->new_session to 2 instead of
6169 using a local variable.
6170 [Lutz Jaenicke, Bodo Moeller]
6172 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
6173 if the SSL_R_LENGTH_MISMATCH error is detected.
6174 [Geoff Thorpe, Bodo Moeller]
6176 *) New 'shared_ldflag' column in Configure platform table.
6179 *) Fix EVP_CIPHER_mode macro.
6180 ["Dan S. Camper" <dan@bti.net>]
6182 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
6183 type, we must throw them away by setting rr->length to 0.
6184 [D P Chang <dpc@qualys.com>]
6186 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
6188 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
6189 <Dominikus.Scherkl@biodata.com>. (The previous implementation
6190 worked incorrectly for those cases where range = 10..._2 and
6191 3*range is two bits longer than range.)
6194 *) Only add signing time to PKCS7 structures if it is not already
6198 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
6199 OBJ_ld_ce should be OBJ_id_ce.
6200 Also some ip-pda OIDs in crypto/objects/objects.txt were
6201 incorrect (cf. RFC 3039).
6202 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
6204 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
6205 returns early because it has nothing to do.
6206 [Andy Schneider <andy.schneider@bjss.co.uk>]
6208 *) [In 0.9.6c-engine release:]
6209 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
6210 [Andy Schneider <andy.schneider@bjss.co.uk>]
6212 *) [In 0.9.6c-engine release:]
6213 Add support for Cryptographic Appliance's keyserver technology.
6214 (Use engine 'keyclient')
6215 [Cryptographic Appliances and Geoff Thorpe]
6217 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
6218 is called via tools/c89.sh because arguments have to be
6219 rearranged (all '-L' options must appear before the first object
6221 [Richard Shapiro <rshapiro@abinitio.com>]
6223 *) [In 0.9.6c-engine release:]
6224 Add support for Broadcom crypto accelerator cards, backported
6226 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
6228 *) [In 0.9.6c-engine release:]
6229 Add support for SureWare crypto accelerator cards from
6230 Baltimore Technologies. (Use engine 'sureware')
6231 [Baltimore Technologies and Mark Cox]
6233 *) [In 0.9.6c-engine release:]
6234 Add support for crypto accelerator cards from Accelerated
6235 Encryption Processing, www.aep.ie. (Use engine 'aep')
6236 [AEP Inc. and Mark Cox]
6238 *) Add a configuration entry for gcc on UnixWare.
6239 [Gary Benson <gbenson@redhat.com>]
6241 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
6242 messages are stored in a single piece (fixed-length part and
6243 variable-length part combined) and fix various bugs found on the way.
6246 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
6247 instead. BIO_gethostbyname() does not know what timeouts are
6248 appropriate, so entries would stay in cache even when they have
6250 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
6252 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
6253 faced with a pathologically small ClientHello fragment that does
6254 not contain client_version: Instead of aborting with an error,
6255 simply choose the highest available protocol version (i.e.,
6256 TLS 1.0 unless it is disabled). In practice, ClientHello
6257 messages are never sent like this, but this change gives us
6258 strictly correct behaviour at least for TLS.
6261 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
6262 never resets s->method to s->ctx->method when called from within
6263 one of the SSL handshake functions.
6264 [Bodo Moeller; problem pointed out by Niko Baric]
6266 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
6267 (sent using the client's version number) if client_version is
6268 smaller than the protocol version in use. Also change
6269 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
6270 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
6271 the client will at least see that alert.
6274 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
6278 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
6279 client receives HelloRequest while in a handshake.
6280 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
6282 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
6283 should end in 'break', not 'goto end' which circuments various
6284 cleanups done in state SSL_ST_OK. But session related stuff
6285 must be disabled for SSL_ST_OK in the case that we just sent a
6288 Also avoid some overhead by not calling ssl_init_wbio_buffer()
6289 before just sending a HelloRequest.
6290 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
6292 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
6293 reveal whether illegal block cipher padding was found or a MAC
6294 verification error occured. (Neither SSLerr() codes nor alerts
6295 are directly visible to potential attackers, but the information
6296 may leak via logfiles.)
6298 Similar changes are not required for the SSL 2.0 implementation
6299 because the number of padding bytes is sent in clear for SSL 2.0,
6300 and the extra bytes are just ignored. However ssl/s2_pkt.c
6301 failed to verify that the purported number of padding bytes is in
6305 *) Add OpenUNIX-8 support including shared libraries
6306 (Boyd Lynn Gerber <gerberb@zenez.com>).
6309 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
6310 'wristwatch attack' using huge encoding parameters (cf.
6311 James H. Manger's CRYPTO 2001 paper). Note that the
6312 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
6313 encoding parameters and hence was not vulnerable.
6316 *) BN_sqr() bug fix.
6317 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
6319 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
6320 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
6321 followed by modular reduction.
6322 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
6324 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
6325 equivalent based on BN_pseudo_rand() instead of BN_rand().
6328 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
6329 This function was broken, as the check for a new client hello message
6330 to handle SGC did not allow these large messages.
6331 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
6334 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
6337 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
6338 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
6341 *) Rework the configuration and shared library support for Tru64 Unix.
6342 The configuration part makes use of modern compiler features and
6343 still retains old compiler behavior for those that run older versions
6344 of the OS. The shared library support part includes a variant that
6345 uses the RPATH feature, and is available through the special
6346 configuration target "alpha-cc-rpath", which will never be selected
6348 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
6350 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
6351 with the same message size as in ssl3_get_certificate_request().
6352 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
6353 messages might inadvertently be reject as too long.
6354 [Petr Lampa <lampa@fee.vutbr.cz>]
6356 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
6359 *) Modified SSL library such that the verify_callback that has been set
6360 specificly for an SSL object with SSL_set_verify() is actually being
6361 used. Before the change, a verify_callback set with this function was
6362 ignored and the verify_callback() set in the SSL_CTX at the time of
6363 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
6364 to allow the necessary settings.
6367 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
6368 explicitly to NULL, as at least on Solaris 8 this seems not always to be
6369 done automatically (in contradiction to the requirements of the C
6370 standard). This made problems when used from OpenSSH.
6373 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
6374 dh->length and always used
6376 BN_rand_range(priv_key, dh->p).
6378 BN_rand_range() is not necessary for Diffie-Hellman, and this
6379 specific range makes Diffie-Hellman unnecessarily inefficient if
6380 dh->length (recommended exponent length) is much smaller than the
6381 length of dh->p. We could use BN_rand_range() if the order of
6382 the subgroup was stored in the DH structure, but we only have
6387 BN_rand(priv_key, l, ...)
6389 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
6395 RSA_eay_public_encrypt
6396 RSA_eay_private_decrypt
6397 RSA_eay_private_encrypt (signing)
6398 RSA_eay_public_decrypt (signature verification)
6400 (default implementations for RSA_public_encrypt,
6401 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
6402 always reject numbers >= n.
6405 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
6406 to synchronize access to 'locking_thread'. This is necessary on
6407 systems where access to 'locking_thread' (an 'unsigned long'
6408 variable) is not atomic.
6411 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
6412 *before* setting the 'crypto_lock_rand' flag. The previous code had
6413 a race condition if 0 is a valid thread ID.
6414 [Travis Vitek <vitek@roguewave.com>]
6416 *) Add support for shared libraries under Irix.
6417 [Albert Chin-A-Young <china@thewrittenword.com>]
6419 *) Add configuration option to build on Linux on both big-endian and
6421 [Ralf Baechle <ralf@uni-koblenz.de>]
6423 *) Add the possibility to create shared libraries on HP-UX.
6426 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
6428 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
6429 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
6430 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
6431 PRNG state recovery was possible based on the output of
6432 one PRNG request appropriately sized to gain knowledge on
6433 'md' followed by enough consecutive 1-byte PRNG requests
6434 to traverse all of 'state'.
6436 1. When updating 'md_local' (the current thread's copy of 'md')
6437 during PRNG output generation, hash all of the previous
6438 'md_local' value, not just the half used for PRNG output.
6440 2. Make the number of bytes from 'state' included into the hash
6441 independent from the number of PRNG bytes requested.
6443 The first measure alone would be sufficient to avoid
6444 Markku-Juhani's attack. (Actually it had never occurred
6445 to me that the half of 'md_local' used for chaining was the
6446 half from which PRNG output bytes were taken -- I had always
6447 assumed that the secret half would be used.) The second
6448 measure makes sure that additional data from 'state' is never
6449 mixed into 'md_local' in small portions; this heuristically
6450 further strengthens the PRNG.
6453 *) Fix crypto/bn/asm/mips3.s.
6456 *) When only the key is given to "enc", the IV is undefined. Print out
6457 an error message in this case.
6460 *) Handle special case when X509_NAME is empty in X509 printing routines.
6463 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
6464 positive and less than q.
6467 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
6468 used: it isn't thread safe and the add_lock_callback should handle
6470 [Paul Rose <Paul.Rose@bridge.com>]
6472 *) Verify that incoming data obeys the block size in
6473 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
6477 [Ulf Möller, Bodo Möller]
6479 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
6480 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
6481 when fixing the server behaviour for backwards-compatible 'client
6482 hello' messages. (Note that the attack is impractical against
6483 SSL 3.0 and TLS 1.0 anyway because length and version checking
6484 means that the probability of guessing a valid ciphertext is
6485 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
6488 Before 0.9.5, the countermeasure (hide the error by generating a
6489 random 'decryption result') did not work properly because
6490 ERR_clear_error() was missing, meaning that SSL_get_error() would
6491 detect the supposedly ignored error.
6493 Both problems are now fixed.
6496 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
6497 (previously it was 1024).
6500 *) Fix for compatibility mode trust settings: ignore trust settings
6501 unless some valid trust or reject settings are present.
6504 *) Fix for blowfish EVP: its a variable length cipher.
6507 *) Fix various bugs related to DSA S/MIME verification. Handle missing
6508 parameters in DSA public key structures and return an error in the
6509 DSA routines if parameters are absent.
6512 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
6513 in the current directory if neither $RANDFILE nor $HOME was set.
6514 RAND_file_name() in 0.9.6a returned NULL in this case. This has
6515 caused some confusion to Windows users who haven't defined $HOME.
6516 Thus RAND_file_name() is changed again: e_os.h can define a
6517 DEFAULT_HOME, which will be used if $HOME is not set.
6518 For Windows, we use "C:"; on other platforms, we still require
6519 environment variables.
6521 *) Move 'if (!initialized) RAND_poll()' into regions protected by
6522 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
6523 having multiple threads call RAND_poll() concurrently.
6526 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
6527 combination of a flag and a thread ID variable.
6528 Otherwise while one thread is in ssleay_rand_bytes (which sets the
6529 flag), *other* threads can enter ssleay_add_bytes without obeying
6530 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
6531 that they do not hold after the first thread unsets add_do_not_lock).
6534 *) Change bctest again: '-x' expressions are not available in all
6538 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
6540 *) Fix a couple of memory leaks in PKCS7_dataDecode()
6541 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
6543 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
6544 the default extension for executables, if any. Also, make the perl
6545 scripts that use symlink() to test if it really exists and use "cp"
6546 if it doesn't. All this made OpenSSL compilable and installable in
6550 *) Fix for asn1_GetSequence() for indefinite length constructed data.
6551 If SEQUENCE is length is indefinite just set c->slen to the total
6552 amount of data available.
6553 [Steve Henson, reported by shige@FreeBSD.org]
6554 [This change does not apply to 0.9.7.]
6556 *) Change bctest to avoid here-documents inside command substitution
6557 (workaround for FreeBSD /bin/sh bug).
6558 For compatibility with Ultrix, avoid shell functions (introduced
6559 in the bctest version that searches along $PATH).
6562 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6563 with des_encrypt() defined on some operating systems, like Solaris
6567 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6568 On the Importance of Eliminating Errors in Cryptographic
6569 Computations, J. Cryptology 14 (2001) 2, 101-119,
6570 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6573 *) MIPS assembler BIGNUM division bug fix.
6576 *) Disabled incorrect Alpha assembler code.
6579 *) Fix PKCS#7 decode routines so they correctly update the length
6580 after reading an EOC for the EXPLICIT tag.
6582 [This change does not apply to 0.9.7.]
6584 *) Fix bug in PKCS#12 key generation routines. This was triggered
6585 if a 3DES key was generated with a 0 initial byte. Include
6586 PKCS12_BROKEN_KEYGEN compilation option to retain the old
6587 (but broken) behaviour.
6590 *) Enhance bctest to search for a working bc along $PATH and print
6592 [Tim Rice <tim@multitalents.net> via Richard Levitte]
6594 *) Fix memory leaks in err.c: free err_data string if necessary;
6595 don't write to the wrong index in ERR_set_error_data.
6598 *) Implement ssl23_peek (analogous to ssl23_read), which previously
6602 *) Replace rdtsc with _emit statements for VC++ version 5.
6603 [Jeremy Cooper <jeremy@baymoo.org>]
6605 *) Make it possible to reuse SSLv2 sessions.
6608 *) In copy_email() check for >= 0 as a return value for
6609 X509_NAME_get_index_by_NID() since 0 is a valid index.
6610 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
6612 *) Avoid coredump with unsupported or invalid public keys by checking if
6613 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
6614 PKCS7_verify() fails with non detached data.
6617 *) Don't use getenv in library functions when run as setuid/setgid.
6618 New function OPENSSL_issetugid().
6621 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
6622 due to incorrect handling of multi-threading:
6624 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
6626 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
6628 3. Count how many times MemCheck_off() has been called so that
6629 nested use can be treated correctly. This also avoids
6630 inband-signalling in the previous code (which relied on the
6631 assumption that thread ID 0 is impossible).
6634 *) Add "-rand" option also to s_client and s_server.
6637 *) Fix CPU detection on Irix 6.x.
6638 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
6639 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6641 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
6644 [This change does not apply to 0.9.7.]
6646 *) Use the cached encoding of an X509_NAME structure rather than
6647 copying it. This is apparently the reason for the libsafe "errors"
6648 but the code is actually correct.
6651 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
6652 Bleichenbacher's DSA attack.
6653 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
6654 to be set and top=0 forces the highest bit to be set; top=-1 is new
6655 and leaves the highest bit random.
6656 [Ulf Moeller, Bodo Moeller]
6658 *) In the NCONF_...-based implementations for CONF_... queries
6659 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
6660 a temporary CONF structure with the data component set to NULL
6661 (which gives segmentation faults in lh_retrieve).
6662 Instead, use NULL for the CONF pointer in CONF_get_string and
6663 CONF_get_number (which may use environment variables) and directly
6664 return NULL from CONF_get_section.
6667 *) Fix potential buffer overrun for EBCDIC.
6670 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
6671 keyUsage if basicConstraints absent for a CA.
6674 *) Make SMIME_write_PKCS7() write mail header values with a format that
6675 is more generally accepted (no spaces before the semicolon), since
6676 some programs can't parse those values properly otherwise. Also make
6677 sure BIO's that break lines after each write do not create invalid
6681 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
6682 macros previously used would not encode an empty SEQUENCE OF
6683 and break the signature.
6685 [This change does not apply to 0.9.7.]
6687 *) Zero the premaster secret after deriving the master secret in
6691 *) Add some EVP_add_digest_alias registrations (as found in
6692 OpenSSL_add_all_digests()) to SSL_library_init()
6693 aka OpenSSL_add_ssl_algorithms(). This provides improved
6694 compatibility with peers using X.509 certificates
6695 with unconventional AlgorithmIdentifier OIDs.
6698 *) Fix for Irix with NO_ASM.
6699 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6701 *) ./config script fixes.
6702 [Ulf Moeller, Richard Levitte]
6704 *) Fix 'openssl passwd -1'.
6707 *) Change PKCS12_key_gen_asc() so it can cope with non null
6708 terminated strings whose length is passed in the passlen
6709 parameter, for example from PEM callbacks. This was done
6710 by adding an extra length parameter to asc2uni().
6711 [Steve Henson, reported by <oddissey@samsung.co.kr>]
6713 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
6714 call failed, free the DSA structure.
6717 *) Fix to uni2asc() to cope with zero length Unicode strings.
6718 These are present in some PKCS#12 files.
6721 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
6722 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
6723 when writing a 32767 byte record.
6724 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
6726 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
6727 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
6729 (RSA objects have a reference count access to which is protected
6730 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
6731 so they are meant to be shared between threads.)
6732 [Bodo Moeller, Geoff Thorpe; original patch submitted by
6733 "Reddie, Steven" <Steven.Reddie@ca.com>]
6735 *) Fix a deadlock in CRYPTO_mem_leaks().
6738 *) Use better test patterns in bntest.
6741 *) rand_win.c fix for Borland C.
6744 *) BN_rshift bugfix for n == 0.
6747 *) Add a 'bctest' script that checks for some known 'bc' bugs
6748 so that 'make test' does not abort just because 'bc' is broken.
6751 *) Store verify_result within SSL_SESSION also for client side to
6752 avoid potential security hole. (Re-used sessions on the client side
6753 always resulted in verify_result==X509_V_OK, not using the original
6754 result of the server certificate verification.)
6757 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
6758 SSL3_RT_APPLICATION_DATA, return 0.
6759 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
6763 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
6764 releases, have been re-implemented by renaming the previous
6765 implementations of ssl2_read and ssl3_read to ssl2_read_internal
6766 and ssl3_read_internal, respectively, and adding 'peek' parameters
6767 to them. The new ssl[23]_{read,peek} functions are calls to
6768 ssl[23]_read_internal with the 'peek' flag set appropriately.
6769 A 'peek' parameter has also been added to ssl3_read_bytes, which
6770 does the actual work for ssl3_read_internal.
6773 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
6774 the method-specific "init()" handler. Also clean up ex_data after
6775 calling the method-specific "finish()" handler. Previously, this was
6776 happening the other way round.
6779 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
6780 The previous value, 12, was not always sufficient for BN_mod_exp().
6783 *) Make sure that shared libraries get the internal name engine with
6784 the full version number and not just 0. This should mark the
6785 shared libraries as not backward compatible. Of course, this should
6786 be changed again when we can guarantee backward binary compatibility.
6789 *) Fix typo in get_cert_by_subject() in by_dir.c
6790 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
6792 *) Rework the system to generate shared libraries:
6794 - Make note of the expected extension for the shared libraries and
6795 if there is a need for symbolic links from for example libcrypto.so.0
6796 to libcrypto.so.0.9.7. There is extended info in Configure for
6799 - Make as few rebuilds of the shared libraries as possible.
6801 - Still avoid linking the OpenSSL programs with the shared libraries.
6803 - When installing, install the shared libraries separately from the
6807 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
6809 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
6810 and not in SSL_clear because the latter is also used by the
6811 accept/connect functions; previously, the settings made by
6812 SSL_set_read_ahead would be lost during the handshake.
6813 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
6815 *) Correct util/mkdef.pl to be selective about disabled algorithms.
6816 Previously, it would create entries for disableed algorithms no
6820 *) Added several new manual pages for SSL_* function.
6823 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
6825 *) In ssl23_get_client_hello, generate an error message when faced
6826 with an initial SSL 3.0/TLS record that is too small to contain the
6827 first two bytes of the ClientHello message, i.e. client_version.
6828 (Note that this is a pathologic case that probably has never happened
6829 in real life.) The previous approach was to use the version number
6830 from the record header as a substitute; but our protocol choice
6831 should not depend on that one because it is not authenticated
6832 by the Finished messages.
6835 *) More robust randomness gathering functions for Windows.
6836 [Jeffrey Altman <jaltman@columbia.edu>]
6838 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
6839 not set then we don't setup the error code for issuer check errors
6840 to avoid possibly overwriting other errors which the callback does
6841 handle. If an application does set the flag then we assume it knows
6842 what it is doing and can handle the new informational codes
6846 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
6847 a general "ANY" type, as such it should be able to decode anything
6848 including tagged types. However it didn't check the class so it would
6849 wrongly interpret tagged types in the same way as their universal
6850 counterpart and unknown types were just rejected. Changed so that the
6851 tagged and unknown types are handled in the same way as a SEQUENCE:
6852 that is the encoding is stored intact. There is also a new type
6853 "V_ASN1_OTHER" which is used when the class is not universal, in this
6854 case we have no idea what the actual type is so we just lump them all
6858 *) On VMS, stdout may very well lead to a file that is written to
6859 in a record-oriented fashion. That means that every write() will
6860 write a separate record, which will be read separately by the
6861 programs trying to read from it. This can be very confusing.
6863 The solution is to put a BIO filter in the way that will buffer
6864 text until a linefeed is reached, and then write everything a
6865 line at a time, so every record written will be an actual line,
6866 not chunks of lines and not (usually doesn't happen, but I've
6867 seen it once) several lines in one record. BIO_f_linebuffer() is
6870 Currently, it's a VMS-only method, because that's where it has
6871 been tested well enough.
6874 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
6875 it can return incorrect results.
6876 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
6877 but it was in 0.9.6-beta[12].)
6880 *) Disable the check for content being present when verifying detached
6881 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
6882 include zero length content when signing messages.
6885 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
6886 BIO_ctrl (for BIO pairs).
6889 *) Add DSO method for VMS.
6892 *) Bug fix: Montgomery multiplication could produce results with the
6896 *) Add RPM specification openssl.spec and modify it to build three
6897 packages. The default package contains applications, application
6898 documentation and run-time libraries. The devel package contains
6899 include files, static libraries and function documentation. The
6900 doc package contains the contents of the doc directory. The original
6901 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
6904 *) Add a large number of documentation files for many SSL routines.
6905 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6907 *) Add a configuration entry for Sony News 4.
6908 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
6910 *) Don't set the two most significant bits to one when generating a
6911 random number < q in the DSA library.
6914 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
6915 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
6916 the underlying transport is blocking) if a handshake took place.
6917 (The default behaviour is needed by applications such as s_client
6918 and s_server that use select() to determine when to use SSL_read;
6919 but for applications that know in advance when to expect data, it
6920 just makes things more complicated.)
6923 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
6927 *) Add a few more EBCDIC conditionals that make `req' and `x509'
6928 work better on such systems.
6929 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6931 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
6932 Update PKCS12_parse() so it copies the friendlyName and the
6933 keyid to the certificates aux info.
6936 *) Fix bug in PKCS7_verify() which caused an infinite loop
6937 if there was more than one signature.
6938 [Sven Uszpelkat <su@celocom.de>]
6940 *) Major change in util/mkdef.pl to include extra information
6941 about each symbol, as well as presentig variables as well
6942 as functions. This change means that there's n more need
6943 to rebuild the .num files when some algorithms are excluded.
6946 *) Allow the verify time to be set by an application,
6947 rather than always using the current time.
6950 *) Phase 2 verify code reorganisation. The certificate
6951 verify code now looks up an issuer certificate by a
6952 number of criteria: subject name, authority key id
6953 and key usage. It also verifies self signed certificates
6954 by the same criteria. The main comparison function is
6955 X509_check_issued() which performs these checks.
6957 Lot of changes were necessary in order to support this
6958 without completely rewriting the lookup code.
6960 Authority and subject key identifier are now cached.
6962 The LHASH 'certs' is X509_STORE has now been replaced
6963 by a STACK_OF(X509_OBJECT). This is mainly because an
6964 LHASH can't store or retrieve multiple objects with
6965 the same hash value.
6967 As a result various functions (which were all internal
6968 use only) have changed to handle the new X509_STORE
6969 structure. This will break anything that messed round
6970 with X509_STORE internally.
6972 The functions X509_STORE_add_cert() now checks for an
6973 exact match, rather than just subject name.
6975 The X509_STORE API doesn't directly support the retrieval
6976 of multiple certificates matching a given criteria, however
6977 this can be worked round by performing a lookup first
6978 (which will fill the cache with candidate certificates)
6979 and then examining the cache for matches. This is probably
6980 the best we can do without throwing out X509_LOOKUP
6981 entirely (maybe later...).
6983 The X509_VERIFY_CTX structure has been enhanced considerably.
6985 All certificate lookup operations now go via a get_issuer()
6986 callback. Although this currently uses an X509_STORE it
6987 can be replaced by custom lookups. This is a simple way
6988 to bypass the X509_STORE hackery necessary to make this
6989 work and makes it possible to use more efficient techniques
6990 in future. A very simple version which uses a simple
6991 STACK for its trusted certificate store is also provided
6992 using X509_STORE_CTX_trusted_stack().
6994 The verify_cb() and verify() callbacks now have equivalents
6995 in the X509_STORE_CTX structure.
6997 X509_STORE_CTX also has a 'flags' field which can be used
6998 to customise the verify behaviour.
7001 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
7002 excludes S/MIME capabilities.
7005 *) When a certificate request is read in keep a copy of the
7006 original encoding of the signed data and use it when outputing
7007 again. Signatures then use the original encoding rather than
7008 a decoded, encoded version which may cause problems if the
7009 request is improperly encoded.
7012 *) For consistency with other BIO_puts implementations, call
7013 buffer_write(b, ...) directly in buffer_puts instead of calling
7016 In BIO_puts, increment b->num_write as in BIO_write.
7017 [Peter.Sylvester@EdelWeb.fr]
7019 *) Fix BN_mul_word for the case where the word is 0. (We have to use
7020 BN_zero, we may not return a BIGNUM with an array consisting of
7024 *) Avoid calling abort() from within the library when problems are
7025 detected, except if preprocessor symbols have been defined
7026 (such as REF_CHECK, BN_DEBUG etc.).
7029 *) New openssl application 'rsautl'. This utility can be
7030 used for low level RSA operations. DER public key
7031 BIO/fp routines also added.
7034 *) New Configure entry and patches for compiling on QNX 4.
7035 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
7037 *) A demo state-machine implementation was sponsored by
7038 Nuron (http://www.nuron.com/) and is now available in
7039 demos/state_machine.
7042 *) New options added to the 'dgst' utility for signature
7043 generation and verification.
7046 *) Unrecognized PKCS#7 content types are now handled via a
7047 catch all ASN1_TYPE structure. This allows unsupported
7048 types to be stored as a "blob" and an application can
7049 encode and decode it manually.
7052 *) Fix various signed/unsigned issues to make a_strex.c
7054 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
7056 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
7057 length if passed a buffer. ASN1_INTEGER_to_BN failed
7058 if passed a NULL BN and its argument was negative.
7059 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
7061 *) Modification to PKCS#7 encoding routines to output definite
7062 length encoding. Since currently the whole structures are in
7063 memory there's not real point in using indefinite length
7064 constructed encoding. However if OpenSSL is compiled with
7065 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
7068 *) Added BIO_vprintf() and BIO_vsnprintf().
7071 *) Added more prefixes to parse for in the the strings written
7072 through a logging bio, to cover all the levels that are available
7073 through syslog. The prefixes are now:
7075 PANIC, EMERG, EMR => LOG_EMERG
7076 ALERT, ALR => LOG_ALERT
7077 CRIT, CRI => LOG_CRIT
7078 ERROR, ERR => LOG_ERR
7079 WARNING, WARN, WAR => LOG_WARNING
7080 NOTICE, NOTE, NOT => LOG_NOTICE
7081 INFO, INF => LOG_INFO
7082 DEBUG, DBG => LOG_DEBUG
7084 and as before, if none of those prefixes are present at the
7085 beginning of the string, LOG_ERR is chosen.
7087 On Win32, the LOG_* levels are mapped according to this:
7089 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
7090 LOG_WARNING => EVENTLOG_WARNING_TYPE
7091 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
7095 *) Made it possible to reconfigure with just the configuration
7096 argument "reconf" or "reconfigure". The command line arguments
7097 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
7098 and are retrieved from there when reconfiguring.
7102 [Assar Westerlund <assar@sics.se>, Richard Levitte]
7104 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
7107 *) The obj_dat.pl script was messing up the sorting of object
7108 names. The reason was that it compared the quoted version
7109 of strings as a result "OCSP" > "OCSP Signing" because
7110 " > SPACE. Changed script to store unquoted versions of
7111 names and add quotes on output. It was also omitting some
7112 names from the lookup table if they were given a default
7113 value (that is if SN is missing it is given the same
7114 value as LN and vice versa), these are now added on the
7115 grounds that if an object has a name we should be able to
7116 look it up. Finally added warning output when duplicate
7117 short or long names are found.
7120 *) Changes needed for Tandem NSK.
7121 [Scott Uroff <scott@xypro.com>]
7123 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
7124 RSA_padding_check_SSLv23(), special padding was never detected
7125 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
7126 version rollback attacks was not effective.
7128 In s23_clnt.c, don't use special rollback-attack detection padding
7129 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
7130 client; similarly, in s23_srvr.c, don't do the rollback check if
7131 SSL 2.0 is the only protocol enabled in the server.
7134 *) Make it possible to get hexdumps of unprintable data with 'openssl
7135 asn1parse'. By implication, the functions ASN1_parse_dump() and
7136 BIO_dump_indent() are added.
7139 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
7140 these print out strings and name structures based on various
7141 flags including RFC2253 support and proper handling of
7142 multibyte characters. Added options to the 'x509' utility
7143 to allow the various flags to be set.
7146 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
7147 Also change the functions X509_cmp_current_time() and
7148 X509_gmtime_adj() work with an ASN1_TIME structure,
7149 this will enable certificates using GeneralizedTime in validity
7150 dates to be checked.
7153 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
7154 negative public key encodings) on by default,
7155 NO_NEG_PUBKEY_BUG can be set to disable it.
7158 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
7159 content octets. An i2c_ASN1_OBJECT is unnecessary because
7160 the encoding can be trivially obtained from the structure.
7163 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
7164 not read locks (CRYPTO_r_[un]lock).
7167 *) A first attempt at creating official support for shared
7168 libraries through configuration. I've kept it so the
7169 default is static libraries only, and the OpenSSL programs
7170 are always statically linked for now, but there are
7171 preparations for dynamic linking in place.
7172 This has been tested on Linux and Tru64.
7175 *) Randomness polling function for Win9x, as described in:
7176 Peter Gutmann, Software Generation of Practically Strong
7180 *) Fix so PRNG is seeded in req if using an already existing
7184 *) New options to smime application. -inform and -outform
7185 allow alternative formats for the S/MIME message including
7186 PEM and DER. The -content option allows the content to be
7187 specified separately. This should allow things like Netscape
7188 form signing output easier to verify.
7191 *) Fix the ASN1 encoding of tags using the 'long form'.
7194 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
7195 STRING types. These convert content octets to and from the
7196 underlying type. The actual tag and length octets are
7197 already assumed to have been read in and checked. These
7198 are needed because all other string types have virtually
7199 identical handling apart from the tag. By having versions
7200 of the ASN1 functions that just operate on content octets
7201 IMPLICIT tagging can be handled properly. It also allows
7202 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
7203 and ASN1_INTEGER are identical apart from the tag.
7206 *) Change the handling of OID objects as follows:
7208 - New object identifiers are inserted in objects.txt, following
7209 the syntax given in objects.README.
7210 - objects.pl is used to process obj_mac.num and create a new
7212 - obj_dat.pl is used to create a new obj_dat.h, using the data in
7215 This is currently kind of a hack, and the perl code in objects.pl
7216 isn't very elegant, but it works as I intended. The simplest way
7217 to check that it worked correctly is to look in obj_dat.h and
7218 check the array nid_objs and make sure the objects haven't moved
7219 around (this is important!). Additions are OK, as well as
7220 consistent name changes.
7223 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
7226 *) Addition of the command line parameter '-rand file' to 'openssl req'.
7227 The given file adds to whatever has already been seeded into the
7228 random pool through the RANDFILE configuration file option or
7229 environment variable, or the default random state file.
7232 *) mkstack.pl now sorts each macro group into lexical order.
7233 Previously the output order depended on the order the files
7234 appeared in the directory, resulting in needless rewriting
7238 *) Patches to make OpenSSL compile under Win32 again. Mostly
7239 work arounds for the VC++ problem that it treats func() as
7240 func(void). Also stripped out the parts of mkdef.pl that
7241 added extra typesafe functions: these no longer exist.
7244 *) Reorganisation of the stack code. The macros are now all
7245 collected in safestack.h . Each macro is defined in terms of
7246 a "stack macro" of the form SKM_<name>(type, a, b). The
7247 DEBUG_SAFESTACK is now handled in terms of function casts,
7248 this has the advantage of retaining type safety without the
7249 use of additional functions. If DEBUG_SAFESTACK is not defined
7250 then the non typesafe macros are used instead. Also modified the
7251 mkstack.pl script to handle the new form. Needs testing to see
7252 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
7253 the default if no major problems. Similar behaviour for ASN1_SET_OF
7254 and PKCS12_STACK_OF.
7257 *) When some versions of IIS use the 'NET' form of private key the
7258 key derivation algorithm is different. Normally MD5(password) is
7259 used as a 128 bit RC4 key. In the modified case
7260 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
7261 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
7262 as the old Netscape_RSA functions except they have an additional
7263 'sgckey' parameter which uses the modified algorithm. Also added
7264 an -sgckey command line option to the rsa utility. Thanks to
7265 Adrian Peck <bertie@ncipher.com> for posting details of the modified
7266 algorithm to openssl-dev.
7269 *) The evp_local.h macros were using 'c.##kname' which resulted in
7270 invalid expansion on some systems (SCO 5.0.5 for example).
7271 Corrected to 'c.kname'.
7272 [Phillip Porch <root@theporch.com>]
7274 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
7275 a STACK of email addresses from a certificate or request, these look
7276 in the subject name and the subject alternative name extensions and
7277 omit any duplicate addresses.
7280 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
7281 This makes DSA verification about 2 % faster.
7284 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
7285 (meaning that now 2^5 values will be precomputed, which is only 4 KB
7286 plus overhead for 1024 bit moduli).
7287 This makes exponentiations about 0.5 % faster for 1024 bit
7288 exponents (as measured by "openssl speed rsa2048").
7291 *) Rename memory handling macros to avoid conflicts with other
7293 Malloc => OPENSSL_malloc
7294 Malloc_locked => OPENSSL_malloc_locked
7295 Realloc => OPENSSL_realloc
7296 Free => OPENSSL_free
7299 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
7300 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
7303 *) CygWin32 support.
7304 [John Jarvie <jjarvie@newsguy.com>]
7306 *) The type-safe stack code has been rejigged. It is now only compiled
7307 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
7308 by default all type-specific stack functions are "#define"d back to
7309 standard stack functions. This results in more streamlined output
7310 but retains the type-safety checking possibilities of the original
7314 *) The STACK code has been cleaned up, and certain type declarations
7315 that didn't make a lot of sense have been brought in line. This has
7316 also involved a cleanup of sorts in safestack.h to more correctly
7317 map type-safe stack functions onto their plain stack counterparts.
7318 This work has also resulted in a variety of "const"ifications of
7319 lots of the code, especially "_cmp" operations which should normally
7320 be prototyped with "const" parameters anyway.
7323 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
7324 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
7325 (The PRNG state consists of two parts, the large pool 'state' and 'md',
7326 where all of 'md' is used each time the PRNG is used, but 'state'
7327 is used only indexed by a cyclic counter. As entropy may not be
7328 well distributed from the beginning, 'md' is important as a
7329 chaining variable. However, the output function chains only half
7330 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
7331 all of 'md', and seeding with STATE_SIZE dummy bytes will result
7332 in all of 'state' being rewritten, with the new values depending
7333 on virtually all of 'md'. This overcomes the 80 bit limitation.)
7336 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
7337 the handshake is continued after ssl_verify_cert_chain();
7338 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
7339 can lead to 'unexplainable' connection aborts later.
7340 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
7342 *) Major EVP API cipher revision.
7343 Add hooks for extra EVP features. This allows various cipher
7344 parameters to be set in the EVP interface. Support added for variable
7345 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
7346 setting of RC2 and RC5 parameters.
7348 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
7351 Remove lots of duplicated code from the EVP library. For example *every*
7352 cipher init() function handles the 'iv' in the same way according to the
7353 cipher mode. They also all do nothing if the 'key' parameter is NULL and
7354 for CFB and OFB modes they zero ctx->num.
7356 New functionality allows removal of S/MIME code RC2 hack.
7358 Most of the routines have the same form and so can be declared in terms
7361 By shifting this to the top level EVP_CipherInit() it can be removed from
7362 all individual ciphers. If the cipher wants to handle IVs or keys
7363 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
7366 Change lots of functions like EVP_EncryptUpdate() to now return a
7367 value: although software versions of the algorithms cannot fail
7368 any installed hardware versions can.
7371 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
7372 this option is set, tolerate broken clients that send the negotiated
7373 protocol version number instead of the requested protocol version
7377 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
7378 i.e. non-zero for export ciphersuites, zero otherwise.
7379 Previous versions had this flag inverted, inconsistent with
7380 rsa_tmp_cb (..._TMP_RSA_CB).
7381 [Bodo Moeller; problem reported by Amit Chopra]
7383 *) Add missing DSA library text string. Work around for some IIS
7384 key files with invalid SEQUENCE encoding.
7387 *) Add a document (doc/standards.txt) that list all kinds of standards
7388 and so on that are implemented in OpenSSL.
7391 *) Enhance c_rehash script. Old version would mishandle certificates
7392 with the same subject name hash and wouldn't handle CRLs at all.
7393 Added -fingerprint option to crl utility, to support new c_rehash
7397 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
7400 *) Fix for SSL server purpose checking. Server checking was
7401 rejecting certificates which had extended key usage present
7402 but no ssl client purpose.
7403 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
7405 *) Make PKCS#12 code work with no password. The PKCS#12 spec
7406 is a little unclear about how a blank password is handled.
7407 Since the password in encoded as a BMPString with terminating
7408 double NULL a zero length password would end up as just the
7409 double NULL. However no password at all is different and is
7410 handled differently in the PKCS#12 key generation code. NS
7411 treats a blank password as zero length. MSIE treats it as no
7412 password on export: but it will try both on import. We now do
7413 the same: PKCS12_parse() tries zero length and no password if
7414 the password is set to "" or NULL (NULL is now a valid password:
7415 it wasn't before) as does the pkcs12 application.
7418 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
7419 perror when PEM_read_bio_X509_REQ fails, the error message must
7420 be obtained from the error queue.
7423 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
7424 it in ERR_remove_state if appropriate, and change ERR_get_state
7425 accordingly to avoid race conditions (this is necessary because
7426 thread_hash is no longer constant once set).
7429 *) Bugfix for linux-elf makefile.one.
7432 *) RSA_get_default_method() will now cause a default
7433 RSA_METHOD to be chosen if one doesn't exist already.
7434 Previously this was only set during a call to RSA_new()
7435 or RSA_new_method(NULL) meaning it was possible for
7436 RSA_get_default_method() to return NULL.
7439 *) Added native name translation to the existing DSO code
7440 that will convert (if the flag to do so is set) filenames
7441 that are sufficiently small and have no path information
7442 into a canonical native form. Eg. "blah" converted to
7443 "libblah.so" or "blah.dll" etc.
7446 *) New function ERR_error_string_n(e, buf, len) which is like
7447 ERR_error_string(e, buf), but writes at most 'len' bytes
7448 including the 0 terminator. For ERR_error_string_n, 'buf'
7450 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
7452 *) CONF library reworked to become more general. A new CONF
7453 configuration file reader "class" is implemented as well as a
7454 new functions (NCONF_*, for "New CONF") to handle it. The now
7455 old CONF_* functions are still there, but are reimplemented to
7456 work in terms of the new functions. Also, a set of functions
7457 to handle the internal storage of the configuration data is
7458 provided to make it easier to write new configuration file
7459 reader "classes" (I can definitely see something reading a
7460 configuration file in XML format, for example), called _CONF_*,
7461 or "the configuration storage API"...
7463 The new configuration file reading functions are:
7465 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
7466 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
7468 NCONF_default, NCONF_WIN32
7470 NCONF_dump_fp, NCONF_dump_bio
7472 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
7473 NCONF_new creates a new CONF object. This works in the same way
7474 as other interfaces in OpenSSL, like the BIO interface.
7475 NCONF_dump_* dump the internal storage of the configuration file,
7476 which is useful for debugging. All other functions take the same
7477 arguments as the old CONF_* functions wth the exception of the
7478 first that must be a `CONF *' instead of a `LHASH *'.
7480 To make it easer to use the new classes with the old CONF_* functions,
7481 the function CONF_set_default_method is provided.
7484 *) Add '-tls1' option to 'openssl ciphers', which was already
7485 mentioned in the documentation but had not been implemented.
7486 (This option is not yet really useful because even the additional
7487 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
7490 *) Initial DSO code added into libcrypto for letting OpenSSL (and
7491 OpenSSL-based applications) load shared libraries and bind to
7492 them in a portable way.
7493 [Geoff Thorpe, with contributions from Richard Levitte]
7495 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
7497 *) Make sure _lrotl and _lrotr are only used with MSVC.
7499 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
7500 (the default implementation of RAND_status).
7502 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
7503 to '-clrext' (= clear extensions), as intended and documented.
7504 [Bodo Moeller; inconsistency pointed out by Michael Attili
7507 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
7508 was larger than the MD block size.
7509 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
7511 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
7512 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
7513 using the passed key: if the passed key was a private key the result
7514 of X509_print(), for example, would be to print out all the private key
7518 *) des_quad_cksum() byte order bug fix.
7519 [Ulf Möller, using the problem description in krb4-0.9.7, where
7520 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
7522 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
7524 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
7526 *) For easily testing in shell scripts whether some command
7527 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
7528 returns with exit code 0 iff no command of the given name is available.
7529 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
7530 the output goes to stdout and nothing is printed to stderr.
7531 Additional arguments are always ignored.
7533 Since for each cipher there is a command of the same name,
7534 the 'no-cipher' compilation switches can be tested this way.
7536 ('openssl no-XXX' is not able to detect pseudo-commands such
7537 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
7540 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
7543 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
7544 is set; it will be thrown away anyway because each handshake creates
7546 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
7547 to parameters -- in previous versions (since OpenSSL 0.9.3) the
7548 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
7549 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7552 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7553 'Q' and 'R' lose their special meanings (quit/renegotiate).
7554 This is part of what -quiet does; unlike -quiet, -ign_eof
7555 does not suppress any output.
7558 *) Add compatibility options to the purpose and trust code. The
7559 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7560 accepts a certificate or CA, this was the previous behaviour,
7561 with all the associated security issues.
7563 X509_TRUST_COMPAT is the old trust behaviour: only and
7564 automatically trust self signed roots in certificate store. A
7565 new trust setting X509_TRUST_DEFAULT is used to specify that
7566 a purpose has no associated trust setting and it should instead
7567 use the value in the default purpose.
7570 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7571 and fix a memory leak.
7574 *) In util/mkerr.pl (which implements 'make errors'), preserve
7575 reason strings from the previous version of the .c file, as
7576 the default to have only downcase letters (and digits) in
7577 automatically generated reasons codes is not always appropriate.
7580 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7581 using strerror. Previously, ERR_reason_error_string() returned
7582 library names as reason strings for SYSerr; but SYSerr is a special
7583 case where small numbers are errno values, not library numbers.
7586 *) Add '-dsaparam' option to 'openssl dhparam' application. This
7587 converts DSA parameters into DH parameters. (When creating parameters,
7588 DSA_generate_parameters is used.)
7591 *) Include 'length' (recommended exponent length) in C code generated
7592 by 'openssl dhparam -C'.
7595 *) The second argument to set_label in perlasm was already being used
7596 so couldn't be used as a "file scope" flag. Moved to third argument
7600 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
7601 instead of RAND_bytes for encryption IVs and salts.
7604 *) Include RAND_status() into RAND_METHOD instead of implementing
7605 it only for md_rand.c Otherwise replacing the PRNG by calling
7606 RAND_set_rand_method would be impossible.
7609 *) Don't let DSA_generate_key() enter an infinite loop if the random
7610 number generation fails.
7613 *) New 'rand' application for creating pseudo-random output.
7616 *) Added configuration support for Linux/IA64
7617 [Rolf Haberrecker <rolf@suse.de>]
7619 *) Assembler module support for Mingw32.
7622 *) Shared library support for HPUX (in shlib/).
7623 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
7625 *) Shared library support for Solaris gcc.
7626 [Lutz Behnke <behnke@trustcenter.de>]
7628 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
7630 *) PKCS7_encrypt() was adding text MIME headers twice because they
7631 were added manually and by SMIME_crlf_copy().
7634 *) In bntest.c don't call BN_rand with zero bits argument.
7635 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
7637 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
7638 case was implemented. This caused BN_div_recp() to fail occasionally.
7641 *) Add an optional second argument to the set_label() in the perl
7642 assembly language builder. If this argument exists and is set
7643 to 1 it signals that the assembler should use a symbol whose
7644 scope is the entire file, not just the current function. This
7645 is needed with MASM which uses the format label:: for this scope.
7646 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
7648 *) Change the ASN1 types so they are typedefs by default. Before
7649 almost all types were #define'd to ASN1_STRING which was causing
7650 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
7654 *) Change names of new functions to the new get1/get0 naming
7655 convention: After 'get1', the caller owns a reference count
7656 and has to call ..._free; 'get0' returns a pointer to some
7657 data structure without incrementing reference counters.
7658 (Some of the existing 'get' functions increment a reference
7659 counter, some don't.)
7660 Similarly, 'set1' and 'add1' functions increase reference
7661 counters or duplicate objects.
7664 *) Allow for the possibility of temp RSA key generation failure:
7665 the code used to assume it always worked and crashed on failure.
7668 *) Fix potential buffer overrun problem in BIO_printf().
7669 [Ulf Möller, using public domain code by Patrick Powell; problem
7670 pointed out by David Sacerdote <das33@cornell.edu>]
7672 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
7673 RAND_egd() and RAND_status(). In the command line application,
7674 the EGD socket can be specified like a seed file using RANDFILE
7678 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
7679 Some CAs (e.g. Verisign) distribute certificates in this form.
7682 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
7683 list to exclude them. This means that no special compilation option
7684 is needed to use anonymous DH: it just needs to be included in the
7688 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
7689 EVP_MD_type. The old functionality is available in a new macro called
7690 EVP_MD_md(). Change code that uses it and update docs.
7693 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
7694 where the 'void *' argument is replaced by a function pointer argument.
7695 Previously 'void *' was abused to point to functions, which works on
7696 many platforms, but is not correct. As these functions are usually
7697 called by macros defined in OpenSSL header files, most source code
7698 should work without changes.
7701 *) <openssl/opensslconf.h> (which is created by Configure) now contains
7702 sections with information on -D... compiler switches used for
7703 compiling the library so that applications can see them. To enable
7704 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
7705 must be defined. E.g.,
7706 #define OPENSSL_ALGORITHM_DEFINES
7707 #include <openssl/opensslconf.h>
7708 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
7709 [Richard Levitte, Ulf and Bodo Möller]
7711 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
7715 *) Change the 'other' type in certificate aux info to a STACK_OF
7716 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
7717 the required ASN1 format: arbitrary types determined by an OID.
7720 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
7721 argument to 'req'. This is not because the function is newer or
7722 better than others it just uses the work 'NEW' in the certificate
7723 request header lines. Some software needs this.
7726 *) Reorganise password command line arguments: now passwords can be
7727 obtained from various sources. Delete the PEM_cb function and make
7728 it the default behaviour: i.e. if the callback is NULL and the
7729 usrdata argument is not NULL interpret it as a null terminated pass
7730 phrase. If usrdata and the callback are NULL then the pass phrase
7731 is prompted for as usual.
7734 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
7735 the support is automatically enabled. The resulting binaries will
7736 autodetect the card and use it if present.
7737 [Ben Laurie and Compaq Inc.]
7739 *) Work around for Netscape hang bug. This sends certificate request
7740 and server done in one record. Since this is perfectly legal in the
7741 SSL/TLS protocol it isn't a "bug" option and is on by default. See
7742 the bugs/SSLv3 entry for more info.
7745 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
7748 *) Add -rand argument to smime and pkcs12 applications and read/write
7752 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
7755 *) Add command line password options to the remaining applications.
7758 *) Bug fix for BN_div_recp() for numerators with an even number of
7762 *) More tests in bntest.c, and changed test_bn output.
7765 *) ./config recognizes MacOS X now.
7768 *) Bug fix for BN_div() when the first words of num and divsor are
7769 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
7772 *) Add support for various broken PKCS#8 formats, and command line
7773 options to produce them.
7776 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
7777 get temporary BIGNUMs from a BN_CTX.
7780 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
7784 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
7785 include a #define from the old name to the new. The original intent
7786 was that statically linked binaries could for example just call
7787 SSLeay_add_all_ciphers() to just add ciphers to the table and not
7788 link with digests. This never worked becayse SSLeay_add_all_digests()
7789 and SSLeay_add_all_ciphers() were in the same source file so calling
7790 one would link with the other. They are now in separate source files.
7793 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
7796 *) Use a less unusual form of the Miller-Rabin primality test (it used
7797 a binary algorithm for exponentiation integrated into the Miller-Rabin
7798 loop, our standard modexp algorithms are faster).
7801 *) Support for the EBCDIC character set completed.
7802 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
7804 *) Source code cleanups: use const where appropriate, eliminate casts,
7805 use void * instead of char * in lhash.
7808 *) Bugfix: ssl3_send_server_key_exchange was not restartable
7809 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
7810 this the server could overwrite ephemeral keys that the client
7814 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
7815 using 50 iterations of the Rabin-Miller test.
7817 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
7818 iterations of the Rabin-Miller test as required by the appendix
7819 to FIPS PUB 186[-1]) instead of DSA_is_prime.
7820 As BN_is_prime_fasttest includes trial division, DSA parameter
7821 generation becomes much faster.
7823 This implies a change for the callback functions in DSA_is_prime
7824 and DSA_generate_parameters: The callback function is called once
7825 for each positive witness in the Rabin-Miller test, not just
7826 occasionally in the inner loop; and the parameters to the
7827 callback function now provide an iteration count for the outer
7828 loop rather than for the current invocation of the inner loop.
7829 DSA_generate_parameters additionally can call the callback
7830 function with an 'iteration count' of -1, meaning that a
7831 candidate has passed the trial division test (when q is generated
7832 from an application-provided seed, trial division is skipped).
7835 *) New function BN_is_prime_fasttest that optionally does trial
7836 division before starting the Rabin-Miller test and has
7837 an additional BN_CTX * argument (whereas BN_is_prime always
7838 has to allocate at least one BN_CTX).
7839 'callback(1, -1, cb_arg)' is called when a number has passed the
7840 trial division stage.
7843 *) Fix for bug in CRL encoding. The validity dates weren't being handled
7847 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
7850 *) New function BN_pseudo_rand().
7853 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
7854 bignum version of BN_from_montgomery() with the working code from
7855 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
7859 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
7860 made it impossible to use the same SSL_SESSION data structure in
7861 SSL2 clients in multiple threads.
7864 *) The return value of RAND_load_file() no longer counts bytes obtained
7865 by stat(). RAND_load_file(..., -1) is new and uses the complete file
7866 to seed the PRNG (previously an explicit byte count was required).
7867 [Ulf Möller, Bodo Möller]
7869 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
7870 used (char *) instead of (void *) and had casts all over the place.
7873 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
7876 *) Retain source code compatibility for BN_prime_checks macro:
7877 BN_is_prime(..., BN_prime_checks, ...) now uses
7878 BN_prime_checks_for_size to determine the appropriate number of
7879 Rabin-Miller iterations.
7882 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
7883 DH_CHECK_P_NOT_SAFE_PRIME.
7884 (Check if this is true? OpenPGP calls them "strong".)
7887 *) Merge the functionality of "dh" and "gendh" programs into a new program
7888 "dhparam". The old programs are retained for now but will handle DH keys
7889 (instead of parameters) in future.
7892 *) Make the ciphers, s_server and s_client programs check the return values
7893 when a new cipher list is set.
7896 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
7897 ciphers. Before when the 56bit ciphers were enabled the sorting was
7900 The syntax for the cipher sorting has been extended to support sorting by
7901 cipher-strength (using the strength_bits hard coded in the tables).
7902 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
7904 Fix a bug in the cipher-command parser: when supplying a cipher command
7905 string with an "undefined" symbol (neither command nor alphanumeric
7906 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
7907 an error is flagged.
7909 Due to the strength-sorting extension, the code of the
7910 ssl_create_cipher_list() function was completely rearranged. I hope that
7911 the readability was also increased :-)
7912 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7914 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
7915 for the first serial number and places 2 in the serial number file. This
7916 avoids problems when the root CA is created with serial number zero and
7917 the first user certificate has the same issuer name and serial number
7921 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
7922 the new code. Add documentation for this stuff.
7925 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
7926 X509_*() to X509at_*() on the grounds that they don't handle X509
7927 structures and behave in an analagous way to the X509v3 functions:
7928 they shouldn't be called directly but wrapper functions should be used
7931 So we also now have some wrapper functions that call the X509at functions
7932 when passed certificate requests. (TO DO: similar things can be done with
7933 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
7934 things. Some of these need some d2i or i2d and print functionality
7935 because they handle more complex structures.)
7938 *) Add missing #ifndefs that caused missing symbols when building libssl
7939 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
7940 NO_RSA in ssl/s2*.c.
7941 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
7943 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
7944 has a return value which indicates the quality of the random data
7945 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
7946 error queue. New function RAND_pseudo_bytes() generates output that is
7947 guaranteed to be unique but not unpredictable. RAND_add is like
7948 RAND_seed, but takes an extra argument for an entropy estimate
7949 (RAND_seed always assumes full entropy).
7952 *) Do more iterations of Rabin-Miller probable prime test (specifically,
7953 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
7954 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
7955 in crypto/bn/bn_prime.c for the complete table). This guarantees a
7956 false-positive rate of at most 2^-80 for random input.
7959 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
7962 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
7963 in the 0.9.5 release), this returns the chain
7964 from an X509_CTX structure with a dup of the stack and all
7965 the X509 reference counts upped: so the stack will exist
7966 after X509_CTX_cleanup() has been called. Modify pkcs12.c
7969 Also make SSL_SESSION_print() print out the verify return
7973 *) Add manpage for the pkcs12 command. Also change the default
7974 behaviour so MAC iteration counts are used unless the new
7975 -nomaciter option is used. This improves file security and
7976 only older versions of MSIE (4.0 for example) need it.
7979 *) Honor the no-xxx Configure options when creating .DEF files.
7982 *) Add PKCS#10 attributes to field table: challengePassword,
7983 unstructuredName and unstructuredAddress. These are taken from
7984 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
7985 international characters are used.
7987 More changes to X509_ATTRIBUTE code: allow the setting of types
7988 based on strings. Remove the 'loc' parameter when adding
7989 attributes because these will be a SET OF encoding which is sorted
7993 *) Initial changes to the 'req' utility to allow request generation
7994 automation. This will allow an application to just generate a template
7995 file containing all the field values and have req construct the
7998 Initial support for X509_ATTRIBUTE handling. Stacks of these are
7999 used all over the place including certificate requests and PKCS#7
8000 structures. They are currently handled manually where necessary with
8001 some primitive wrappers for PKCS#7. The new functions behave in a
8002 manner analogous to the X509 extension functions: they allow
8003 attributes to be looked up by NID and added.
8005 Later something similar to the X509V3 code would be desirable to
8006 automatically handle the encoding, decoding and printing of the
8007 more complex types. The string types like challengePassword can
8008 be handled by the string table functions.
8010 Also modified the multi byte string table handling. Now there is
8011 a 'global mask' which masks out certain types. The table itself
8012 can use the flag STABLE_NO_MASK to ignore the mask setting: this
8013 is useful when for example there is only one permissible type
8014 (as in countryName) and using the mask might result in no valid
8018 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
8019 SSL_get_peer_finished to allow applications to obtain the latest
8020 Finished messages sent to the peer or expected from the peer,
8021 respectively. (SSL_get_peer_finished is usually the Finished message
8022 actually received from the peer, otherwise the protocol will be aborted.)
8024 As the Finished message are message digests of the complete handshake
8025 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
8026 be used for external authentication procedures when the authentication
8027 provided by SSL/TLS is not desired or is not enough.
8030 *) Enhanced support for Alpha Linux is added. Now ./config checks if
8031 the host supports BWX extension and if Compaq C is present on the
8032 $PATH. Just exploiting of the BWX extension results in 20-30%
8033 performance kick for some algorithms, e.g. DES and RC4 to mention
8034 a couple. Compaq C in turn generates ~20% faster code for MD5 and
8038 *) Add support for MS "fast SGC". This is arguably a violation of the
8039 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
8040 weak crypto and after checking the certificate is SGC a second one
8041 with strong crypto. MS SGC stops the first handshake after receiving
8042 the server certificate message and sends a second client hello. Since
8043 a server will typically do all the time consuming operations before
8044 expecting any further messages from the client (server key exchange
8045 is the most expensive) there is little difference between the two.
8047 To get OpenSSL to support MS SGC we have to permit a second client
8048 hello message after we have sent server done. In addition we have to
8049 reset the MAC if we do get this second client hello.
8052 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
8053 if a DER encoded private key is RSA or DSA traditional format. Changed
8054 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
8055 format DER encoded private key. Newer code should use PKCS#8 format which
8056 has the key type encoded in the ASN1 structure. Added DER private key
8057 support to pkcs8 application.
8060 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
8061 ciphersuites has been selected (as required by the SSL 3/TLS 1
8062 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
8063 is set, we interpret this as a request to violate the specification
8064 (the worst that can happen is a handshake failure, and 'correct'
8065 behaviour would result in a handshake failure anyway).
8068 *) In SSL_CTX_add_session, take into account that there might be multiple
8069 SSL_SESSION structures with the same session ID (e.g. when two threads
8070 concurrently obtain them from an external cache).
8071 The internal cache can handle only one SSL_SESSION with a given ID,
8072 so if there's a conflict, we now throw out the old one to achieve
8076 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
8077 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
8078 some routines that use cipher OIDs: some ciphers do not have OIDs
8079 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
8083 *) Simplify the trust setting structure and code. Now we just have
8084 two sequences of OIDs for trusted and rejected settings. These will
8085 typically have values the same as the extended key usage extension
8086 and any application specific purposes.
8088 The trust checking code now has a default behaviour: it will just
8089 check for an object with the same NID as the passed id. Functions can
8090 be provided to override either the default behaviour or the behaviour
8091 for a given id. SSL client, server and email already have functions
8092 in place for compatibility: they check the NID and also return "trusted"
8093 if the certificate is self signed.
8096 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
8097 traditional format into an EVP_PKEY structure.
8100 *) Add a password callback function PEM_cb() which either prompts for
8101 a password if usr_data is NULL or otherwise assumes it is a null
8102 terminated password. Allow passwords to be passed on command line
8103 environment or config files in a few more utilities.
8106 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
8107 keys. Add some short names for PKCS#8 PBE algorithms and allow them
8108 to be specified on the command line for the pkcs8 and pkcs12 utilities.
8109 Update documentation.
8112 *) Support for ASN1 "NULL" type. This could be handled before by using
8113 ASN1_TYPE but there wasn't any function that would try to read a NULL
8114 and produce an error if it couldn't. For compatibility we also have
8115 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
8116 don't allocate anything because they don't need to.
8119 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
8121 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
8123 *) Rebuild of the memory allocation routines used by OpenSSL code and
8124 possibly others as well. The purpose is to make an interface that
8125 provide hooks so anyone can build a separate set of allocation and
8126 deallocation routines to be used by OpenSSL, for example memory
8127 pool implementations, or something else, which was previously hard
8128 since Malloc(), Realloc() and Free() were defined as macros having
8129 the values malloc, realloc and free, respectively (except for Win32
8130 compilations). The same is provided for memory debugging code.
8131 OpenSSL already comes with functionality to find memory leaks, but
8132 this gives people a chance to debug other memory problems.
8134 With these changes, a new set of functions and macros have appeared:
8136 CRYPTO_set_mem_debug_functions() [F]
8137 CRYPTO_get_mem_debug_functions() [F]
8138 CRYPTO_dbg_set_options() [F]
8139 CRYPTO_dbg_get_options() [F]
8140 CRYPTO_malloc_debug_init() [M]
8142 The memory debug functions are NULL by default, unless the library
8143 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
8144 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
8145 gives the standard debugging functions that come with OpenSSL) or
8146 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
8147 provided by the library user) must be used. When the standard
8148 debugging functions are used, CRYPTO_dbg_set_options can be used to
8149 request additional information:
8150 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
8151 the CRYPTO_MDEBUG_xxx macro when compiling the library.
8153 Also, things like CRYPTO_set_mem_functions will always give the
8154 expected result (the new set of functions is used for allocation
8155 and deallocation) at all times, regardless of platform and compiler
8158 To finish it up, some functions that were never use in any other
8159 way than through macros have a new API and new semantic:
8162 CRYPTO_dbg_realloc()
8165 All macros of value have retained their old syntax.
8166 [Richard Levitte and Bodo Moeller]
8168 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
8169 ordering of SMIMECapabilities wasn't in "strength order" and there
8170 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
8174 *) Some ASN1 types with illegal zero length encoding (INTEGER,
8175 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
8176 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
8178 *) Merge in my S/MIME library for OpenSSL. This provides a simple
8179 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
8180 functionality to handle multipart/signed properly) and a utility
8181 called 'smime' to call all this stuff. This is based on code I
8182 originally wrote for Celo who have kindly allowed it to be
8183 included in OpenSSL.
8186 *) Add variants des_set_key_checked and des_set_key_unchecked of
8187 des_set_key (aka des_key_sched). Global variable des_check_key
8188 decides which of these is called by des_set_key; this way
8189 des_check_key behaves as it always did, but applications and
8190 the library itself, which was buggy for des_check_key == 1,
8191 have a cleaner way to pick the version they need.
8194 *) New function PKCS12_newpass() which changes the password of a
8198 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
8199 dynamic mix. In both cases the ids can be used as an index into the
8200 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
8201 functions so they accept a list of the field values and the
8202 application doesn't need to directly manipulate the X509_TRUST
8206 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
8210 *) Modify the way the V3 extension code looks up extensions. This now
8211 works in a similar way to the object code: we have some "standard"
8212 extensions in a static table which is searched with OBJ_bsearch()
8213 and the application can add dynamic ones if needed. The file
8214 crypto/x509v3/ext_dat.h now has the info: this file needs to be
8215 updated whenever a new extension is added to the core code and kept
8216 in ext_nid order. There is a simple program 'tabtest.c' which checks
8217 this. New extensions are not added too often so this file can readily
8218 be maintained manually.
8220 There are two big advantages in doing things this way. The extensions
8221 can be looked up immediately and no longer need to be "added" using
8222 X509V3_add_standard_extensions(): this function now does nothing.
8223 [Side note: I get *lots* of email saying the extension code doesn't
8224 work because people forget to call this function]
8225 Also no dynamic allocation is done unless new extensions are added:
8226 so if we don't add custom extensions there is no need to call
8227 X509V3_EXT_cleanup().
8230 *) Modify enc utility's salting as follows: make salting the default. Add a
8231 magic header, so unsalted files fail gracefully instead of just decrypting
8232 to garbage. This is because not salting is a big security hole, so people
8233 should be discouraged from doing it.
8236 *) Fixes and enhancements to the 'x509' utility. It allowed a message
8237 digest to be passed on the command line but it only used this
8238 parameter when signing a certificate. Modified so all relevant
8239 operations are affected by the digest parameter including the
8240 -fingerprint and -x509toreq options. Also -x509toreq choked if a
8241 DSA key was used because it didn't fix the digest.
8244 *) Initial certificate chain verify code. Currently tests the untrusted
8245 certificates for consistency with the verify purpose (which is set
8246 when the X509_STORE_CTX structure is set up) and checks the pathlength.
8248 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
8249 this is because it will reject chains with invalid extensions whereas
8250 every previous version of OpenSSL and SSLeay made no checks at all.
8252 Trust code: checks the root CA for the relevant trust settings. Trust
8253 settings have an initial value consistent with the verify purpose: e.g.
8254 if the verify purpose is for SSL client use it expects the CA to be
8255 trusted for SSL client use. However the default value can be changed to
8256 permit custom trust settings: one example of this would be to only trust
8257 certificates from a specific "secure" set of CAs.
8259 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
8260 which should be used for version portability: especially since the
8261 verify structure is likely to change more often now.
8263 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
8264 to set them. If not set then assume SSL clients will verify SSL servers
8267 Two new options to the verify program: -untrusted allows a set of
8268 untrusted certificates to be passed in and -purpose which sets the
8269 intended purpose of the certificate. If a purpose is set then the
8270 new chain verify code is used to check extension consistency.
8273 *) Support for the authority information access extension.
8276 *) Modify RSA and DSA PEM read routines to transparently handle
8277 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
8278 public keys in a format compatible with certificate
8279 SubjectPublicKeyInfo structures. Unfortunately there were already
8280 functions called *_PublicKey_* which used various odd formats so
8281 these are retained for compatibility: however the DSA variants were
8282 never in a public release so they have been deleted. Changed dsa/rsa
8283 utilities to handle the new format: note no releases ever handled public
8284 keys so we should be OK.
8286 The primary motivation for this change is to avoid the same fiasco
8287 that dogs private keys: there are several incompatible private key
8288 formats some of which are standard and some OpenSSL specific and
8289 require various evil hacks to allow partial transparent handling and
8290 even then it doesn't work with DER formats. Given the option anything
8291 other than PKCS#8 should be dumped: but the other formats have to
8292 stay in the name of compatibility.
8294 With public keys and the benefit of hindsight one standard format
8295 is used which works with EVP_PKEY, RSA or DSA structures: though
8296 it clearly returns an error if you try to read the wrong kind of key.
8298 Added a -pubkey option to the 'x509' utility to output the public key.
8299 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
8300 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
8301 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
8302 that do the same as the EVP_PKEY_assign_*() except they up the
8303 reference count of the added key (they don't "swallow" the
8307 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
8308 CRLs would fail if the file contained no certificates or no CRLs:
8309 added a new function to read in both types and return the number
8310 read: this means that if none are read it will be an error. The
8311 DER versions of the certificate and CRL reader would always fail
8312 because it isn't possible to mix certificates and CRLs in DER format
8313 without choking one or the other routine. Changed this to just read
8314 a certificate: this is the best we can do. Also modified the code
8315 in apps/verify.c to take notice of return codes: it was previously
8316 attempting to read in certificates from NULL pointers and ignoring
8317 any errors: this is one reason why the cert and CRL reader seemed
8318 to work. It doesn't check return codes from the default certificate
8319 routines: these may well fail if the certificates aren't installed.
8322 *) Code to support otherName option in GeneralName.
8325 *) First update to verify code. Change the verify utility
8326 so it warns if it is passed a self signed certificate:
8327 for consistency with the normal behaviour. X509_verify
8328 has been modified to it will now verify a self signed
8329 certificate if *exactly* the same certificate appears
8330 in the store: it was previously impossible to trust a
8331 single self signed certificate. This means that:
8332 openssl verify ss.pem
8333 now gives a warning about a self signed certificate but
8334 openssl verify -CAfile ss.pem ss.pem
8338 *) For servers, store verify_result in SSL_SESSION data structure
8339 (and add it to external session representation).
8340 This is needed when client certificate verifications fails,
8341 but an application-provided verification callback (set by
8342 SSL_CTX_set_cert_verify_callback) allows accepting the session
8343 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
8344 but returns 1): When the session is reused, we have to set
8345 ssl->verify_result to the appropriate error code to avoid
8347 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
8349 *) Fix a bug in the new PKCS#7 code: it didn't consider the
8350 case in PKCS7_dataInit() where the signed PKCS7 structure
8351 didn't contain any existing data because it was being created.
8352 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
8354 *) Add a salt to the key derivation routines in enc.c. This
8355 forms the first 8 bytes of the encrypted file. Also add a
8356 -S option to allow a salt to be input on the command line.
8359 *) New function X509_cmp(). Oddly enough there wasn't a function
8360 to compare two certificates. We do this by working out the SHA1
8361 hash and comparing that. X509_cmp() will be needed by the trust
8365 *) SSL_get1_session() is like SSL_get_session(), but increments
8366 the reference count in the SSL_SESSION returned.
8367 [Geoff Thorpe <geoff@eu.c2.net>]
8369 *) Fix for 'req': it was adding a null to request attributes.
8370 Also change the X509_LOOKUP and X509_INFO code to handle
8371 certificate auxiliary information.
8374 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
8378 *) Add the possibility to add extra information to the memory leak
8379 detecting output, to form tracebacks, showing from where each
8380 allocation was originated: CRYPTO_push_info("constant string") adds
8381 the string plus current file name and line number to a per-thread
8382 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
8383 is like calling CYRPTO_pop_info() until the stack is empty.
8384 Also updated memory leak detection code to be multi-thread-safe.
8387 *) Add options -text and -noout to pkcs7 utility and delete the
8388 encryption options which never did anything. Update docs.
8391 *) Add options to some of the utilities to allow the pass phrase
8392 to be included on either the command line (not recommended on
8393 OSes like Unix) or read from the environment. Update the
8394 manpages and fix a few bugs.
8397 *) Add a few manpages for some of the openssl commands.
8400 *) Fix the -revoke option in ca. It was freeing up memory twice,
8401 leaking and not finding already revoked certificates.
8404 *) Extensive changes to support certificate auxiliary information.
8405 This involves the use of X509_CERT_AUX structure and X509_AUX
8406 functions. An X509_AUX function such as PEM_read_X509_AUX()
8407 can still read in a certificate file in the usual way but it
8408 will also read in any additional "auxiliary information". By
8409 doing things this way a fair degree of compatibility can be
8410 retained: existing certificates can have this information added
8411 using the new 'x509' options.
8413 Current auxiliary information includes an "alias" and some trust
8414 settings. The trust settings will ultimately be used in enhanced
8415 certificate chain verification routines: currently a certificate
8416 can only be trusted if it is self signed and then it is trusted
8420 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
8421 The problem was that one of the replacement routines had not been working
8422 since SSLeay releases. For now the offending routine has been replaced
8423 with non-optimised assembler. Even so, this now gives around 95%
8424 performance improvement for 1024 bit RSA signs.
8427 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
8428 handling. Most clients have the effective key size in bits equal to
8429 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
8430 A few however don't do this and instead use the size of the decrypted key
8431 to determine the RC2 key length and the AlgorithmIdentifier to determine
8432 the effective key length. In this case the effective key length can still
8433 be 40 bits but the key length can be 168 bits for example. This is fixed
8434 by manually forcing an RC2 key into the EVP_PKEY structure because the
8435 EVP code can't currently handle unusual RC2 key sizes: it always assumes
8436 the key length and effective key length are equal.
8439 *) Add a bunch of functions that should simplify the creation of
8440 X509_NAME structures. Now you should be able to do:
8441 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
8442 and have it automatically work out the correct field type and fill in
8443 the structures. The more adventurous can try:
8444 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
8445 and it will (hopefully) work out the correct multibyte encoding.
8448 *) Change the 'req' utility to use the new field handling and multibyte
8449 copy routines. Before the DN field creation was handled in an ad hoc
8450 way in req, ca, and x509 which was rather broken and didn't support
8451 BMPStrings or UTF8Strings. Since some software doesn't implement
8452 BMPStrings or UTF8Strings yet, they can be enabled using the config file
8453 using the dirstring_type option. See the new comment in the default
8454 openssl.cnf for more info.
8457 *) Make crypto/rand/md_rand.c more robust:
8458 - Assure unique random numbers after fork().
8459 - Make sure that concurrent threads access the global counter and
8460 md serializably so that we never lose entropy in them
8461 or use exactly the same state in multiple threads.
8462 Access to the large state is not always serializable because
8463 the additional locking could be a performance killer, and
8464 md should be large enough anyway.
8467 *) New file apps/app_rand.c with commonly needed functionality
8468 for handling the random seed file.
8470 Use the random seed file in some applications that previously did not:
8472 dsaparam -genkey (which also ignored its '-rand' option),
8475 x509 (when signing).
8476 Except on systems with /dev/urandom, it is crucial to have a random
8477 seed file at least for key creation, DSA signing, and for DH exchanges;
8478 for RSA signatures we could do without one.
8480 gendh and gendsa (unlike genrsa) used to read only the first byte
8481 of each file listed in the '-rand' option. The function as previously
8482 found in genrsa is now in app_rand.c and is used by all programs
8483 that support '-rand'.
8486 *) In RAND_write_file, use mode 0600 for creating files;
8487 don't just chmod when it may be too late.
8490 *) Report an error from X509_STORE_load_locations
8491 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
8494 *) New function ASN1_mbstring_copy() this copies a string in either
8495 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
8496 into an ASN1_STRING type. A mask of permissible types is passed
8497 and it chooses the "minimal" type to use or an error if not type
8501 *) Add function equivalents to the various macros in asn1.h. The old
8502 macros are retained with an M_ prefix. Code inside the library can
8503 use the M_ macros. External code (including the openssl utility)
8504 should *NOT* in order to be "shared library friendly".
8507 *) Add various functions that can check a certificate's extensions
8508 to see if it usable for various purposes such as SSL client,
8509 server or S/MIME and CAs of these types. This is currently
8510 VERY EXPERIMENTAL but will ultimately be used for certificate chain
8511 verification. Also added a -purpose flag to x509 utility to
8512 print out all the purposes.
8515 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
8519 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
8520 for, obtain and decode and extension and obtain its critical flag.
8521 This allows all the necessary extension code to be handled in a
8522 single function call.
8525 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
8526 platforms. See crypto/rc4/rc4_enc.c for further details.
8529 *) New -noout option to asn1parse. This causes no output to be produced
8530 its main use is when combined with -strparse and -out to extract data
8531 from a file (which may not be in ASN.1 format).
8534 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
8535 when producing the local key id.
8536 [Richard Levitte <levitte@stacken.kth.se>]
8538 *) New option -dhparam in s_server. This allows a DH parameter file to be
8539 stated explicitly. If it is not stated then it tries the first server
8540 certificate file. The previous behaviour hard coded the filename
8544 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
8545 a public key to be input or output. For example:
8546 openssl rsa -in key.pem -pubout -out pubkey.pem
8547 Also added necessary DSA public key functions to handle this.
8550 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
8551 in the message. This was handled by allowing
8552 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8553 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8555 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8556 to the end of the strings whereas this didn't. This would cause problems
8557 if strings read with d2i_ASN1_bytes() were later modified.
8558 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8560 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8561 data and it contains EOF it will end up returning an error. This is
8562 caused by input 46 bytes long. The cause is due to the way base64
8563 BIOs find the start of base64 encoded data. They do this by trying a
8564 trial decode on each line until they find one that works. When they
8565 do a flag is set and it starts again knowing it can pass all the
8566 data directly through the decoder. Unfortunately it doesn't reset
8567 the context it uses. This means that if EOF is reached an attempt
8568 is made to pass two EOFs through the context and this causes the
8569 resulting error. This can also cause other problems as well. As is
8570 usual with these problems it takes *ages* to find and the fix is
8571 trivial: move one line.
8572 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8574 *) Ugly workaround to get s_client and s_server working under Windows. The
8575 old code wouldn't work because it needed to select() on sockets and the
8576 tty (for keypresses and to see if data could be written). Win32 only
8577 supports select() on sockets so we select() with a 1s timeout on the
8578 sockets and then see if any characters are waiting to be read, if none
8579 are present then we retry, we also assume we can always write data to
8580 the tty. This isn't nice because the code then blocks until we've
8581 received a complete line of data and it is effectively polling the
8582 keyboard at 1s intervals: however it's quite a bit better than not
8583 working at all :-) A dedicated Windows application might handle this
8584 with an event loop for example.
8587 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
8588 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
8589 will be called when RSA_sign() and RSA_verify() are used. This is useful
8590 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
8591 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
8592 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
8593 This necessitated the support of an extra signature type NID_md5_sha1
8594 for SSL signatures and modifications to the SSL library to use it instead
8595 of calling RSA_public_decrypt() and RSA_private_encrypt().
8598 *) Add new -verify -CAfile and -CApath options to the crl program, these
8599 will lookup a CRL issuers certificate and verify the signature in a
8600 similar way to the verify program. Tidy up the crl program so it
8601 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
8602 less strict. It will now permit CRL extensions even if it is not
8603 a V2 CRL: this will allow it to tolerate some broken CRLs.
8606 *) Initialize all non-automatic variables each time one of the openssl
8607 sub-programs is started (this is necessary as they may be started
8608 multiple times from the "OpenSSL>" prompt).
8609 [Lennart Bang, Bodo Moeller]
8611 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
8612 removing all other RSA functionality (this is what NO_RSA does). This
8613 is so (for example) those in the US can disable those operations covered
8614 by the RSA patent while allowing storage and parsing of RSA keys and RSA
8618 *) Non-copying interface to BIO pairs.
8619 (still largely untested)
8622 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
8623 ASCII string. This was handled independently in various places before.
8626 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
8627 UTF8 strings a character at a time.
8630 *) Use client_version from client hello to select the protocol
8631 (s23_srvr.c) and for RSA client key exchange verification
8632 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
8635 *) Add various utility functions to handle SPKACs, these were previously
8636 handled by poking round in the structure internals. Added new function
8637 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
8638 print, verify and generate SPKACs. Based on an original idea from
8639 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
8642 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
8645 *) Allow the config file extension section to be overwritten on the
8646 command line. Based on an original idea from Massimiliano Pala
8647 <madwolf@comune.modena.it>. The new option is called -extensions
8648 and can be applied to ca, req and x509. Also -reqexts to override
8649 the request extensions in req and -crlexts to override the crl extensions
8653 *) Add new feature to the SPKAC handling in ca. Now you can include
8654 the same field multiple times by preceding it by "XXXX." for example:
8657 this is the same syntax as used in the req config file.
8660 *) Allow certificate extensions to be added to certificate requests. These
8661 are specified in a 'req_extensions' option of the req section of the
8662 config file. They can be printed out with the -text option to req but
8663 are otherwise ignored at present.
8666 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
8667 data read consists of only the final block it would not decrypted because
8668 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
8669 A misplaced 'break' also meant the decrypted final block might not be
8670 copied until the next read.
8673 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
8674 a few extra parameters to the DH structure: these will be useful if
8675 for example we want the value of 'q' or implement X9.42 DH.
8678 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
8679 provides hooks that allow the default DSA functions or functions on a
8680 "per key" basis to be replaced. This allows hardware acceleration and
8681 hardware key storage to be handled without major modification to the
8682 library. Also added low level modexp hooks and CRYPTO_EX structure and
8683 associated functions.
8686 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
8687 as "read only": it can't be written to and the buffer it points to will
8688 not be freed. Reading from a read only BIO is much more efficient than
8689 a normal memory BIO. This was added because there are several times when
8690 an area of memory needs to be read from a BIO. The previous method was
8691 to create a memory BIO and write the data to it, this results in two
8692 copies of the data and an O(n^2) reading algorithm. There is a new
8693 function BIO_new_mem_buf() which creates a read only memory BIO from
8694 an area of memory. Also modified the PKCS#7 routines to use read only
8698 *) Bugfix: ssl23_get_client_hello did not work properly when called in
8699 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
8700 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
8701 but a retry condition occured while trying to read the rest.
8704 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
8705 NID_pkcs7_encrypted by default: this was wrong since this should almost
8706 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
8707 the encrypted data type: this is a more sensible place to put it and it
8708 allows the PKCS#12 code to be tidied up that duplicated this
8712 *) Changed obj_dat.pl script so it takes its input and output files on
8713 the command line. This should avoid shell escape redirection problems
8717 *) Initial support for certificate extension requests, these are included
8718 in things like Xenroll certificate requests. Included functions to allow
8719 extensions to be obtained and added.
8722 *) -crlf option to s_client and s_server for sending newlines as
8723 CRLF (as required by many protocols).
8726 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
8728 *) Install libRSAglue.a when OpenSSL is built with RSAref.
8729 [Ralf S. Engelschall]
8731 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
8732 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
8734 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
8738 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
8739 DH parameters/keys (q is lost during that conversion, but the resulting
8740 DH parameters contain its length).
8742 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
8743 much faster than DH_generate_parameters (which creates parameters
8744 where p = 2*q + 1), and also the smaller q makes DH computations
8745 much more efficient (160-bit exponentiation instead of 1024-bit
8746 exponentiation); so this provides a convenient way to support DHE
8747 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
8748 utter importance to use
8749 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8751 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8752 when such DH parameters are used, because otherwise small subgroup
8753 attacks may become possible!
8756 *) Avoid memory leak in i2d_DHparams.
8759 *) Allow the -k option to be used more than once in the enc program:
8760 this allows the same encrypted message to be read by multiple recipients.
8763 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
8764 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
8765 it will always use the numerical form of the OID, even if it has a short
8769 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
8770 method only got called if p,q,dmp1,dmq1,iqmp components were present,
8771 otherwise bn_mod_exp was called. In the case of hardware keys for example
8772 no private key components need be present and it might store extra data
8773 in the RSA structure, which cannot be accessed from bn_mod_exp.
8774 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
8775 private key operations.
8778 *) Added support for SPARC Linux.
8781 *) pem_password_cb function type incompatibly changed from
8782 typedef int pem_password_cb(char *buf, int size, int rwflag);
8784 ....(char *buf, int size, int rwflag, void *userdata);
8785 so that applications can pass data to their callbacks:
8786 The PEM[_ASN1]_{read,write}... functions and macros now take an
8787 additional void * argument, which is just handed through whenever
8788 the password callback is called.
8789 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
8791 New function SSL_CTX_set_default_passwd_cb_userdata.
8793 Compatibility note: As many C implementations push function arguments
8794 onto the stack in reverse order, the new library version is likely to
8795 interoperate with programs that have been compiled with the old
8796 pem_password_cb definition (PEM_whatever takes some data that
8797 happens to be on the stack as its last argument, and the callback
8798 just ignores this garbage); but there is no guarantee whatsoever that
8801 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
8802 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
8803 problems not only on Windows, but also on some Unix platforms.
8804 To avoid problematic command lines, these definitions are now in an
8805 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
8806 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
8809 *) MIPS III/IV assembler module is reimplemented.
8812 *) More DES library cleanups: remove references to srand/rand and
8813 delete an unused file.
8816 *) Add support for the the free Netwide assembler (NASM) under Win32,
8817 since not many people have MASM (ml) and it can be hard to obtain.
8818 This is currently experimental but it seems to work OK and pass all
8819 the tests. Check out INSTALL.W32 for info.
8822 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
8823 without temporary keys kept an extra copy of the server key,
8824 and connections with temporary keys did not free everything in case
8828 *) New function RSA_check_key and new openssl rsa option -check
8829 for verifying the consistency of RSA keys.
8830 [Ulf Moeller, Bodo Moeller]
8832 *) Various changes to make Win32 compile work:
8833 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
8834 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
8835 comparison" warnings.
8836 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
8839 *) Add a debugging option to PKCS#5 v2 key generation function: when
8840 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
8841 derived keys are printed to stderr.
8844 *) Copy the flags in ASN1_STRING_dup().
8845 [Roman E. Pavlov <pre@mo.msk.ru>]
8847 *) The x509 application mishandled signing requests containing DSA
8848 keys when the signing key was also DSA and the parameters didn't match.
8850 It was supposed to omit the parameters when they matched the signing key:
8851 the verifying software was then supposed to automatically use the CA's
8852 parameters if they were absent from the end user certificate.
8854 Omitting parameters is no longer recommended. The test was also
8855 the wrong way round! This was probably due to unusual behaviour in
8856 EVP_cmp_parameters() which returns 1 if the parameters match.
8857 This meant that parameters were omitted when they *didn't* match and
8858 the certificate was useless. Certificates signed with 'ca' didn't have
8860 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
8862 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
8863 The interface is as follows:
8864 Applications can use
8865 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
8866 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
8867 "off" is now the default.
8868 The library internally uses
8869 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
8870 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
8871 to disable memory-checking temporarily.
8873 Some inconsistent states that previously were possible (and were
8874 even the default) are now avoided.
8876 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
8877 with each memory chunk allocated; this is occasionally more helpful
8878 than just having a counter.
8880 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
8882 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
8886 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
8887 which largely parallels "options", but is for changing API behaviour,
8888 whereas "options" are about protocol behaviour.
8889 Initial "mode" flags are:
8891 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
8892 a single record has been written.
8893 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
8894 retries use the same buffer location.
8895 (But all of the contents must be
8899 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
8902 *) Fix problems with no-hmac etc.
8903 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
8905 *) New functions RSA_get_default_method(), RSA_set_method() and
8906 RSA_get_method(). These allows replacement of RSA_METHODs without having
8907 to mess around with the internals of an RSA structure.
8910 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
8911 Also really enable memory leak checks in openssl.c and in some
8913 [Chad C. Mulligan, Bodo Moeller]
8915 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
8916 up the length of negative integers. This has now been simplified to just
8917 store the length when it is first determined and use it later, rather
8918 than trying to keep track of where data is copied and updating it to
8920 [Steve Henson, reported by Brien Wheeler
8921 <bwheeler@authentica-security.com>]
8923 *) Add a new function PKCS7_signatureVerify. This allows the verification
8924 of a PKCS#7 signature but with the signing certificate passed to the
8925 function itself. This contrasts with PKCS7_dataVerify which assumes the
8926 certificate is present in the PKCS#7 structure. This isn't always the
8927 case: certificates can be omitted from a PKCS#7 structure and be
8928 distributed by "out of band" means (such as a certificate database).
8931 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
8932 function prototypes in pem.h, also change util/mkdef.pl to add the
8933 necessary function names.
8936 *) mk1mf.pl (used by Windows builds) did not properly read the
8937 options set by Configure in the top level Makefile, and Configure
8938 was not even able to write more than one option correctly.
8939 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
8942 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
8943 file to be loaded from a BIO or FILE pointer. The BIO version will
8944 for example allow memory BIOs to contain config info.
8947 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
8948 Whoever hopes to achieve shared-library compatibility across versions
8949 must use this, not the compile-time macro.
8950 (Exercise 0.9.4: Which is the minimum library version required by
8952 Note: All this applies only to multi-threaded programs, others don't
8956 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
8957 through a BIO pair triggered the default case, i.e.
8958 SSLerr(...,SSL_R_UNKNOWN_STATE).
8961 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
8962 can use the SSL library even if none of the specific BIOs is
8966 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
8967 for the encoded length.
8968 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
8970 *) Add initial documentation of the X509V3 functions.
8973 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
8974 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
8975 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
8976 secure PKCS#8 private key format with a high iteration count.
8979 *) Fix determination of Perl interpreter: A perl or perl5
8980 _directory_ in $PATH was also accepted as the interpreter.
8981 [Ralf S. Engelschall]
8983 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
8984 wrong with it but it was very old and did things like calling
8985 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
8989 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
8990 to use the new extension code.
8993 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
8994 with macros. This should make it easier to change their form, add extra
8995 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
8999 *) Add to configuration table a new entry that can specify an alternative
9000 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
9001 according to Mark Crispin <MRC@Panda.COM>.
9005 *) DES CBC did not update the IV. Weird.
9008 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
9009 Changing the behaviour of the former might break existing programs --
9010 where IV updating is needed, des_ncbc_encrypt can be used.
9013 *) When bntest is run from "make test" it drives bc to check its
9014 calculations, as well as internally checking them. If an internal check
9015 fails, it needs to cause bc to give a non-zero result or make test carries
9016 on without noticing the failure. Fixed.
9019 *) DES library cleanups.
9022 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
9023 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
9024 ciphers. NOTE: although the key derivation function has been verified
9025 against some published test vectors it has not been extensively tested
9026 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
9030 *) Instead of "mkdir -p", which is not fully portable, use new
9031 Perl script "util/mkdir-p.pl".
9034 *) Rewrite the way password based encryption (PBE) is handled. It used to
9035 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
9036 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
9037 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
9038 the 'parameter' field of the AlgorithmIdentifier is passed to the
9039 underlying key generation function so it must do its own ASN1 parsing.
9040 This has also changed the EVP_PBE_CipherInit() function which now has a
9041 'parameter' argument instead of literal salt and iteration count values
9042 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
9045 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
9046 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
9047 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
9048 KEY" because this clashed with PKCS#8 unencrypted string. Since this
9049 value was just used as a "magic string" and not used directly its
9050 value doesn't matter.
9053 *) Introduce some semblance of const correctness to BN. Shame C doesn't
9057 *) "linux-sparc64" configuration (ultrapenguin).
9058 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
9059 "linux-sparc" configuration.
9060 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
9062 *) config now generates no-xxx options for missing ciphers.
9065 *) Support the EBCDIC character set (work in progress).
9066 File ebcdic.c not yet included because it has a different license.
9067 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9069 *) Support BS2000/OSD-POSIX.
9070 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9072 *) Make callbacks for key generation use void * instead of char *.
9075 *) Make S/MIME samples compile (not yet tested).
9078 *) Additional typesafe stacks.
9081 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
9085 Changes between 0.9.3 and 0.9.3a [29 May 1999]
9087 *) New configuration variant "sco5-gcc".
9089 *) Updated some demos.
9090 [Sean O Riordain, Wade Scholine]
9092 *) Add missing BIO_free at exit of pkcs12 application.
9095 *) Fix memory leak in conf.c.
9098 *) Updates for Win32 to assembler version of MD5.
9101 *) Set #! path to perl in apps/der_chop to where we found it
9102 instead of using a fixed path.
9105 *) SHA library changes for irix64-mips4-cc.
9108 *) Improvements for VMS support.
9112 Changes between 0.9.2b and 0.9.3 [24 May 1999]
9114 *) Bignum library bug fix. IRIX 6 passes "make test" now!
9115 This also avoids the problems with SC4.2 and unpatched SC5.
9116 [Andy Polyakov <appro@fy.chalmers.se>]
9118 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
9119 These are required because of the typesafe stack would otherwise break
9120 existing code. If old code used a structure member which used to be STACK
9121 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
9122 sk_num or sk_value it would produce an error because the num, data members
9123 are not present in STACK_OF. Now it just produces a warning. sk_set
9124 replaces the old method of assigning a value to sk_value
9125 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
9126 that does this will no longer work (and should use sk_set instead) but
9127 this could be regarded as a "questionable" behaviour anyway.
9130 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
9131 correctly handle encrypted S/MIME data.
9134 *) Change type of various DES function arguments from des_cblock
9135 (which means, in function argument declarations, pointer to char)
9136 to des_cblock * (meaning pointer to array with 8 char elements),
9137 which allows the compiler to do more typechecking; it was like
9138 that back in SSLeay, but with lots of ugly casts.
9140 Introduce new type const_des_cblock.
9143 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
9144 problems: find RecipientInfo structure that matches recipient certificate
9145 and initialise the ASN1 structures properly based on passed cipher.
9148 *) Belatedly make the BN tests actually check the results.
9151 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
9152 to and from BNs: it was completely broken. New compilation option
9153 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
9154 key elements as negative integers.
9157 *) Reorganize and speed up MD5.
9158 [Andy Polyakov <appro@fy.chalmers.se>]
9161 [Richard Levitte <richard@levitte.org>]
9163 *) New option -out to asn1parse to allow the parsed structure to be
9164 output to a file. This is most useful when combined with the -strparse
9165 option to examine the output of things like OCTET STRINGS.
9168 *) Make SSL library a little more fool-proof by not requiring any longer
9169 that SSL_set_{accept,connect}_state be called before
9170 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
9171 in many applications because usually everything *appeared* to work as
9172 intended anyway -- now it really works as intended).
9175 *) Move openssl.cnf out of lib/.
9178 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
9179 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
9180 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
9181 [Ralf S. Engelschall]
9183 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
9184 handle PKCS#7 enveloped data properly.
9185 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
9187 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
9188 copying pointers. The cert_st handling is changed by this in
9189 various ways (and thus what used to be known as ctx->default_cert
9190 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
9191 any longer when s->cert does not give us what we need).
9192 ssl_cert_instantiate becomes obsolete by this change.
9193 As soon as we've got the new code right (possibly it already is?),
9194 we have solved a couple of bugs of the earlier code where s->cert
9195 was used as if it could not have been shared with other SSL structures.
9197 Note that using the SSL API in certain dirty ways now will result
9198 in different behaviour than observed with earlier library versions:
9199 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
9200 does not influence s as it used to.
9202 In order to clean up things more thoroughly, inside SSL_SESSION
9203 we don't use CERT any longer, but a new structure SESS_CERT
9204 that holds per-session data (if available); currently, this is
9205 the peer's certificate chain and, for clients, the server's certificate
9206 and temporary key. CERT holds only those values that can have
9207 meaningful defaults in an SSL_CTX.
9210 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
9211 from the internal representation. Various PKCS#7 fixes: remove some
9212 evil casts and set the enc_dig_alg field properly based on the signing
9216 *) Allow PKCS#12 password to be set from the command line or the
9217 environment. Let 'ca' get its config file name from the environment
9218 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
9222 *) Allow certificate policies extension to use an IA5STRING for the
9223 organization field. This is contrary to the PKIX definition but
9224 VeriSign uses it and IE5 only recognises this form. Document 'x509'
9228 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
9229 without disallowing inline assembler and the like for non-pedantic builds.
9232 *) Support Borland C++ builder.
9233 [Janez Jere <jj@void.si>, modified by Ulf Möller]
9238 *) SHA-1 cleanups and performance enhancements.
9239 [Andy Polyakov <appro@fy.chalmers.se>]
9241 *) Sparc v8plus assembler for the bignum library.
9242 [Andy Polyakov <appro@fy.chalmers.se>]
9244 *) Accept any -xxx and +xxx compiler options in Configure.
9247 *) Update HPUX configuration.
9250 *) Add missing sk_<type>_unshift() function to safestack.h
9251 [Ralf S. Engelschall]
9253 *) New function SSL_CTX_use_certificate_chain_file that sets the
9254 "extra_cert"s in addition to the certificate. (This makes sense
9255 only for "PEM" format files, as chains as a whole are not
9259 *) Support verify_depth from the SSL API.
9260 x509_vfy.c had what can be considered an off-by-one-error:
9261 Its depth (which was not part of the external interface)
9262 was actually counting the number of certificates in a chain;
9263 now it really counts the depth.
9266 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
9267 instead of X509err, which often resulted in confusing error
9268 messages since the error codes are not globally unique
9269 (e.g. an alleged error in ssl3_accept when a certificate
9270 didn't match the private key).
9272 *) New function SSL_CTX_set_session_id_context that allows to set a default
9273 value (so that you don't need SSL_set_session_id_context for each
9274 connection using the SSL_CTX).
9277 *) OAEP decoding bug fix.
9280 *) Support INSTALL_PREFIX for package builders, as proposed by
9284 *) New Configure options "threads" and "no-threads". For systems
9285 where the proper compiler options are known (currently Solaris
9286 and Linux), "threads" is the default.
9289 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
9292 *) Install various scripts to $(OPENSSLDIR)/misc, not to
9293 $(INSTALLTOP)/bin -- they shouldn't clutter directories
9294 such as /usr/local/bin.
9297 *) "make linux-shared" to build shared libraries.
9298 [Niels Poppe <niels@netbox.org>]
9300 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
9303 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
9304 extension adding in x509 utility.
9307 *) Remove NOPROTO sections and error code comments.
9310 *) Partial rewrite of the DEF file generator to now parse the ANSI
9314 *) New Configure options --prefix=DIR and --openssldir=DIR.
9317 *) Complete rewrite of the error code script(s). It is all now handled
9318 by one script at the top level which handles error code gathering,
9319 header rewriting and C source file generation. It should be much better
9320 than the old method: it now uses a modified version of Ulf's parser to
9321 read the ANSI prototypes in all header files (thus the old K&R definitions
9322 aren't needed for error creation any more) and do a better job of
9323 translating function codes into names. The old 'ASN1 error code imbedded
9324 in a comment' is no longer necessary and it doesn't use .err files which
9325 have now been deleted. Also the error code call doesn't have to appear all
9326 on one line (which resulted in some large lines...).
9329 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
9332 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
9333 0 (which usually indicates a closed connection), but continue reading.
9336 *) Fix some race conditions.
9339 *) Add support for CRL distribution points extension. Add Certificate
9340 Policies and CRL distribution points documentation.
9343 *) Move the autogenerated header file parts to crypto/opensslconf.h.
9346 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
9347 8 of keying material. Merlin has also confirmed interop with this fix
9348 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
9349 [Merlin Hughes <merlin@baltimore.ie>]
9351 *) Fix lots of warnings.
9352 [Richard Levitte <levitte@stacken.kth.se>]
9354 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
9355 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
9356 [Richard Levitte <levitte@stacken.kth.se>]
9358 *) Fix problems with sizeof(long) == 8.
9359 [Andy Polyakov <appro@fy.chalmers.se>]
9361 *) Change functions to ANSI C.
9364 *) Fix typos in error codes.
9365 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
9367 *) Remove defunct assembler files from Configure.
9370 *) SPARC v8 assembler BIGNUM implementation.
9371 [Andy Polyakov <appro@fy.chalmers.se>]
9373 *) Support for Certificate Policies extension: both print and set.
9374 Various additions to support the r2i method this uses.
9377 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
9378 return a const string when you are expecting an allocated buffer.
9381 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
9382 types DirectoryString and DisplayText.
9385 *) Add code to allow r2i extensions to access the configuration database,
9386 add an LHASH database driver and add several ctx helper functions.
9389 *) Fix an evil bug in bn_expand2() which caused various BN functions to
9390 fail when they extended the size of a BIGNUM.
9393 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
9394 support typesafe stack.
9397 *) Fix typo in SSL_[gs]et_options().
9398 [Nils Frostberg <nils@medcom.se>]
9400 *) Delete various functions and files that belonged to the (now obsolete)
9401 old X509V3 handling code.
9404 *) New Configure option "rsaref".
9407 *) Don't auto-generate pem.h.
9410 *) Introduce type-safe ASN.1 SETs.
9413 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
9414 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
9416 *) Introduce type-safe STACKs. This will almost certainly break lots of code
9417 that links with OpenSSL (well at least cause lots of warnings), but fear
9418 not: the conversion is trivial, and it eliminates loads of evil casts. A
9419 few STACKed things have been converted already. Feel free to convert more.
9420 In the fullness of time, I'll do away with the STACK type altogether.
9423 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
9424 specified in <certfile> by updating the entry in the index.txt file.
9425 This way one no longer has to edit the index.txt file manually for
9426 revoking a certificate. The -revoke option does the gory details now.
9427 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
9429 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
9430 `-text' option at all and this way the `-noout -text' combination was
9431 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
9432 [Ralf S. Engelschall]
9434 *) Make sure a corresponding plain text error message exists for the
9435 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
9436 verify callback function determined that a certificate was revoked.
9437 [Ralf S. Engelschall]
9439 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
9440 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
9441 all available cipers including rc5, which was forgotten until now.
9442 In order to let the testing shell script know which algorithms
9443 are available, a new (up to now undocumented) command
9444 "openssl list-cipher-commands" is used.
9447 *) Bugfix: s_client occasionally would sleep in select() when
9448 it should have checked SSL_pending() first.
9451 *) New functions DSA_do_sign and DSA_do_verify to provide access to
9452 the raw DSA values prior to ASN.1 encoding.
9455 *) Tweaks to Configure
9456 [Niels Poppe <niels@netbox.org>]
9458 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
9462 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
9465 *) New config option to avoid instructions that are illegal on the 80386.
9466 The default code is faster, but requires at least a 486.
9469 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
9470 SSL2_SERVER_VERSION (not used at all) macros, which are now the
9471 same as SSL2_VERSION anyway.
9474 *) New "-showcerts" option for s_client.
9477 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
9478 application. Various cleanups and fixes.
9481 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
9482 modify error routines to work internally. Add error codes and PBE init
9483 to library startup routines.
9486 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
9487 packing functions to asn1 and evp. Changed function names and error
9488 codes along the way.
9491 *) PKCS12 integration: and so it begins... First of several patches to
9492 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
9493 objects to objects.h
9496 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
9497 and display support for Thawte strong extranet extension.
9500 *) Add LinuxPPC support.
9501 [Jeff Dubrule <igor@pobox.org>]
9503 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
9504 bn_div_words in alpha.s.
9505 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
9507 *) Make sure the RSA OAEP test is skipped under -DRSAref because
9508 OAEP isn't supported when OpenSSL is built with RSAref.
9509 [Ulf Moeller <ulf@fitug.de>]
9511 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
9512 so they no longer are missing under -DNOPROTO.
9513 [Soren S. Jorvang <soren@t.dk>]
9516 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
9518 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
9519 doesn't work when the session is reused. Coming soon!
9522 *) Fix a security hole, that allows sessions to be reused in the wrong
9523 context thus bypassing client cert protection! All software that uses
9524 client certs and session caches in multiple contexts NEEDS PATCHING to
9525 allow session reuse! A fuller solution is in the works.
9526 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
9528 *) Some more source tree cleanups (removed obsolete files
9529 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
9530 permission on "config" script to be executable) and a fix for the INSTALL
9532 [Ulf Moeller <ulf@fitug.de>]
9534 *) Remove some legacy and erroneous uses of malloc, free instead of
9536 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
9538 *) Make rsa_oaep_test return non-zero on error.
9539 [Ulf Moeller <ulf@fitug.de>]
9541 *) Add support for native Solaris shared libraries. Configure
9542 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
9543 if someone would make that last step automatic.
9544 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
9546 *) ctx_size was not built with the right compiler during "make links". Fixed.
9549 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
9550 except NULL ciphers". This means the default cipher list will no longer
9551 enable NULL ciphers. They need to be specifically enabled e.g. with
9552 the string "DEFAULT:eNULL".
9555 *) Fix to RSA private encryption routines: if p < q then it would
9556 occasionally produce an invalid result. This will only happen with
9557 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9560 *) Be less restrictive and allow also `perl util/perlpath.pl
9561 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9562 because this way one can also use an interpreter named `perl5' (which is
9563 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9564 installed as `perl').
9565 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9567 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9568 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9570 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9571 advapi32.lib to Win32 build and change the pem test comparision
9572 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9573 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9574 and crypto/des/ede_cbcm_enc.c.
9577 *) DES quad checksum was broken on big-endian architectures. Fixed.
9580 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9581 Win32 test batch file so it (might) work again. The Win32 test batch file
9582 is horrible: I feel ill....
9585 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
9586 in e_os.h. Audit of header files to check ANSI and non ANSI
9587 sections: 10 functions were absent from non ANSI section and not exported
9588 from Windows DLLs. Fixed up libeay.num for new functions.
9591 *) Make `openssl version' output lines consistent.
9592 [Ralf S. Engelschall]
9594 *) Fix Win32 symbol export lists for BIO functions: Added
9595 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
9596 to ms/libeay{16,32}.def.
9597 [Ralf S. Engelschall]
9599 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
9600 fine under Unix and passes some trivial tests I've now added. But the
9601 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
9602 added to make sure no one expects that this stuff really works in the
9603 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
9604 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
9606 [Ralf S. Engelschall]
9608 *) Fix the generation of two part addresses in perl.
9609 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
9611 *) Add config entry for Linux on MIPS.
9612 [John Tobey <jtobey@channel1.com>]
9614 *) Make links whenever Configure is run, unless we are on Windoze.
9617 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
9618 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
9622 *) Add a useful kludge to allow package maintainers to specify compiler and
9623 other platforms details on the command line without having to patch the
9624 Configure script everytime: One now can use ``perl Configure
9625 <id>:<details>'', i.e. platform ids are allowed to have details appended
9626 to them (seperated by colons). This is treated as there would be a static
9627 pre-configured entry in Configure's %table under key <id> with value
9628 <details> and ``perl Configure <id>'' is called. So, when you want to
9629 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
9630 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
9631 now, which overrides the FreeBSD-elf entry on-the-fly.
9632 [Ralf S. Engelschall]
9634 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
9637 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
9638 on the `perl Configure ...' command line. This way one can compile
9639 OpenSSL libraries with Position Independent Code (PIC) which is needed
9640 for linking it into DSOs.
9641 [Ralf S. Engelschall]
9643 *) Remarkably, export ciphers were totally broken and no-one had noticed!
9647 *) Cleaned up the LICENSE document: The official contact for any license
9648 questions now is the OpenSSL core team under openssl-core@openssl.org.
9649 And add a paragraph about the dual-license situation to make sure people
9650 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
9651 to the OpenSSL toolkit.
9652 [Ralf S. Engelschall]
9654 *) General source tree makefile cleanups: Made `making xxx in yyy...'
9655 display consistent in the source tree and replaced `/bin/rm' by `rm'.
9656 Additonally cleaned up the `make links' target: Remove unnecessary
9657 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
9658 to speed processing and no longer clutter the display with confusing
9659 stuff. Instead only the actually done links are displayed.
9660 [Ralf S. Engelschall]
9662 *) Permit null encryption ciphersuites, used for authentication only. It used
9663 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
9664 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
9668 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
9669 signed attributes when verifying signatures (this would break them),
9670 the detached data encoding was wrong and public keys obtained using
9671 X509_get_pubkey() weren't freed.
9674 *) Add text documentation for the BUFFER functions. Also added a work around
9675 to a Win95 console bug. This was triggered by the password read stuff: the
9676 last character typed gets carried over to the next fread(). If you were
9677 generating a new cert request using 'req' for example then the last
9678 character of the passphrase would be CR which would then enter the first
9682 *) Added the new `Includes OpenSSL Cryptography Software' button as
9683 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
9684 button and can be used by applications based on OpenSSL to show the
9685 relationship to the OpenSSL project.
9686 [Ralf S. Engelschall]
9688 *) Remove confusing variables in function signatures in files
9689 ssl/ssl_lib.c and ssl/ssl.h.
9690 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9692 *) Don't install bss_file.c under PREFIX/include/
9693 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9695 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
9696 functions that return function pointers and has support for NT specific
9697 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
9698 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
9699 unsigned to signed types: this was killing the Win32 compile.
9702 *) Add new certificate file to stack functions,
9703 SSL_add_dir_cert_subjects_to_stack() and
9704 SSL_add_file_cert_subjects_to_stack(). These largely supplant
9705 SSL_load_client_CA_file(), and can be used to add multiple certs easily
9706 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
9707 This means that Apache-SSL and similar packages don't have to mess around
9708 to add as many CAs as they want to the preferred list.
9711 *) Experiment with doxygen documentation. Currently only partially applied to
9713 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
9714 openssl.doxy as the configuration file.
9717 *) Get rid of remaining C++-style comments which strict C compilers hate.
9718 [Ralf S. Engelschall, pointed out by Carlos Amengual]
9720 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
9721 compiled in by default: it has problems with large keys.
9724 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
9725 DH private keys and/or callback functions which directly correspond to
9726 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
9727 is needed for applications which have to configure certificates on a
9728 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
9730 For the RSA certificate situation is makes no difference, but
9731 for the DSA certificate situation this fixes the "no shared cipher"
9732 problem where the OpenSSL cipher selection procedure failed because the
9733 temporary keys were not overtaken from the context and the API provided
9734 no way to reconfigure them.
9735 The new functions now let applications reconfigure the stuff and they
9736 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
9737 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
9738 non-public-API function ssl_cert_instantiate() is used as a helper
9739 function and also to reduce code redundancy inside ssl_rsa.c.
9740 [Ralf S. Engelschall]
9742 *) Move s_server -dcert and -dkey options out of the undocumented feature
9743 area because they are useful for the DSA situation and should be
9744 recognized by the users.
9745 [Ralf S. Engelschall]
9747 *) Fix the cipher decision scheme for export ciphers: the export bits are
9748 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
9749 SSL_EXP_MASK. So, the original variable has to be used instead of the
9750 already masked variable.
9751 [Richard Levitte <levitte@stacken.kth.se>]
9753 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
9754 [Richard Levitte <levitte@stacken.kth.se>]
9756 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
9757 from `int' to `unsigned int' because it's a length and initialized by
9758 EVP_DigestFinal() which expects an `unsigned int *'.
9759 [Richard Levitte <levitte@stacken.kth.se>]
9761 *) Don't hard-code path to Perl interpreter on shebang line of Configure
9762 script. Instead use the usual Shell->Perl transition trick.
9763 [Ralf S. Engelschall]
9765 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
9766 (in addition to RSA certificates) to match the behaviour of `openssl dsa
9767 -noout -modulus' as it's already the case for `openssl rsa -noout
9768 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
9769 currently the public key is printed (a decision which was already done by
9770 `openssl dsa -modulus' in the past) which serves a similar purpose.
9771 Additionally the NO_RSA no longer completely removes the whole -modulus
9772 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
9774 [Ralf S. Engelschall]
9776 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
9777 BIO. See the source (crypto/evp/bio_ok.c) for more info.
9778 [Arne Ansper <arne@ats.cyber.ee>]
9780 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
9781 to be added. Now both 'req' and 'ca' can use new objects defined in the
9785 *) Add cool BIO that does syslog (or event log on NT).
9786 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
9788 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
9789 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
9790 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
9791 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
9794 *) Add preliminary config info for new extension code.
9797 *) Make RSA_NO_PADDING really use no padding.
9798 [Ulf Moeller <ulf@fitug.de>]
9800 *) Generate errors when private/public key check is done.
9803 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
9804 for some CRL extensions and new objects added.
9807 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
9808 key usage extension and fuller support for authority key id.
9811 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
9812 padding method for RSA, which is recommended for new applications in PKCS
9813 #1 v2.0 (RFC 2437, October 1998).
9814 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
9815 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
9816 against Bleichbacher's attack on RSA.
9817 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
9820 *) Updates to the new SSL compression code
9821 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9823 *) Fix so that the version number in the master secret, when passed
9824 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
9825 (because the server will not accept higher), that the version number
9826 is 0x03,0x01, not 0x03,0x00
9827 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9829 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
9830 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
9831 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
9834 *) Support for RAW extensions where an arbitrary extension can be
9835 created by including its DER encoding. See apps/openssl.cnf for
9839 *) Make sure latest Perl versions don't interpret some generated C array
9840 code as Perl array code in the crypto/err/err_genc.pl script.
9841 [Lars Weber <3weber@informatik.uni-hamburg.de>]
9843 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
9844 not many people have the assembler. Various Win32 compilation fixes and
9845 update to the INSTALL.W32 file with (hopefully) more accurate Win32
9849 *) Modify configure script 'Configure' to automatically create crypto/date.h
9850 file under Win32 and also build pem.h from pem.org. New script
9851 util/mkfiles.pl to create the MINFO file on environments that can't do a
9852 'make files': perl util/mkfiles.pl >MINFO should work.
9855 *) Major rework of DES function declarations, in the pursuit of correctness
9856 and purity. As a result, many evil casts evaporated, and some weirdness,
9857 too. You may find this causes warnings in your code. Zapping your evil
9858 casts will probably fix them. Mostly.
9861 *) Fix for a typo in asn1.h. Bug fix to object creation script
9862 obj_dat.pl. It considered a zero in an object definition to mean
9863 "end of object": none of the objects in objects.h have any zeros
9864 so it wasn't spotted.
9865 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
9867 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
9868 Masking (CBCM). In the absence of test vectors, the best I have been able
9869 to do is check that the decrypt undoes the encrypt, so far. Send me test
9870 vectors if you have them.
9873 *) Correct calculation of key length for export ciphers (too much space was
9874 allocated for null ciphers). This has not been tested!
9877 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
9878 message is now correct (it understands "crypto" and "ssl" on its
9879 command line). There is also now an "update" option. This will update
9880 the util/ssleay.num and util/libeay.num files with any new functions.
9882 perl util/mkdef.pl crypto ssl update
9883 it will update them.
9886 *) Overhauled the Perl interface (perl/*):
9887 - ported BN stuff to OpenSSL's different BN library
9888 - made the perl/ source tree CVS-aware
9889 - renamed the package from SSLeay to OpenSSL (the files still contain
9890 their history because I've copied them in the repository)
9891 - removed obsolete files (the test scripts will be replaced
9892 by better Test::Harness variants in the future)
9893 [Ralf S. Engelschall]
9895 *) First cut for a very conservative source tree cleanup:
9896 1. merge various obsolete readme texts into doc/ssleay.txt
9897 where we collect the old documents and readme texts.
9898 2. remove the first part of files where I'm already sure that we no
9899 longer need them because of three reasons: either they are just temporary
9900 files which were left by Eric or they are preserved original files where
9901 I've verified that the diff is also available in the CVS via "cvs diff
9902 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
9903 the crypto/md/ stuff).
9904 [Ralf S. Engelschall]
9906 *) More extension code. Incomplete support for subject and issuer alt
9907 name, issuer and authority key id. Change the i2v function parameters
9908 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
9909 what that's for :-) Fix to ASN1 macro which messed up
9910 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
9913 *) Preliminary support for ENUMERATED type. This is largely copied from the
9917 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
9918 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9920 *) Make sure `make rehash' target really finds the `openssl' program.
9921 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9923 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
9924 like to hear about it if this slows down other processors.
9927 *) Add CygWin32 platform information to Configure script.
9928 [Alan Batie <batie@aahz.jf.intel.com>]
9930 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
9931 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
9933 *) New program nseq to manipulate netscape certificate sequences
9936 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
9940 *) Fixes to BN code. Previously the default was to define BN_RECURSION
9941 but the BN code had some problems that would cause failures when
9942 doing certificate verification and some other functions.
9943 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9945 *) Add ASN1 and PEM code to support netscape certificate sequences.
9948 *) Add ASN1 and PEM code to support netscape certificate sequences.
9951 *) Add several PKIX and private extended key usage OIDs.
9954 *) Modify the 'ca' program to handle the new extension code. Modify
9955 openssl.cnf for new extension format, add comments.
9958 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
9959 and add a sample to openssl.cnf so req -x509 now adds appropriate
9963 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
9964 error code, add initial support to X509_print() and x509 application.
9967 *) Takes a deep breath and start addding X509 V3 extension support code. Add
9968 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
9969 stuff is currently isolated and isn't even compiled yet.
9972 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
9973 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
9974 Removed the versions check from X509 routines when loading extensions:
9975 this allows certain broken certificates that don't set the version
9976 properly to be processed.
9979 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
9980 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
9981 can still be regenerated with "make depend".
9984 *) Spelling mistake in C version of CAST-128.
9985 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
9987 *) Changes to the error generation code. The perl script err-code.pl
9988 now reads in the old error codes and retains the old numbers, only
9989 adding new ones if necessary. It also only changes the .err files if new
9990 codes are added. The makefiles have been modified to only insert errors
9991 when needed (to avoid needlessly modifying header files). This is done
9992 by only inserting errors if the .err file is newer than the auto generated
9993 C file. To rebuild all the error codes from scratch (the old behaviour)
9994 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
9995 or delete all the .err files.
9998 *) CAST-128 was incorrectly implemented for short keys. The C version has
9999 been fixed, but is untested. The assembler versions are also fixed, but
10000 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
10001 to regenerate it if needed.
10002 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
10003 Hagino <itojun@kame.net>]
10005 *) File was opened incorrectly in randfile.c.
10006 [Ulf Möller <ulf@fitug.de>]
10008 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
10009 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
10010 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
10011 al: it's just almost always a UTCTime. Note this patch adds new error
10012 codes so do a "make errors" if there are problems.
10015 *) Correct Linux 1 recognition in config.
10016 [Ulf Möller <ulf@fitug.de>]
10018 *) Remove pointless MD5 hash when using DSA keys in ca.
10019 [Anonymous <nobody@replay.com>]
10021 *) Generate an error if given an empty string as a cert directory. Also
10022 generate an error if handed NULL (previously returned 0 to indicate an
10023 error, but didn't set one).
10024 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
10026 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
10029 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
10030 parameters. This was causing a warning which killed off the Win32 compile.
10033 *) Remove C++ style comments from crypto/bn/bn_local.h.
10034 [Neil Costigan <neil.costigan@celocom.com>]
10036 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
10037 based on a text string, looking up short and long names and finally
10038 "dot" format. The "dot" format stuff didn't work. Added new function
10039 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
10040 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
10041 OID is not part of the table.
10044 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
10045 X509_LOOKUP_by_alias().
10048 *) Sort openssl functions by name.
10051 *) Get the gendsa program working (hopefully) and add it to app list. Remove
10052 encryption from sample DSA keys (in case anyone is interested the password
10056 *) Make _all_ *_free functions accept a NULL pointer.
10057 [Frans Heymans <fheymans@isaserver.be>]
10059 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
10061 [Anonymous <nobody@replay.com>]
10063 *) s_server should send the CAfile as acceptable CAs, not its own cert.
10064 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10066 *) Don't blow it for numeric -newkey arguments to apps/req.
10067 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10069 *) Temp key "for export" tests were wrong in s3_srvr.c.
10070 [Anonymous <nobody@replay.com>]
10072 *) Add prototype for temp key callback functions
10073 SSL_CTX_set_tmp_{rsa,dh}_callback().
10076 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
10077 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
10080 *) X509_name_add_entry() freed the wrong thing after an error.
10081 [Arne Ansper <arne@ats.cyber.ee>]
10083 *) rsa_eay.c would attempt to free a NULL context.
10084 [Arne Ansper <arne@ats.cyber.ee>]
10086 *) BIO_s_socket() had a broken should_retry() on Windoze.
10087 [Arne Ansper <arne@ats.cyber.ee>]
10089 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
10090 [Arne Ansper <arne@ats.cyber.ee>]
10092 *) Make sure the already existing X509_STORE->depth variable is initialized
10093 in X509_STORE_new(), but document the fact that this variable is still
10094 unused in the certificate verification process.
10095 [Ralf S. Engelschall]
10097 *) Fix the various library and apps files to free up pkeys obtained from
10098 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
10101 *) Fix reference counting in X509_PUBKEY_get(). This makes
10102 demos/maurice/example2.c work, amongst others, probably.
10103 [Steve Henson and Ben Laurie]
10105 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
10106 `openssl' and second, the shortcut symlinks for the `openssl <command>'
10107 are no longer created. This way we have a single and consistent command
10108 line interface `openssl <command>', similar to `cvs <command>'.
10109 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
10111 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
10112 BIT STRING wrapper always have zero unused bits.
10115 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
10118 *) Make the top-level INSTALL documentation easier to understand.
10121 *) Makefiles updated to exit if an error occurs in a sub-directory
10122 make (including if user presses ^C) [Paul Sutton]
10124 *) Make Montgomery context stuff explicit in RSA data structure.
10127 *) Fix build order of pem and err to allow for generated pem.h.
10130 *) Fix renumbering bug in X509_NAME_delete_entry().
10133 *) Enhanced the err-ins.pl script so it makes the error library number
10134 global and can add a library name. This is needed for external ASN1 and
10135 other error libraries.
10138 *) Fixed sk_insert which never worked properly.
10141 *) Fix ASN1 macros so they can handle indefinite length construted
10142 EXPLICIT tags. Some non standard certificates use these: they can now
10146 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
10147 into a single doc/ssleay.txt bundle. This way the information is still
10148 preserved but no longer messes up this directory. Now it's new room for
10149 the new set of documenation files.
10150 [Ralf S. Engelschall]
10152 *) SETs were incorrectly DER encoded. This was a major pain, because they
10153 shared code with SEQUENCEs, which aren't coded the same. This means that
10154 almost everything to do with SETs or SEQUENCEs has either changed name or
10155 number of arguments.
10156 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
10158 *) Fix test data to work with the above.
10161 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
10162 was already fixed by Eric for 0.9.1 it seems.
10163 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
10165 *) Autodetect FreeBSD3.
10168 *) Fix various bugs in Configure. This affects the following platforms:
10172 unixware-2.0-pentium
10176 *) Eliminate generated files from CVS. Reorder tests to regenerate files
10177 before they are needed.
10180 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
10184 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
10186 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
10187 changed SSLeay to OpenSSL in version strings.
10188 [Ralf S. Engelschall]
10190 *) Some fixups to the top-level documents.
10193 *) Fixed the nasty bug where rsaref.h was not found under compile-time
10194 because the symlink to include/ was missing.
10195 [Ralf S. Engelschall]
10197 *) Incorporated the popular no-RSA/DSA-only patches
10198 which allow to compile a RSA-free SSLeay.
10199 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
10201 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
10202 when "ssleay" is still not found.
10203 [Ralf S. Engelschall]
10205 *) Added more platforms to Configure: Cray T3E, HPUX 11,
10206 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
10208 *) Updated the README file.
10209 [Ralf S. Engelschall]
10211 *) Added various .cvsignore files in the CVS repository subdirs
10212 to make a "cvs update" really silent.
10213 [Ralf S. Engelschall]
10215 *) Recompiled the error-definition header files and added
10216 missing symbols to the Win32 linker tables.
10217 [Ralf S. Engelschall]
10219 *) Cleaned up the top-level documents;
10220 o new files: CHANGES and LICENSE
10221 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
10222 o merged COPYRIGHT into LICENSE
10223 o removed obsolete TODO file
10224 o renamed MICROSOFT to INSTALL.W32
10225 [Ralf S. Engelschall]
10227 *) Removed dummy files from the 0.9.1b source tree:
10228 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
10229 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
10230 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
10231 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
10232 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
10233 [Ralf S. Engelschall]
10235 *) Added various platform portability fixes.
10238 *) The Genesis of the OpenSSL rpject:
10239 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
10240 Young and Tim J. Hudson created while they were working for C2Net until
10242 [The OpenSSL Project]
10245 Changes between 0.9.0b and 0.9.1b [not released]
10247 *) Updated a few CA certificates under certs/
10250 *) Changed some BIGNUM api stuff.
10253 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
10254 DGUX x86, Linux Alpha, etc.
10257 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
10258 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
10262 *) Add -strparse option to asn1pars program which parses nested
10264 [Dr Stephen Henson <shenson@bigfoot.com>]
10266 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
10269 *) DSA fix for "ca" program.
10272 *) Added "-genkey" option to "dsaparam" program.
10275 *) Added RIPE MD160 (rmd160) message digest.
10278 *) Added -a (all) option to "ssleay version" command.
10281 *) Added PLATFORM define which is the id given to Configure.
10284 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
10287 *) Extended the ASN.1 parser routines.
10290 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
10293 *) Added a BN_CTX to the BN library.
10296 *) Fixed the weak key values in DES library
10299 *) Changed API in EVP library for cipher aliases.
10302 *) Added support for RC2/64bit cipher.
10305 *) Converted the lhash library to the crypto/mem.c functions.
10308 *) Added more recognized ASN.1 object ids.
10311 *) Added more RSA padding checks for SSL/TLS.
10314 *) Added BIO proxy/filter functionality.
10317 *) Added extra_certs to SSL_CTX which can be used
10318 send extra CA certificates to the client in the CA cert chain sending
10319 process. It can be configured with SSL_CTX_add_extra_chain_cert().
10322 *) Now Fortezza is denied in the authentication phase because
10323 this is key exchange mechanism is not supported by SSLeay at all.
10326 *) Additional PKCS1 checks.
10329 *) Support the string "TLSv1" for all TLS v1 ciphers.
10332 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
10333 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
10336 *) Fixed a few memory leaks.
10339 *) Fixed various code and comment typos.
10342 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
10343 bytes sent in the client random.
10344 [Edward Bishop <ebishop@spyglass.com>]