5 Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
7 *) Segmentation fault in ASN1_TYPE_cmp fix
9 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
10 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
11 certificate signature algorithm consistency this can be used to crash any
12 certificate verification operation and exploited in a DoS attack. Any
13 application which performs certificate verification is vulnerable including
14 OpenSSL clients and servers which enable client authentication.
18 *) ASN.1 structure reuse memory corruption fix
20 Reusing a structure in ASN.1 parsing may allow an attacker to cause
21 memory corruption via an invalid write. Such reuse is and has been
22 strongly discouraged and is believed to be rare.
24 Applications that parse structures containing CHOICE or ANY DEFINED BY
25 components may be affected. Certificate parsing (d2i_X509 and related
26 functions) are however not affected. OpenSSL clients and servers are
31 *) PKCS7 NULL pointer dereferences fix
33 The PKCS#7 parsing code does not handle missing outer ContentInfo
34 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
35 missing content and trigger a NULL pointer dereference on parsing.
37 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
38 otherwise parse PKCS#7 structures from untrusted sources are
39 affected. OpenSSL clients and servers are not affected.
41 This issue was reported to OpenSSL by Michal Zalewski (Google).
45 *) DoS via reachable assert in SSLv2 servers fix
47 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
48 servers that both support SSLv2 and enable export cipher suites by sending
49 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
51 This issue was discovered by Sean Burford (Google) and Emilia Käsper
52 (OpenSSL development team).
56 *) Use After Free following d2i_ECPrivatekey error fix
58 A malformed EC private key file consumed via the d2i_ECPrivateKey function
59 could cause a use after free condition. This, in turn, could cause a double
60 free in several private key parsing functions (such as d2i_PrivateKey
61 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
62 for applications that receive EC private keys from untrusted
63 sources. This scenario is considered rare.
65 This issue was discovered by the BoringSSL project and fixed in their
70 *) X509_to_X509_REQ NULL pointer deref fix
72 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
73 the certificate key is invalid. This function is rarely used in practice.
75 This issue was discovered by Brian Carpenter.
79 *) Removed the export ciphers from the DEFAULT ciphers
82 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
84 *) Build fixes for the Windows and OpenVMS platforms
85 [Matt Caswell and Richard Levitte]
87 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
89 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
90 message can cause a segmentation fault in OpenSSL due to a NULL pointer
91 dereference. This could lead to a Denial Of Service attack. Thanks to
92 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
96 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
97 dtls1_buffer_record function under certain conditions. In particular this
98 could occur if an attacker sent repeated DTLS records with the same
99 sequence number but for the next epoch. The memory leak could be exploited
100 by an attacker in a Denial of Service attack through memory exhaustion.
101 Thanks to Chris Mueller for reporting this issue.
105 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
106 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
107 method would be set to NULL which could later result in a NULL pointer
108 dereference. Thanks to Frank Schmirler for reporting this issue.
112 *) Abort handshake if server key exchange message is omitted for ephemeral
115 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
116 reporting this issue.
120 *) Remove non-export ephemeral RSA code on client and server. This code
121 violated the TLS standard by allowing the use of temporary RSA keys in
122 non-export ciphersuites and could be used by a server to effectively
123 downgrade the RSA key length used to a value smaller than the server
124 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
125 INRIA or reporting this issue.
129 *) Fixed issue where DH client certificates are accepted without verification.
130 An OpenSSL server will accept a DH certificate for client authentication
131 without the certificate verify message. This effectively allows a client to
132 authenticate without the use of a private key. This only affects servers
133 which trust a client certificate authority which issues certificates
134 containing DH keys: these are extremely rare and hardly ever encountered.
135 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
140 *) Ensure that the session ID context of an SSL is updated when its
141 SSL_CTX is updated via SSL_set_SSL_CTX.
143 The session ID context is typically set from the parent SSL_CTX,
144 and can vary with the CTX.
147 *) Fix various certificate fingerprint issues.
149 By using non-DER or invalid encodings outside the signed portion of a
150 certificate the fingerprint can be changed without breaking the signature.
151 Although no details of the signed portion of the certificate can be changed
152 this can cause problems with some applications: e.g. those using the
153 certificate fingerprint for blacklists.
155 1. Reject signatures with non zero unused bits.
157 If the BIT STRING containing the signature has non zero unused bits reject
158 the signature. All current signature algorithms require zero unused bits.
160 2. Check certificate algorithm consistency.
162 Check the AlgorithmIdentifier inside TBS matches the one in the
163 certificate signature. NB: this will result in signature failure
164 errors for some broken certificates.
166 Thanks to Konrad Kraszewski from Google for reporting this issue.
168 3. Check DSA/ECDSA signatures use DER.
170 Reencode DSA/ECDSA signatures and compare with the original received
171 signature. Return an error if there is a mismatch.
173 This will reject various cases including garbage after signature
174 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
175 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
176 (negative or with leading zeroes).
178 Further analysis was conducted and fixes were developed by Stephen Henson
179 of the OpenSSL core team.
184 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
185 results on some platforms, including x86_64. This bug occurs at random
186 with a very low probability, and is not known to be exploitable in any
187 way, though its exact impact is difficult to determine. Thanks to Pieter
188 Wuille (Blockstream) who reported this issue and also suggested an initial
189 fix. Further analysis was conducted by the OpenSSL development team and
190 Adam Langley of Google. The final fix was developed by Andy Polyakov of
191 the OpenSSL core team.
195 *) Do not resume sessions on the server if the negotiated protocol
196 version does not match the session's version. Resuming with a different
197 version, while not strictly forbidden by the RFC, is of questionable
198 sanity and breaks all known clients.
199 [David Benjamin, Emilia Käsper]
201 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
202 early CCS messages during renegotiation. (Note that because
203 renegotiation is encrypted, this early CCS was not exploitable.)
206 *) Tighten client-side session ticket handling during renegotiation:
207 ensure that the client only accepts a session ticket if the server sends
208 the extension anew in the ServerHello. Previously, a TLS client would
209 reuse the old extension state and thus accept a session ticket if one was
210 announced in the initial ServerHello.
212 Similarly, ensure that the client requires a session ticket if one
213 was advertised in the ServerHello. Previously, a TLS client would
214 ignore a missing NewSessionTicket message.
217 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
221 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
222 sends a carefully crafted handshake message, to cause OpenSSL to fail
223 to free up to 64k of memory causing a memory leak. This could be
224 exploited in a Denial Of Service attack. This issue affects OpenSSL
225 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
226 whether SRTP is used or configured. Implementations of OpenSSL that
227 have been compiled with OPENSSL_NO_SRTP defined are not affected.
229 The fix was developed by the OpenSSL team.
233 *) Session Ticket Memory Leak.
235 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
236 integrity of that ticket is first verified. In the event of a session
237 ticket integrity check failing, OpenSSL will fail to free memory
238 causing a memory leak. By sending a large number of invalid session
239 tickets an attacker could exploit this issue in a Denial Of Service
244 *) Build option no-ssl3 is incomplete.
246 When OpenSSL is configured with "no-ssl3" as a build option, servers
247 could accept and complete a SSL 3.0 handshake, and clients could be
248 configured to send them.
250 [Akamai and the OpenSSL team]
252 *) Add support for TLS_FALLBACK_SCSV.
253 Client applications doing fallback retries should call
254 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
256 [Adam Langley, Bodo Moeller]
258 *) Add additional DigestInfo checks.
260 Reencode DigestInto in DER and check against the original when
261 verifying RSA signature: this will reject any improperly encoded
262 DigestInfo structures.
264 Note: this is a precautionary measure and no attacks are currently known.
268 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
270 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
271 SRP code can be overrun an internal buffer. Add sanity check that
272 g, A, B < N to SRP code.
274 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
275 Group for discovering this issue.
279 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
280 TLS 1.0 instead of higher protocol versions when the ClientHello message
281 is badly fragmented. This allows a man-in-the-middle attacker to force a
282 downgrade to TLS 1.0 even if both the server and the client support a
283 higher protocol version, by modifying the client's TLS records.
285 Thanks to David Benjamin and Adam Langley (Google) for discovering and
286 researching this issue.
290 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
291 to a denial of service attack. A malicious server can crash the client
292 with a null pointer dereference (read) by specifying an anonymous (EC)DH
293 ciphersuite and sending carefully crafted handshake messages.
295 Thanks to Felix Gröbert (Google) for discovering and researching this
300 *) By sending carefully crafted DTLS packets an attacker could cause openssl
301 to leak memory. This can be exploited through a Denial of Service attack.
302 Thanks to Adam Langley for discovering and researching this issue.
306 *) An attacker can force openssl to consume large amounts of memory whilst
307 processing DTLS handshake messages. This can be exploited through a
308 Denial of Service attack.
309 Thanks to Adam Langley for discovering and researching this issue.
313 *) An attacker can force an error condition which causes openssl to crash
314 whilst processing DTLS packets due to memory being freed twice. This
315 can be exploited through a Denial of Service attack.
316 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
321 *) If a multithreaded client connects to a malicious server using a resumed
322 session and the server sends an ec point format extension it could write
323 up to 255 bytes to freed memory.
325 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
330 *) A malicious server can crash an OpenSSL client with a null pointer
331 dereference (read) by specifying an SRP ciphersuite even though it was not
332 properly negotiated with the client. This can be exploited through a
333 Denial of Service attack.
335 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
336 discovering and researching this issue.
340 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
341 X509_name_oneline, X509_name_print_ex et al. to leak some information
342 from the stack. Applications may be affected if they echo pretty printing
343 output to the attacker.
345 Thanks to Ivan Fratric (Google) for discovering this issue.
347 [Emilia Käsper, and Steve Henson]
349 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
350 for corner cases. (Certain input points at infinity could lead to
351 bogus results, with non-infinity inputs mapped to infinity too.)
354 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
356 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
357 handshake can force the use of weak keying material in OpenSSL
358 SSL/TLS clients and servers.
360 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
361 researching this issue. (CVE-2014-0224)
362 [KIKUCHI Masashi, Steve Henson]
364 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
365 OpenSSL DTLS client the code can be made to recurse eventually crashing
368 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
370 [Imre Rad, Steve Henson]
372 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
373 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
374 client or server. This is potentially exploitable to run arbitrary
375 code on a vulnerable client or server.
377 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
378 [Jüri Aedla, Steve Henson]
380 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
381 are subject to a denial of service attack.
383 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
384 this issue. (CVE-2014-3470)
385 [Felix Gröbert, Ivan Fratric, Steve Henson]
387 *) Harmonize version and its documentation. -f flag is used to display
389 [mancha <mancha1@zoho.com>]
391 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
393 [mancha <mancha1@zoho.com>]
395 *) Fix some double frees. These are not thought to be exploitable.
396 [mancha <mancha1@zoho.com>]
398 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
400 *) A missing bounds check in the handling of the TLS heartbeat extension
401 can be used to reveal up to 64k of memory to a connected client or
404 Thanks for Neel Mehta of Google Security for discovering this bug and to
405 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
406 preparing the fix (CVE-2014-0160)
407 [Adam Langley, Bodo Moeller]
409 *) Fix for the attack described in the paper "Recovering OpenSSL
410 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
411 by Yuval Yarom and Naomi Benger. Details can be obtained from:
412 http://eprint.iacr.org/2014/140
414 Thanks to Yuval Yarom and Naomi Benger for discovering this
415 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
416 [Yuval Yarom and Naomi Benger]
418 *) TLS pad extension: draft-agl-tls-padding-03
420 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
421 TLS client Hello record length value would otherwise be > 255 and
422 less that 512 pad with a dummy extension containing zeroes so it
423 is at least 512 bytes long.
425 [Adam Langley, Steve Henson]
427 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
429 *) Fix for TLS record tampering bug. A carefully crafted invalid
430 handshake could crash OpenSSL with a NULL pointer exception.
431 Thanks to Anton Johansson for reporting this issues.
434 *) Keep original DTLS digest and encryption contexts in retransmission
435 structures so we can use the previous session parameters if they need
436 to be resent. (CVE-2013-6450)
439 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
440 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
441 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
442 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
443 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
444 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
445 [Rob Stradling, Adam Langley]
447 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
449 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
450 supporting platforms or when small records were transferred.
451 [Andy Polyakov, Steve Henson]
453 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
455 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
457 This addresses the flaw in CBC record processing discovered by
458 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
459 at: http://www.isg.rhul.ac.uk/tls/
461 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
462 Security Group at Royal Holloway, University of London
463 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
464 Emilia Käsper for the initial patch.
466 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
468 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
469 ciphersuites which can be exploited in a denial of service attack.
470 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
471 and detecting this bug and to Wolfgang Ettlinger
472 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
476 *) Return an error when checking OCSP signatures when key is NULL.
477 This fixes a DoS attack. (CVE-2013-0166)
480 *) Make openssl verify return errors.
481 [Chris Palmer <palmer@google.com> and Ben Laurie]
483 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
484 the right response is stapled. Also change SSL_get_certificate()
485 so it returns the certificate actually sent.
486 See http://rt.openssl.org/Ticket/Display.html?id=2836.
487 [Rob Stradling <rob.stradling@comodo.com>]
489 *) Fix possible deadlock when decoding public keys.
492 *) Don't use TLS 1.0 record version number in initial client hello
496 Changes between 1.0.1b and 1.0.1c [10 May 2012]
498 *) Sanity check record length before skipping explicit IV in TLS
499 1.2, 1.1 and DTLS to fix DoS attack.
501 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
502 fuzzing as a service testing platform.
506 *) Initialise tkeylen properly when encrypting CMS messages.
507 Thanks to Solar Designer of Openwall for reporting this issue.
510 *) In FIPS mode don't try to use composite ciphers as they are not
514 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
516 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
517 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
518 mean any application compiled against OpenSSL 1.0.0 headers setting
519 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
520 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
521 0x10000000L Any application which was previously compiled against
522 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
523 will need to be recompiled as a result. Letting be results in
524 inability to disable specifically TLS 1.1 and in client context,
525 in unlike event, limit maximum offered version to TLS 1.0 [see below].
528 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
529 disable just protocol X, but all protocols above X *if* there are
530 protocols *below* X still enabled. In more practical terms it means
531 that if application wants to disable TLS1.0 in favor of TLS1.1 and
532 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
533 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
537 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
539 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
540 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
541 in CRYPTO_realloc_clean.
543 Thanks to Tavis Ormandy, Google Security Team, for discovering this
544 issue and to Adam Langley <agl@chromium.org> for fixing it.
546 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
548 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
551 *) Workarounds for some broken servers that "hang" if a client hello
552 record length exceeds 255 bytes.
554 1. Do not use record version number > TLS 1.0 in initial client
555 hello: some (but not all) hanging servers will now work.
556 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
557 the number of ciphers sent in the client hello. This should be
558 set to an even number, such as 50, for example by passing:
559 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
560 Most broken servers should now work.
561 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
562 TLS 1.2 client support entirely.
565 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
568 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
570 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
571 STRING form instead of a DigestInfo.
574 *) The format used for MDC2 RSA signatures is inconsistent between EVP
575 and the RSA_sign/RSA_verify functions. This was made more apparent when
576 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
577 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
578 the correct format in RSA_verify so both forms transparently work.
581 *) Some servers which support TLS 1.0 can choke if we initially indicate
582 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
583 encrypted premaster secret. As a workaround use the maximum pemitted
584 client version in client hello, this should keep such servers happy
585 and still work with previous versions of OpenSSL.
588 *) Add support for TLS/DTLS heartbeats.
589 [Robin Seggelmann <seggelmann@fh-muenster.de>]
591 *) Add support for SCTP.
592 [Robin Seggelmann <seggelmann@fh-muenster.de>]
594 *) Improved PRNG seeding for VOS.
595 [Paul Green <Paul.Green@stratus.com>]
597 *) Extensive assembler packs updates, most notably:
599 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
600 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
601 - x86_64: bit-sliced AES implementation;
602 - ARM: NEON support, contemporary platforms optimizations;
603 - s390x: z196 support;
604 - *: GHASH and GF(2^m) multiplication implementations;
608 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
609 (removal of unnecessary code)
610 [Peter Sylvester <peter.sylvester@edelweb.fr>]
612 *) Add TLS key material exporter from RFC 5705.
615 *) Add DTLS-SRTP negotiation from RFC 5764.
618 *) Add Next Protocol Negotiation,
619 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
620 disabled with a no-npn flag to config or Configure. Code donated
622 [Adam Langley <agl@google.com> and Ben Laurie]
624 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
625 NIST-P256, NIST-P521, with constant-time single point multiplication on
626 typical inputs. Compiler support for the nonstandard type __uint128_t is
627 required to use this (present in gcc 4.4 and later, for 64-bit builds).
628 Code made available under Apache License version 2.0.
630 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
631 line to include this in your build of OpenSSL, and run "make depend" (or
632 "make update"). This enables the following EC_METHODs:
634 EC_GFp_nistp224_method()
635 EC_GFp_nistp256_method()
636 EC_GFp_nistp521_method()
638 EC_GROUP_new_by_curve_name() will automatically use these (while
639 EC_GROUP_new_curve_GFp() currently prefers the more flexible
641 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
643 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
644 all platforms. Move ssize_t definition from e_os.h to the public
645 header file e_os2.h as it now appears in public header file cms.h
648 *) New -sigopt option to the ca, req and x509 utilities. Additional
649 signature parameters can be passed using this option and in
653 *) Add RSA PSS signing function. This will generate and set the
654 appropriate AlgorithmIdentifiers for PSS based on those in the
655 corresponding EVP_MD_CTX structure. No application support yet.
658 *) Support for companion algorithm specific ASN1 signing routines.
659 New function ASN1_item_sign_ctx() signs a pre-initialised
660 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
661 the appropriate parameters.
664 *) Add new algorithm specific ASN1 verification initialisation function
665 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
666 handling will be the same no matter what EVP_PKEY_METHOD is used.
667 Add a PSS handler to support verification of PSS signatures: checked
668 against a number of sample certificates.
671 *) Add signature printing for PSS. Add PSS OIDs.
672 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
674 *) Add algorithm specific signature printing. An individual ASN1 method
675 can now print out signatures instead of the standard hex dump.
677 More complex signatures (e.g. PSS) can print out more meaningful
678 information. Include DSA version that prints out the signature
682 *) Password based recipient info support for CMS library: implementing
686 *) Split password based encryption into PBES2 and PBKDF2 functions. This
687 neatly separates the code into cipher and PBE sections and is required
688 for some algorithms that split PBES2 into separate pieces (such as
692 *) Session-handling fixes:
693 - Fix handling of connections that are resuming with a session ID,
694 but also support Session Tickets.
695 - Fix a bug that suppressed issuing of a new ticket if the client
696 presented a ticket with an expired session.
697 - Try to set the ticket lifetime hint to something reasonable.
698 - Make tickets shorter by excluding irrelevant information.
699 - On the client side, don't ignore renewed tickets.
700 [Adam Langley, Bodo Moeller (Google)]
702 *) Fix PSK session representation.
705 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
707 This work was sponsored by Intel.
710 *) Add GCM support to TLS library. Some custom code is needed to split
711 the IV between the fixed (from PRF) and explicit (from TLS record)
712 portions. This adds all GCM ciphersuites supported by RFC5288 and
713 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
714 add a special AESGCM string for GCM only.
717 *) Expand range of ctrls for AES GCM. Permit setting invocation
718 field on decrypt and retrieval of invocation field only on encrypt.
721 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
722 As required by RFC5289 these ciphersuites cannot be used if for
723 versions of TLS earlier than 1.2.
726 *) For FIPS capable OpenSSL interpret a NULL default public key method
727 as unset and return the appopriate default but do *not* set the default.
728 This means we can return the appopriate method in applications that
729 swicth between FIPS and non-FIPS modes.
732 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
733 ENGINE is used then we cannot handle that in the FIPS module so we
734 keep original code iff non-FIPS operations are allowed.
737 *) Add -attime option to openssl utilities.
738 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
740 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
743 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
744 FIPS EC methods unconditionally for now.
747 *) New build option no-ec2m to disable characteristic 2 code.
750 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
751 all cases can be covered as some introduce binary incompatibilities.
754 *) Redirect RSA operations to FIPS module including keygen,
755 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
758 *) Add similar low level API blocking to ciphers.
761 *) Low level digest APIs are not approved in FIPS mode: any attempt
762 to use these will cause a fatal error. Applications that *really* want
763 to use them can use the private_* version instead.
766 *) Redirect cipher operations to FIPS module for FIPS builds.
769 *) Redirect digest operations to FIPS module for FIPS builds.
772 *) Update build system to add "fips" flag which will link in fipscanister.o
773 for static and shared library builds embedding a signature if needed.
776 *) Output TLS supported curves in preference order instead of numerical
777 order. This is currently hardcoded for the highest order curves first.
778 This should be configurable so applications can judge speed vs strength.
781 *) Add TLS v1.2 server support for client authentication.
784 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
788 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
789 FIPS modules versions.
792 *) Add TLS v1.2 client side support for client authentication. Keep cache
793 of handshake records longer as we don't know the hash algorithm to use
794 until after the certificate request message is received.
797 *) Initial TLS v1.2 client support. Add a default signature algorithms
798 extension including all the algorithms we support. Parse new signature
799 format in client key exchange. Relax some ECC signing restrictions for
800 TLS v1.2 as indicated in RFC5246.
803 *) Add server support for TLS v1.2 signature algorithms extension. Switch
804 to new signature format when needed using client digest preference.
805 All server ciphersuites should now work correctly in TLS v1.2. No client
806 support yet and no support for client certificates.
809 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
810 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
811 ciphersuites. At present only RSA key exchange ciphersuites work with
812 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
813 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
814 and version checking.
817 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
818 with this defined it will not be affected by any changes to ssl internal
819 structures. Add several utility functions to allow openssl application
820 to work with OPENSSL_NO_SSL_INTERN defined.
824 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
826 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
829 *) Permit abbreviated handshakes when renegotiating using the function
830 SSL_renegotiate_abbreviated().
831 [Robin Seggelmann <seggelmann@fh-muenster.de>]
833 *) Add call to ENGINE_register_all_complete() to
834 ENGINE_load_builtin_engines(), so some implementations get used
835 automatically instead of needing explicit application support.
838 *) Add support for TLS key exporter as described in RFC5705.
839 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
841 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
842 a few changes are required:
844 Add SSL_OP_NO_TLSv1_1 flag.
846 Update version checking logic to handle version 1.1.
847 Add explicit IV handling (ported from DTLS code).
848 Add command line options to s_client/s_server.
851 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
853 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
854 in CMS and PKCS7 code. When RSA decryption fails use a random key for
855 content decryption and always return the same error. Note: this attack
856 needs on average 2^20 messages so it only affects automated senders. The
857 old behaviour can be reenabled in the CMS code by setting the
858 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
859 an MMA defence is not necessary.
860 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
861 this issue. (CVE-2012-0884)
864 *) Fix CVE-2011-4619: make sure we really are receiving a
865 client hello before rejecting multiple SGC restarts. Thanks to
866 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
869 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
871 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
872 Thanks to Antonio Martin, Enterprise Secure Access Research and
873 Development, Cisco Systems, Inc. for discovering this bug and
874 preparing a fix. (CVE-2012-0050)
877 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
879 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
880 of the Vaudenay padding oracle attack on CBC mode encryption
881 which enables an efficient plaintext recovery attack against
882 the OpenSSL implementation of DTLS. Their attack exploits timing
883 differences arising during decryption processing. A research
884 paper describing this attack can be found at:
885 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
886 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
887 Security Group at Royal Holloway, University of London
888 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
889 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
890 for preparing the fix. (CVE-2011-4108)
891 [Robin Seggelmann, Michael Tuexen]
893 *) Clear bytes used for block padding of SSL 3.0 records.
895 [Adam Langley (Google)]
897 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
898 Kadianakis <desnacked@gmail.com> for discovering this issue and
899 Adam Langley for preparing the fix. (CVE-2011-4619)
900 [Adam Langley (Google)]
902 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
903 [Andrey Kulikov <amdeich@gmail.com>]
905 *) Prevent malformed RFC3779 data triggering an assertion failure.
906 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
907 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
908 [Rob Austein <sra@hactrn.net>]
910 *) Improved PRNG seeding for VOS.
911 [Paul Green <Paul.Green@stratus.com>]
913 *) Fix ssl_ciph.c set-up race.
914 [Adam Langley (Google)]
916 *) Fix spurious failures in ecdsatest.c.
917 [Emilia Käsper (Google)]
919 *) Fix the BIO_f_buffer() implementation (which was mixing different
920 interpretations of the '..._len' fields).
921 [Adam Langley (Google)]
923 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
924 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
925 threads won't reuse the same blinding coefficients.
927 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
928 lock to call BN_BLINDING_invert_ex, and avoids one use of
929 BN_BLINDING_update for each BN_BLINDING structure (previously,
930 the last update always remained unused).
931 [Emilia Käsper (Google)]
933 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
934 [Bob Buckholz (Google)]
936 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
938 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
939 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
940 [Kaspar Brand <ossl@velox.ch>]
942 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
943 for multi-threaded use of ECDH. (CVE-2011-3210)
944 [Adam Langley (Google)]
946 *) Fix x509_name_ex_d2i memory leak on bad inputs.
949 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
950 signature public key algorithm by using OID xref utilities instead.
951 Before this you could only use some ECC ciphersuites with SHA1 only.
954 *) Add protection against ECDSA timing attacks as mentioned in the paper
955 by Billy Bob Brumley and Nicola Tuveri, see:
957 http://eprint.iacr.org/2011/232.pdf
959 [Billy Bob Brumley and Nicola Tuveri]
961 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
963 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
964 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
966 *) Fix bug in string printing code: if *any* escaping is enabled we must
967 escape the escape character (backslash) or the resulting string is
971 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
973 *) Disable code workaround for ancient and obsolete Netscape browsers
974 and servers: an attacker can use it in a ciphersuite downgrade attack.
975 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
978 *) Fixed J-PAKE implementation error, originally discovered by
979 Sebastien Martini, further info and confirmation from Stefan
980 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
983 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
985 *) Fix extension code to avoid race conditions which can result in a buffer
986 overrun vulnerability: resumed sessions must not be modified as they can
987 be shared by multiple threads. CVE-2010-3864
990 *) Fix WIN32 build system to correctly link an ENGINE directory into
994 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
996 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
998 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
1000 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
1002 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
1003 context. The operation can be customised via the ctrl mechanism in
1004 case ENGINEs want to include additional functionality.
1007 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
1010 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
1011 output hashes compatible with older versions of OpenSSL.
1012 [Willy Weisz <weisz@vcpc.univie.ac.at>]
1014 *) Fix compression algorithm handling: if resuming a session use the
1015 compression algorithm of the resumed session instead of determining
1016 it from client hello again. Don't allow server to change algorithm.
1019 *) Add load_crls() function to apps tidying load_certs() too. Add option
1020 to verify utility to allow additional CRLs to be included.
1023 *) Update OCSP request code to permit adding custom headers to the request:
1024 some responders need this.
1027 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
1029 [Julia Lawall <julia@diku.dk>]
1031 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
1032 needlessly dereferenced structures, used obsolete functions and
1033 didn't handle all updated verify codes correctly.
1036 *) Disable MD2 in the default configuration.
1039 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
1040 indicate the initial BIO being pushed or popped. This makes it possible
1041 to determine whether the BIO is the one explicitly called or as a result
1042 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
1043 it handles reference counts correctly and doesn't zero out the I/O bio
1044 when it is not being explicitly popped. WARNING: applications which
1045 included workarounds for the old buggy behaviour will need to be modified
1046 or they could free up already freed BIOs.
1049 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
1050 renaming to all platforms (within the 0.9.8 branch, this was
1051 done conditionally on Netware platforms to avoid a name clash).
1052 [Guenter <lists@gknw.net>]
1054 *) Add ECDHE and PSK support to DTLS.
1055 [Michael Tuexen <tuexen@fh-muenster.de>]
1057 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
1061 *) Add "missing" function EVP_MD_flags() (without this the only way to
1062 retrieve a digest flags is by accessing the structure directly. Update
1063 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
1064 or cipher is registered as in the "from" argument. Print out all
1065 registered digests in the dgst usage message instead of manually
1066 attempting to work them out.
1069 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
1070 this allows the use of compression and extensions. Change default cipher
1071 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
1072 by default unless an application cipher string requests it.
1075 *) Alter match criteria in PKCS12_parse(). It used to try to use local
1076 key ids to find matching certificates and keys but some PKCS#12 files
1077 don't follow the (somewhat unwritten) rules and this strategy fails.
1078 Now just gather all certificates together and the first private key
1079 then look for the first certificate that matches the key.
1082 *) Support use of registered digest and cipher names for dgst and cipher
1083 commands instead of having to add each one as a special case. So now
1090 openssl dgst -sha256 foo
1092 and this works for ENGINE based algorithms too.
1096 *) Update Gost ENGINE to support parameter files.
1097 [Victor B. Wagner <vitus@cryptocom.ru>]
1099 *) Support GeneralizedTime in ca utility.
1100 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
1102 *) Enhance the hash format used for certificate directory links. The new
1103 form uses the canonical encoding (meaning equivalent names will work
1104 even if they aren't identical) and uses SHA1 instead of MD5. This form
1105 is incompatible with the older format and as a result c_rehash should
1106 be used to rebuild symbolic links.
1109 *) Make PKCS#8 the default write format for private keys, replacing the
1110 traditional format. This form is standardised, more secure and doesn't
1111 include an implicit MD5 dependency.
1114 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
1115 committed to OpenSSL should pass this lot as a minimum.
1118 *) Add session ticket override functionality for use by EAP-FAST.
1119 [Jouni Malinen <j@w1.fi>]
1121 *) Modify HMAC functions to return a value. Since these can be implemented
1122 in an ENGINE errors can occur.
1125 *) Type-checked OBJ_bsearch_ex.
1128 *) Type-checked OBJ_bsearch. Also some constification necessitated
1129 by type-checking. Still to come: TXT_DB, bsearch(?),
1130 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
1134 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
1135 seconds to a tm structure directly, instead of going through OS
1136 specific date routines. This avoids any issues with OS routines such
1137 as the year 2038 bug. New *_adj() functions for ASN1 time structures
1138 and X509_time_adj_ex() to cover the extended range. The existing
1139 X509_time_adj() is still usable and will no longer have any date issues.
1142 *) Delta CRL support. New use deltas option which will attempt to locate
1143 and search any appropriate delta CRLs available.
1145 This work was sponsored by Google.
1148 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
1149 code and add additional score elements. Validate alternate CRL paths
1150 as part of the CRL checking and indicate a new error "CRL path validation
1151 error" in this case. Applications wanting additional details can use
1152 the verify callback and check the new "parent" field. If this is not
1153 NULL CRL path validation is taking place. Existing applications wont
1154 see this because it requires extended CRL support which is off by
1157 This work was sponsored by Google.
1160 *) Support for freshest CRL extension.
1162 This work was sponsored by Google.
1165 *) Initial indirect CRL support. Currently only supported in the CRLs
1166 passed directly and not via lookup. Process certificate issuer
1167 CRL entry extension and lookup CRL entries by bother issuer name
1168 and serial number. Check and process CRL issuer entry in IDP extension.
1170 This work was sponsored by Google.
1173 *) Add support for distinct certificate and CRL paths. The CRL issuer
1174 certificate is validated separately in this case. Only enabled if
1175 an extended CRL support flag is set: this flag will enable additional
1176 CRL functionality in future.
1178 This work was sponsored by Google.
1181 *) Add support for policy mappings extension.
1183 This work was sponsored by Google.
1186 *) Fixes to pathlength constraint, self issued certificate handling,
1187 policy processing to align with RFC3280 and PKITS tests.
1189 This work was sponsored by Google.
1192 *) Support for name constraints certificate extension. DN, email, DNS
1193 and URI types are currently supported.
1195 This work was sponsored by Google.
1198 *) To cater for systems that provide a pointer-based thread ID rather
1199 than numeric, deprecate the current numeric thread ID mechanism and
1200 replace it with a structure and associated callback type. This
1201 mechanism allows a numeric "hash" to be extracted from a thread ID in
1202 either case, and on platforms where pointers are larger than 'long',
1203 mixing is done to help ensure the numeric 'hash' is usable even if it
1204 can't be guaranteed unique. The default mechanism is to use "&errno"
1205 as a pointer-based thread ID to distinguish between threads.
1207 Applications that want to provide their own thread IDs should now use
1208 CRYPTO_THREADID_set_callback() to register a callback that will call
1209 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
1211 Note that ERR_remove_state() is now deprecated, because it is tied
1212 to the assumption that thread IDs are numeric. ERR_remove_state(0)
1213 to free the current thread's error state should be replaced by
1214 ERR_remove_thread_state(NULL).
1216 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
1217 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
1218 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
1219 application was previously providing a numeric thread callback that
1220 was inappropriate for distinguishing threads, then uniqueness might
1221 have been obtained with &errno that happened immediately in the
1222 intermediate development versions of OpenSSL; this is no longer the
1223 case, the numeric thread callback will now override the automatic use
1225 [Geoff Thorpe, with help from Bodo Moeller]
1227 *) Initial support for different CRL issuing certificates. This covers a
1228 simple case where the self issued certificates in the chain exist and
1229 the real CRL issuer is higher in the existing chain.
1231 This work was sponsored by Google.
1234 *) Removed effectively defunct crypto/store from the build.
1237 *) Revamp of STACK to provide stronger type-checking. Still to come:
1238 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
1239 ASN1_STRING, CONF_VALUE.
1242 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
1243 RAM on SSL connections. This option can save about 34k per idle SSL.
1246 *) Revamp of LHASH to provide stronger type-checking. Still to come:
1247 STACK, TXT_DB, bsearch, qsort.
1250 *) Initial support for Cryptographic Message Syntax (aka CMS) based
1251 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
1252 support for data, signedData, compressedData, digestedData and
1253 encryptedData, envelopedData types included. Scripts to check against
1254 RFC4134 examples draft and interop and consistency checks of many
1255 content types and variants.
1258 *) Add options to enc utility to support use of zlib compression BIO.
1261 *) Extend mk1mf to support importing of options and assembly language
1262 files from Configure script, currently only included in VC-WIN32.
1263 The assembly language rules can now optionally generate the source
1264 files from the associated perl scripts.
1267 *) Implement remaining functionality needed to support GOST ciphersuites.
1268 Interop testing has been performed using CryptoPro implementations.
1269 [Victor B. Wagner <vitus@cryptocom.ru>]
1271 *) s390x assembler pack.
1274 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
1278 *) Implement Opaque PRF Input TLS extension as specified in
1279 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
1280 official specification yet and no extension type assignment by
1281 IANA exists, this extension (for now) will have to be explicitly
1282 enabled when building OpenSSL by providing the extension number
1283 to use. For example, specify an option
1285 -DTLSEXT_TYPE_opaque_prf_input=0x9527
1287 to the "config" or "Configure" script to enable the extension,
1288 assuming extension number 0x9527 (which is a completely arbitrary
1289 and unofficial assignment based on the MD5 hash of the Internet
1290 Draft). Note that by doing so, you potentially lose
1291 interoperability with other TLS implementations since these might
1292 be using the same extension number for other purposes.
1294 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
1295 opaque PRF input value to use in the handshake. This will create
1296 an interal copy of the length-'len' string at 'src', and will
1297 return non-zero for success.
1299 To get more control and flexibility, provide a callback function
1302 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
1303 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
1307 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
1310 Callback function 'cb' will be called in handshakes, and is
1311 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
1312 Argument 'arg' is for application purposes (the value as given to
1313 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
1314 be provided to the callback function). The callback function
1315 has to return non-zero to report success: usually 1 to use opaque
1316 PRF input just if possible, or 2 to enforce use of the opaque PRF
1317 input. In the latter case, the library will abort the handshake
1318 if opaque PRF input is not successfully negotiated.
1320 Arguments 'peerinput' and 'len' given to the callback function
1321 will always be NULL and 0 in the case of a client. A server will
1322 see the client's opaque PRF input through these variables if
1323 available (NULL and 0 otherwise). Note that if the server
1324 provides an opaque PRF input, the length must be the same as the
1325 length of the client's opaque PRF input.
1327 Note that the callback function will only be called when creating
1328 a new session (session resumption can resume whatever was
1329 previously negotiated), and will not be called in SSL 2.0
1330 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
1331 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
1332 for applications that need to enforce opaque PRF input.
1336 *) Update ssl code to support digests other than SHA1+MD5 for handshake
1339 [Victor B. Wagner <vitus@cryptocom.ru>]
1341 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1342 RFC4507bis. The encrypted ticket format is an encrypted encoded
1343 SSL_SESSION structure, that way new session features are automatically
1346 If a client application caches session in an SSL_SESSION structure
1347 support is transparent because tickets are now stored in the encoded
1350 The SSL_CTX structure automatically generates keys for ticket
1351 protection in servers so again support should be possible
1352 with no application modification.
1354 If a client or server wishes to disable RFC4507 support then the option
1355 SSL_OP_NO_TICKET can be set.
1357 Add a TLS extension debugging callback to allow the contents of any client
1358 or server extensions to be examined.
1360 This work was sponsored by Google.
1363 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
1364 OpenSSL should now compile cleanly on gcc 4.2
1365 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
1367 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
1368 support including streaming MAC support: this is required for GOST
1369 ciphersuite support.
1370 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
1372 *) Add option -stream to use PKCS#7 streaming in smime utility. New
1373 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
1374 to output in BER and PEM format.
1377 *) Experimental support for use of HMAC via EVP_PKEY interface. This
1378 allows HMAC to be handled via the EVP_DigestSign*() interface. The
1379 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
1380 ENGINE support for HMAC keys which are unextractable. New -mac and
1381 -macopt options to dgst utility.
1384 *) New option -sigopt to dgst utility. Update dgst to use
1385 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
1386 alternative signing paramaters such as X9.31 or PSS in the dgst
1390 *) Change ssl_cipher_apply_rule(), the internal function that does
1391 the work each time a ciphersuite string requests enabling
1392 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
1393 removing ("!foo+bar") a class of ciphersuites: Now it maintains
1394 the order of disabled ciphersuites such that those ciphersuites
1395 that most recently went from enabled to disabled not only stay
1396 in order with respect to each other, but also have higher priority
1397 than other disabled ciphersuites the next time ciphersuites are
1400 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
1401 the same ciphersuites as with "HIGH" alone, but in a specific
1402 order where the PSK ciphersuites come first (since they are the
1403 most recently disabled ciphersuites when "HIGH" is parsed).
1405 Also, change ssl_create_cipher_list() (using this new
1406 funcionality) such that between otherwise identical
1407 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
1411 *) Change ssl_create_cipher_list() so that it automatically
1412 arranges the ciphersuites in reasonable order before starting
1413 to process the rule string. Thus, the definition for "DEFAULT"
1414 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
1415 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
1416 This makes it much easier to arrive at a reasonable default order
1417 in applications for which anonymous ciphers are OK (meaning
1418 that you can't actually use DEFAULT).
1419 [Bodo Moeller; suggested by Victor Duchovni]
1421 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
1422 processing) into multiple integers instead of setting
1423 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
1424 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
1425 (These masks as well as the individual bit definitions are hidden
1426 away into the non-exported interface ssl/ssl_locl.h, so this
1427 change to the definition of the SSL_CIPHER structure shouldn't
1428 affect applications.) This give us more bits for each of these
1429 categories, so there is no longer a need to coagulate AES128 and
1430 AES256 into a single algorithm bit, and to coagulate Camellia128
1431 and Camellia256 into a single algorithm bit, which has led to all
1434 Thus, among other things, the kludge introduced in 0.9.7m and
1435 0.9.8e for masking out AES256 independently of AES128 or masking
1436 out Camellia256 independently of AES256 is not needed here in 0.9.9.
1438 With the change, we also introduce new ciphersuite aliases that
1439 so far were missing: "AES128", "AES256", "CAMELLIA128", and
1443 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
1444 Use the leftmost N bytes of the signature input if the input is
1445 larger than the prime q (with N being the size in bytes of q).
1448 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
1449 it yet and it is largely untested.
1452 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
1455 *) Initial incomplete changes to avoid need for function casts in OpenSSL
1456 some compilers (gcc 4.2 and later) reject their use. Safestack is
1457 reimplemented. Update ASN1 to avoid use of legacy functions.
1460 *) Win32/64 targets are linked with Winsock2.
1463 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
1464 to external functions. This can be used to increase CRL handling
1465 efficiency especially when CRLs are very large by (for example) storing
1466 the CRL revoked certificates in a database.
1469 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
1470 new CRLs added to a directory can be used. New command line option
1471 -verify_return_error to s_client and s_server. This causes real errors
1472 to be returned by the verify callback instead of carrying on no matter
1473 what. This reflects the way a "real world" verify callback would behave.
1476 *) GOST engine, supporting several GOST algorithms and public key formats.
1477 Kindly donated by Cryptocom.
1480 *) Partial support for Issuing Distribution Point CRL extension. CRLs
1481 partitioned by DP are handled but no indirect CRL or reason partitioning
1482 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
1483 selected via a scoring technique which handles IDP and AKID in CRLs.
1486 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
1487 will ultimately be used for all verify operations: this will remove the
1488 X509_STORE dependency on certificate verification and allow alternative
1489 lookup methods. X509_STORE based implementations of these two callbacks.
1492 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
1493 Modify get_crl() to find a valid (unexpired) CRL if possible.
1496 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
1497 this would be called X509_CRL_cmp() but that name is already used by
1498 a function that just compares CRL issuer names. Cache several CRL
1499 extensions in X509_CRL structure and cache CRLDP in X509.
1502 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
1503 this maps equivalent X509_NAME structures into a consistent structure.
1504 Name comparison can then be performed rapidly using memcmp().
1507 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
1511 *) Allow digests to supply their own micalg string for S/MIME type using
1512 the ctrl EVP_MD_CTRL_MICALG.
1515 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
1516 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
1517 ctrl. It can then customise the structure before and/or after signing
1521 *) New function OBJ_add_sigid() to allow application defined signature OIDs
1522 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
1523 to free up any added signature OIDs.
1526 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
1527 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
1528 digest and cipher tables. New options added to openssl utility:
1529 list-message-digest-algorithms and list-cipher-algorithms.
1532 *) Change the array representation of binary polynomials: the list
1533 of degrees of non-zero coefficients is now terminated with -1.
1534 Previously it was terminated with 0, which was also part of the
1535 value; thus, the array representation was not applicable to
1536 polynomials where t^0 has coefficient zero. This change makes
1537 the array representation useful in a more general context.
1540 *) Various modifications and fixes to SSL/TLS cipher string
1541 handling. For ECC, the code now distinguishes between fixed ECDH
1542 with RSA certificates on the one hand and with ECDSA certificates
1543 on the other hand, since these are separate ciphersuites. The
1544 unused code for Fortezza ciphersuites has been removed.
1546 For consistency with EDH, ephemeral ECDH is now called "EECDH"
1547 (not "ECDHE"). For consistency with the code for DH
1548 certificates, use of ECDH certificates is now considered ECDH
1549 authentication, not RSA or ECDSA authentication (the latter is
1550 merely the CA's signing algorithm and not actively used in the
1553 The temporary ciphersuite alias "ECCdraft" is no longer
1554 available, and ECC ciphersuites are no longer excluded from "ALL"
1555 and "DEFAULT". The following aliases now exist for RFC 4492
1556 ciphersuites, most of these by analogy with the DH case:
1558 kECDHr - ECDH cert, signed with RSA
1559 kECDHe - ECDH cert, signed with ECDSA
1560 kECDH - ECDH cert (signed with either RSA or ECDSA)
1561 kEECDH - ephemeral ECDH
1562 ECDH - ECDH cert or ephemeral ECDH
1568 AECDH - anonymous ECDH
1569 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
1573 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
1574 Use correct micalg parameters depending on digest(s) in signed message.
1577 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
1578 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
1581 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
1582 an engine to register a method. Add ENGINE lookups for methods and
1583 functional reference processing.
1586 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
1587 EVP_{Sign,Verify}* which allow an application to customise the signature
1591 *) New -resign option to smime utility. This adds one or more signers
1592 to an existing PKCS#7 signedData structure. Also -md option to use an
1593 alternative message digest algorithm for signing.
1596 *) Tidy up PKCS#7 routines and add new functions to make it easier to
1597 create PKCS7 structures containing multiple signers. Update smime
1598 application to support multiple signers.
1601 *) New -macalg option to pkcs12 utility to allow setting of an alternative
1605 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
1606 Reorganize PBE internals to lookup from a static table using NIDs,
1607 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
1608 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
1609 PRF which will be automatically used with PBES2.
1612 *) Replace the algorithm specific calls to generate keys in "req" with the
1616 *) Update PKCS#7 enveloped data routines to use new API. This is now
1617 supported by any public key method supporting the encrypt operation. A
1618 ctrl is added to allow the public key algorithm to examine or modify
1619 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
1623 *) Add a ctrl to asn1 method to allow a public key algorithm to express
1624 a default digest type to use. In most cases this will be SHA1 but some
1625 algorithms (such as GOST) need to specify an alternative digest. The
1626 return value indicates how strong the prefernce is 1 means optional and
1627 2 is mandatory (that is it is the only supported type). Modify
1628 ASN1_item_sign() to accept a NULL digest argument to indicate it should
1629 use the default md. Update openssl utilities to use the default digest
1630 type for signing if it is not explicitly indicated.
1633 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
1634 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
1635 signing method from the key type. This effectively removes the link
1636 between digests and public key types.
1639 *) Add an OID cross reference table and utility functions. Its purpose is to
1640 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
1641 rsaEncryption. This will allow some of the algorithm specific hackery
1642 needed to use the correct OID to be removed.
1645 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
1646 structures for PKCS7_sign(). They are now set up by the relevant public
1650 *) Add provisional EC pkey method with support for ECDSA and ECDH.
1653 *) Add support for key derivation (agreement) in the API, DH method and
1657 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
1658 public and private key formats. As a side effect these add additional
1659 command line functionality not previously available: DSA signatures can be
1660 generated and verified using pkeyutl and DH key support and generation in
1665 [Oliver Tappe <zooey@hirschkaefer.de>]
1667 *) New make target "install_html_docs" installs HTML renditions of the
1669 [Oliver Tappe <zooey@hirschkaefer.de>]
1671 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
1672 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
1673 support key and parameter generation and add initial key generation
1674 functionality for RSA.
1677 *) Add functions for main EVP_PKEY_method operations. The undocumented
1678 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
1679 EVP_PKEY_{encrypt,decrypt}_old.
1682 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
1683 key API, doesn't do much yet.
1686 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
1687 public key algorithms. New option to openssl utility:
1688 "list-public-key-algorithms" to print out info.
1691 *) Implement the Supported Elliptic Curves Extension for
1692 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1695 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
1696 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
1699 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
1700 utilities such as rsa, dsa, dsaparam etc except they process any key
1704 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1705 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1706 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
1710 *) Initial support for pluggable public key ASN1.
1711 De-spaghettify the public key ASN1 handling. Move public and private
1712 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
1713 algorithm specific handling to a single module within the relevant
1714 algorithm directory. Add functions to allow (near) opaque processing
1715 of public and private key structures.
1718 *) Implement the Supported Point Formats Extension for
1719 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1722 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
1723 for the psk identity [hint] and the psk callback functions to the
1724 SSL_SESSION, SSL and SSL_CTX structure.
1727 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
1731 SSL_CTX_use_psk_identity_hint
1732 SSL_get_psk_identity_hint
1733 SSL_get_psk_identity
1734 SSL_use_psk_identity_hint
1736 [Mika Kousa and Pasi Eronen of Nokia Corporation]
1738 *) Add RFC 3161 compliant time stamp request creation, response generation
1739 and response verification functionality.
1740 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
1742 *) Add initial support for TLS extensions, specifically for the server_name
1743 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1744 have new members for a host name. The SSL data structure has an
1745 additional member SSL_CTX *initial_ctx so that new sessions can be
1746 stored in that context to allow for session resumption, even after the
1747 SSL has been switched to a new SSL_CTX in reaction to a client's
1748 server_name extension.
1750 New functions (subject to change):
1752 SSL_get_servername()
1753 SSL_get_servername_type()
1756 New CTRL codes and macros (subject to change):
1758 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1759 - SSL_CTX_set_tlsext_servername_callback()
1760 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1761 - SSL_CTX_set_tlsext_servername_arg()
1762 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1764 openssl s_client has a new '-servername ...' option.
1766 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1767 '-key2 ...', '-servername_fatal' (subject to change). This allows
1768 testing the HostName extension for a specific single host name ('-cert'
1769 and '-key' remain fallbacks for handshakes without HostName
1770 negotiation). If the unrecogninzed_name alert has to be sent, this by
1771 default is a warning; it becomes fatal with the '-servername_fatal'
1774 [Peter Sylvester, Remy Allais, Christophe Renou]
1776 *) Whirlpool hash implementation is added.
1779 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1780 bn(64,32). Because of instruction set limitations it doesn't have
1781 any negative impact on performance. This was done mostly in order
1782 to make it possible to share assembler modules, such as bn_mul_mont
1783 implementations, between 32- and 64-bit builds without hassle.
1786 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1787 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1791 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1792 dedicated Montgomery multiplication procedure, is introduced.
1793 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1794 "64-bit" performance on certain 32-bit targets.
1797 *) New option SSL_OP_NO_COMP to disable use of compression selectively
1798 in SSL structures. New SSL ctrl to set maximum send fragment size.
1799 Save memory by seeting the I/O buffer sizes dynamically instead of
1800 using the maximum available value.
1803 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
1804 in addition to the text details.
1807 *) Very, very preliminary EXPERIMENTAL support for printing of general
1808 ASN1 structures. This currently produces rather ugly output and doesn't
1809 handle several customised structures at all.
1812 *) Integrated support for PVK file format and some related formats such
1813 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
1814 these in the 'rsa' and 'dsa' utilities.
1817 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
1820 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
1821 place for the (very old) "NETSCAPE" format certificates which are now
1822 handled using new ASN1 code equivalents.
1825 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
1826 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
1827 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
1830 *) Modify CRL distribution points extension code to print out previously
1831 unsupported fields. Enhance extension setting code to allow setting of
1835 *) Add print and set support for Issuing Distribution Point CRL extension.
1838 *) Change 'Configure' script to enable Camellia by default.
1841 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
1843 *) When rejecting SSL/TLS records due to an incorrect version number, never
1844 update s->server with a new major version number. As of
1845 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
1846 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
1847 the previous behavior could result in a read attempt at NULL when
1848 receiving specific incorrect SSL/TLS records once record payload
1849 protection is active. (CVE-2010-0740)
1850 [Bodo Moeller, Adam Langley <agl@chromium.org>]
1852 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
1853 could be crashed if the relevant tables were not present (e.g. chrooted).
1854 [Tomas Hoger <thoger@redhat.com>]
1856 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
1858 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
1859 [Martin Olsson, Neel Mehta]
1861 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
1862 accommodate for stack sorting, always a write lock!).
1865 *) On some versions of WIN32 Heap32Next is very slow. This can cause
1866 excessive delays in the RAND_poll(): over a minute. As a workaround
1867 include a time check in the inner Heap32Next loop too.
1870 *) The code that handled flushing of data in SSL/TLS originally used the
1871 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
1872 the problem outlined in PR#1949. The fix suggested there however can
1873 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
1874 of Apache). So instead simplify the code to flush unconditionally.
1875 This should be fine since flushing with no data to flush is a no op.
1878 *) Handle TLS versions 2.0 and later properly and correctly use the
1879 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
1880 off ancient servers have a habit of sticking around for a while...
1883 *) Modify compression code so it frees up structures without using the
1884 ex_data callbacks. This works around a problem where some applications
1885 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
1886 restarting) then use compression (e.g. SSL with compression) later.
1887 This results in significant per-connection memory leaks and
1888 has caused some security issues including CVE-2008-1678 and
1892 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
1893 change when encrypting or decrypting.
1896 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
1897 connect and renegotiate with servers which do not support RI.
1898 Until RI is more widely deployed this option is enabled by default.
1901 *) Add "missing" ssl ctrls to clear options and mode.
1904 *) If client attempts to renegotiate and doesn't support RI respond with
1905 a no_renegotiation alert as required by RFC5746. Some renegotiating
1906 TLS clients will continue a connection gracefully when they receive
1907 the alert. Unfortunately OpenSSL mishandled this alert and would hang
1908 waiting for a server hello which it will never receive. Now we treat a
1909 received no_renegotiation alert as a fatal error. This is because
1910 applications requesting a renegotiation might well expect it to succeed
1911 and would have no code in place to handle the server denying it so the
1912 only safe thing to do is to terminate the connection.
1915 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
1916 peer supports secure renegotiation and 0 otherwise. Print out peer
1917 renegotiation support in s_client/s_server.
1920 *) Replace the highly broken and deprecated SPKAC certification method with
1921 the updated NID creation version. This should correctly handle UTF8.
1924 *) Implement RFC5746. Re-enable renegotiation but require the extension
1925 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
1926 turns out to be a bad idea. It has been replaced by
1927 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
1928 SSL_CTX_set_options(). This is really not recommended unless you
1929 know what you are doing.
1930 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
1932 *) Fixes to stateless session resumption handling. Use initial_ctx when
1933 issuing and attempting to decrypt tickets in case it has changed during
1934 servername handling. Use a non-zero length session ID when attempting
1935 stateless session resumption: this makes it possible to determine if
1936 a resumption has occurred immediately after receiving server hello
1937 (several places in OpenSSL subtly assume this) instead of later in
1941 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
1942 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
1943 fixes for a few places where the return code is not checked
1945 [Julia Lawall <julia@diku.dk>]
1947 *) Add --strict-warnings option to Configure script to include devteam
1948 warnings in other configurations.
1951 *) Add support for --libdir option and LIBDIR variable in makefiles. This
1952 makes it possible to install openssl libraries in locations which
1953 have names other than "lib", for example "/usr/lib64" which some
1955 [Steve Henson, based on patch from Jeremy Utley]
1957 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
1958 X690 8.9.12 and can produce some misleading textual output of OIDs.
1959 [Steve Henson, reported by Dan Kaminsky]
1961 *) Delete MD2 from algorithm tables. This follows the recommendation in
1962 several standards that it is not used in new applications due to
1963 several cryptographic weaknesses. For binary compatibility reasons
1964 the MD2 API is still compiled in by default.
1967 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
1971 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
1972 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
1974 [Guenter <lists@gknw.net>]
1976 *) Fix the server certificate chain building code to use X509_verify_cert(),
1977 it used to have an ad-hoc builder which was unable to cope with anything
1978 other than a simple chain.
1979 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
1981 *) Don't check self signed certificate signatures in X509_verify_cert()
1982 by default (a flag can override this): it just wastes time without
1983 adding any security. As a useful side effect self signed root CAs
1984 with non-FIPS digests are now usable in FIPS mode.
1987 *) In dtls1_process_out_of_seq_message() the check if the current message
1988 is already buffered was missing. For every new message was memory
1989 allocated, allowing an attacker to perform an denial of service attack
1990 with sending out of seq handshake messages until there is no memory
1991 left. Additionally every future messege was buffered, even if the
1992 sequence number made no sense and would be part of another handshake.
1993 So only messages with sequence numbers less than 10 in advance will be
1994 buffered. (CVE-2009-1378)
1995 [Robin Seggelmann, discovered by Daniel Mentz]
1997 *) Records are buffered if they arrive with a future epoch to be
1998 processed after finishing the corresponding handshake. There is
1999 currently no limitation to this buffer allowing an attacker to perform
2000 a DOS attack with sending records with future epochs until there is no
2001 memory left. This patch adds the pqueue_size() function to detemine
2002 the size of a buffer and limits the record buffer to 100 entries.
2004 [Robin Seggelmann, discovered by Daniel Mentz]
2006 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
2007 parent structure is freed. (CVE-2009-1379)
2010 *) Handle non-blocking I/O properly in SSL_shutdown() call.
2011 [Darryl Miles <darryl-mailinglists@netbauds.net>]
2014 [Ilya O. <vrghost@gmail.com>]
2016 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
2018 *) Disable renegotiation completely - this fixes a severe security
2019 problem (CVE-2009-3555) at the cost of breaking all
2020 renegotiation. Renegotiation can be re-enabled by setting
2021 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
2022 run-time. This is really not recommended unless you know what
2026 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
2028 *) Don't set val to NULL when freeing up structures, it is freed up by
2029 underlying code. If sizeof(void *) > sizeof(long) this can result in
2030 zeroing past the valid field. (CVE-2009-0789)
2031 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
2033 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
2034 checked correctly. This would allow some invalid signed attributes to
2035 appear to verify correctly. (CVE-2009-0591)
2036 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2038 *) Reject UniversalString and BMPString types with invalid lengths. This
2039 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
2040 a legal length. (CVE-2009-0590)
2043 *) Set S/MIME signing as the default purpose rather than setting it
2044 unconditionally. This allows applications to override it at the store
2048 *) Permit restricted recursion of ASN1 strings. This is needed in practice
2049 to handle some structures.
2052 *) Improve efficiency of mem_gets: don't search whole buffer each time
2054 [Jeremy Shapiro <jnshapir@us.ibm.com>]
2056 *) New -hex option for openssl rand.
2059 *) Print out UTF8String and NumericString when parsing ASN1.
2062 *) Support NumericString type for name components.
2065 *) Allow CC in the environment to override the automatically chosen
2066 compiler. Note that nothing is done to ensure flags work with the
2070 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
2072 *) Properly check EVP_VerifyFinal() and similar return values
2074 [Ben Laurie, Bodo Moeller, Google Security Team]
2076 *) Enable TLS extensions by default.
2079 *) Allow the CHIL engine to be loaded, whether the application is
2080 multithreaded or not. (This does not release the developer from the
2081 obligation to set up the dynamic locking callbacks.)
2082 [Sander Temme <sander@temme.net>]
2084 *) Use correct exit code if there is an error in dgst command.
2085 [Steve Henson; problem pointed out by Roland Dirlewanger]
2087 *) Tweak Configure so that you need to say "experimental-jpake" to enable
2088 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
2091 *) Add experimental JPAKE support, including demo authentication in
2092 s_client and s_server.
2095 *) Set the comparison function in v3_addr_canonize().
2096 [Rob Austein <sra@hactrn.net>]
2098 *) Add support for XMPP STARTTLS in s_client.
2099 [Philip Paeps <philip@freebsd.org>]
2101 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
2102 to ensure that even with this option, only ciphersuites in the
2103 server's preference list will be accepted. (Note that the option
2104 applies only when resuming a session, so the earlier behavior was
2105 just about the algorithm choice for symmetric cryptography.)
2108 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
2110 *) Fix NULL pointer dereference if a DTLS server received
2111 ChangeCipherSpec as first record (CVE-2009-1386).
2114 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
2115 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
2118 *) The fix in 0.9.8c that supposedly got rid of unsafe
2119 double-checked locking was incomplete for RSA blinding,
2120 addressing just one layer of what turns out to have been
2121 doubly unsafe triple-checked locking.
2123 So now fix this for real by retiring the MONT_HELPER macro
2124 in crypto/rsa/rsa_eay.c.
2126 [Bodo Moeller; problem pointed out by Marius Schilder]
2128 *) Various precautionary measures:
2130 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
2132 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
2133 (NB: This would require knowledge of the secret session ticket key
2134 to exploit, in which case you'd be SOL either way.)
2136 - Change bn_nist.c so that it will properly handle input BIGNUMs
2137 outside the expected range.
2139 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
2142 [Neel Mehta, Bodo Moeller]
2144 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
2145 the load fails. Useful for distros.
2146 [Ben Laurie and the FreeBSD team]
2148 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
2151 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
2154 *) Expand ENGINE to support engine supplied SSL client certificate functions.
2156 This work was sponsored by Logica.
2159 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
2160 keystores. Support for SSL/TLS client authentication too.
2161 Not compiled unless enable-capieng specified to Configure.
2163 This work was sponsored by Logica.
2166 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
2167 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
2168 attribute creation routines such as certifcate requests and PKCS#12
2172 Changes between 0.9.8g and 0.9.8h [28 May 2008]
2174 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
2175 handshake which could lead to a cilent crash as found using the
2176 Codenomicon TLS test suite (CVE-2008-1672)
2177 [Steve Henson, Mark Cox]
2179 *) Fix double free in TLS server name extensions which could lead to
2180 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
2183 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
2185 Clear the error queue to ensure that error entries left from
2186 older function calls do not interfere with the correct operation.
2187 [Lutz Jaenicke, Erik de Castro Lopo]
2189 *) Remove root CA certificates of commercial CAs:
2191 The OpenSSL project does not recommend any specific CA and does not
2192 have any policy with respect to including or excluding any CA.
2193 Therefore it does not make any sense to ship an arbitrary selection
2194 of root CA certificates with the OpenSSL software.
2197 *) RSA OAEP patches to fix two separate invalid memory reads.
2198 The first one involves inputs when 'lzero' is greater than
2199 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
2200 before the beginning of from). The second one involves inputs where
2201 the 'db' section contains nothing but zeroes (there is a one-byte
2202 invalid read after the end of 'db').
2203 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2205 *) Partial backport from 0.9.9-dev:
2207 Introduce bn_mul_mont (dedicated Montgomery multiplication
2208 procedure) as a candidate for BIGNUM assembler implementation.
2209 While 0.9.9-dev uses assembler for various architectures, only
2210 x86_64 is available by default here in the 0.9.8 branch, and
2211 32-bit x86 is available through a compile-time setting.
2213 To try the 32-bit x86 assembler implementation, use Configure
2214 option "enable-montasm" (which exists only for this backport).
2216 As "enable-montasm" for 32-bit x86 disclaims code stability
2217 anyway, in this constellation we activate additional code
2218 backported from 0.9.9-dev for further performance improvements,
2219 namely BN_from_montgomery_word. (To enable this otherwise,
2220 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
2222 [Andy Polyakov (backport partially by Bodo Moeller)]
2224 *) Add TLS session ticket callback. This allows an application to set
2225 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
2226 values. This is useful for key rollover for example where several key
2227 sets may exist with different names.
2230 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
2231 This was broken until now in 0.9.8 releases, such that the only way
2232 a registered ENGINE could be used (assuming it initialises
2233 successfully on the host) was to explicitly set it as the default
2234 for the relevant algorithms. This is in contradiction with 0.9.7
2235 behaviour and the documentation. With this fix, when an ENGINE is
2236 registered into a given algorithm's table of implementations, the
2237 'uptodate' flag is reset so that auto-discovery will be used next
2238 time a new context for that algorithm attempts to select an
2240 [Ian Lister (tweaked by Geoff Thorpe)]
2242 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
2243 implemention in the following ways:
2245 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
2248 Lack of BER streaming support means one pass streaming processing is
2249 only supported if data is detached: setting the streaming flag is
2250 ignored for embedded content.
2252 CMS support is disabled by default and must be explicitly enabled
2253 with the enable-cms configuration option.
2256 *) Update the GMP engine glue to do direct copies between BIGNUM and
2257 mpz_t when openssl and GMP use the same limb size. Otherwise the
2258 existing "conversion via a text string export" trick is still used.
2259 [Paul Sheer <paulsheer@gmail.com>]
2261 *) Zlib compression BIO. This is a filter BIO which compressed and
2262 uncompresses any data passed through it.
2265 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
2266 RFC3394 compatible AES key wrapping.
2269 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
2270 sets string data without copying. X509_ALGOR_set0() and
2271 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
2272 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
2273 from an X509_ATTRIBUTE structure optionally checking it occurs only
2274 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
2278 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
2279 to get the expected BN_FLG_CONSTTIME behavior.
2280 [Bodo Moeller (Google)]
2284 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
2285 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
2286 - added some more tests to do_tests.pl
2287 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
2288 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
2289 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
2290 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
2291 - various changes to netware.pl to enable gcc-cross builds on Win32
2293 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
2294 - various changes to fix missing prototype warnings
2295 - fixed x86nasm.pl to create correct asm files for NASM COFF output
2296 - added AES, WHIRLPOOL and CPUID assembler code to build files
2297 - added missing AES assembler make rules to mk1mf.pl
2298 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
2299 [Guenter Knauf <eflash@gmx.net>]
2301 *) Implement certificate status request TLS extension defined in RFC3546.
2302 A client can set the appropriate parameters and receive the encoded
2303 OCSP response via a callback. A server can query the supplied parameters
2304 and set the encoded OCSP response in the callback. Add simplified examples
2305 to s_client and s_server.
2308 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
2310 *) Fix various bugs:
2311 + Binary incompatibility of ssl_ctx_st structure
2312 + DTLS interoperation with non-compliant servers
2313 + Don't call get_session_cb() without proposed session
2314 + Fix ia64 assembler code
2315 [Andy Polyakov, Steve Henson]
2317 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
2319 *) DTLS Handshake overhaul. There were longstanding issues with
2320 OpenSSL DTLS implementation, which were making it impossible for
2321 RFC 4347 compliant client to communicate with OpenSSL server.
2322 Unfortunately just fixing these incompatibilities would "cut off"
2323 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
2324 server keeps tolerating non RFC compliant syntax. The opposite is
2325 not true, 0.9.8f client can not communicate with earlier server.
2326 This update even addresses CVE-2007-4995.
2329 *) Changes to avoid need for function casts in OpenSSL: some compilers
2330 (gcc 4.2 and later) reject their use.
2331 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
2334 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2335 RFC4507bis. The encrypted ticket format is an encrypted encoded
2336 SSL_SESSION structure, that way new session features are automatically
2339 If a client application caches session in an SSL_SESSION structure
2340 support is transparent because tickets are now stored in the encoded
2343 The SSL_CTX structure automatically generates keys for ticket
2344 protection in servers so again support should be possible
2345 with no application modification.
2347 If a client or server wishes to disable RFC4507 support then the option
2348 SSL_OP_NO_TICKET can be set.
2350 Add a TLS extension debugging callback to allow the contents of any client
2351 or server extensions to be examined.
2353 This work was sponsored by Google.
2356 *) Add initial support for TLS extensions, specifically for the server_name
2357 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2358 have new members for a host name. The SSL data structure has an
2359 additional member SSL_CTX *initial_ctx so that new sessions can be
2360 stored in that context to allow for session resumption, even after the
2361 SSL has been switched to a new SSL_CTX in reaction to a client's
2362 server_name extension.
2364 New functions (subject to change):
2366 SSL_get_servername()
2367 SSL_get_servername_type()
2370 New CTRL codes and macros (subject to change):
2372 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2373 - SSL_CTX_set_tlsext_servername_callback()
2374 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2375 - SSL_CTX_set_tlsext_servername_arg()
2376 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2378 openssl s_client has a new '-servername ...' option.
2380 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2381 '-key2 ...', '-servername_fatal' (subject to change). This allows
2382 testing the HostName extension for a specific single host name ('-cert'
2383 and '-key' remain fallbacks for handshakes without HostName
2384 negotiation). If the unrecogninzed_name alert has to be sent, this by
2385 default is a warning; it becomes fatal with the '-servername_fatal'
2388 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
2390 *) Add AES and SSE2 assembly language support to VC++ build.
2393 *) Mitigate attack on final subtraction in Montgomery reduction.
2396 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
2397 (which previously caused an internal error).
2400 *) Squeeze another 10% out of IGE mode when in != out.
2403 *) AES IGE mode speedup.
2404 [Dean Gaudet (Google)]
2406 *) Add the Korean symmetric 128-bit cipher SEED (see
2407 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
2408 add SEED ciphersuites from RFC 4162:
2410 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
2411 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
2412 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
2413 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
2415 To minimize changes between patchlevels in the OpenSSL 0.9.8
2416 series, SEED remains excluded from compilation unless OpenSSL
2417 is configured with 'enable-seed'.
2418 [KISA, Bodo Moeller]
2420 *) Mitigate branch prediction attacks, which can be practical if a
2421 single processor is shared, allowing a spy process to extract
2422 information. For detailed background information, see
2423 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
2424 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
2425 and Necessary Software Countermeasures"). The core of the change
2426 are new versions BN_div_no_branch() and
2427 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
2428 respectively, which are slower, but avoid the security-relevant
2429 conditional branches. These are automatically called by BN_div()
2430 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
2431 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
2432 remove a conditional branch.
2434 BN_FLG_CONSTTIME is the new name for the previous
2435 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
2436 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
2437 in the exponent causes BN_mod_exp_mont() to use the alternative
2438 implementation in BN_mod_exp_mont_consttime().) The old name
2439 remains as a deprecated alias.
2441 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
2442 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
2443 constant-time implementations for more than just exponentiation.
2444 Here too the old name is kept as a deprecated alias.
2446 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2447 the BN_BLINDING structure gets an independent copy of the
2448 modulus. This means that the previous "BIGNUM *m" argument to
2449 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2450 essentially becomes "const BIGNUM *m", although we can't actually
2451 change this in the header file before 0.9.9. It allows
2452 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2453 enable BN_FLG_CONSTTIME.
2455 [Matthew D Wood (Intel Corp)]
2457 *) In the SSL/TLS server implementation, be strict about session ID
2458 context matching (which matters if an application uses a single
2459 external cache for different purposes). Previously,
2460 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2461 set. This did ensure strict client verification, but meant that,
2462 with applications using a single external cache for quite
2463 different requirements, clients could circumvent ciphersuite
2464 restrictions for a given session ID context by starting a session
2465 in a different context.
2468 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2469 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2470 authentication-only ciphersuites.
2473 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
2474 not complete and could lead to a possible single byte overflow
2475 (CVE-2007-5135) [Ben Laurie]
2477 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
2479 *) Since AES128 and AES256 (and similarly Camellia128 and
2480 Camellia256) share a single mask bit in the logic of
2481 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2482 kludge to work properly if AES128 is available and AES256 isn't
2483 (or if Camellia128 is available and Camellia256 isn't).
2486 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
2487 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
2488 When a point or a seed is encoded in a BIT STRING, we need to
2489 prevent the removal of trailing zero bits to get the proper DER
2490 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
2491 of a NamedBitList, for which trailing 0 bits need to be removed.)
2494 *) Have SSL/TLS server implementation tolerate "mismatched" record
2495 protocol version while receiving ClientHello even if the
2496 ClientHello is fragmented. (The server can't insist on the
2497 particular protocol version it has chosen before the ServerHello
2498 message has informed the client about his choice.)
2501 *) Add RFC 3779 support.
2502 [Rob Austein for ARIN, Ben Laurie]
2504 *) Load error codes if they are not already present instead of using a
2505 static variable. This allows them to be cleanly unloaded and reloaded.
2506 Improve header file function name parsing.
2509 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
2510 or CAPABILITY handshake as required by RFCs.
2513 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
2515 *) Introduce limits to prevent malicious keys being able to
2516 cause a denial of service. (CVE-2006-2940)
2517 [Steve Henson, Bodo Moeller]
2519 *) Fix ASN.1 parsing of certain invalid structures that can result
2520 in a denial of service. (CVE-2006-2937) [Steve Henson]
2522 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2523 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2525 *) Fix SSL client code which could crash if connecting to a
2526 malicious SSLv2 server. (CVE-2006-4343)
2527 [Tavis Ormandy and Will Drewry, Google Security Team]
2529 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
2530 match only those. Before that, "AES256-SHA" would be interpreted
2531 as a pattern and match "AES128-SHA" too (since AES128-SHA got
2532 the same strength classification in 0.9.7h) as we currently only
2533 have a single AES bit in the ciphersuite description bitmap.
2534 That change, however, also applied to ciphersuite strings such as
2535 "RC4-MD5" that intentionally matched multiple ciphersuites --
2536 namely, SSL 2.0 ciphersuites in addition to the more common ones
2537 from SSL 3.0/TLS 1.0.
2539 So we change the selection algorithm again: Naming an explicit
2540 ciphersuite selects this one ciphersuite, and any other similar
2541 ciphersuite (same bitmap) from *other* protocol versions.
2542 Thus, "RC4-MD5" again will properly select both the SSL 2.0
2543 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
2545 Since SSL 2.0 does not have any ciphersuites for which the
2546 128/256 bit distinction would be relevant, this works for now.
2547 The proper fix will be to use different bits for AES128 and
2548 AES256, which would have avoided the problems from the beginning;
2549 however, bits are scarce, so we can only do this in a new release
2550 (not just a patchlevel) when we can change the SSL_CIPHER
2551 definition to split the single 'unsigned long mask' bitmap into
2552 multiple values to extend the available space.
2556 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2558 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2559 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2561 *) Add AES IGE and biIGE modes.
2564 *) Change the Unix randomness entropy gathering to use poll() when
2565 possible instead of select(), since the latter has some
2566 undesirable limitations.
2567 [Darryl Miles via Richard Levitte and Bodo Moeller]
2569 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2570 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2571 cannot be implicitly activated as part of, e.g., the "AES" alias.
2572 However, please upgrade to OpenSSL 0.9.9[-dev] for
2573 non-experimental use of the ECC ciphersuites to get TLS extension
2574 support, which is required for curve and point format negotiation
2575 to avoid potential handshake problems.
2578 *) Disable rogue ciphersuites:
2580 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2581 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2582 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2584 The latter two were purportedly from
2585 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2588 Also deactivate the remaining ciphersuites from
2589 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2590 unofficial, and the ID has long expired.
2593 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2594 dual-core machines) and other potential thread-safety issues.
2597 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
2598 versions), which is now available for royalty-free use
2599 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
2600 Also, add Camellia TLS ciphersuites from RFC 4132.
2602 To minimize changes between patchlevels in the OpenSSL 0.9.8
2603 series, Camellia remains excluded from compilation unless OpenSSL
2604 is configured with 'enable-camellia'.
2607 *) Disable the padding bug check when compression is in use. The padding
2608 bug check assumes the first packet is of even length, this is not
2609 necessarily true if compresssion is enabled and can result in false
2610 positives causing handshake failure. The actual bug test is ancient
2611 code so it is hoped that implementations will either have fixed it by
2612 now or any which still have the bug do not support compression.
2615 Changes between 0.9.8a and 0.9.8b [04 May 2006]
2617 *) When applying a cipher rule check to see if string match is an explicit
2618 cipher suite and only match that one cipher suite if it is.
2621 *) Link in manifests for VC++ if needed.
2622 [Austin Ziegler <halostatue@gmail.com>]
2624 *) Update support for ECC-based TLS ciphersuites according to
2625 draft-ietf-tls-ecc-12.txt with proposed changes (but without
2626 TLS extensions, which are supported starting with the 0.9.9
2627 branch, not in the OpenSSL 0.9.8 branch).
2630 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
2631 opaque EVP_CIPHER_CTX handling.
2634 *) Fixes and enhancements to zlib compression code. We now only use
2635 "zlib1.dll" and use the default __cdecl calling convention on Win32
2636 to conform with the standards mentioned here:
2637 http://www.zlib.net/DLL_FAQ.txt
2638 Static zlib linking now works on Windows and the new --with-zlib-include
2639 --with-zlib-lib options to Configure can be used to supply the location
2640 of the headers and library. Gracefully handle case where zlib library
2644 *) Several fixes and enhancements to the OID generation code. The old code
2645 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
2646 handle numbers larger than ULONG_MAX, truncated printing and had a
2647 non standard OBJ_obj2txt() behaviour.
2650 *) Add support for building of engines under engine/ as shared libraries
2651 under VC++ build system.
2654 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
2655 Hopefully, we will not see any false combination of paths any more.
2658 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
2660 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2661 (part of SSL_OP_ALL). This option used to disable the
2662 countermeasure against man-in-the-middle protocol-version
2663 rollback in the SSL 2.0 server implementation, which is a bad
2664 idea. (CVE-2005-2969)
2666 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2667 for Information Security, National Institute of Advanced Industrial
2668 Science and Technology [AIST], Japan)]
2670 *) Add two function to clear and return the verify parameter flags.
2673 *) Keep cipherlists sorted in the source instead of sorting them at
2674 runtime, thus removing the need for a lock.
2677 *) Avoid some small subgroup attacks in Diffie-Hellman.
2678 [Nick Mathewson and Ben Laurie]
2680 *) Add functions for well-known primes.
2683 *) Extended Windows CE support.
2684 [Satoshi Nakamura and Andy Polyakov]
2686 *) Initialize SSL_METHOD structures at compile time instead of during
2687 runtime, thus removing the need for a lock.
2690 *) Make PKCS7_decrypt() work even if no certificate is supplied by
2691 attempting to decrypt each encrypted key in turn. Add support to
2695 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
2697 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
2700 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
2703 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
2704 key into the same file any more.
2707 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
2710 *) Add -utf8 command line and config file option to 'ca'.
2711 [Stefan <stf@udoma.org]
2713 *) Removed the macro des_crypt(), as it seems to conflict with some
2714 libraries. Use DES_crypt().
2717 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
2718 involves renaming the source and generated shared-libs for
2719 both. The engines will accept the corrected or legacy ids
2720 ('ncipher' and '4758_cca' respectively) when binding. NB,
2721 this only applies when building 'shared'.
2722 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
2724 *) Add attribute functions to EVP_PKEY structure. Modify
2725 PKCS12_create() to recognize a CSP name attribute and
2726 use it. Make -CSP option work again in pkcs12 utility.
2729 *) Add new functionality to the bn blinding code:
2730 - automatic re-creation of the BN_BLINDING parameters after
2731 a fixed number of uses (currently 32)
2732 - add new function for parameter creation
2733 - introduce flags to control the update behaviour of the
2734 BN_BLINDING parameters
2735 - hide BN_BLINDING structure
2736 Add a second BN_BLINDING slot to the RSA structure to improve
2737 performance when a single RSA object is shared among several
2741 *) Add support for DTLS.
2742 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
2744 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
2745 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
2748 *) Remove buggy and incompletet DH cert support from
2749 ssl/ssl_rsa.c and ssl/s3_both.c
2752 *) Use SHA-1 instead of MD5 as the default digest algorithm for
2753 the apps/openssl applications.
2756 *) Compile clean with "-Wall -Wmissing-prototypes
2757 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
2758 DEBUG_SAFESTACK must also be set.
2761 *) Change ./Configure so that certain algorithms can be disabled by default.
2762 The new counterpiece to "no-xxx" is "enable-xxx".
2764 The patented RC5 and MDC2 algorithms will now be disabled unless
2765 "enable-rc5" and "enable-mdc2", respectively, are specified.
2767 (IDEA remains enabled despite being patented. This is because IDEA
2768 is frequently required for interoperability, and there is no license
2769 fee for non-commercial use. As before, "no-idea" can be used to
2770 avoid this algorithm.)
2774 *) Add processing of proxy certificates (see RFC 3820). This work was
2775 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
2776 EGEE (Enabling Grids for E-science in Europe).
2779 *) RC4 performance overhaul on modern architectures/implementations, such
2780 as Intel P4, IA-64 and AMD64.
2783 *) New utility extract-section.pl. This can be used specify an alternative
2784 section number in a pod file instead of having to treat each file as
2785 a separate case in Makefile. This can be done by adding two lines to the
2788 =for comment openssl_section:XXX
2790 The blank line is mandatory.
2794 *) New arguments -certform, -keyform and -pass for s_client and s_server
2795 to allow alternative format key and certificate files and passphrase
2799 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
2800 update associated structures and add various utility functions.
2802 Add new policy related verify parameters, include policy checking in
2803 standard verify code. Enhance 'smime' application with extra parameters
2804 to support policy checking and print out.
2807 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
2808 Nehemiah processors. These extensions support AES encryption in hardware
2809 as well as RNG (though RNG support is currently disabled).
2810 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
2812 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
2815 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
2816 [Andy Polyakov and a number of other people]
2818 *) Improved PowerPC platform support. Most notably BIGNUM assembler
2819 implementation contributed by IBM.
2820 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
2822 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
2823 exponent rather than 'unsigned long'. There is a corresponding change to
2824 the new 'rsa_keygen' element of the RSA_METHOD structure.
2825 [Jelte Jansen, Geoff Thorpe]
2827 *) Functionality for creating the initial serial number file is now
2828 moved from CA.pl to the 'ca' utility with a new option -create_serial.
2830 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
2831 number file to 1, which is bound to cause problems. To avoid
2832 the problems while respecting compatibility between different 0.9.7
2833 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
2834 CA.pl for serial number initialization. With the new release 0.9.8,
2835 we can fix the problem directly in the 'ca' utility.)
2838 *) Reduced header interdepencies by declaring more opaque objects in
2839 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
2840 give fewer recursive includes, which could break lazy source code - so
2841 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
2842 developers should define this symbol when building and using openssl to
2843 ensure they track the recommended behaviour, interfaces, [etc], but
2844 backwards-compatible behaviour prevails when this isn't defined.
2847 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
2850 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
2851 This will generate a random key of the appropriate length based on the
2852 cipher context. The EVP_CIPHER can provide its own random key generation
2853 routine to support keys of a specific form. This is used in the des and
2854 3des routines to generate a key of the correct parity. Update S/MIME
2855 code to use new functions and hence generate correct parity DES keys.
2856 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
2857 valid (weak or incorrect parity).
2860 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
2861 as looking them up. This is useful when the verified structure may contain
2862 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
2863 present unless the new PKCS7_NO_CRL flag is asserted.
2866 *) Extend ASN1 oid configuration module. It now additionally accepts the
2869 shortName = some long name, 1.2.3.4
2872 *) Reimplemented the BN_CTX implementation. There is now no more static
2873 limitation on the number of variables it can handle nor the depth of the
2874 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
2875 information can now expand as required, and rather than having a single
2876 static array of bignums, BN_CTX now uses a linked-list of such arrays
2877 allowing it to expand on demand whilst maintaining the usefulness of
2878 BN_CTX's "bundling".
2881 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
2882 to allow all RSA operations to function using a single BN_CTX.
2885 *) Preliminary support for certificate policy evaluation and checking. This
2886 is initially intended to pass the tests outlined in "Conformance Testing
2887 of Relying Party Client Certificate Path Processing Logic" v1.07.
2890 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
2891 remained unused and not that useful. A variety of other little bignum
2892 tweaks and fixes have also been made continuing on from the audit (see
2896 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
2897 associated ASN1, EVP and SSL functions and old ASN1 macros.
2900 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
2901 and this should never fail. So the return value from the use of
2902 BN_set_word() (which can fail due to needless expansion) is now deprecated;
2903 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
2906 *) BN_CTX_get() should return zero-valued bignums, providing the same
2907 initialised value as BN_new().
2908 [Geoff Thorpe, suggested by Ulf Möller]
2910 *) Support for inhibitAnyPolicy certificate extension.
2913 *) An audit of the BIGNUM code is underway, for which debugging code is
2914 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
2915 is considered valid when processing BIGNUMs, and causes execution to
2916 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
2917 further steps are taken to deliberately pollute unused data in BIGNUM
2918 structures to try and expose faulty code further on. For now, openssl will
2919 (in its default mode of operation) continue to tolerate the inconsistent
2920 forms that it has tolerated in the past, but authors and packagers should
2921 consider trying openssl and their own applications when compiled with
2922 these debugging symbols defined. It will help highlight potential bugs in
2923 their own code, and will improve the test coverage for OpenSSL itself. At
2924 some point, these tighter rules will become openssl's default to improve
2925 maintainability, though the assert()s and other overheads will remain only
2926 in debugging configurations. See bn.h for more details.
2927 [Geoff Thorpe, Nils Larsch, Ulf Möller]
2929 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
2930 that can only be obtained through BN_CTX_new() (which implicitly
2931 initialises it). The presence of this function only made it possible
2932 to overwrite an existing structure (and cause memory leaks).
2935 *) Because of the callback-based approach for implementing LHASH as a
2936 template type, lh_insert() adds opaque objects to hash-tables and
2937 lh_doall() or lh_doall_arg() are typically used with a destructor callback
2938 to clean up those corresponding objects before destroying the hash table
2939 (and losing the object pointers). So some over-zealous constifications in
2940 LHASH have been relaxed so that lh_insert() does not take (nor store) the
2941 objects as "const" and the lh_doall[_arg] callback wrappers are not
2942 prototyped to have "const" restrictions on the object pointers they are
2943 given (and so aren't required to cast them away any more).
2946 *) The tmdiff.h API was so ugly and minimal that our own timing utility
2947 (speed) prefers to use its own implementation. The two implementations
2948 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
2949 its object type properly exposed (MS_TM) instead of casting to/from "char
2950 *". This may still change yet if someone realises MS_TM and "ms_time_***"
2951 aren't necessarily the greatest nomenclatures - but this is what was used
2952 internally to the implementation so I've used that for now.
2955 *) Ensure that deprecated functions do not get compiled when
2956 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
2957 the self-tests were still using deprecated key-generation functions so
2958 these have been updated also.
2961 *) Reorganise PKCS#7 code to separate the digest location functionality
2962 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
2963 New function PKCS7_set_digest() to set the digest type for PKCS#7
2964 digestedData type. Add additional code to correctly generate the
2965 digestedData type and add support for this type in PKCS7 initialization
2969 *) New function PKCS7_set0_type_other() this initializes a PKCS7
2970 structure of type "other".
2973 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
2974 sure the loop does correctly stop and breaking ("division by zero")
2975 modulus operations are not performed. The (pre-generated) prime
2976 table crypto/bn/bn_prime.h was already correct, but it could not be
2977 re-generated on some platforms because of the "division by zero"
2978 situation in the script.
2979 [Ralf S. Engelschall]
2981 *) Update support for ECC-based TLS ciphersuites according to
2982 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
2983 SHA-1 now is only used for "small" curves (where the
2984 representation of a field element takes up to 24 bytes); for
2985 larger curves, the field element resulting from ECDH is directly
2986 used as premaster secret.
2987 [Douglas Stebila (Sun Microsystems Laboratories)]
2989 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
2990 curve secp160r1 to the tests.
2991 [Douglas Stebila (Sun Microsystems Laboratories)]
2993 *) Add the possibility to load symbols globally with DSO.
2994 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
2996 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
2997 control of the error stack.
3000 *) Add support for STORE in ENGINE.
3003 *) Add the STORE type. The intention is to provide a common interface
3004 to certificate and key stores, be they simple file-based stores, or
3005 HSM-type store, or LDAP stores, or...
3006 NOTE: The code is currently UNTESTED and isn't really used anywhere.
3009 *) Add a generic structure called OPENSSL_ITEM. This can be used to
3010 pass a list of arguments to any function as well as provide a way
3011 for a function to pass data back to the caller.
3014 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
3015 works like BUF_strdup() but can be used to duplicate a portion of
3016 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
3020 *) Add the function sk_find_ex() which works like sk_find(), but will
3021 return an index to an element even if an exact match couldn't be
3022 found. The index is guaranteed to point at the element where the
3023 searched-for key would be inserted to preserve sorting order.
3026 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
3027 takes an extra flags argument for optional functionality. Currently,
3028 the following flags are defined:
3030 OBJ_BSEARCH_VALUE_ON_NOMATCH
3031 This one gets OBJ_bsearch_ex() to return a pointer to the first
3032 element where the comparing function returns a negative or zero
3035 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
3036 This one gets OBJ_bsearch_ex() to return a pointer to the first
3037 element where the comparing function returns zero. This is useful
3038 if there are more than one element where the comparing function
3042 *) Make it possible to create self-signed certificates with 'openssl ca'
3043 in such a way that the self-signed certificate becomes part of the
3044 CA database and uses the same mechanisms for serial number generation
3045 as all other certificate signing. The new flag '-selfsign' enables
3046 this functionality. Adapt CA.sh and CA.pl.in.
3049 *) Add functionality to check the public key of a certificate request
3050 against a given private. This is useful to check that a certificate
3051 request can be signed by that key (self-signing).
3054 *) Make it possible to have multiple active certificates with the same
3055 subject in the CA index file. This is done only if the keyword
3056 'unique_subject' is set to 'no' in the main CA section (default
3057 if 'CA_default') of the configuration file. The value is saved
3058 with the database itself in a separate index attribute file,
3059 named like the index file with '.attr' appended to the name.
3062 *) Generate muti valued AVAs using '+' notation in config files for
3066 *) Support for nameConstraints certificate extension.
3069 *) Support for policyConstraints certificate extension.
3072 *) Support for policyMappings certificate extension.
3075 *) Make sure the default DSA_METHOD implementation only uses its
3076 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
3077 and change its own handlers to be NULL so as to remove unnecessary
3078 indirection. This lets alternative implementations fallback to the
3079 default implementation more easily.
3082 *) Support for directoryName in GeneralName related extensions
3086 *) Make it possible to link applications using Makefile.shared.
3087 Make that possible even when linking against static libraries!
3090 *) Support for single pass processing for S/MIME signing. This now
3091 means that S/MIME signing can be done from a pipe, in addition
3092 cleartext signing (multipart/signed type) is effectively streaming
3093 and the signed data does not need to be all held in memory.
3095 This is done with a new flag PKCS7_STREAM. When this flag is set
3096 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
3097 is done after the data is output (and digests calculated) in
3098 SMIME_write_PKCS7().
3101 *) Add full support for -rpath/-R, both in shared libraries and
3102 applications, at least on the platforms where it's known how
3106 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
3107 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
3108 will now compute a table of multiples of the generator that
3109 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
3110 faster (notably in the case of a single point multiplication,
3111 scalar * generator).
3112 [Nils Larsch, Bodo Moeller]
3114 *) IPv6 support for certificate extensions. The various extensions
3115 which use the IP:a.b.c.d can now take IPv6 addresses using the
3116 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
3120 *) Added an ENGINE that implements RSA by performing private key
3121 exponentiations with the GMP library. The conversions to and from
3122 GMP's mpz_t format aren't optimised nor are any montgomery forms
3123 cached, and on x86 it appears OpenSSL's own performance has caught up.
3124 However there are likely to be other architectures where GMP could
3125 provide a boost. This ENGINE is not built in by default, but it can be
3126 specified at Configure time and should be accompanied by the necessary
3127 linker additions, eg;
3128 ./config -DOPENSSL_USE_GMP -lgmp
3131 *) "openssl engine" will not display ENGINE/DSO load failure errors when
3132 testing availability of engines with "-t" - the old behaviour is
3133 produced by increasing the feature's verbosity with "-tt".
3136 *) ECDSA routines: under certain error conditions uninitialized BN objects
3137 could be freed. Solution: make sure initialization is performed early
3138 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
3142 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
3143 and DH_METHOD (eg. by ENGINE implementations) to override the normal
3144 software implementations. For DSA and DH, parameter generation can
3145 also be overriden by providing the appropriate method callbacks.
3148 *) Change the "progress" mechanism used in key-generation and
3149 primality testing to functions that take a new BN_GENCB pointer in
3150 place of callback/argument pairs. The new API functions have "_ex"
3151 postfixes and the older functions are reimplemented as wrappers for
3152 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
3153 declarations of the old functions to help (graceful) attempts to
3154 migrate to the new functions. Also, the new key-generation API
3155 functions operate on a caller-supplied key-structure and return
3156 success/failure rather than returning a key or NULL - this is to
3157 help make "keygen" another member function of RSA_METHOD etc.
3159 Example for using the new callback interface:
3161 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
3165 BN_GENCB_set(&my_cb, my_callback, my_arg);
3167 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
3168 /* For the meaning of a, b in calls to my_callback(), see the
3169 * documentation of the function that calls the callback.
3170 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
3171 * my_callback should return 1 if it wants BN_is_prime_ex()
3172 * to continue, or 0 to stop.
3177 *) Change the ZLIB compression method to be stateful, and make it
3178 available to TLS with the number defined in
3179 draft-ietf-tls-compression-04.txt.
3182 *) Add the ASN.1 structures and functions for CertificatePair, which
3183 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
3185 CertificatePair ::= SEQUENCE {
3186 forward [0] Certificate OPTIONAL,
3187 reverse [1] Certificate OPTIONAL,
3188 -- at least one of the pair shall be present -- }
3190 Also implement the PEM functions to read and write certificate
3191 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
3193 This needed to be defined, mostly for the sake of the LDAP
3194 attribute crossCertificatePair, but may prove useful elsewhere as
3198 *) Make it possible to inhibit symlinking of shared libraries in
3199 Makefile.shared, for Cygwin's sake.
3202 *) Extend the BIGNUM API by creating a function
3203 void BN_set_negative(BIGNUM *a, int neg);
3204 and a macro that behave like
3205 int BN_is_negative(const BIGNUM *a);
3207 to avoid the need to access 'a->neg' directly in applications.
3210 *) Implement fast modular reduction for pseudo-Mersenne primes
3211 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
3212 EC_GROUP_new_curve_GFp() will now automatically use this
3214 [Nils Larsch <nla@trustcenter.de>]
3216 *) Add new lock type (CRYPTO_LOCK_BN).
3219 *) Change the ENGINE framework to automatically load engines
3220 dynamically from specific directories unless they could be
3221 found to already be built in or loaded. Move all the
3222 current engines except for the cryptodev one to a new
3224 The engines in engines/ are built as shared libraries if
3225 the "shared" options was given to ./Configure or ./config.
3226 Otherwise, they are inserted in libcrypto.a.
3227 /usr/local/ssl/engines is the default directory for dynamic
3228 engines, but that can be overriden at configure time through
3229 the usual use of --prefix and/or --openssldir, and at run
3230 time with the environment variable OPENSSL_ENGINES.
3231 [Geoff Thorpe and Richard Levitte]
3233 *) Add Makefile.shared, a helper makefile to build shared
3234 libraries. Addapt Makefile.org.
3237 *) Add version info to Win32 DLLs.
3238 [Peter 'Luna' Runestig" <peter@runestig.com>]
3240 *) Add new 'medium level' PKCS#12 API. Certificates and keys
3241 can be added using this API to created arbitrary PKCS#12
3242 files while avoiding the low level API.
3244 New options to PKCS12_create(), key or cert can be NULL and
3245 will then be omitted from the output file. The encryption
3246 algorithm NIDs can be set to -1 for no encryption, the mac
3247 iteration count can be set to 0 to omit the mac.
3249 Enhance pkcs12 utility by making the -nokeys and -nocerts
3250 options work when creating a PKCS#12 file. New option -nomac
3251 to omit the mac, NONE can be set for an encryption algorithm.
3252 New code is modified to use the enhanced PKCS12_create()
3253 instead of the low level API.
3256 *) Extend ASN1 encoder to support indefinite length constructed
3257 encoding. This can output sequences tags and octet strings in
3258 this form. Modify pk7_asn1.c to support indefinite length
3259 encoding. This is experimental and needs additional code to
3260 be useful, such as an ASN1 bio and some enhanced streaming
3263 Extend template encode functionality so that tagging is passed
3264 down to the template encoder.
3267 *) Let 'openssl req' fail if an argument to '-newkey' is not
3268 recognized instead of using RSA as a default.
3271 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
3272 As these are not official, they are not included in "ALL";
3273 the "ECCdraft" ciphersuite group alias can be used to select them.
3274 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
3276 *) Add ECDH engine support.
3277 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
3279 *) Add ECDH in new directory crypto/ecdh/.
3280 [Douglas Stebila (Sun Microsystems Laboratories)]
3282 *) Let BN_rand_range() abort with an error after 100 iterations
3283 without success (which indicates a broken PRNG).
3286 *) Change BN_mod_sqrt() so that it verifies that the input value
3287 is really the square of the return value. (Previously,
3288 BN_mod_sqrt would show GIGO behaviour.)
3291 *) Add named elliptic curves over binary fields from X9.62, SECG,
3292 and WAP/WTLS; add OIDs that were still missing.
3294 [Sheueling Chang Shantz and Douglas Stebila
3295 (Sun Microsystems Laboratories)]
3297 *) Extend the EC library for elliptic curves over binary fields
3298 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
3301 EC_GF2m_simple_method
3305 EC_GROUP_new_curve_GF2m
3306 EC_GROUP_set_curve_GF2m
3307 EC_GROUP_get_curve_GF2m
3308 EC_POINT_set_affine_coordinates_GF2m
3309 EC_POINT_get_affine_coordinates_GF2m
3310 EC_POINT_set_compressed_coordinates_GF2m
3312 Point compression for binary fields is disabled by default for
3313 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
3316 As binary polynomials are represented as BIGNUMs, various members
3317 of the EC_GROUP and EC_POINT data structures can be shared
3318 between the implementations for prime fields and binary fields;
3319 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
3320 are essentially identical to their ..._GFp counterparts.
3321 (For simplicity, the '..._GFp' prefix has been dropped from
3322 various internal method names.)
3324 An internal 'field_div' method (similar to 'field_mul' and
3325 'field_sqr') has been added; this is used only for binary fields.
3327 [Sheueling Chang Shantz and Douglas Stebila
3328 (Sun Microsystems Laboratories)]
3330 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
3331 through methods ('mul', 'precompute_mult').
3333 The generic implementations (now internally called 'ec_wNAF_mul'
3334 and 'ec_wNAF_precomputed_mult') remain the default if these
3335 methods are undefined.
3337 [Sheueling Chang Shantz and Douglas Stebila
3338 (Sun Microsystems Laboratories)]
3340 *) New function EC_GROUP_get_degree, which is defined through
3341 EC_METHOD. For curves over prime fields, this returns the bit
3342 length of the modulus.
3344 [Sheueling Chang Shantz and Douglas Stebila
3345 (Sun Microsystems Laboratories)]
3347 *) New functions EC_GROUP_dup, EC_POINT_dup.
3348 (These simply call ..._new and ..._copy).
3350 [Sheueling Chang Shantz and Douglas Stebila
3351 (Sun Microsystems Laboratories)]
3353 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
3354 Polynomials are represented as BIGNUMs (where the sign bit is not
3355 used) in the following functions [macros]:
3358 BN_GF2m_sub [= BN_GF2m_add]
3359 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
3360 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
3361 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
3363 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
3364 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
3365 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
3366 BN_GF2m_cmp [= BN_ucmp]
3368 (Note that only the 'mod' functions are actually for fields GF(2^m).
3369 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
3371 For some functions, an the irreducible polynomial defining a
3372 field can be given as an 'unsigned int[]' with strictly
3373 decreasing elements giving the indices of those bits that are set;
3374 i.e., p[] represents the polynomial
3375 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
3377 p[0] > p[1] > ... > p[k] = 0.
3378 This applies to the following functions:
3383 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
3384 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
3386 BN_GF2m_mod_sqrt_arr
3387 BN_GF2m_mod_solve_quad_arr
3391 Conversion can be performed by the following functions:
3396 bntest.c has additional tests for binary polynomial arithmetic.
3398 Two implementations for BN_GF2m_mod_div() are available.
3399 The default algorithm simply uses BN_GF2m_mod_inv() and
3400 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
3401 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
3402 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
3404 [Sheueling Chang Shantz and Douglas Stebila
3405 (Sun Microsystems Laboratories)]
3407 *) Add new error code 'ERR_R_DISABLED' that can be used when some
3408 functionality is disabled at compile-time.
3409 [Douglas Stebila <douglas.stebila@sun.com>]
3411 *) Change default behaviour of 'openssl asn1parse' so that more
3412 information is visible when viewing, e.g., a certificate:
3414 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
3415 mode the content of non-printable OCTET STRINGs is output in a
3416 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
3417 avoid the appearance of a printable string.
3418 [Nils Larsch <nla@trustcenter.de>]
3420 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
3422 EC_GROUP_set_asn1_flag()
3423 EC_GROUP_get_asn1_flag()
3424 EC_GROUP_set_point_conversion_form()
3425 EC_GROUP_get_point_conversion_form()
3426 These control ASN1 encoding details:
3427 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
3428 has been set to OPENSSL_EC_NAMED_CURVE.
3429 - Points are encoded in uncompressed form by default; options for
3430 asn1_for are as for point2oct, namely
3431 POINT_CONVERSION_COMPRESSED
3432 POINT_CONVERSION_UNCOMPRESSED
3433 POINT_CONVERSION_HYBRID
3435 Also add 'seed' and 'seed_len' members to EC_GROUP with access
3438 EC_GROUP_get0_seed()
3439 EC_GROUP_get_seed_len()
3440 This is used only for ASN1 purposes (so far).
3441 [Nils Larsch <nla@trustcenter.de>]
3443 *) Add 'field_type' member to EC_METHOD, which holds the NID
3444 of the appropriate field type OID. The new function
3445 EC_METHOD_get_field_type() returns this value.
3446 [Nils Larsch <nla@trustcenter.de>]
3451 EC_POINT_point2hex()
3452 EC_POINT_hex2point()
3453 providing useful interfaces to EC_POINT_point2oct() and
3454 EC_POINT_oct2point().
3455 [Nils Larsch <nla@trustcenter.de>]
3457 *) Change internals of the EC library so that the functions
3458 EC_GROUP_set_generator()
3459 EC_GROUP_get_generator()
3460 EC_GROUP_get_order()
3461 EC_GROUP_get_cofactor()
3462 are implemented directly in crypto/ec/ec_lib.c and not dispatched
3463 to methods, which would lead to unnecessary code duplication when
3464 adding different types of curves.
3465 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
3467 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
3468 arithmetic, and such that modified wNAFs are generated
3469 (which avoid length expansion in many cases).
3472 *) Add a function EC_GROUP_check_discriminant() (defined via
3473 EC_METHOD) that verifies that the curve discriminant is non-zero.
3475 Add a function EC_GROUP_check() that makes some sanity tests
3476 on a EC_GROUP, its generator and order. This includes
3477 EC_GROUP_check_discriminant().
3478 [Nils Larsch <nla@trustcenter.de>]
3480 *) Add ECDSA in new directory crypto/ecdsa/.
3482 Add applications 'openssl ecparam' and 'openssl ecdsa'
3483 (these are based on 'openssl dsaparam' and 'openssl dsa').
3485 ECDSA support is also included in various other files across the
3486 library. Most notably,
3487 - 'openssl req' now has a '-newkey ecdsa:file' option;
3488 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
3489 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
3490 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
3491 them suitable for ECDSA where domain parameters must be
3492 extracted before the specific public key;
3493 - ECDSA engine support has been added.
3494 [Nils Larsch <nla@trustcenter.de>]
3496 *) Include some named elliptic curves, and add OIDs from X9.62,
3497 SECG, and WAP/WTLS. Each curve can be obtained from the new
3499 EC_GROUP_new_by_curve_name(),
3500 and the list of available named curves can be obtained with
3501 EC_get_builtin_curves().
3502 Also add a 'curve_name' member to EC_GROUP objects, which can be
3504 EC_GROUP_set_curve_name()
3505 EC_GROUP_get_curve_name()
3506 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
3508 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3509 was actually never needed) and in BN_mul(). The removal in BN_mul()
3510 required a small change in bn_mul_part_recursive() and the addition
3511 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3512 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3513 bn_sub_words() and bn_add_words() except they take arrays with
3517 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
3519 *) Cleanse PEM buffers before freeing them since they may contain
3521 [Benjamin Bennett <ben@psc.edu>]
3523 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3524 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3525 authentication-only ciphersuites.
3528 *) Since AES128 and AES256 share a single mask bit in the logic of
3529 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3530 kludge to work properly if AES128 is available and AES256 isn't.
3533 *) Expand security boundary to match 1.1.1 module.
3536 *) Remove redundant features: hash file source, editing of test vectors
3537 modify fipsld to use external fips_premain.c signature.
3540 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
3541 run algorithm test programs.
3544 *) Make algorithm test programs more tolerant of whitespace.
3547 *) Have SSL/TLS server implementation tolerate "mismatched" record
3548 protocol version while receiving ClientHello even if the
3549 ClientHello is fragmented. (The server can't insist on the
3550 particular protocol version it has chosen before the ServerHello
3551 message has informed the client about his choice.)
3554 *) Load error codes if they are not already present instead of using a
3555 static variable. This allows them to be cleanly unloaded and reloaded.
3558 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3560 *) Introduce limits to prevent malicious keys being able to
3561 cause a denial of service. (CVE-2006-2940)
3562 [Steve Henson, Bodo Moeller]
3564 *) Fix ASN.1 parsing of certain invalid structures that can result
3565 in a denial of service. (CVE-2006-2937) [Steve Henson]
3567 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3568 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3570 *) Fix SSL client code which could crash if connecting to a
3571 malicious SSLv2 server. (CVE-2006-4343)
3572 [Tavis Ormandy and Will Drewry, Google Security Team]
3574 *) Change ciphersuite string processing so that an explicit
3575 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3576 will no longer include "AES128-SHA"), and any other similar
3577 ciphersuite (same bitmap) from *other* protocol versions (so that
3578 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3579 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3580 changes from 0.9.8b and 0.9.8d.
3583 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
3585 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3586 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3588 *) Change the Unix randomness entropy gathering to use poll() when
3589 possible instead of select(), since the latter has some
3590 undesirable limitations.
3591 [Darryl Miles via Richard Levitte and Bodo Moeller]
3593 *) Disable rogue ciphersuites:
3595 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3596 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3597 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3599 The latter two were purportedly from
3600 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3603 Also deactive the remaining ciphersuites from
3604 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3605 unofficial, and the ID has long expired.
3608 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3609 dual-core machines) and other potential thread-safety issues.
3612 Changes between 0.9.7i and 0.9.7j [04 May 2006]
3614 *) Adapt fipsld and the build system to link against the validated FIPS
3615 module in FIPS mode.
3618 *) Fixes for VC++ 2005 build under Windows.
3621 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
3622 from a Windows bash shell such as MSYS. It is autodetected from the
3623 "config" script when run from a VC++ environment. Modify standard VC++
3624 build to use fipscanister.o from the GNU make build.
3627 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
3629 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
3630 The value now differs depending on if you build for FIPS or not.
3631 BEWARE! A program linked with a shared FIPSed libcrypto can't be
3632 safely run with a non-FIPSed libcrypto, as it may crash because of
3633 the difference induced by this change.
3636 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
3638 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3639 (part of SSL_OP_ALL). This option used to disable the
3640 countermeasure against man-in-the-middle protocol-version
3641 rollback in the SSL 2.0 server implementation, which is a bad
3642 idea. (CVE-2005-2969)
3644 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3645 for Information Security, National Institute of Advanced Industrial
3646 Science and Technology [AIST], Japan)]
3648 *) Minimal support for X9.31 signatures and PSS padding modes. This is
3649 mainly for FIPS compliance and not fully integrated at this stage.
3652 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
3653 the exponentiation using a fixed-length exponent. (Otherwise,
3654 the information leaked through timing could expose the secret key
3655 after many signatures; cf. Bleichenbacher's attack on DSA with
3659 *) Make a new fixed-window mod_exp implementation the default for
3660 RSA, DSA, and DH private-key operations so that the sequence of
3661 squares and multiplies and the memory access pattern are
3662 independent of the particular secret key. This will mitigate
3663 cache-timing and potential related attacks.
3665 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
3666 and this is automatically used by BN_mod_exp_mont() if the new flag
3667 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
3668 will use this BN flag for private exponents unless the flag
3669 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
3670 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
3672 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
3674 *) Change the client implementation for SSLv23_method() and
3675 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
3676 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
3677 (Previously, the SSL 2.0 backwards compatible Client Hello
3678 message format would be used even with SSL_OP_NO_SSLv2.)
3681 *) Add support for smime-type MIME parameter in S/MIME messages which some
3685 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
3686 a threadsafe manner. Modify rsa code to use new function and add calls
3687 to dsa and dh code (which had race conditions before).
3690 *) Include the fixed error library code in the C error file definitions
3691 instead of fixing them up at runtime. This keeps the error code
3692 structures constant.
3695 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
3697 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
3700 *) Fixes for newer kerberos headers. NB: the casts are needed because
3701 the 'length' field is signed on one version and unsigned on another
3702 with no (?) obvious way to tell the difference, without these VC++
3703 complains. Also the "definition" of FAR (blank) is no longer included
3704 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
3705 some needed definitions.
3708 *) Undo Cygwin change.
3711 *) Added support for proxy certificates according to RFC 3820.
3712 Because they may be a security thread to unaware applications,
3713 they must be explicitely allowed in run-time. See
3714 docs/HOWTO/proxy_certificates.txt for further information.
3717 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
3719 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
3720 server and client random values. Previously
3721 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
3722 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
3724 This change has negligible security impact because:
3726 1. Server and client random values still have 24 bytes of pseudo random
3729 2. Server and client random values are sent in the clear in the initial
3732 3. The master secret is derived using the premaster secret (48 bytes in
3733 size for static RSA ciphersuites) as well as client server and random
3736 The OpenSSL team would like to thank the UK NISCC for bringing this issue
3739 [Stephen Henson, reported by UK NISCC]
3741 *) Use Windows randomness collection on Cygwin.
3744 *) Fix hang in EGD/PRNGD query when communication socket is closed
3745 prematurely by EGD/PRNGD.
3746 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
3748 *) Prompt for pass phrases when appropriate for PKCS12 input format.
3751 *) Back-port of selected performance improvements from development
3752 branch, as well as improved support for PowerPC platforms.
3755 *) Add lots of checks for memory allocation failure, error codes to indicate
3756 failure and freeing up memory if a failure occurs.
3757 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
3759 *) Add new -passin argument to dgst.
3762 *) Perform some character comparisons of different types in X509_NAME_cmp:
3763 this is needed for some certificates that reencode DNs into UTF8Strings
3764 (in violation of RFC3280) and can't or wont issue name rollover
3768 *) Make an explicit check during certificate validation to see that
3769 the CA setting in each certificate on the chain is correct. As a
3770 side effect always do the following basic checks on extensions,
3771 not just when there's an associated purpose to the check:
3773 - if there is an unhandled critical extension (unless the user
3774 has chosen to ignore this fault)
3775 - if the path length has been exceeded (if one is set at all)
3776 - that certain extensions fit the associated purpose (if one has
3780 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
3782 *) Avoid a race condition when CRLs are checked in a multi threaded
3783 environment. This would happen due to the reordering of the revoked
3784 entries during signature checking and serial number lookup. Now the
3785 encoding is cached and the serial number sort performed under a lock.
3786 Add new STACK function sk_is_sorted().
3789 *) Add Delta CRL to the extension code.
3792 *) Various fixes to s3_pkt.c so alerts are sent properly.
3793 [David Holmes <d.holmes@f5.com>]
3795 *) Reduce the chances of duplicate issuer name and serial numbers (in
3796 violation of RFC3280) using the OpenSSL certificate creation utilities.
3797 This is done by creating a random 64 bit value for the initial serial
3798 number when a serial number file is created or when a self signed
3799 certificate is created using 'openssl req -x509'. The initial serial
3800 number file is created using 'openssl x509 -next_serial' in CA.pl
3801 rather than being initialized to 1.
3804 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
3806 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3807 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3808 [Joe Orton, Steve Henson]
3810 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
3812 [Joe Orton, Steve Henson]
3814 *) Make it possible to have multiple active certificates with the same
3815 subject in the CA index file. This is done only if the keyword
3816 'unique_subject' is set to 'no' in the main CA section (default
3817 if 'CA_default') of the configuration file. The value is saved
3818 with the database itself in a separate index attribute file,
3819 named like the index file with '.attr' appended to the name.
3822 *) X509 verify fixes. Disable broken certificate workarounds when
3823 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
3824 keyUsage extension present. Don't accept CRLs with unhandled critical
3825 extensions: since verify currently doesn't process CRL extensions this
3826 rejects a CRL with *any* critical extensions. Add new verify error codes
3830 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
3831 A clarification of RFC2560 will require the use of OCTET STRINGs and
3832 some implementations cannot handle the current raw format. Since OpenSSL
3833 copies and compares OCSP nonces as opaque blobs without any attempt at
3834 parsing them this should not create any compatibility issues.
3837 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
3838 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
3839 this HMAC (and other) operations are several times slower than OpenSSL
3843 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
3844 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3846 *) Use the correct content when signing type "other".
3849 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
3851 *) Fix various bugs revealed by running the NISCC test suite:
3853 Stop out of bounds reads in the ASN1 code when presented with
3854 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3856 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
3858 If verify callback ignores invalid public key errors don't try to check
3859 certificate signature with the NULL public key.
3863 *) New -ignore_err option in ocsp application to stop the server
3864 exiting on the first error in a request.
3867 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3868 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3872 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3873 extra data after the compression methods not only for TLS 1.0
3874 but also for SSL 3.0 (as required by the specification).
3875 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3877 *) Change X509_certificate_type() to mark the key as exported/exportable
3878 when it's 512 *bits* long, not 512 bytes.
3881 *) Change AES_cbc_encrypt() so it outputs exact multiple of
3882 blocks during encryption.
3885 *) Various fixes to base64 BIO and non blocking I/O. On write
3886 flushes were not handled properly if the BIO retried. On read
3887 data was not being buffered properly and had various logic bugs.
3888 This also affects blocking I/O when the data being decoded is a
3892 *) Various S/MIME bugfixes and compatibility changes:
3893 output correct application/pkcs7 MIME type if
3894 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
3895 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
3896 of files as .eml work). Correctly handle very long lines in MIME
3900 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
3902 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3903 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3904 a protocol version number mismatch like a decryption error
3905 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3908 *) Turn on RSA blinding by default in the default implementation
3909 to avoid a timing attack. Applications that don't want it can call
3910 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3911 They would be ill-advised to do so in most cases.
3912 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3914 *) Change RSA blinding code so that it works when the PRNG is not
3915 seeded (in this case, the secret RSA exponent is abused as
3916 an unpredictable seed -- if it is not unpredictable, there
3917 is no point in blinding anyway). Make RSA blinding thread-safe
3918 by remembering the creator's thread ID in rsa->blinding and
3919 having all other threads use local one-time blinding factors
3920 (this requires more computation than sharing rsa->blinding, but
3921 avoids excessive locking; and if an RSA object is not shared
3922 between threads, blinding will still be very fast).
3925 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
3926 ENGINE as defaults for all supported algorithms irrespective of
3927 the 'flags' parameter. 'flags' is now honoured, so applications
3928 should make sure they are passing it correctly.
3931 *) Target "mingw" now allows native Windows code to be generated in
3932 the Cygwin environment as well as with the MinGW compiler.
3935 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
3937 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3938 via timing by performing a MAC computation even if incorrrect
3939 block cipher padding has been found. This is a countermeasure
3940 against active attacks where the attacker has to distinguish
3941 between bad padding and a MAC verification error. (CVE-2003-0078)
3943 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3944 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3945 Martin Vuagnoux (EPFL, Ilion)]
3947 *) Make the no-err option work as intended. The intention with no-err
3948 is not to have the whole error stack handling routines removed from
3949 libcrypto, it's only intended to remove all the function name and
3950 reason texts, thereby removing some of the footprint that may not
3951 be interesting if those errors aren't displayed anyway.
3953 NOTE: it's still possible for any application or module to have it's
3954 own set of error texts inserted. The routines are there, just not
3955 used by default when no-err is given.
3958 *) Add support for FreeBSD on IA64.
3959 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
3961 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
3962 Kerberos function mit_des_cbc_cksum(). Before this change,
3963 the value returned by DES_cbc_cksum() was like the one from
3964 mit_des_cbc_cksum(), except the bytes were swapped.
3965 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
3967 *) Allow an application to disable the automatic SSL chain building.
3968 Before this a rather primitive chain build was always performed in
3969 ssl3_output_cert_chain(): an application had no way to send the
3970 correct chain if the automatic operation produced an incorrect result.
3972 Now the chain builder is disabled if either:
3974 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
3976 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
3978 The reasoning behind this is that an application would not want the
3979 auto chain building to take place if extra chain certificates are
3980 present and it might also want a means of sending no additional
3981 certificates (for example the chain has two certificates and the
3985 *) Add the possibility to build without the ENGINE framework.
3986 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3988 *) Under Win32 gmtime() can return NULL: check return value in
3989 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
3992 *) DSA routines: under certain error conditions uninitialized BN objects
3993 could be freed. Solution: make sure initialization is performed early
3994 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
3995 Nils Larsch <nla@trustcenter.de> via PR#459)
3998 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
3999 checked on reconnect on the client side, therefore session resumption
4000 could still fail with a "ssl session id is different" error. This
4001 behaviour is masked when SSL_OP_ALL is used due to
4002 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
4003 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4004 followup to PR #377.
4007 *) IA-32 assembler support enhancements: unified ELF targets, support
4008 for SCO/Caldera platforms, fix for Cygwin shared build.
4011 *) Add support for FreeBSD on sparc64. As a consequence, support for
4012 FreeBSD on non-x86 processors is separate from x86 processors on
4013 the config script, much like the NetBSD support.
4014 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
4016 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
4018 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
4021 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
4022 code (06) was taken as the first octet of the session ID and the last
4023 octet was ignored consequently. As a result SSLv2 client side session
4024 caching could not have worked due to the session ID mismatch between
4026 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4030 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
4031 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
4035 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
4036 seems that in spite of existing for more than a year, many application
4037 author have done nothing to provide the necessary callbacks, which
4038 means that this particular engine will not work properly anywhere.
4039 This is a very unfortunate situation which forces us, in the name
4040 of usability, to give the hw_ncipher.c a static lock, which is part
4042 NOTE: This is for the 0.9.7 series ONLY. This hack will never
4043 appear in 0.9.8 or later. We EXPECT application authors to have
4044 dealt properly with this when 0.9.8 is released (unless we actually
4045 make such changes in the libcrypto locking code that changes will
4046 have to be made anyway).
4049 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
4050 octets have been read, EOF or an error occurs. Without this change
4051 some truncated ASN1 structures will not produce an error.
4054 *) Disable Heimdal support, since it hasn't been fully implemented.
4055 Still give the possibility to force the use of Heimdal, but with
4056 warnings and a request that patches get sent to openssl-dev.
4059 *) Add the VC-CE target, introduce the WINCE sysname, and add
4060 INSTALL.WCE and appropriate conditionals to make it build.
4061 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4063 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
4064 cygssl-x.y.z.dll, where x, y and z are the major, minor and
4065 edit numbers of the version.
4066 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4068 *) Introduce safe string copy and catenation functions
4069 (BUF_strlcpy() and BUF_strlcat()).
4070 [Ben Laurie (CHATS) and Richard Levitte]
4072 *) Avoid using fixed-size buffers for one-line DNs.
4073 [Ben Laurie (CHATS)]
4075 *) Add BUF_MEM_grow_clean() to avoid information leakage when
4076 resizing buffers containing secrets, and use where appropriate.
4077 [Ben Laurie (CHATS)]
4079 *) Avoid using fixed size buffers for configuration file location.
4080 [Ben Laurie (CHATS)]
4082 *) Avoid filename truncation for various CA files.
4083 [Ben Laurie (CHATS)]
4085 *) Use sizeof in preference to magic numbers.
4086 [Ben Laurie (CHATS)]
4088 *) Avoid filename truncation in cert requests.
4089 [Ben Laurie (CHATS)]
4091 *) Add assertions to check for (supposedly impossible) buffer
4093 [Ben Laurie (CHATS)]
4095 *) Don't cache truncated DNS entries in the local cache (this could
4096 potentially lead to a spoofing attack).
4097 [Ben Laurie (CHATS)]
4099 *) Fix various buffers to be large enough for hex/decimal
4100 representations in a platform independent manner.
4101 [Ben Laurie (CHATS)]
4103 *) Add CRYPTO_realloc_clean() to avoid information leakage when
4104 resizing buffers containing secrets, and use where appropriate.
4105 [Ben Laurie (CHATS)]
4107 *) Add BIO_indent() to avoid much slightly worrying code to do
4109 [Ben Laurie (CHATS)]
4111 *) Convert sprintf()/BIO_puts() to BIO_printf().
4112 [Ben Laurie (CHATS)]
4114 *) buffer_gets() could terminate with the buffer only half
4116 [Ben Laurie (CHATS)]
4118 *) Add assertions to prevent user-supplied crypto functions from
4119 overflowing internal buffers by having large block sizes, etc.
4120 [Ben Laurie (CHATS)]
4122 *) New OPENSSL_assert() macro (similar to assert(), but enabled
4124 [Ben Laurie (CHATS)]
4126 *) Eliminate unused copy of key in RC4.
4127 [Ben Laurie (CHATS)]
4129 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
4130 [Ben Laurie (CHATS)]
4132 *) Fix off-by-one error in EGD path.
4133 [Ben Laurie (CHATS)]
4135 *) If RANDFILE path is too long, ignore instead of truncating.
4136 [Ben Laurie (CHATS)]
4138 *) Eliminate unused and incorrectly sized X.509 structure
4140 [Ben Laurie (CHATS)]
4142 *) Eliminate unused and dangerous function knumber().
4143 [Ben Laurie (CHATS)]
4145 *) Eliminate unused and dangerous structure, KSSL_ERR.
4146 [Ben Laurie (CHATS)]
4148 *) Protect against overlong session ID context length in an encoded
4149 session object. Since these are local, this does not appear to be
4151 [Ben Laurie (CHATS)]
4153 *) Change from security patch (see 0.9.6e below) that did not affect
4154 the 0.9.6 release series:
4156 Remote buffer overflow in SSL3 protocol - an attacker could
4157 supply an oversized master key in Kerberos-enabled versions.
4159 [Ben Laurie (CHATS)]
4161 *) Change the SSL kerb5 codes to match RFC 2712.
4164 *) Make -nameopt work fully for req and add -reqopt switch.
4165 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
4167 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
4168 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
4170 *) Make sure tests can be performed even if the corresponding algorithms
4171 have been removed entirely. This was also the last step to make
4172 OpenSSL compilable with DJGPP under all reasonable conditions.
4173 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
4175 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
4176 to allow version independent disabling of normally unselected ciphers,
4177 which may be activated as a side-effect of selecting a single cipher.
4179 (E.g., cipher list string "RSA" enables ciphersuites that are left
4180 out of "ALL" because they do not provide symmetric encryption.
4181 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
4182 [Lutz Jaenicke, Bodo Moeller]
4184 *) Add appropriate support for separate platform-dependent build
4185 directories. The recommended way to make a platform-dependent
4186 build directory is the following (tested on Linux), maybe with
4189 # Place yourself outside of the OpenSSL source tree. In
4190 # this example, the environment variable OPENSSL_SOURCE
4191 # is assumed to contain the absolute OpenSSL source directory.
4192 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
4193 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
4194 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
4195 mkdir -p `dirname $F`
4196 ln -s $OPENSSL_SOURCE/$F $F
4199 To be absolutely sure not to disturb the source tree, a "make clean"
4200 is a good thing. If it isn't successfull, don't worry about it,
4201 it probably means the source directory is very clean.
4204 *) Make sure any ENGINE control commands make local copies of string
4205 pointers passed to them whenever necessary. Otherwise it is possible
4206 the caller may have overwritten (or deallocated) the original string
4207 data when a later ENGINE operation tries to use the stored values.
4208 [Götz Babin-Ebell <babinebell@trustcenter.de>]
4210 *) Improve diagnostics in file reading and command-line digests.
4211 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
4213 *) Add AES modes CFB and OFB to the object database. Correct an
4214 error in AES-CFB decryption.
4217 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
4218 allows existing EVP_CIPHER_CTX structures to be reused after
4219 calling EVP_*Final(). This behaviour is used by encryption
4220 BIOs and some applications. This has the side effect that
4221 applications must explicitly clean up cipher contexts with
4222 EVP_CIPHER_CTX_cleanup() or they will leak memory.
4225 *) Check the values of dna and dnb in bn_mul_recursive before calling
4226 bn_mul_comba (a non zero value means the a or b arrays do not contain
4227 n2 elements) and fallback to bn_mul_normal if either is not zero.
4230 *) Fix escaping of non-ASCII characters when using the -subj option
4231 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
4234 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
4235 form for "surname", serialNumber has no short form.
4236 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
4237 therefore remove "mail" short name for "internet 7".
4238 The OID for unique identifiers in X509 certificates is
4239 x500UniqueIdentifier, not uniqueIdentifier.
4240 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
4243 *) Add an "init" command to the ENGINE config module and auto initialize
4244 ENGINEs. Without any "init" command the ENGINE will be initialized
4245 after all ctrl commands have been executed on it. If init=1 the
4246 ENGINE is initailized at that point (ctrls before that point are run
4247 on the uninitialized ENGINE and after on the initialized one). If
4248 init=0 then the ENGINE will not be iniatialized at all.
4251 *) Fix the 'app_verify_callback' interface so that the user-defined
4252 argument is actually passed to the callback: In the
4253 SSL_CTX_set_cert_verify_callback() prototype, the callback
4254 declaration has been changed from
4257 int (*cb)(X509_STORE_CTX *,void *);
4258 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
4259 i=s->ctx->app_verify_callback(&ctx)
4260 has been changed into
4261 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
4263 To update applications using SSL_CTX_set_cert_verify_callback(),
4264 a dummy argument can be added to their callback functions.
4265 [D. K. Smetters <smetters@parc.xerox.com>]
4267 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
4268 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
4270 *) Add and OPENSSL_LOAD_CONF define which will cause
4271 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
4272 This allows older applications to transparently support certain
4273 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
4274 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
4275 load the config file and OPENSSL_add_all_algorithms_conf() which will
4276 always load it have also been added.
4279 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
4280 Adjust NIDs and EVP layer.
4281 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4283 *) Config modules support in openssl utility.
4285 Most commands now load modules from the config file,
4286 though in a few (such as version) this isn't done
4287 because it couldn't be used for anything.
4289 In the case of ca and req the config file used is
4290 the same as the utility itself: that is the -config
4291 command line option can be used to specify an
4295 *) Move default behaviour from OPENSSL_config(). If appname is NULL
4296 use "openssl_conf" if filename is NULL use default openssl config file.
4299 *) Add an argument to OPENSSL_config() to allow the use of an alternative
4300 config section name. Add a new flag to tolerate a missing config file
4301 and move code to CONF_modules_load_file().
4304 *) Support for crypto accelerator cards from Accelerated Encryption
4305 Processing, www.aep.ie. (Use engine 'aep')
4306 The support was copied from 0.9.6c [engine] and adapted/corrected
4307 to work with the new engine framework.
4308 [AEP Inc. and Richard Levitte]
4310 *) Support for SureWare crypto accelerator cards from Baltimore
4311 Technologies. (Use engine 'sureware')
4312 The support was copied from 0.9.6c [engine] and adapted
4313 to work with the new engine framework.
4316 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
4317 make the newer ENGINE framework commands for the CHIL engine work.
4318 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
4320 *) Make it possible to produce shared libraries on ReliantUNIX.
4321 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
4323 *) Add the configuration target debug-linux-ppro.
4324 Make 'openssl rsa' use the general key loading routines
4325 implemented in apps.c, and make those routines able to
4326 handle the key format FORMAT_NETSCAPE and the variant
4328 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4330 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4331 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4333 *) Add -keyform to rsautl, and document -engine.
4334 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
4336 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
4337 BIO_R_NO_SUCH_FILE error code rather than the generic
4338 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
4341 *) Add new functions
4343 ERR_peek_last_error_line
4344 ERR_peek_last_error_line_data.
4345 These are similar to
4348 ERR_peek_error_line_data,
4349 but report on the latest error recorded rather than the first one
4350 still in the error queue.
4351 [Ben Laurie, Bodo Moeller]
4353 *) default_algorithms option in ENGINE config module. This allows things
4355 default_algorithms = ALL
4356 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
4359 *) Prelminary ENGINE config module.
4362 *) New experimental application configuration code.
4365 *) Change the AES code to follow the same name structure as all other
4366 symmetric ciphers, and behave the same way. Move everything to
4367 the directory crypto/aes, thereby obsoleting crypto/rijndael.
4368 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4370 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
4371 [Ben Laurie and Theo de Raadt]
4373 *) Add option to output public keys in req command.
4374 [Massimiliano Pala madwolf@openca.org]
4376 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
4377 (up to about 10% better than before for P-192 and P-224).
4380 *) New functions/macros
4382 SSL_CTX_set_msg_callback(ctx, cb)
4383 SSL_CTX_set_msg_callback_arg(ctx, arg)
4384 SSL_set_msg_callback(ssl, cb)
4385 SSL_set_msg_callback_arg(ssl, arg)
4387 to request calling a callback function
4389 void cb(int write_p, int version, int content_type,
4390 const void *buf, size_t len, SSL *ssl, void *arg)
4392 whenever a protocol message has been completely received
4393 (write_p == 0) or sent (write_p == 1). Here 'version' is the
4394 protocol version according to which the SSL library interprets
4395 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
4396 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
4397 the content type as defined in the SSL 3.0/TLS 1.0 protocol
4398 specification (change_cipher_spec(20), alert(21), handshake(22)).
4399 'buf' and 'len' point to the actual message, 'ssl' to the
4400 SSL object, and 'arg' is the application-defined value set by
4401 SSL[_CTX]_set_msg_callback_arg().
4403 'openssl s_client' and 'openssl s_server' have new '-msg' options
4404 to enable a callback that displays all protocol messages.
4407 *) Change the shared library support so shared libraries are built as
4408 soon as the corresponding static library is finished, and thereby get
4409 openssl and the test programs linked against the shared library.
4410 This still only happens when the keyword "shard" has been given to
4411 the configuration scripts.
4413 NOTE: shared library support is still an experimental thing, and
4414 backward binary compatibility is still not guaranteed.
4415 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
4417 *) Add support for Subject Information Access extension.
4418 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4420 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
4421 additional bytes when new memory had to be allocated, not just
4422 when reusing an existing buffer.
4425 *) New command line and configuration option 'utf8' for the req command.
4426 This allows field values to be specified as UTF8 strings.
4429 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
4430 runs for the former and machine-readable output for the latter.
4433 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
4434 of the e-mail address in the DN (i.e., it will go into a certificate
4435 extension only). The new configuration file option 'email_in_dn = no'
4436 has the same effect.
4437 [Massimiliano Pala madwolf@openca.org]
4439 *) Change all functions with names starting with des_ to be starting
4440 with DES_ instead. Add wrappers that are compatible with libdes,
4441 but are named _ossl_old_des_*. Finally, add macros that map the
4442 des_* symbols to the corresponding _ossl_old_des_* if libdes
4443 compatibility is desired. If OpenSSL 0.9.6c compatibility is
4444 desired, the des_* symbols will be mapped to DES_*, with one
4447 Since we provide two compatibility mappings, the user needs to
4448 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
4449 compatibility is desired. The default (i.e., when that macro
4450 isn't defined) is OpenSSL 0.9.6c compatibility.
4452 There are also macros that enable and disable the support of old
4453 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
4454 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
4455 are defined, the default will apply: to support the old des routines.
4457 In either case, one must include openssl/des.h to get the correct
4458 definitions. Do not try to just include openssl/des_old.h, that
4461 NOTE: This is a major break of an old API into a new one. Software
4462 authors are encouraged to switch to the DES_ style functions. Some
4463 time in the future, des_old.h and the libdes compatibility functions
4464 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
4465 default), and then completely removed.
4468 *) Test for certificates which contain unsupported critical extensions.
4469 If such a certificate is found during a verify operation it is
4470 rejected by default: this behaviour can be overridden by either
4471 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
4472 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
4473 X509_supported_extension() has also been added which returns 1 if a
4474 particular extension is supported.
4477 *) Modify the behaviour of EVP cipher functions in similar way to digests
4478 to retain compatibility with existing code.
4481 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
4482 compatibility with existing code. In particular the 'ctx' parameter does
4483 not have to be to be initialized before the call to EVP_DigestInit() and
4484 it is tidied up after a call to EVP_DigestFinal(). New function
4485 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
4486 EVP_MD_CTX_copy() changed to not require the destination to be
4487 initialized valid and new function EVP_MD_CTX_copy_ex() added which
4488 requires the destination to be valid.
4490 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
4491 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
4494 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
4495 so that complete 'Handshake' protocol structures are kept in memory
4496 instead of overwriting 'msg_type' and 'length' with 'body' data.
4499 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
4500 [Massimo Santin via Richard Levitte]
4502 *) Major restructuring to the underlying ENGINE code. This includes
4503 reduction of linker bloat, separation of pure "ENGINE" manipulation
4504 (initialisation, etc) from functionality dealing with implementations
4505 of specific crypto iterfaces. This change also introduces integrated
4506 support for symmetric ciphers and digest implementations - so ENGINEs
4507 can now accelerate these by providing EVP_CIPHER and EVP_MD
4508 implementations of their own. This is detailed in crypto/engine/README
4509 as it couldn't be adequately described here. However, there are a few
4510 API changes worth noting - some RSA, DSA, DH, and RAND functions that
4511 were changed in the original introduction of ENGINE code have now
4512 reverted back - the hooking from this code to ENGINE is now a good
4513 deal more passive and at run-time, operations deal directly with
4514 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
4515 dereferencing through an ENGINE pointer any more. Also, the ENGINE
4516 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
4517 they were not being used by the framework as there is no concept of a
4518 BIGNUM_METHOD and they could not be generalised to the new
4519 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
4520 ENGINE_cpy() has been removed as it cannot be consistently defined in
4524 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
4527 *) Change mkdef.pl to sort symbols that get the same entry number,
4528 and make sure the automatically generated functions ERR_load_*
4529 become part of libeay.num as well.
4532 *) New function SSL_renegotiate_pending(). This returns true once
4533 renegotiation has been requested (either SSL_renegotiate() call
4534 or HelloRequest/ClientHello receveived from the peer) and becomes
4535 false once a handshake has been completed.
4536 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
4537 sends a HelloRequest, but does not ensure that a handshake takes
4538 place. SSL_renegotiate_pending() is useful for checking if the
4539 client has followed the request.)
4542 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
4543 By default, clients may request session resumption even during
4544 renegotiation (if session ID contexts permit); with this option,
4545 session resumption is possible only in the first handshake.
4547 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4548 more bits available for options that should not be part of
4549 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4552 *) Add some demos for certificate and certificate request creation.
4555 *) Make maximum certificate chain size accepted from the peer application
4556 settable (SSL*_get/set_max_cert_list()), as proposed by
4557 "Douglas E. Engert" <deengert@anl.gov>.
4560 *) Add support for shared libraries for Unixware-7
4561 (Boyd Lynn Gerber <gerberb@zenez.com>).
4564 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4565 be done prior to destruction. Use this to unload error strings from
4566 ENGINEs that load their own error strings. NB: This adds two new API
4567 functions to "get" and "set" this destroy handler in an ENGINE.
4570 *) Alter all existing ENGINE implementations (except "openssl" and
4571 "openbsd") to dynamically instantiate their own error strings. This
4572 makes them more flexible to be built both as statically-linked ENGINEs
4573 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4574 Also, add stub code to each that makes building them as self-contained
4575 shared-libraries easier (see README.ENGINE).
4578 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4579 implementations into applications that are completely implemented in
4580 self-contained shared-libraries. The "dynamic" ENGINE exposes control
4581 commands that can be used to configure what shared-library to load and
4582 to control aspects of the way it is handled. Also, made an update to
4583 the README.ENGINE file that brings its information up-to-date and
4584 provides some information and instructions on the "dynamic" ENGINE
4585 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
4588 *) Make it possible to unload ranges of ERR strings with a new
4589 "ERR_unload_strings" function.
4592 *) Add a copy() function to EVP_MD.
4595 *) Make EVP_MD routines take a context pointer instead of just the
4596 md_data void pointer.
4599 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
4600 that the digest can only process a single chunk of data
4601 (typically because it is provided by a piece of
4602 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
4603 is only going to provide a single chunk of data, and hence the
4604 framework needn't accumulate the data for oneshot drivers.
4607 *) As with "ERR", make it possible to replace the underlying "ex_data"
4608 functions. This change also alters the storage and management of global
4609 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
4610 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
4611 index counters. The API functions that use this state have been changed
4612 to take a "class_index" rather than pointers to the class's local STACK
4613 and counter, and there is now an API function to dynamically create new
4614 classes. This centralisation allows us to (a) plug a lot of the
4615 thread-safety problems that existed, and (b) makes it possible to clean
4616 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
4617 such data would previously have always leaked in application code and
4618 workarounds were in place to make the memory debugging turn a blind eye
4619 to it. Application code that doesn't use this new function will still
4620 leak as before, but their memory debugging output will announce it now
4621 rather than letting it slide.
4623 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
4624 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
4625 has a return value to indicate success or failure.
4628 *) Make it possible to replace the underlying "ERR" functions such that the
4629 global state (2 LHASH tables and 2 locks) is only used by the "default"
4630 implementation. This change also adds two functions to "get" and "set"
4631 the implementation prior to it being automatically set the first time
4632 any other ERR function takes place. Ie. an application can call "get",
4633 pass the return value to a module it has just loaded, and that module
4634 can call its own "set" function using that value. This means the
4635 module's "ERR" operations will use (and modify) the error state in the
4636 application and not in its own statically linked copy of OpenSSL code.
4639 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
4640 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
4641 the operation, and provides a more encapsulated way for external code
4642 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
4643 to use these functions rather than manually incrementing the counts.
4645 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
4648 *) Add EVP test program.
4651 *) Add symmetric cipher support to ENGINE. Expect the API to change!
4654 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
4655 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
4656 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
4657 These allow a CRL to be built without having to access X509_CRL fields
4658 directly. Modify 'ca' application to use new functions.
4661 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
4662 bug workarounds. Rollback attack detection is a security feature.
4663 The problem will only arise on OpenSSL servers when TLSv1 is not
4664 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
4665 Software authors not wanting to support TLSv1 will have special reasons
4666 for their choice and can explicitly enable this option.
4667 [Bodo Moeller, Lutz Jaenicke]
4669 *) Rationalise EVP so it can be extended: don't include a union of
4670 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
4671 (similar to those existing for EVP_CIPHER_CTX).
4676 EVP_MD_CTX_init(&md); /* new function call */
4677 EVP_DigestInit(&md, EVP_sha1());
4678 EVP_DigestUpdate(&md, in, len);
4679 EVP_DigestFinal(&md, out, NULL);
4680 EVP_MD_CTX_cleanup(&md); /* new function call */
4684 *) Make DES key schedule conform to the usual scheme, as well as
4685 correcting its structure. This means that calls to DES functions
4686 now have to pass a pointer to a des_key_schedule instead of a
4687 plain des_key_schedule (which was actually always a pointer
4690 des_key_schedule ks;
4692 des_set_key_checked(..., &ks);
4693 des_ncbc_encrypt(..., &ks, ...);
4695 (Note that a later change renames 'des_...' into 'DES_...'.)
4698 *) Initial reduction of linker bloat: the use of some functions, such as
4699 PEM causes large amounts of unused functions to be linked in due to
4700 poor organisation. For example pem_all.c contains every PEM function
4701 which has a knock on effect of linking in large amounts of (unused)
4702 ASN1 code. Grouping together similar functions and splitting unrelated
4703 functions prevents this.
4706 *) Cleanup of EVP macros.
4709 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
4710 correct _ecb suffix.
4713 *) Add initial OCSP responder support to ocsp application. The
4714 revocation information is handled using the text based index
4715 use by the ca application. The responder can either handle
4716 requests generated internally, supplied in files (for example
4717 via a CGI script) or using an internal minimal server.
4720 *) Add configuration choices to get zlib compression for TLS.
4723 *) Changes to Kerberos SSL for RFC 2712 compliance:
4724 1. Implemented real KerberosWrapper, instead of just using
4725 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
4726 2. Implemented optional authenticator field of KerberosWrapper.
4728 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
4729 and authenticator structs; see crypto/krb5/.
4731 Generalized Kerberos calls to support multiple Kerberos libraries.
4732 [Vern Staats <staatsvr@asc.hpc.mil>,
4733 Jeffrey Altman <jaltman@columbia.edu>
4734 via Richard Levitte]
4736 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
4737 already does with RSA. testdsa.h now has 'priv_key/pub_key'
4738 values for each of the key sizes rather than having just
4739 parameters (and 'speed' generating keys each time).
4742 *) Speed up EVP routines.
4745 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
4746 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
4747 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
4748 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
4750 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
4751 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
4752 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
4755 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
4757 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
4760 *) Added the OS2-EMX target.
4761 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
4763 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
4764 to support NCONF routines in extension code. New function CONF_set_nconf()
4765 to allow functions which take an NCONF to also handle the old LHASH
4766 structure: this means that the old CONF compatible routines can be
4767 retained (in particular wrt extensions) without having to duplicate the
4768 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
4771 *) Enhance the general user interface with mechanisms for inner control
4772 and with possibilities to have yes/no kind of prompts.
4775 *) Change all calls to low level digest routines in the library and
4776 applications to use EVP. Add missing calls to HMAC_cleanup() and
4777 don't assume HMAC_CTX can be copied using memcpy().
4778 [Verdon Walker <VWalker@novell.com>, Steve Henson]
4780 *) Add the possibility to control engines through control names but with
4781 arbitrary arguments instead of just a string.
4782 Change the key loaders to take a UI_METHOD instead of a callback
4783 function pointer. NOTE: this breaks binary compatibility with earlier
4784 versions of OpenSSL [engine].
4785 Adapt the nCipher code for these new conditions and add a card insertion
4789 *) Enhance the general user interface with mechanisms to better support
4790 dialog box interfaces, application-defined prompts, the possibility
4791 to use defaults (for example default passwords from somewhere else)
4792 and interrupts/cancellations.
4795 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
4796 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
4799 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
4800 tidy up some unnecessarily weird code in 'sk_new()').
4801 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
4803 *) Change the key loading routines for ENGINEs to use the same kind
4804 callback (pem_password_cb) as all other routines that need this
4808 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
4809 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
4810 than this minimum value is recommended.
4813 *) New random seeder for OpenVMS, using the system process statistics
4814 that are easily reachable.
4817 *) Windows apparently can't transparently handle global
4818 variables defined in DLLs. Initialisations such as:
4820 const ASN1_ITEM *it = &ASN1_INTEGER_it;
4822 wont compile. This is used by the any applications that need to
4823 declare their own ASN1 modules. This was fixed by adding the option
4824 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
4825 needed for static libraries under Win32.
4828 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
4829 setting of purpose and trust fields. New X509_STORE trust and
4830 purpose functions and tidy up setting in other SSL functions.
4833 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
4834 structure. These are inherited by X509_STORE_CTX when it is
4835 initialised. This allows various defaults to be set in the
4836 X509_STORE structure (such as flags for CRL checking and custom
4837 purpose or trust settings) for functions which only use X509_STORE_CTX
4838 internally such as S/MIME.
4840 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
4841 trust settings if they are not set in X509_STORE. This allows X509_STORE
4842 purposes and trust (in S/MIME for example) to override any set by default.
4844 Add command line options for CRL checking to smime, s_client and s_server
4848 *) Initial CRL based revocation checking. If the CRL checking flag(s)
4849 are set then the CRL is looked up in the X509_STORE structure and
4850 its validity and signature checked, then if the certificate is found
4851 in the CRL the verify fails with a revoked error.
4853 Various new CRL related callbacks added to X509_STORE_CTX structure.
4855 Command line options added to 'verify' application to support this.
4857 This needs some additional work, such as being able to handle multiple
4858 CRLs with different times, extension based lookup (rather than just
4859 by subject name) and ultimately more complete V2 CRL extension
4863 *) Add a general user interface API (crypto/ui/). This is designed
4864 to replace things like des_read_password and friends (backward
4865 compatibility functions using this new API are provided).
4866 The purpose is to remove prompting functions from the DES code
4867 section as well as provide for prompting through dialog boxes in
4868 a window system and the like.
4871 *) Add "ex_data" support to ENGINE so implementations can add state at a
4872 per-structure level rather than having to store it globally.
4875 *) Make it possible for ENGINE structures to be copied when retrieved by
4876 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
4877 This causes the "original" ENGINE structure to act like a template,
4878 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
4879 operational state can be localised to each ENGINE structure, despite the
4880 fact they all share the same "methods". New ENGINE structures returned in
4881 this case have no functional references and the return value is the single
4882 structural reference. This matches the single structural reference returned
4883 by ENGINE_by_id() normally, when it is incremented on the pre-existing
4887 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
4888 needs to match any other type at all we need to manually clear the
4892 *) Changes to the "openssl engine" utility to include;
4893 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
4894 about an ENGINE's available control commands.
4895 - executing control commands from command line arguments using the
4896 '-pre' and '-post' switches. '-post' is only used if '-t' is
4897 specified and the ENGINE is successfully initialised. The syntax for
4898 the individual commands are colon-separated, for example;
4899 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
4902 *) New dynamic control command support for ENGINEs. ENGINEs can now
4903 declare their own commands (numbers), names (strings), descriptions,
4904 and input types for run-time discovery by calling applications. A
4905 subset of these commands are implicitly classed as "executable"
4906 depending on their input type, and only these can be invoked through
4907 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
4908 can be based on user input, config files, etc). The distinction is
4909 that "executable" commands cannot return anything other than a boolean
4910 result and can only support numeric or string input, whereas some
4911 discoverable commands may only be for direct use through
4912 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
4913 pointers, or other custom uses. The "executable" commands are to
4914 support parameterisations of ENGINE behaviour that can be
4915 unambiguously defined by ENGINEs and used consistently across any
4916 OpenSSL-based application. Commands have been added to all the
4917 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
4918 control over shared-library paths without source code alterations.
4921 *) Changed all ENGINE implementations to dynamically allocate their
4922 ENGINEs rather than declaring them statically. Apart from this being
4923 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
4924 this also allows the implementations to compile without using the
4925 internal engine_int.h header.
4928 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
4929 'const' value. Any code that should be able to modify a RAND_METHOD
4930 should already have non-const pointers to it (ie. they should only
4931 modify their own ones).
4934 *) Made a variety of little tweaks to the ENGINE code.
4935 - "atalla" and "ubsec" string definitions were moved from header files
4936 to C code. "nuron" string definitions were placed in variables
4937 rather than hard-coded - allowing parameterisation of these values
4938 later on via ctrl() commands.
4939 - Removed unused "#if 0"'d code.
4940 - Fixed engine list iteration code so it uses ENGINE_free() to release
4941 structural references.
4942 - Constified the RAND_METHOD element of ENGINE structures.
4943 - Constified various get/set functions as appropriate and added
4944 missing functions (including a catch-all ENGINE_cpy that duplicates
4945 all ENGINE values onto a new ENGINE except reference counts/state).
4946 - Removed NULL parameter checks in get/set functions. Setting a method
4947 or function to NULL is a way of cancelling out a previously set
4948 value. Passing a NULL ENGINE parameter is just plain stupid anyway
4949 and doesn't justify the extra error symbols and code.
4950 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
4951 flags from engine_int.h to engine.h.
4952 - Changed prototypes for ENGINE handler functions (init(), finish(),
4953 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
4956 *) Implement binary inversion algorithm for BN_mod_inverse in addition
4957 to the algorithm using long division. The binary algorithm can be
4958 used only if the modulus is odd. On 32-bit systems, it is faster
4959 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
4960 roughly 5-15% for 256-bit moduli), so we use it only for moduli
4961 up to 450 bits. In 64-bit environments, the binary algorithm
4962 appears to be advantageous for much longer moduli; here we use it
4963 for moduli up to 2048 bits.
4966 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
4967 could not support the combine flag in choice fields.
4970 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
4971 extensions from a certificate request to the certificate.
4974 *) Allow multiple 'certopt' and 'nameopt' options to be separated
4975 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
4976 file: this allows the display of the certificate about to be
4977 signed to be customised, to allow certain fields to be included
4978 or excluded and extension details. The old system didn't display
4979 multicharacter strings properly, omitted fields not in the policy
4980 and couldn't display additional details such as extensions.
4983 *) Function EC_POINTs_mul for multiple scalar multiplication
4984 of an arbitrary number of elliptic curve points
4985 \sum scalars[i]*points[i],
4986 optionally including the generator defined for the EC_GROUP:
4987 scalar*generator + \sum scalars[i]*points[i].
4989 EC_POINT_mul is a simple wrapper function for the typical case
4990 that the point list has just one item (besides the optional
4994 *) First EC_METHODs for curves over GF(p):
4996 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
4997 operations and provides various method functions that can also
4998 operate with faster implementations of modular arithmetic.
5000 EC_GFp_mont_method() reuses most functions that are part of
5001 EC_GFp_simple_method, but uses Montgomery arithmetic.
5003 [Bodo Moeller; point addition and point doubling
5004 implementation directly derived from source code provided by
5005 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
5007 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
5008 crypto/ec/ec_lib.c):
5010 Curves are EC_GROUP objects (with an optional group generator)
5011 based on EC_METHODs that are built into the library.
5013 Points are EC_POINT objects based on EC_GROUP objects.
5015 Most of the framework would be able to handle curves over arbitrary
5016 finite fields, but as there are no obvious types for fields other
5017 than GF(p), some functions are limited to that for now.
5020 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
5021 that the file contains a complete HTTP response.
5024 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
5025 change the def and num file printf format specifier from "%-40sXXX"
5026 to "%-39s XXX". The latter will always guarantee a space after the
5027 field while the former will cause them to run together if the field
5028 is 40 of more characters long.
5031 *) Constify the cipher and digest 'method' functions and structures
5032 and modify related functions to take constant EVP_MD and EVP_CIPHER
5036 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
5037 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
5040 *) Modify EVP_Digest*() routines so they now return values. Although the
5041 internal software routines can never fail additional hardware versions
5045 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
5047 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
5048 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
5051 ERR_R_NESTED_ASN1_ERROR
5053 ERR_R_MISSING_ASN1_EOS
5054 were 4 .. 9, conflicting with
5055 ERR_LIB_RSA (= ERR_R_RSA_LIB)
5057 ERR_LIB_PEM (= ERR_R_PEM_LIB).
5058 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
5060 Add new error code 'ERR_R_INTERNAL_ERROR'.
5063 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
5067 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
5068 sets the subject name for a new request or supersedes the
5069 subject name in a given request. Formats that can be parsed are
5070 'CN=Some Name, OU=myOU, C=IT'
5072 'CN=Some Name/OU=myOU/C=IT'.
5074 Add options '-batch' and '-verbose' to 'openssl req'.
5075 [Massimiliano Pala <madwolf@hackmasters.net>]
5077 *) Introduce the possibility to access global variables through
5078 functions on platform were that's the best way to handle exporting
5079 global variables in shared libraries. To enable this functionality,
5080 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
5081 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
5082 is normally done by Configure or something similar).
5084 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
5085 in the source file (foo.c) like this:
5087 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
5088 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
5090 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
5091 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
5093 OPENSSL_DECLARE_GLOBAL(int,foo);
5094 #define foo OPENSSL_GLOBAL_REF(foo)
5095 OPENSSL_DECLARE_GLOBAL(double,bar);
5096 #define bar OPENSSL_GLOBAL_REF(bar)
5098 The #defines are very important, and therefore so is including the
5099 header file everywhere where the defined globals are used.
5101 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
5102 of ASN.1 items, but that structure is a bit different.
5104 The largest change is in util/mkdef.pl which has been enhanced with
5105 better and easier to understand logic to choose which symbols should
5106 go into the Windows .def files as well as a number of fixes and code
5107 cleanup (among others, algorithm keywords are now sorted
5108 lexicographically to avoid constant rewrites).
5111 *) In BN_div() keep a copy of the sign of 'num' before writing the
5112 result to 'rm' because if rm==num the value will be overwritten
5113 and produce the wrong result if 'num' is negative: this caused
5114 problems with BN_mod() and BN_nnmod().
5117 *) Function OCSP_request_verify(). This checks the signature on an
5118 OCSP request and verifies the signer certificate. The signer
5119 certificate is just checked for a generic purpose and OCSP request
5123 *) Add OCSP_check_validity() function to check the validity of OCSP
5124 responses. OCSP responses are prepared in real time and may only
5125 be a few seconds old. Simply checking that the current time lies
5126 between thisUpdate and nextUpdate max reject otherwise valid responses
5127 caused by either OCSP responder or client clock inaccuracy. Instead
5128 we allow thisUpdate and nextUpdate to fall within a certain period of
5129 the current time. The age of the response can also optionally be
5130 checked. Two new options -validity_period and -status_age added to
5134 *) If signature or public key algorithm is unrecognized print out its
5135 OID rather that just UNKNOWN.
5138 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
5139 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
5140 ID to be generated from the issuer certificate alone which can then be
5141 passed to OCSP_id_issuer_cmp().
5144 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
5145 ASN1 modules to export functions returning ASN1_ITEM pointers
5146 instead of the ASN1_ITEM structures themselves. This adds several
5147 new macros which allow the underlying ASN1 function/structure to
5148 be accessed transparently. As a result code should not use ASN1_ITEM
5149 references directly (such as &X509_it) but instead use the relevant
5150 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
5151 use of the new ASN1 code on platforms where exporting structures
5152 is problematical (for example in shared libraries) but exporting
5153 functions returning pointers to structures is not.
5156 *) Add support for overriding the generation of SSL/TLS session IDs.
5157 These callbacks can be registered either in an SSL_CTX or per SSL.
5158 The purpose of this is to allow applications to control, if they wish,
5159 the arbitrary values chosen for use as session IDs, particularly as it
5160 can be useful for session caching in multiple-server environments. A
5161 command-line switch for testing this (and any client code that wishes
5162 to use such a feature) has been added to "s_server".
5163 [Geoff Thorpe, Lutz Jaenicke]
5165 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
5166 of the form '#if defined(...) || defined(...) || ...' and
5167 '#if !defined(...) && !defined(...) && ...'. This also avoids
5168 the growing number of special cases it was previously handling.
5171 *) Make all configuration macros available for application by making
5172 sure they are available in opensslconf.h, by giving them names starting
5173 with "OPENSSL_" to avoid conflicts with other packages and by making
5174 sure e_os2.h will cover all platform-specific cases together with
5176 Additionally, it is now possible to define configuration/platform-
5177 specific names (called "system identities"). In the C code, these
5178 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
5179 macro with the name beginning with "OPENSSL_SYS_", which is determined
5180 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
5184 *) New option -set_serial to 'req' and 'x509' this allows the serial
5185 number to use to be specified on the command line. Previously self
5186 signed certificates were hard coded with serial number 0 and the
5187 CA options of 'x509' had to use a serial number in a file which was
5191 *) New options to 'ca' utility to support V2 CRL entry extensions.
5192 Currently CRL reason, invalidity date and hold instruction are
5193 supported. Add new CRL extensions to V3 code and some new objects.
5196 *) New function EVP_CIPHER_CTX_set_padding() this is used to
5197 disable standard block padding (aka PKCS#5 padding) in the EVP
5198 API, which was previously mandatory. This means that the data is
5199 not padded in any way and so the total length much be a multiple
5200 of the block size, otherwise an error occurs.
5203 *) Initial (incomplete) OCSP SSL support.
5206 *) New function OCSP_parse_url(). This splits up a URL into its host,
5207 port and path components: primarily to parse OCSP URLs. New -url
5208 option to ocsp utility.
5211 *) New nonce behavior. The return value of OCSP_check_nonce() now
5212 reflects the various checks performed. Applications can decide
5213 whether to tolerate certain situations such as an absent nonce
5214 in a response when one was present in a request: the ocsp application
5215 just prints out a warning. New function OCSP_add1_basic_nonce()
5216 this is to allow responders to include a nonce in a response even if
5217 the request is nonce-less.
5220 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
5221 skipped when using openssl x509 multiple times on a single input file,
5222 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
5225 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
5226 set string type: to handle setting ASN1_TIME structures. Fix ca
5227 utility to correctly initialize revocation date of CRLs.
5230 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
5231 the clients preferred ciphersuites and rather use its own preferences.
5232 Should help to work around M$ SGC (Server Gated Cryptography) bug in
5233 Internet Explorer by ensuring unchanged hash method during stepup.
5234 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
5237 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
5238 to aes and add a new 'exist' option to print out symbols that don't
5242 *) Additional options to ocsp utility to allow flags to be set and
5243 additional certificates supplied.
5246 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
5247 OCSP client a number of certificate to only verify the response
5251 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
5252 handle the new API. Currently only ECB, CBC modes supported. Add new
5255 Add TLS AES ciphersuites as described in RFC3268, "Advanced
5256 Encryption Standard (AES) Ciphersuites for Transport Layer
5257 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
5258 not enabled by default and were not part of the "ALL" ciphersuite
5259 alias because they were not yet official; they could be
5260 explicitly requested by specifying the "AESdraft" ciphersuite
5261 group alias. In the final release of OpenSSL 0.9.7, the group
5262 alias is called "AES" and is part of "ALL".)
5263 [Ben Laurie, Steve Henson, Bodo Moeller]
5265 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
5266 request to response.
5269 *) Functions for OCSP responders. OCSP_request_onereq_count(),
5270 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
5271 extract information from a certificate request. OCSP_response_create()
5272 creates a response and optionally adds a basic response structure.
5273 OCSP_basic_add1_status() adds a complete single response to a basic
5274 response and returns the OCSP_SINGLERESP structure just added (to allow
5275 extensions to be included for example). OCSP_basic_add1_cert() adds a
5276 certificate to a basic response and OCSP_basic_sign() signs a basic
5277 response with various flags. New helper functions ASN1_TIME_check()
5278 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
5279 (converts ASN1_TIME to GeneralizedTime).
5282 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
5283 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
5284 structure from a certificate. X509_pubkey_digest() digests the public_key
5285 contents: this is used in various key identifiers.
5288 *) Make sk_sort() tolerate a NULL argument.
5289 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
5291 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
5292 passed by the function are trusted implicitly. If any of them signed the
5293 response then it is assumed to be valid and is not verified.
5296 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
5297 to data. This was previously part of the PKCS7 ASN1 code. This
5298 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
5299 [Steve Henson, reported by Kenneth R. Robinette
5300 <support@securenetterm.com>]
5302 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
5303 routines: without these tracing memory leaks is very painful.
5304 Fix leaks in PKCS12 and PKCS7 routines.
5307 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
5308 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
5309 effectively meant GeneralizedTime would never be used. Now it
5310 is initialised to -1 but X509_time_adj() now has to check the value
5311 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
5312 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
5313 [Steve Henson, reported by Kenneth R. Robinette
5314 <support@securenetterm.com>]
5316 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
5317 result in a zero length in the ASN1_INTEGER structure which was
5318 not consistent with the structure when d2i_ASN1_INTEGER() was used
5319 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
5320 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
5321 where it did not print out a minus for negative ASN1_INTEGER.
5324 *) Add summary printout to ocsp utility. The various functions which
5325 convert status values to strings have been renamed to:
5326 OCSP_response_status_str(), OCSP_cert_status_str() and
5327 OCSP_crl_reason_str() and are no longer static. New options
5328 to verify nonce values and to disable verification. OCSP response
5329 printout format cleaned up.
5332 *) Add additional OCSP certificate checks. These are those specified
5333 in RFC2560. This consists of two separate checks: the CA of the
5334 certificate being checked must either be the OCSP signer certificate
5335 or the issuer of the OCSP signer certificate. In the latter case the
5336 OCSP signer certificate must contain the OCSP signing extended key
5337 usage. This check is performed by attempting to match the OCSP
5338 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
5339 in the OCSP_CERTID structures of the response.
5342 *) Initial OCSP certificate verification added to OCSP_basic_verify()
5343 and related routines. This uses the standard OpenSSL certificate
5344 verify routines to perform initial checks (just CA validity) and
5345 to obtain the certificate chain. Then additional checks will be
5346 performed on the chain. Currently the root CA is checked to see
5347 if it is explicitly trusted for OCSP signing. This is used to set
5348 a root CA as a global signing root: that is any certificate that
5349 chains to that CA is an acceptable OCSP signing certificate.
5352 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
5353 extensions from a separate configuration file.
5354 As when reading extensions from the main configuration file,
5355 the '-extensions ...' option may be used for specifying the
5357 [Massimiliano Pala <madwolf@comune.modena.it>]
5359 *) New OCSP utility. Allows OCSP requests to be generated or
5360 read. The request can be sent to a responder and the output
5361 parsed, outputed or printed in text form. Not complete yet:
5362 still needs to check the OCSP response validity.
5365 *) New subcommands for 'openssl ca':
5366 'openssl ca -status <serial>' prints the status of the cert with
5367 the given serial number (according to the index file).
5368 'openssl ca -updatedb' updates the expiry status of certificates
5370 [Massimiliano Pala <madwolf@comune.modena.it>]
5372 *) New '-newreq-nodes' command option to CA.pl. This is like
5373 '-newreq', but calls 'openssl req' with the '-nodes' option
5374 so that the resulting key is not encrypted.
5375 [Damien Miller <djm@mindrot.org>]
5377 *) New configuration for the GNU Hurd.
5378 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
5380 *) Initial code to implement OCSP basic response verify. This
5381 is currently incomplete. Currently just finds the signer's
5382 certificate and verifies the signature on the response.
5385 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
5386 value of OPENSSLDIR. This is available via the new '-d' option
5387 to 'openssl version', and is also included in 'openssl version -a'.
5390 *) Allowing defining memory allocation callbacks that will be given
5391 file name and line number information in additional arguments
5392 (a const char* and an int). The basic functionality remains, as
5393 well as the original possibility to just replace malloc(),
5394 realloc() and free() by functions that do not know about these
5395 additional arguments. To register and find out the current
5396 settings for extended allocation functions, the following
5397 functions are provided:
5399 CRYPTO_set_mem_ex_functions
5400 CRYPTO_set_locked_mem_ex_functions
5401 CRYPTO_get_mem_ex_functions
5402 CRYPTO_get_locked_mem_ex_functions
5404 These work the same way as CRYPTO_set_mem_functions and friends.
5405 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
5406 extended allocation function is enabled.
5407 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
5408 a conventional allocation function is enabled.
5409 [Richard Levitte, Bodo Moeller]
5411 *) Finish off removing the remaining LHASH function pointer casts.
5412 There should no longer be any prototype-casting required when using
5413 the LHASH abstraction, and any casts that remain are "bugs". See
5414 the callback types and macros at the head of lhash.h for details
5415 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
5418 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
5419 If /dev/[u]random devices are not available or do not return enough
5420 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
5422 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
5423 /etc/entropy will be queried once each in this sequence, quering stops
5424 when enough entropy was collected without querying more sockets.
5427 *) Change the Unix RAND_poll() variant to be able to poll several
5428 random devices, as specified by DEVRANDOM, until a sufficient amount
5429 of data has been collected. We spend at most 10 ms on each file
5430 (select timeout) and read in non-blocking mode. DEVRANDOM now
5431 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
5432 (previously it was just the string "/dev/urandom"), so on typical
5433 platforms the 10 ms delay will never occur.
5434 Also separate out the Unix variant to its own file, rand_unix.c.
5435 For VMS, there's a currently-empty rand_vms.c.
5438 *) Move OCSP client related routines to ocsp_cl.c. These
5439 provide utility functions which an application needing
5440 to issue a request to an OCSP responder and analyse the
5441 response will typically need: as opposed to those which an
5442 OCSP responder itself would need which will be added later.
5444 OCSP_request_sign() signs an OCSP request with an API similar
5445 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
5446 response. OCSP_response_get1_basic() extracts basic response
5447 from response. OCSP_resp_find_status(): finds and extracts status
5448 information from an OCSP_CERTID structure (which will be created
5449 when the request structure is built). These are built from lower
5450 level functions which work on OCSP_SINGLERESP structures but
5451 wont normally be used unless the application wishes to examine
5452 extensions in the OCSP response for example.
5454 Replace nonce routines with a pair of functions.
5455 OCSP_request_add1_nonce() adds a nonce value and optionally
5456 generates a random value. OCSP_check_nonce() checks the
5457 validity of the nonce in an OCSP response.
5460 *) Change function OCSP_request_add() to OCSP_request_add0_id().
5461 This doesn't copy the supplied OCSP_CERTID and avoids the
5462 need to free up the newly created id. Change return type
5463 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
5464 This can then be used to add extensions to the request.
5465 Deleted OCSP_request_new(), since most of its functionality
5466 is now in OCSP_REQUEST_new() (and the case insensitive name
5467 clash) apart from the ability to set the request name which
5468 will be added elsewhere.
5471 *) Update OCSP API. Remove obsolete extensions argument from
5472 various functions. Extensions are now handled using the new
5473 OCSP extension code. New simple OCSP HTTP function which
5474 can be used to send requests and parse the response.
5477 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
5478 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
5479 uses the special reorder version of SET OF to sort the attributes
5480 and reorder them to match the encoded order. This resolves a long
5481 standing problem: a verify on a PKCS7 structure just after signing
5482 it used to fail because the attribute order did not match the
5483 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
5484 it uses the received order. This is necessary to tolerate some broken
5485 software that does not order SET OF. This is handled by encoding
5486 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
5487 to produce the required SET OF.
5490 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
5491 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
5492 files to get correct declarations of the ASN.1 item variables.
5495 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
5496 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
5497 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
5498 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
5499 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
5500 ASN1_ITEM and no wrapper functions.
5503 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
5504 replace the old function pointer based I/O routines. Change most of
5505 the *_d2i_bio() and *_d2i_fp() functions to use these.
5508 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
5509 lines, recognice more "algorithms" that can be deselected, and make
5510 it complain about algorithm deselection that isn't recognised.
5513 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
5514 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
5515 to use new functions. Add NO_ASN1_OLD which can be set to remove
5516 some old style ASN1 functions: this can be used to determine if old
5517 code will still work when these eventually go away.
5520 *) New extension functions for OCSP structures, these follow the
5521 same conventions as certificates and CRLs.
5524 *) New function X509V3_add1_i2d(). This automatically encodes and
5525 adds an extension. Its behaviour can be customised with various
5526 flags to append, replace or delete. Various wrappers added for
5527 certifcates and CRLs.
5530 *) Fix to avoid calling the underlying ASN1 print routine when
5531 an extension cannot be parsed. Correct a typo in the
5532 OCSP_SERVICELOC extension. Tidy up print OCSP format.
5535 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
5536 entries for variables.
5539 *) Add functionality to apps/openssl.c for detecting locking
5540 problems: As the program is single-threaded, all we have
5541 to do is register a locking callback using an array for
5542 storing which locks are currently held by the program.
5545 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
5546 SSL_get_ex_data_X509_STORE_idx(), which is used in
5547 ssl_verify_cert_chain() and thus can be called at any time
5548 during TLS/SSL handshakes so that thread-safety is essential.
5549 Unfortunately, the ex_data design is not at all suited
5550 for multi-threaded use, so it probably should be abolished.
5553 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5554 [Broadcom, tweaked and integrated by Geoff Thorpe]
5556 *) Move common extension printing code to new function
5557 X509V3_print_extensions(). Reorganise OCSP print routines and
5558 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5561 *) New function X509_signature_print() to remove duplication in some
5565 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5566 set (this was treated exactly the same as SET OF previously). This
5567 is used to reorder the STACK representing the structure to match the
5568 encoding. This will be used to get round a problem where a PKCS7
5569 structure which was signed could not be verified because the STACK
5570 order did not reflect the encoded order.
5573 *) Reimplement the OCSP ASN1 module using the new code.
5576 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5577 for its ASN1 operations. The old style function pointers still exist
5578 for now but they will eventually go away.
5581 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
5582 completely replaces the old ASN1 functionality with a table driven
5583 encoder and decoder which interprets an ASN1_ITEM structure describing
5584 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
5585 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
5586 has also been converted to the new form.
5589 *) Change BN_mod_exp_recp so that negative moduli are tolerated
5590 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
5591 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
5592 for negative moduli.
5595 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
5596 of not touching the result's sign bit.
5599 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
5603 *) Changed the LHASH code to use prototypes for callbacks, and created
5604 macros to declare and implement thin (optionally static) functions
5605 that provide type-safety and avoid function pointer casting for the
5606 type-specific callbacks.
5609 *) Added Kerberos Cipher Suites to be used with TLS, as written in
5611 [Veers Staats <staatsvr@asc.hpc.mil>,
5612 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
5614 *) Reformat the FAQ so the different questions and answers can be divided
5615 in sections depending on the subject.
5618 *) Have the zlib compression code load ZLIB.DLL dynamically under
5622 *) New function BN_mod_sqrt for computing square roots modulo a prime
5623 (using the probabilistic Tonelli-Shanks algorithm unless
5624 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
5625 be handled deterministically).
5626 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5628 *) Make BN_mod_inverse faster by explicitly handling small quotients
5629 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
5630 512 bits], about 30% for larger ones [1024 or 2048 bits].)
5633 *) New function BN_kronecker.
5636 *) Fix BN_gcd so that it works on negative inputs; the result is
5637 positive unless both parameters are zero.
5638 Previously something reasonably close to an infinite loop was
5639 possible because numbers could be growing instead of shrinking
5640 in the implementation of Euclid's algorithm.
5643 *) Fix BN_is_word() and BN_is_one() macros to take into account the
5644 sign of the number in question.
5646 Fix BN_is_word(a,w) to work correctly for w == 0.
5648 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
5649 because its test if the absolute value of 'a' equals 'w'.
5650 Note that BN_abs_is_word does *not* handle w == 0 reliably;
5651 it exists mostly for use in the implementations of BN_is_zero(),
5652 BN_is_one(), and BN_is_word().
5655 *) New function BN_swap.
5658 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
5659 the exponentiation functions are more likely to produce reasonable
5660 results on negative inputs.
5663 *) Change BN_mod_mul so that the result is always non-negative.
5664 Previously, it could be negative if one of the factors was negative;
5665 I don't think anyone really wanted that behaviour.
5668 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
5669 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
5670 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
5671 and add new functions:
5680 BN_mod_lshift1_quick
5684 These functions always generate non-negative results.
5686 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
5687 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
5689 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
5690 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
5691 be reduced modulo m.
5692 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5695 The following entry accidentily appeared in the CHANGES file
5696 distributed with OpenSSL 0.9.7. The modifications described in
5697 it do *not* apply to OpenSSL 0.9.7.
5699 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5700 was actually never needed) and in BN_mul(). The removal in BN_mul()
5701 required a small change in bn_mul_part_recursive() and the addition
5702 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5703 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5704 bn_sub_words() and bn_add_words() except they take arrays with
5709 *) In 'openssl passwd', verify passwords read from the terminal
5710 unless the '-salt' option is used (which usually means that
5711 verification would just waste user's time since the resulting
5712 hash is going to be compared with some given password hash)
5713 or the new '-noverify' option is used.
5715 This is an incompatible change, but it does not affect
5716 non-interactive use of 'openssl passwd' (passwords on the command
5717 line, '-stdin' option, '-in ...' option) and thus should not
5721 *) Remove all references to RSAref, since there's no more need for it.
5724 *) Make DSO load along a path given through an environment variable
5725 (SHLIB_PATH) with shl_load().
5728 *) Constify the ENGINE code as a result of BIGNUM constification.
5729 Also constify the RSA code and most things related to it. In a
5730 few places, most notable in the depth of the ASN.1 code, ugly
5731 casts back to non-const were required (to be solved at a later
5735 *) Make it so the openssl application has all engines loaded by default.
5738 *) Constify the BIGNUM routines a little more.
5741 *) Add the following functions:
5743 ENGINE_load_cswift()
5745 ENGINE_load_atalla()
5747 ENGINE_load_builtin_engines()
5749 That way, an application can itself choose if external engines that
5750 are built-in in OpenSSL shall ever be used or not. The benefit is
5751 that applications won't have to be linked with libdl or other dso
5752 libraries unless it's really needed.
5754 Changed 'openssl engine' to load all engines on demand.
5755 Changed the engine header files to avoid the duplication of some
5756 declarations (they differed!).
5759 *) 'openssl engine' can now list capabilities.
5762 *) Better error reporting in 'openssl engine'.
5765 *) Never call load_dh_param(NULL) in s_server.
5768 *) Add engine application. It can currently list engines by name and
5769 identity, and test if they are actually available.
5772 *) Improve RPM specification file by forcing symbolic linking and making
5773 sure the installed documentation is also owned by root.root.
5774 [Damien Miller <djm@mindrot.org>]
5776 *) Give the OpenSSL applications more possibilities to make use of
5777 keys (public as well as private) handled by engines.
5780 *) Add OCSP code that comes from CertCo.
5783 *) Add VMS support for the Rijndael code.
5786 *) Added untested support for Nuron crypto accelerator.
5789 *) Add support for external cryptographic devices. This code was
5790 previously distributed separately as the "engine" branch.
5791 [Geoff Thorpe, Richard Levitte]
5793 *) Rework the filename-translation in the DSO code. It is now possible to
5794 have far greater control over how a "name" is turned into a filename
5795 depending on the operating environment and any oddities about the
5796 different shared library filenames on each system.
5799 *) Support threads on FreeBSD-elf in Configure.
5802 *) Fix for SHA1 assembly problem with MASM: it produces
5803 warnings about corrupt line number information when assembling
5804 with debugging information. This is caused by the overlapping
5806 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
5809 NCONF_get_number() has no error checking at all. As a replacement,
5810 NCONF_get_number_e() is defined (_e for "error checking") and is
5811 promoted strongly. The old NCONF_get_number is kept around for
5812 binary backward compatibility.
5813 Make it possible for methods to load from something other than a BIO,
5814 by providing a function pointer that is given a name instead of a BIO.
5815 For example, this could be used to load configuration data from an
5819 *) Fix for non blocking accept BIOs. Added new I/O special reason
5820 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
5821 with non blocking I/O was not possible because no retry code was
5822 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
5826 *) Added the beginnings of Rijndael support.
5829 *) Fix for bug in DirectoryString mask setting. Add support for
5830 X509_NAME_print_ex() in 'req' and X509_print_ex() function
5831 to allow certificate printing to more controllable, additional
5832 'certopt' option to 'x509' to allow new printing options to be
5836 *) Clean old EAY MD5 hack from e_os.h.
5839 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
5841 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5842 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5843 [Joe Orton, Steve Henson]
5845 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
5847 *) Fix additional bug revealed by the NISCC test suite:
5849 Stop bug triggering large recursion when presented with
5850 certain ASN.1 tags (CVE-2003-0851)
5853 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
5855 *) Fix various bugs revealed by running the NISCC test suite:
5857 Stop out of bounds reads in the ASN1 code when presented with
5858 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5860 If verify callback ignores invalid public key errors don't try to check
5861 certificate signature with the NULL public key.
5865 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5866 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5870 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5871 extra data after the compression methods not only for TLS 1.0
5872 but also for SSL 3.0 (as required by the specification).
5873 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5875 *) Change X509_certificate_type() to mark the key as exported/exportable
5876 when it's 512 *bits* long, not 512 bytes.
5879 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
5881 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5882 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5883 a protocol version number mismatch like a decryption error
5884 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5887 *) Turn on RSA blinding by default in the default implementation
5888 to avoid a timing attack. Applications that don't want it can call
5889 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5890 They would be ill-advised to do so in most cases.
5891 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5893 *) Change RSA blinding code so that it works when the PRNG is not
5894 seeded (in this case, the secret RSA exponent is abused as
5895 an unpredictable seed -- if it is not unpredictable, there
5896 is no point in blinding anyway). Make RSA blinding thread-safe
5897 by remembering the creator's thread ID in rsa->blinding and
5898 having all other threads use local one-time blinding factors
5899 (this requires more computation than sharing rsa->blinding, but
5900 avoids excessive locking; and if an RSA object is not shared
5901 between threads, blinding will still be very fast).
5904 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
5906 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5907 via timing by performing a MAC computation even if incorrrect
5908 block cipher padding has been found. This is a countermeasure
5909 against active attacks where the attacker has to distinguish
5910 between bad padding and a MAC verification error. (CVE-2003-0078)
5912 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5913 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5914 Martin Vuagnoux (EPFL, Ilion)]
5916 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
5918 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
5919 memory from it's contents. This is done with a counter that will
5920 place alternating values in each byte. This can be used to solve
5921 two issues: 1) the removal of calls to memset() by highly optimizing
5922 compilers, and 2) cleansing with other values than 0, since those can
5923 be read through on certain media, for example a swap space on disk.
5926 *) Bugfix: client side session caching did not work with external caching,
5927 because the session->cipher setting was not restored when reloading
5928 from the external cache. This problem was masked, when
5929 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
5930 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
5933 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
5934 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
5935 [Zeev Lieber <zeev-l@yahoo.com>]
5937 *) Undo an undocumented change introduced in 0.9.6e which caused
5938 repeated calls to OpenSSL_add_all_ciphers() and
5939 OpenSSL_add_all_digests() to be ignored, even after calling
5943 *) Change the default configuration reader to deal with last line not
5944 being properly terminated.
5947 *) Change X509_NAME_cmp() so it applies the special rules on handling
5948 DN values that are of type PrintableString, as well as RDNs of type
5949 emailAddress where the value has the type ia5String.
5950 [stefank@valicert.com via Richard Levitte]
5952 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
5953 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
5954 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
5955 the bitwise-OR of the two for use by the majority of applications
5956 wanting this behaviour, and update the docs. The documented
5957 behaviour and actual behaviour were inconsistent and had been
5958 changing anyway, so this is more a bug-fix than a behavioural
5960 [Geoff Thorpe, diagnosed by Nadav Har'El]
5962 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
5963 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
5966 *) Fix initialization code race conditions in
5967 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
5968 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
5969 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
5970 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
5971 ssl2_get_cipher_by_char(),
5972 ssl3_get_cipher_by_char().
5973 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
5975 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
5976 the cached sessions are flushed, as the remove_cb() might use ex_data
5977 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
5978 (see [openssl.org #212]).
5979 [Geoff Thorpe, Lutz Jaenicke]
5981 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
5982 length, instead of the encoding length to d2i_ASN1_OBJECT.
5985 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
5987 *) [In 0.9.6g-engine release:]
5988 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
5989 [Lynn Gazis <lgazis@rainbow.com>]
5991 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
5993 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
5994 and get fix the header length calculation.
5995 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
5996 Alon Kantor <alonk@checkpoint.com> (and others),
5999 *) Use proper error handling instead of 'assertions' in buffer
6000 overflow checks added in 0.9.6e. This prevents DoS (the
6001 assertions could call abort()).
6002 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
6004 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
6006 *) Add various sanity checks to asn1_get_length() to reject
6007 the ASN1 length bytes if they exceed sizeof(long), will appear
6008 negative or the content length exceeds the length of the
6010 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6012 *) Fix cipher selection routines: ciphers without encryption had no flags
6013 for the cipher strength set and where therefore not handled correctly
6014 by the selection routines (PR #130).
6017 *) Fix EVP_dsa_sha macro.
6021 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
6022 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
6023 that was added in OpenSSL 0.9.6d.
6025 As the countermeasure turned out to be incompatible with some
6026 broken SSL implementations, the new option is part of SSL_OP_ALL.
6027 SSL_OP_ALL is usually employed when compatibility with weird SSL
6028 implementations is desired (e.g. '-bugs' option to 's_client' and
6029 's_server'), so the new option is automatically set in many
6033 *) Changes in security patch:
6035 Changes marked "(CHATS)" were sponsored by the Defense Advanced
6036 Research Projects Agency (DARPA) and Air Force Research Laboratory,
6037 Air Force Materiel Command, USAF, under agreement number
6040 *) Add various sanity checks to asn1_get_length() to reject
6041 the ASN1 length bytes if they exceed sizeof(long), will appear
6042 negative or the content length exceeds the length of the
6043 supplied buffer. (CVE-2002-0659)
6044 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6046 *) Assertions for various potential buffer overflows, not known to
6048 [Ben Laurie (CHATS)]
6050 *) Various temporary buffers to hold ASCII versions of integers were
6051 too small for 64 bit platforms. (CVE-2002-0655)
6052 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
6054 *) Remote buffer overflow in SSL3 protocol - an attacker could
6055 supply an oversized session ID to a client. (CVE-2002-0656)
6056 [Ben Laurie (CHATS)]
6058 *) Remote buffer overflow in SSL2 protocol - an attacker could
6059 supply an oversized client master key. (CVE-2002-0656)
6060 [Ben Laurie (CHATS)]
6062 Changes between 0.9.6c and 0.9.6d [9 May 2002]
6064 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
6065 encoded as NULL) with id-dsa-with-sha1.
6066 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
6068 *) Check various X509_...() return values in apps/req.c.
6069 [Nils Larsch <nla@trustcenter.de>]
6071 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
6072 an end-of-file condition would erronously be flagged, when the CRLF
6073 was just at the end of a processed block. The bug was discovered when
6074 processing data through a buffering memory BIO handing the data to a
6075 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
6076 <ptsekov@syntrex.com> and Nedelcho Stanev.
6079 *) Implement a countermeasure against a vulnerability recently found
6080 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
6081 before application data chunks to avoid the use of known IVs
6082 with data potentially chosen by the attacker.
6085 *) Fix length checks in ssl3_get_client_hello().
6088 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
6089 to prevent ssl3_read_internal() from incorrectly assuming that
6090 ssl3_read_bytes() found application data while handshake
6091 processing was enabled when in fact s->s3->in_read_app_data was
6092 merely automatically cleared during the initial handshake.
6093 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
6095 *) Fix object definitions for Private and Enterprise: they were not
6096 recognized in their shortname (=lowercase) representation. Extend
6097 obj_dat.pl to issue an error when using undefined keywords instead
6098 of silently ignoring the problem (Svenning Sorensen
6099 <sss@sss.dnsalias.net>).
6102 *) Fix DH_generate_parameters() so that it works for 'non-standard'
6103 generators, i.e. generators other than 2 and 5. (Previously, the
6104 code did not properly initialise the 'add' and 'rem' values to
6105 BN_generate_prime().)
6107 In the new general case, we do not insist that 'generator' is
6108 actually a primitive root: This requirement is rather pointless;
6109 a generator of the order-q subgroup is just as good, if not
6113 *) Map new X509 verification errors to alerts. Discovered and submitted by
6114 Tom Wu <tom@arcot.com>.
6117 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
6118 returning non-zero before the data has been completely received
6119 when using non-blocking I/O.
6120 [Bodo Moeller; problem pointed out by John Hughes]
6122 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
6123 [Ben Laurie, Lutz Jaenicke]
6125 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
6126 Yoram Zahavi <YoramZ@gilian.com>).
6129 *) Add information about CygWin 1.3 and on, and preserve proper
6130 configuration for the versions before that.
6131 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6133 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
6134 check whether we deal with a copy of a session and do not delete from
6135 the cache in this case. Problem reported by "Izhar Shoshani Levi"
6136 <izhar@checkpoint.com>.
6139 *) Do not store session data into the internal session cache, if it
6140 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
6141 flag is set). Proposed by Aslam <aslam@funk.com>.
6144 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
6148 *) [In 0.9.6d-engine release:]
6149 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6150 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6152 *) Add the configuration target linux-s390x.
6153 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
6155 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
6156 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
6157 variable as an indication that a ClientHello message has been
6158 received. As the flag value will be lost between multiple
6159 invocations of ssl3_accept when using non-blocking I/O, the
6160 function may not be aware that a handshake has actually taken
6161 place, thus preventing a new session from being added to the
6164 To avoid this problem, we now set s->new_session to 2 instead of
6165 using a local variable.
6166 [Lutz Jaenicke, Bodo Moeller]
6168 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
6169 if the SSL_R_LENGTH_MISMATCH error is detected.
6170 [Geoff Thorpe, Bodo Moeller]
6172 *) New 'shared_ldflag' column in Configure platform table.
6175 *) Fix EVP_CIPHER_mode macro.
6176 ["Dan S. Camper" <dan@bti.net>]
6178 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
6179 type, we must throw them away by setting rr->length to 0.
6180 [D P Chang <dpc@qualys.com>]
6182 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
6184 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
6185 <Dominikus.Scherkl@biodata.com>. (The previous implementation
6186 worked incorrectly for those cases where range = 10..._2 and
6187 3*range is two bits longer than range.)
6190 *) Only add signing time to PKCS7 structures if it is not already
6194 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
6195 OBJ_ld_ce should be OBJ_id_ce.
6196 Also some ip-pda OIDs in crypto/objects/objects.txt were
6197 incorrect (cf. RFC 3039).
6198 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
6200 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
6201 returns early because it has nothing to do.
6202 [Andy Schneider <andy.schneider@bjss.co.uk>]
6204 *) [In 0.9.6c-engine release:]
6205 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
6206 [Andy Schneider <andy.schneider@bjss.co.uk>]
6208 *) [In 0.9.6c-engine release:]
6209 Add support for Cryptographic Appliance's keyserver technology.
6210 (Use engine 'keyclient')
6211 [Cryptographic Appliances and Geoff Thorpe]
6213 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
6214 is called via tools/c89.sh because arguments have to be
6215 rearranged (all '-L' options must appear before the first object
6217 [Richard Shapiro <rshapiro@abinitio.com>]
6219 *) [In 0.9.6c-engine release:]
6220 Add support for Broadcom crypto accelerator cards, backported
6222 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
6224 *) [In 0.9.6c-engine release:]
6225 Add support for SureWare crypto accelerator cards from
6226 Baltimore Technologies. (Use engine 'sureware')
6227 [Baltimore Technologies and Mark Cox]
6229 *) [In 0.9.6c-engine release:]
6230 Add support for crypto accelerator cards from Accelerated
6231 Encryption Processing, www.aep.ie. (Use engine 'aep')
6232 [AEP Inc. and Mark Cox]
6234 *) Add a configuration entry for gcc on UnixWare.
6235 [Gary Benson <gbenson@redhat.com>]
6237 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
6238 messages are stored in a single piece (fixed-length part and
6239 variable-length part combined) and fix various bugs found on the way.
6242 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
6243 instead. BIO_gethostbyname() does not know what timeouts are
6244 appropriate, so entries would stay in cache even when they have
6246 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
6248 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
6249 faced with a pathologically small ClientHello fragment that does
6250 not contain client_version: Instead of aborting with an error,
6251 simply choose the highest available protocol version (i.e.,
6252 TLS 1.0 unless it is disabled). In practice, ClientHello
6253 messages are never sent like this, but this change gives us
6254 strictly correct behaviour at least for TLS.
6257 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
6258 never resets s->method to s->ctx->method when called from within
6259 one of the SSL handshake functions.
6260 [Bodo Moeller; problem pointed out by Niko Baric]
6262 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
6263 (sent using the client's version number) if client_version is
6264 smaller than the protocol version in use. Also change
6265 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
6266 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
6267 the client will at least see that alert.
6270 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
6274 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
6275 client receives HelloRequest while in a handshake.
6276 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
6278 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
6279 should end in 'break', not 'goto end' which circuments various
6280 cleanups done in state SSL_ST_OK. But session related stuff
6281 must be disabled for SSL_ST_OK in the case that we just sent a
6284 Also avoid some overhead by not calling ssl_init_wbio_buffer()
6285 before just sending a HelloRequest.
6286 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
6288 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
6289 reveal whether illegal block cipher padding was found or a MAC
6290 verification error occured. (Neither SSLerr() codes nor alerts
6291 are directly visible to potential attackers, but the information
6292 may leak via logfiles.)
6294 Similar changes are not required for the SSL 2.0 implementation
6295 because the number of padding bytes is sent in clear for SSL 2.0,
6296 and the extra bytes are just ignored. However ssl/s2_pkt.c
6297 failed to verify that the purported number of padding bytes is in
6301 *) Add OpenUNIX-8 support including shared libraries
6302 (Boyd Lynn Gerber <gerberb@zenez.com>).
6305 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
6306 'wristwatch attack' using huge encoding parameters (cf.
6307 James H. Manger's CRYPTO 2001 paper). Note that the
6308 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
6309 encoding parameters and hence was not vulnerable.
6312 *) BN_sqr() bug fix.
6313 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
6315 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
6316 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
6317 followed by modular reduction.
6318 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
6320 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
6321 equivalent based on BN_pseudo_rand() instead of BN_rand().
6324 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
6325 This function was broken, as the check for a new client hello message
6326 to handle SGC did not allow these large messages.
6327 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
6330 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
6333 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
6334 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
6337 *) Rework the configuration and shared library support for Tru64 Unix.
6338 The configuration part makes use of modern compiler features and
6339 still retains old compiler behavior for those that run older versions
6340 of the OS. The shared library support part includes a variant that
6341 uses the RPATH feature, and is available through the special
6342 configuration target "alpha-cc-rpath", which will never be selected
6344 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
6346 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
6347 with the same message size as in ssl3_get_certificate_request().
6348 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
6349 messages might inadvertently be reject as too long.
6350 [Petr Lampa <lampa@fee.vutbr.cz>]
6352 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
6355 *) Modified SSL library such that the verify_callback that has been set
6356 specificly for an SSL object with SSL_set_verify() is actually being
6357 used. Before the change, a verify_callback set with this function was
6358 ignored and the verify_callback() set in the SSL_CTX at the time of
6359 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
6360 to allow the necessary settings.
6363 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
6364 explicitly to NULL, as at least on Solaris 8 this seems not always to be
6365 done automatically (in contradiction to the requirements of the C
6366 standard). This made problems when used from OpenSSH.
6369 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
6370 dh->length and always used
6372 BN_rand_range(priv_key, dh->p).
6374 BN_rand_range() is not necessary for Diffie-Hellman, and this
6375 specific range makes Diffie-Hellman unnecessarily inefficient if
6376 dh->length (recommended exponent length) is much smaller than the
6377 length of dh->p. We could use BN_rand_range() if the order of
6378 the subgroup was stored in the DH structure, but we only have
6383 BN_rand(priv_key, l, ...)
6385 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
6391 RSA_eay_public_encrypt
6392 RSA_eay_private_decrypt
6393 RSA_eay_private_encrypt (signing)
6394 RSA_eay_public_decrypt (signature verification)
6396 (default implementations for RSA_public_encrypt,
6397 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
6398 always reject numbers >= n.
6401 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
6402 to synchronize access to 'locking_thread'. This is necessary on
6403 systems where access to 'locking_thread' (an 'unsigned long'
6404 variable) is not atomic.
6407 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
6408 *before* setting the 'crypto_lock_rand' flag. The previous code had
6409 a race condition if 0 is a valid thread ID.
6410 [Travis Vitek <vitek@roguewave.com>]
6412 *) Add support for shared libraries under Irix.
6413 [Albert Chin-A-Young <china@thewrittenword.com>]
6415 *) Add configuration option to build on Linux on both big-endian and
6417 [Ralf Baechle <ralf@uni-koblenz.de>]
6419 *) Add the possibility to create shared libraries on HP-UX.
6422 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
6424 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
6425 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
6426 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
6427 PRNG state recovery was possible based on the output of
6428 one PRNG request appropriately sized to gain knowledge on
6429 'md' followed by enough consecutive 1-byte PRNG requests
6430 to traverse all of 'state'.
6432 1. When updating 'md_local' (the current thread's copy of 'md')
6433 during PRNG output generation, hash all of the previous
6434 'md_local' value, not just the half used for PRNG output.
6436 2. Make the number of bytes from 'state' included into the hash
6437 independent from the number of PRNG bytes requested.
6439 The first measure alone would be sufficient to avoid
6440 Markku-Juhani's attack. (Actually it had never occurred
6441 to me that the half of 'md_local' used for chaining was the
6442 half from which PRNG output bytes were taken -- I had always
6443 assumed that the secret half would be used.) The second
6444 measure makes sure that additional data from 'state' is never
6445 mixed into 'md_local' in small portions; this heuristically
6446 further strengthens the PRNG.
6449 *) Fix crypto/bn/asm/mips3.s.
6452 *) When only the key is given to "enc", the IV is undefined. Print out
6453 an error message in this case.
6456 *) Handle special case when X509_NAME is empty in X509 printing routines.
6459 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
6460 positive and less than q.
6463 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
6464 used: it isn't thread safe and the add_lock_callback should handle
6466 [Paul Rose <Paul.Rose@bridge.com>]
6468 *) Verify that incoming data obeys the block size in
6469 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
6473 [Ulf Möller, Bodo Möller]
6475 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
6476 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
6477 when fixing the server behaviour for backwards-compatible 'client
6478 hello' messages. (Note that the attack is impractical against
6479 SSL 3.0 and TLS 1.0 anyway because length and version checking
6480 means that the probability of guessing a valid ciphertext is
6481 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
6484 Before 0.9.5, the countermeasure (hide the error by generating a
6485 random 'decryption result') did not work properly because
6486 ERR_clear_error() was missing, meaning that SSL_get_error() would
6487 detect the supposedly ignored error.
6489 Both problems are now fixed.
6492 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
6493 (previously it was 1024).
6496 *) Fix for compatibility mode trust settings: ignore trust settings
6497 unless some valid trust or reject settings are present.
6500 *) Fix for blowfish EVP: its a variable length cipher.
6503 *) Fix various bugs related to DSA S/MIME verification. Handle missing
6504 parameters in DSA public key structures and return an error in the
6505 DSA routines if parameters are absent.
6508 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
6509 in the current directory if neither $RANDFILE nor $HOME was set.
6510 RAND_file_name() in 0.9.6a returned NULL in this case. This has
6511 caused some confusion to Windows users who haven't defined $HOME.
6512 Thus RAND_file_name() is changed again: e_os.h can define a
6513 DEFAULT_HOME, which will be used if $HOME is not set.
6514 For Windows, we use "C:"; on other platforms, we still require
6515 environment variables.
6517 *) Move 'if (!initialized) RAND_poll()' into regions protected by
6518 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
6519 having multiple threads call RAND_poll() concurrently.
6522 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
6523 combination of a flag and a thread ID variable.
6524 Otherwise while one thread is in ssleay_rand_bytes (which sets the
6525 flag), *other* threads can enter ssleay_add_bytes without obeying
6526 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
6527 that they do not hold after the first thread unsets add_do_not_lock).
6530 *) Change bctest again: '-x' expressions are not available in all
6534 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
6536 *) Fix a couple of memory leaks in PKCS7_dataDecode()
6537 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
6539 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
6540 the default extension for executables, if any. Also, make the perl
6541 scripts that use symlink() to test if it really exists and use "cp"
6542 if it doesn't. All this made OpenSSL compilable and installable in
6546 *) Fix for asn1_GetSequence() for indefinite length constructed data.
6547 If SEQUENCE is length is indefinite just set c->slen to the total
6548 amount of data available.
6549 [Steve Henson, reported by shige@FreeBSD.org]
6550 [This change does not apply to 0.9.7.]
6552 *) Change bctest to avoid here-documents inside command substitution
6553 (workaround for FreeBSD /bin/sh bug).
6554 For compatibility with Ultrix, avoid shell functions (introduced
6555 in the bctest version that searches along $PATH).
6558 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6559 with des_encrypt() defined on some operating systems, like Solaris
6563 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6564 On the Importance of Eliminating Errors in Cryptographic
6565 Computations, J. Cryptology 14 (2001) 2, 101-119,
6566 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6569 *) MIPS assembler BIGNUM division bug fix.
6572 *) Disabled incorrect Alpha assembler code.
6575 *) Fix PKCS#7 decode routines so they correctly update the length
6576 after reading an EOC for the EXPLICIT tag.
6578 [This change does not apply to 0.9.7.]
6580 *) Fix bug in PKCS#12 key generation routines. This was triggered
6581 if a 3DES key was generated with a 0 initial byte. Include
6582 PKCS12_BROKEN_KEYGEN compilation option to retain the old
6583 (but broken) behaviour.
6586 *) Enhance bctest to search for a working bc along $PATH and print
6588 [Tim Rice <tim@multitalents.net> via Richard Levitte]
6590 *) Fix memory leaks in err.c: free err_data string if necessary;
6591 don't write to the wrong index in ERR_set_error_data.
6594 *) Implement ssl23_peek (analogous to ssl23_read), which previously
6598 *) Replace rdtsc with _emit statements for VC++ version 5.
6599 [Jeremy Cooper <jeremy@baymoo.org>]
6601 *) Make it possible to reuse SSLv2 sessions.
6604 *) In copy_email() check for >= 0 as a return value for
6605 X509_NAME_get_index_by_NID() since 0 is a valid index.
6606 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
6608 *) Avoid coredump with unsupported or invalid public keys by checking if
6609 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
6610 PKCS7_verify() fails with non detached data.
6613 *) Don't use getenv in library functions when run as setuid/setgid.
6614 New function OPENSSL_issetugid().
6617 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
6618 due to incorrect handling of multi-threading:
6620 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
6622 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
6624 3. Count how many times MemCheck_off() has been called so that
6625 nested use can be treated correctly. This also avoids
6626 inband-signalling in the previous code (which relied on the
6627 assumption that thread ID 0 is impossible).
6630 *) Add "-rand" option also to s_client and s_server.
6633 *) Fix CPU detection on Irix 6.x.
6634 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
6635 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6637 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
6640 [This change does not apply to 0.9.7.]
6642 *) Use the cached encoding of an X509_NAME structure rather than
6643 copying it. This is apparently the reason for the libsafe "errors"
6644 but the code is actually correct.
6647 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
6648 Bleichenbacher's DSA attack.
6649 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
6650 to be set and top=0 forces the highest bit to be set; top=-1 is new
6651 and leaves the highest bit random.
6652 [Ulf Moeller, Bodo Moeller]
6654 *) In the NCONF_...-based implementations for CONF_... queries
6655 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
6656 a temporary CONF structure with the data component set to NULL
6657 (which gives segmentation faults in lh_retrieve).
6658 Instead, use NULL for the CONF pointer in CONF_get_string and
6659 CONF_get_number (which may use environment variables) and directly
6660 return NULL from CONF_get_section.
6663 *) Fix potential buffer overrun for EBCDIC.
6666 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
6667 keyUsage if basicConstraints absent for a CA.
6670 *) Make SMIME_write_PKCS7() write mail header values with a format that
6671 is more generally accepted (no spaces before the semicolon), since
6672 some programs can't parse those values properly otherwise. Also make
6673 sure BIO's that break lines after each write do not create invalid
6677 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
6678 macros previously used would not encode an empty SEQUENCE OF
6679 and break the signature.
6681 [This change does not apply to 0.9.7.]
6683 *) Zero the premaster secret after deriving the master secret in
6687 *) Add some EVP_add_digest_alias registrations (as found in
6688 OpenSSL_add_all_digests()) to SSL_library_init()
6689 aka OpenSSL_add_ssl_algorithms(). This provides improved
6690 compatibility with peers using X.509 certificates
6691 with unconventional AlgorithmIdentifier OIDs.
6694 *) Fix for Irix with NO_ASM.
6695 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6697 *) ./config script fixes.
6698 [Ulf Moeller, Richard Levitte]
6700 *) Fix 'openssl passwd -1'.
6703 *) Change PKCS12_key_gen_asc() so it can cope with non null
6704 terminated strings whose length is passed in the passlen
6705 parameter, for example from PEM callbacks. This was done
6706 by adding an extra length parameter to asc2uni().
6707 [Steve Henson, reported by <oddissey@samsung.co.kr>]
6709 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
6710 call failed, free the DSA structure.
6713 *) Fix to uni2asc() to cope with zero length Unicode strings.
6714 These are present in some PKCS#12 files.
6717 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
6718 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
6719 when writing a 32767 byte record.
6720 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
6722 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
6723 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
6725 (RSA objects have a reference count access to which is protected
6726 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
6727 so they are meant to be shared between threads.)
6728 [Bodo Moeller, Geoff Thorpe; original patch submitted by
6729 "Reddie, Steven" <Steven.Reddie@ca.com>]
6731 *) Fix a deadlock in CRYPTO_mem_leaks().
6734 *) Use better test patterns in bntest.
6737 *) rand_win.c fix for Borland C.
6740 *) BN_rshift bugfix for n == 0.
6743 *) Add a 'bctest' script that checks for some known 'bc' bugs
6744 so that 'make test' does not abort just because 'bc' is broken.
6747 *) Store verify_result within SSL_SESSION also for client side to
6748 avoid potential security hole. (Re-used sessions on the client side
6749 always resulted in verify_result==X509_V_OK, not using the original
6750 result of the server certificate verification.)
6753 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
6754 SSL3_RT_APPLICATION_DATA, return 0.
6755 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
6759 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
6760 releases, have been re-implemented by renaming the previous
6761 implementations of ssl2_read and ssl3_read to ssl2_read_internal
6762 and ssl3_read_internal, respectively, and adding 'peek' parameters
6763 to them. The new ssl[23]_{read,peek} functions are calls to
6764 ssl[23]_read_internal with the 'peek' flag set appropriately.
6765 A 'peek' parameter has also been added to ssl3_read_bytes, which
6766 does the actual work for ssl3_read_internal.
6769 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
6770 the method-specific "init()" handler. Also clean up ex_data after
6771 calling the method-specific "finish()" handler. Previously, this was
6772 happening the other way round.
6775 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
6776 The previous value, 12, was not always sufficient for BN_mod_exp().
6779 *) Make sure that shared libraries get the internal name engine with
6780 the full version number and not just 0. This should mark the
6781 shared libraries as not backward compatible. Of course, this should
6782 be changed again when we can guarantee backward binary compatibility.
6785 *) Fix typo in get_cert_by_subject() in by_dir.c
6786 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
6788 *) Rework the system to generate shared libraries:
6790 - Make note of the expected extension for the shared libraries and
6791 if there is a need for symbolic links from for example libcrypto.so.0
6792 to libcrypto.so.0.9.7. There is extended info in Configure for
6795 - Make as few rebuilds of the shared libraries as possible.
6797 - Still avoid linking the OpenSSL programs with the shared libraries.
6799 - When installing, install the shared libraries separately from the
6803 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
6805 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
6806 and not in SSL_clear because the latter is also used by the
6807 accept/connect functions; previously, the settings made by
6808 SSL_set_read_ahead would be lost during the handshake.
6809 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
6811 *) Correct util/mkdef.pl to be selective about disabled algorithms.
6812 Previously, it would create entries for disableed algorithms no
6816 *) Added several new manual pages for SSL_* function.
6819 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
6821 *) In ssl23_get_client_hello, generate an error message when faced
6822 with an initial SSL 3.0/TLS record that is too small to contain the
6823 first two bytes of the ClientHello message, i.e. client_version.
6824 (Note that this is a pathologic case that probably has never happened
6825 in real life.) The previous approach was to use the version number
6826 from the record header as a substitute; but our protocol choice
6827 should not depend on that one because it is not authenticated
6828 by the Finished messages.
6831 *) More robust randomness gathering functions for Windows.
6832 [Jeffrey Altman <jaltman@columbia.edu>]
6834 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
6835 not set then we don't setup the error code for issuer check errors
6836 to avoid possibly overwriting other errors which the callback does
6837 handle. If an application does set the flag then we assume it knows
6838 what it is doing and can handle the new informational codes
6842 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
6843 a general "ANY" type, as such it should be able to decode anything
6844 including tagged types. However it didn't check the class so it would
6845 wrongly interpret tagged types in the same way as their universal
6846 counterpart and unknown types were just rejected. Changed so that the
6847 tagged and unknown types are handled in the same way as a SEQUENCE:
6848 that is the encoding is stored intact. There is also a new type
6849 "V_ASN1_OTHER" which is used when the class is not universal, in this
6850 case we have no idea what the actual type is so we just lump them all
6854 *) On VMS, stdout may very well lead to a file that is written to
6855 in a record-oriented fashion. That means that every write() will
6856 write a separate record, which will be read separately by the
6857 programs trying to read from it. This can be very confusing.
6859 The solution is to put a BIO filter in the way that will buffer
6860 text until a linefeed is reached, and then write everything a
6861 line at a time, so every record written will be an actual line,
6862 not chunks of lines and not (usually doesn't happen, but I've
6863 seen it once) several lines in one record. BIO_f_linebuffer() is
6866 Currently, it's a VMS-only method, because that's where it has
6867 been tested well enough.
6870 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
6871 it can return incorrect results.
6872 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
6873 but it was in 0.9.6-beta[12].)
6876 *) Disable the check for content being present when verifying detached
6877 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
6878 include zero length content when signing messages.
6881 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
6882 BIO_ctrl (for BIO pairs).
6885 *) Add DSO method for VMS.
6888 *) Bug fix: Montgomery multiplication could produce results with the
6892 *) Add RPM specification openssl.spec and modify it to build three
6893 packages. The default package contains applications, application
6894 documentation and run-time libraries. The devel package contains
6895 include files, static libraries and function documentation. The
6896 doc package contains the contents of the doc directory. The original
6897 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
6900 *) Add a large number of documentation files for many SSL routines.
6901 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6903 *) Add a configuration entry for Sony News 4.
6904 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
6906 *) Don't set the two most significant bits to one when generating a
6907 random number < q in the DSA library.
6910 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
6911 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
6912 the underlying transport is blocking) if a handshake took place.
6913 (The default behaviour is needed by applications such as s_client
6914 and s_server that use select() to determine when to use SSL_read;
6915 but for applications that know in advance when to expect data, it
6916 just makes things more complicated.)
6919 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
6923 *) Add a few more EBCDIC conditionals that make `req' and `x509'
6924 work better on such systems.
6925 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6927 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
6928 Update PKCS12_parse() so it copies the friendlyName and the
6929 keyid to the certificates aux info.
6932 *) Fix bug in PKCS7_verify() which caused an infinite loop
6933 if there was more than one signature.
6934 [Sven Uszpelkat <su@celocom.de>]
6936 *) Major change in util/mkdef.pl to include extra information
6937 about each symbol, as well as presentig variables as well
6938 as functions. This change means that there's n more need
6939 to rebuild the .num files when some algorithms are excluded.
6942 *) Allow the verify time to be set by an application,
6943 rather than always using the current time.
6946 *) Phase 2 verify code reorganisation. The certificate
6947 verify code now looks up an issuer certificate by a
6948 number of criteria: subject name, authority key id
6949 and key usage. It also verifies self signed certificates
6950 by the same criteria. The main comparison function is
6951 X509_check_issued() which performs these checks.
6953 Lot of changes were necessary in order to support this
6954 without completely rewriting the lookup code.
6956 Authority and subject key identifier are now cached.
6958 The LHASH 'certs' is X509_STORE has now been replaced
6959 by a STACK_OF(X509_OBJECT). This is mainly because an
6960 LHASH can't store or retrieve multiple objects with
6961 the same hash value.
6963 As a result various functions (which were all internal
6964 use only) have changed to handle the new X509_STORE
6965 structure. This will break anything that messed round
6966 with X509_STORE internally.
6968 The functions X509_STORE_add_cert() now checks for an
6969 exact match, rather than just subject name.
6971 The X509_STORE API doesn't directly support the retrieval
6972 of multiple certificates matching a given criteria, however
6973 this can be worked round by performing a lookup first
6974 (which will fill the cache with candidate certificates)
6975 and then examining the cache for matches. This is probably
6976 the best we can do without throwing out X509_LOOKUP
6977 entirely (maybe later...).
6979 The X509_VERIFY_CTX structure has been enhanced considerably.
6981 All certificate lookup operations now go via a get_issuer()
6982 callback. Although this currently uses an X509_STORE it
6983 can be replaced by custom lookups. This is a simple way
6984 to bypass the X509_STORE hackery necessary to make this
6985 work and makes it possible to use more efficient techniques
6986 in future. A very simple version which uses a simple
6987 STACK for its trusted certificate store is also provided
6988 using X509_STORE_CTX_trusted_stack().
6990 The verify_cb() and verify() callbacks now have equivalents
6991 in the X509_STORE_CTX structure.
6993 X509_STORE_CTX also has a 'flags' field which can be used
6994 to customise the verify behaviour.
6997 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
6998 excludes S/MIME capabilities.
7001 *) When a certificate request is read in keep a copy of the
7002 original encoding of the signed data and use it when outputing
7003 again. Signatures then use the original encoding rather than
7004 a decoded, encoded version which may cause problems if the
7005 request is improperly encoded.
7008 *) For consistency with other BIO_puts implementations, call
7009 buffer_write(b, ...) directly in buffer_puts instead of calling
7012 In BIO_puts, increment b->num_write as in BIO_write.
7013 [Peter.Sylvester@EdelWeb.fr]
7015 *) Fix BN_mul_word for the case where the word is 0. (We have to use
7016 BN_zero, we may not return a BIGNUM with an array consisting of
7020 *) Avoid calling abort() from within the library when problems are
7021 detected, except if preprocessor symbols have been defined
7022 (such as REF_CHECK, BN_DEBUG etc.).
7025 *) New openssl application 'rsautl'. This utility can be
7026 used for low level RSA operations. DER public key
7027 BIO/fp routines also added.
7030 *) New Configure entry and patches for compiling on QNX 4.
7031 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
7033 *) A demo state-machine implementation was sponsored by
7034 Nuron (http://www.nuron.com/) and is now available in
7035 demos/state_machine.
7038 *) New options added to the 'dgst' utility for signature
7039 generation and verification.
7042 *) Unrecognized PKCS#7 content types are now handled via a
7043 catch all ASN1_TYPE structure. This allows unsupported
7044 types to be stored as a "blob" and an application can
7045 encode and decode it manually.
7048 *) Fix various signed/unsigned issues to make a_strex.c
7050 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
7052 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
7053 length if passed a buffer. ASN1_INTEGER_to_BN failed
7054 if passed a NULL BN and its argument was negative.
7055 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
7057 *) Modification to PKCS#7 encoding routines to output definite
7058 length encoding. Since currently the whole structures are in
7059 memory there's not real point in using indefinite length
7060 constructed encoding. However if OpenSSL is compiled with
7061 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
7064 *) Added BIO_vprintf() and BIO_vsnprintf().
7067 *) Added more prefixes to parse for in the the strings written
7068 through a logging bio, to cover all the levels that are available
7069 through syslog. The prefixes are now:
7071 PANIC, EMERG, EMR => LOG_EMERG
7072 ALERT, ALR => LOG_ALERT
7073 CRIT, CRI => LOG_CRIT
7074 ERROR, ERR => LOG_ERR
7075 WARNING, WARN, WAR => LOG_WARNING
7076 NOTICE, NOTE, NOT => LOG_NOTICE
7077 INFO, INF => LOG_INFO
7078 DEBUG, DBG => LOG_DEBUG
7080 and as before, if none of those prefixes are present at the
7081 beginning of the string, LOG_ERR is chosen.
7083 On Win32, the LOG_* levels are mapped according to this:
7085 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
7086 LOG_WARNING => EVENTLOG_WARNING_TYPE
7087 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
7091 *) Made it possible to reconfigure with just the configuration
7092 argument "reconf" or "reconfigure". The command line arguments
7093 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
7094 and are retrieved from there when reconfiguring.
7098 [Assar Westerlund <assar@sics.se>, Richard Levitte]
7100 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
7103 *) The obj_dat.pl script was messing up the sorting of object
7104 names. The reason was that it compared the quoted version
7105 of strings as a result "OCSP" > "OCSP Signing" because
7106 " > SPACE. Changed script to store unquoted versions of
7107 names and add quotes on output. It was also omitting some
7108 names from the lookup table if they were given a default
7109 value (that is if SN is missing it is given the same
7110 value as LN and vice versa), these are now added on the
7111 grounds that if an object has a name we should be able to
7112 look it up. Finally added warning output when duplicate
7113 short or long names are found.
7116 *) Changes needed for Tandem NSK.
7117 [Scott Uroff <scott@xypro.com>]
7119 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
7120 RSA_padding_check_SSLv23(), special padding was never detected
7121 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
7122 version rollback attacks was not effective.
7124 In s23_clnt.c, don't use special rollback-attack detection padding
7125 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
7126 client; similarly, in s23_srvr.c, don't do the rollback check if
7127 SSL 2.0 is the only protocol enabled in the server.
7130 *) Make it possible to get hexdumps of unprintable data with 'openssl
7131 asn1parse'. By implication, the functions ASN1_parse_dump() and
7132 BIO_dump_indent() are added.
7135 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
7136 these print out strings and name structures based on various
7137 flags including RFC2253 support and proper handling of
7138 multibyte characters. Added options to the 'x509' utility
7139 to allow the various flags to be set.
7142 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
7143 Also change the functions X509_cmp_current_time() and
7144 X509_gmtime_adj() work with an ASN1_TIME structure,
7145 this will enable certificates using GeneralizedTime in validity
7146 dates to be checked.
7149 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
7150 negative public key encodings) on by default,
7151 NO_NEG_PUBKEY_BUG can be set to disable it.
7154 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
7155 content octets. An i2c_ASN1_OBJECT is unnecessary because
7156 the encoding can be trivially obtained from the structure.
7159 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
7160 not read locks (CRYPTO_r_[un]lock).
7163 *) A first attempt at creating official support for shared
7164 libraries through configuration. I've kept it so the
7165 default is static libraries only, and the OpenSSL programs
7166 are always statically linked for now, but there are
7167 preparations for dynamic linking in place.
7168 This has been tested on Linux and Tru64.
7171 *) Randomness polling function for Win9x, as described in:
7172 Peter Gutmann, Software Generation of Practically Strong
7176 *) Fix so PRNG is seeded in req if using an already existing
7180 *) New options to smime application. -inform and -outform
7181 allow alternative formats for the S/MIME message including
7182 PEM and DER. The -content option allows the content to be
7183 specified separately. This should allow things like Netscape
7184 form signing output easier to verify.
7187 *) Fix the ASN1 encoding of tags using the 'long form'.
7190 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
7191 STRING types. These convert content octets to and from the
7192 underlying type. The actual tag and length octets are
7193 already assumed to have been read in and checked. These
7194 are needed because all other string types have virtually
7195 identical handling apart from the tag. By having versions
7196 of the ASN1 functions that just operate on content octets
7197 IMPLICIT tagging can be handled properly. It also allows
7198 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
7199 and ASN1_INTEGER are identical apart from the tag.
7202 *) Change the handling of OID objects as follows:
7204 - New object identifiers are inserted in objects.txt, following
7205 the syntax given in objects.README.
7206 - objects.pl is used to process obj_mac.num and create a new
7208 - obj_dat.pl is used to create a new obj_dat.h, using the data in
7211 This is currently kind of a hack, and the perl code in objects.pl
7212 isn't very elegant, but it works as I intended. The simplest way
7213 to check that it worked correctly is to look in obj_dat.h and
7214 check the array nid_objs and make sure the objects haven't moved
7215 around (this is important!). Additions are OK, as well as
7216 consistent name changes.
7219 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
7222 *) Addition of the command line parameter '-rand file' to 'openssl req'.
7223 The given file adds to whatever has already been seeded into the
7224 random pool through the RANDFILE configuration file option or
7225 environment variable, or the default random state file.
7228 *) mkstack.pl now sorts each macro group into lexical order.
7229 Previously the output order depended on the order the files
7230 appeared in the directory, resulting in needless rewriting
7234 *) Patches to make OpenSSL compile under Win32 again. Mostly
7235 work arounds for the VC++ problem that it treats func() as
7236 func(void). Also stripped out the parts of mkdef.pl that
7237 added extra typesafe functions: these no longer exist.
7240 *) Reorganisation of the stack code. The macros are now all
7241 collected in safestack.h . Each macro is defined in terms of
7242 a "stack macro" of the form SKM_<name>(type, a, b). The
7243 DEBUG_SAFESTACK is now handled in terms of function casts,
7244 this has the advantage of retaining type safety without the
7245 use of additional functions. If DEBUG_SAFESTACK is not defined
7246 then the non typesafe macros are used instead. Also modified the
7247 mkstack.pl script to handle the new form. Needs testing to see
7248 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
7249 the default if no major problems. Similar behaviour for ASN1_SET_OF
7250 and PKCS12_STACK_OF.
7253 *) When some versions of IIS use the 'NET' form of private key the
7254 key derivation algorithm is different. Normally MD5(password) is
7255 used as a 128 bit RC4 key. In the modified case
7256 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
7257 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
7258 as the old Netscape_RSA functions except they have an additional
7259 'sgckey' parameter which uses the modified algorithm. Also added
7260 an -sgckey command line option to the rsa utility. Thanks to
7261 Adrian Peck <bertie@ncipher.com> for posting details of the modified
7262 algorithm to openssl-dev.
7265 *) The evp_local.h macros were using 'c.##kname' which resulted in
7266 invalid expansion on some systems (SCO 5.0.5 for example).
7267 Corrected to 'c.kname'.
7268 [Phillip Porch <root@theporch.com>]
7270 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
7271 a STACK of email addresses from a certificate or request, these look
7272 in the subject name and the subject alternative name extensions and
7273 omit any duplicate addresses.
7276 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
7277 This makes DSA verification about 2 % faster.
7280 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
7281 (meaning that now 2^5 values will be precomputed, which is only 4 KB
7282 plus overhead for 1024 bit moduli).
7283 This makes exponentiations about 0.5 % faster for 1024 bit
7284 exponents (as measured by "openssl speed rsa2048").
7287 *) Rename memory handling macros to avoid conflicts with other
7289 Malloc => OPENSSL_malloc
7290 Malloc_locked => OPENSSL_malloc_locked
7291 Realloc => OPENSSL_realloc
7292 Free => OPENSSL_free
7295 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
7296 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
7299 *) CygWin32 support.
7300 [John Jarvie <jjarvie@newsguy.com>]
7302 *) The type-safe stack code has been rejigged. It is now only compiled
7303 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
7304 by default all type-specific stack functions are "#define"d back to
7305 standard stack functions. This results in more streamlined output
7306 but retains the type-safety checking possibilities of the original
7310 *) The STACK code has been cleaned up, and certain type declarations
7311 that didn't make a lot of sense have been brought in line. This has
7312 also involved a cleanup of sorts in safestack.h to more correctly
7313 map type-safe stack functions onto their plain stack counterparts.
7314 This work has also resulted in a variety of "const"ifications of
7315 lots of the code, especially "_cmp" operations which should normally
7316 be prototyped with "const" parameters anyway.
7319 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
7320 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
7321 (The PRNG state consists of two parts, the large pool 'state' and 'md',
7322 where all of 'md' is used each time the PRNG is used, but 'state'
7323 is used only indexed by a cyclic counter. As entropy may not be
7324 well distributed from the beginning, 'md' is important as a
7325 chaining variable. However, the output function chains only half
7326 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
7327 all of 'md', and seeding with STATE_SIZE dummy bytes will result
7328 in all of 'state' being rewritten, with the new values depending
7329 on virtually all of 'md'. This overcomes the 80 bit limitation.)
7332 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
7333 the handshake is continued after ssl_verify_cert_chain();
7334 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
7335 can lead to 'unexplainable' connection aborts later.
7336 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
7338 *) Major EVP API cipher revision.
7339 Add hooks for extra EVP features. This allows various cipher
7340 parameters to be set in the EVP interface. Support added for variable
7341 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
7342 setting of RC2 and RC5 parameters.
7344 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
7347 Remove lots of duplicated code from the EVP library. For example *every*
7348 cipher init() function handles the 'iv' in the same way according to the
7349 cipher mode. They also all do nothing if the 'key' parameter is NULL and
7350 for CFB and OFB modes they zero ctx->num.
7352 New functionality allows removal of S/MIME code RC2 hack.
7354 Most of the routines have the same form and so can be declared in terms
7357 By shifting this to the top level EVP_CipherInit() it can be removed from
7358 all individual ciphers. If the cipher wants to handle IVs or keys
7359 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
7362 Change lots of functions like EVP_EncryptUpdate() to now return a
7363 value: although software versions of the algorithms cannot fail
7364 any installed hardware versions can.
7367 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
7368 this option is set, tolerate broken clients that send the negotiated
7369 protocol version number instead of the requested protocol version
7373 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
7374 i.e. non-zero for export ciphersuites, zero otherwise.
7375 Previous versions had this flag inverted, inconsistent with
7376 rsa_tmp_cb (..._TMP_RSA_CB).
7377 [Bodo Moeller; problem reported by Amit Chopra]
7379 *) Add missing DSA library text string. Work around for some IIS
7380 key files with invalid SEQUENCE encoding.
7383 *) Add a document (doc/standards.txt) that list all kinds of standards
7384 and so on that are implemented in OpenSSL.
7387 *) Enhance c_rehash script. Old version would mishandle certificates
7388 with the same subject name hash and wouldn't handle CRLs at all.
7389 Added -fingerprint option to crl utility, to support new c_rehash
7393 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
7396 *) Fix for SSL server purpose checking. Server checking was
7397 rejecting certificates which had extended key usage present
7398 but no ssl client purpose.
7399 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
7401 *) Make PKCS#12 code work with no password. The PKCS#12 spec
7402 is a little unclear about how a blank password is handled.
7403 Since the password in encoded as a BMPString with terminating
7404 double NULL a zero length password would end up as just the
7405 double NULL. However no password at all is different and is
7406 handled differently in the PKCS#12 key generation code. NS
7407 treats a blank password as zero length. MSIE treats it as no
7408 password on export: but it will try both on import. We now do
7409 the same: PKCS12_parse() tries zero length and no password if
7410 the password is set to "" or NULL (NULL is now a valid password:
7411 it wasn't before) as does the pkcs12 application.
7414 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
7415 perror when PEM_read_bio_X509_REQ fails, the error message must
7416 be obtained from the error queue.
7419 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
7420 it in ERR_remove_state if appropriate, and change ERR_get_state
7421 accordingly to avoid race conditions (this is necessary because
7422 thread_hash is no longer constant once set).
7425 *) Bugfix for linux-elf makefile.one.
7428 *) RSA_get_default_method() will now cause a default
7429 RSA_METHOD to be chosen if one doesn't exist already.
7430 Previously this was only set during a call to RSA_new()
7431 or RSA_new_method(NULL) meaning it was possible for
7432 RSA_get_default_method() to return NULL.
7435 *) Added native name translation to the existing DSO code
7436 that will convert (if the flag to do so is set) filenames
7437 that are sufficiently small and have no path information
7438 into a canonical native form. Eg. "blah" converted to
7439 "libblah.so" or "blah.dll" etc.
7442 *) New function ERR_error_string_n(e, buf, len) which is like
7443 ERR_error_string(e, buf), but writes at most 'len' bytes
7444 including the 0 terminator. For ERR_error_string_n, 'buf'
7446 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
7448 *) CONF library reworked to become more general. A new CONF
7449 configuration file reader "class" is implemented as well as a
7450 new functions (NCONF_*, for "New CONF") to handle it. The now
7451 old CONF_* functions are still there, but are reimplemented to
7452 work in terms of the new functions. Also, a set of functions
7453 to handle the internal storage of the configuration data is
7454 provided to make it easier to write new configuration file
7455 reader "classes" (I can definitely see something reading a
7456 configuration file in XML format, for example), called _CONF_*,
7457 or "the configuration storage API"...
7459 The new configuration file reading functions are:
7461 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
7462 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
7464 NCONF_default, NCONF_WIN32
7466 NCONF_dump_fp, NCONF_dump_bio
7468 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
7469 NCONF_new creates a new CONF object. This works in the same way
7470 as other interfaces in OpenSSL, like the BIO interface.
7471 NCONF_dump_* dump the internal storage of the configuration file,
7472 which is useful for debugging. All other functions take the same
7473 arguments as the old CONF_* functions wth the exception of the
7474 first that must be a `CONF *' instead of a `LHASH *'.
7476 To make it easer to use the new classes with the old CONF_* functions,
7477 the function CONF_set_default_method is provided.
7480 *) Add '-tls1' option to 'openssl ciphers', which was already
7481 mentioned in the documentation but had not been implemented.
7482 (This option is not yet really useful because even the additional
7483 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
7486 *) Initial DSO code added into libcrypto for letting OpenSSL (and
7487 OpenSSL-based applications) load shared libraries and bind to
7488 them in a portable way.
7489 [Geoff Thorpe, with contributions from Richard Levitte]
7491 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
7493 *) Make sure _lrotl and _lrotr are only used with MSVC.
7495 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
7496 (the default implementation of RAND_status).
7498 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
7499 to '-clrext' (= clear extensions), as intended and documented.
7500 [Bodo Moeller; inconsistency pointed out by Michael Attili
7503 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
7504 was larger than the MD block size.
7505 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
7507 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
7508 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
7509 using the passed key: if the passed key was a private key the result
7510 of X509_print(), for example, would be to print out all the private key
7514 *) des_quad_cksum() byte order bug fix.
7515 [Ulf Möller, using the problem description in krb4-0.9.7, where
7516 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
7518 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
7520 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
7522 *) For easily testing in shell scripts whether some command
7523 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
7524 returns with exit code 0 iff no command of the given name is available.
7525 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
7526 the output goes to stdout and nothing is printed to stderr.
7527 Additional arguments are always ignored.
7529 Since for each cipher there is a command of the same name,
7530 the 'no-cipher' compilation switches can be tested this way.
7532 ('openssl no-XXX' is not able to detect pseudo-commands such
7533 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
7536 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
7539 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
7540 is set; it will be thrown away anyway because each handshake creates
7542 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
7543 to parameters -- in previous versions (since OpenSSL 0.9.3) the
7544 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
7545 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7548 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7549 'Q' and 'R' lose their special meanings (quit/renegotiate).
7550 This is part of what -quiet does; unlike -quiet, -ign_eof
7551 does not suppress any output.
7554 *) Add compatibility options to the purpose and trust code. The
7555 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7556 accepts a certificate or CA, this was the previous behaviour,
7557 with all the associated security issues.
7559 X509_TRUST_COMPAT is the old trust behaviour: only and
7560 automatically trust self signed roots in certificate store. A
7561 new trust setting X509_TRUST_DEFAULT is used to specify that
7562 a purpose has no associated trust setting and it should instead
7563 use the value in the default purpose.
7566 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7567 and fix a memory leak.
7570 *) In util/mkerr.pl (which implements 'make errors'), preserve
7571 reason strings from the previous version of the .c file, as
7572 the default to have only downcase letters (and digits) in
7573 automatically generated reasons codes is not always appropriate.
7576 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7577 using strerror. Previously, ERR_reason_error_string() returned
7578 library names as reason strings for SYSerr; but SYSerr is a special
7579 case where small numbers are errno values, not library numbers.
7582 *) Add '-dsaparam' option to 'openssl dhparam' application. This
7583 converts DSA parameters into DH parameters. (When creating parameters,
7584 DSA_generate_parameters is used.)
7587 *) Include 'length' (recommended exponent length) in C code generated
7588 by 'openssl dhparam -C'.
7591 *) The second argument to set_label in perlasm was already being used
7592 so couldn't be used as a "file scope" flag. Moved to third argument
7596 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
7597 instead of RAND_bytes for encryption IVs and salts.
7600 *) Include RAND_status() into RAND_METHOD instead of implementing
7601 it only for md_rand.c Otherwise replacing the PRNG by calling
7602 RAND_set_rand_method would be impossible.
7605 *) Don't let DSA_generate_key() enter an infinite loop if the random
7606 number generation fails.
7609 *) New 'rand' application for creating pseudo-random output.
7612 *) Added configuration support for Linux/IA64
7613 [Rolf Haberrecker <rolf@suse.de>]
7615 *) Assembler module support for Mingw32.
7618 *) Shared library support for HPUX (in shlib/).
7619 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
7621 *) Shared library support for Solaris gcc.
7622 [Lutz Behnke <behnke@trustcenter.de>]
7624 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
7626 *) PKCS7_encrypt() was adding text MIME headers twice because they
7627 were added manually and by SMIME_crlf_copy().
7630 *) In bntest.c don't call BN_rand with zero bits argument.
7631 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
7633 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
7634 case was implemented. This caused BN_div_recp() to fail occasionally.
7637 *) Add an optional second argument to the set_label() in the perl
7638 assembly language builder. If this argument exists and is set
7639 to 1 it signals that the assembler should use a symbol whose
7640 scope is the entire file, not just the current function. This
7641 is needed with MASM which uses the format label:: for this scope.
7642 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
7644 *) Change the ASN1 types so they are typedefs by default. Before
7645 almost all types were #define'd to ASN1_STRING which was causing
7646 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
7650 *) Change names of new functions to the new get1/get0 naming
7651 convention: After 'get1', the caller owns a reference count
7652 and has to call ..._free; 'get0' returns a pointer to some
7653 data structure without incrementing reference counters.
7654 (Some of the existing 'get' functions increment a reference
7655 counter, some don't.)
7656 Similarly, 'set1' and 'add1' functions increase reference
7657 counters or duplicate objects.
7660 *) Allow for the possibility of temp RSA key generation failure:
7661 the code used to assume it always worked and crashed on failure.
7664 *) Fix potential buffer overrun problem in BIO_printf().
7665 [Ulf Möller, using public domain code by Patrick Powell; problem
7666 pointed out by David Sacerdote <das33@cornell.edu>]
7668 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
7669 RAND_egd() and RAND_status(). In the command line application,
7670 the EGD socket can be specified like a seed file using RANDFILE
7674 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
7675 Some CAs (e.g. Verisign) distribute certificates in this form.
7678 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
7679 list to exclude them. This means that no special compilation option
7680 is needed to use anonymous DH: it just needs to be included in the
7684 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
7685 EVP_MD_type. The old functionality is available in a new macro called
7686 EVP_MD_md(). Change code that uses it and update docs.
7689 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
7690 where the 'void *' argument is replaced by a function pointer argument.
7691 Previously 'void *' was abused to point to functions, which works on
7692 many platforms, but is not correct. As these functions are usually
7693 called by macros defined in OpenSSL header files, most source code
7694 should work without changes.
7697 *) <openssl/opensslconf.h> (which is created by Configure) now contains
7698 sections with information on -D... compiler switches used for
7699 compiling the library so that applications can see them. To enable
7700 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
7701 must be defined. E.g.,
7702 #define OPENSSL_ALGORITHM_DEFINES
7703 #include <openssl/opensslconf.h>
7704 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
7705 [Richard Levitte, Ulf and Bodo Möller]
7707 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
7711 *) Change the 'other' type in certificate aux info to a STACK_OF
7712 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
7713 the required ASN1 format: arbitrary types determined by an OID.
7716 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
7717 argument to 'req'. This is not because the function is newer or
7718 better than others it just uses the work 'NEW' in the certificate
7719 request header lines. Some software needs this.
7722 *) Reorganise password command line arguments: now passwords can be
7723 obtained from various sources. Delete the PEM_cb function and make
7724 it the default behaviour: i.e. if the callback is NULL and the
7725 usrdata argument is not NULL interpret it as a null terminated pass
7726 phrase. If usrdata and the callback are NULL then the pass phrase
7727 is prompted for as usual.
7730 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
7731 the support is automatically enabled. The resulting binaries will
7732 autodetect the card and use it if present.
7733 [Ben Laurie and Compaq Inc.]
7735 *) Work around for Netscape hang bug. This sends certificate request
7736 and server done in one record. Since this is perfectly legal in the
7737 SSL/TLS protocol it isn't a "bug" option and is on by default. See
7738 the bugs/SSLv3 entry for more info.
7741 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
7744 *) Add -rand argument to smime and pkcs12 applications and read/write
7748 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
7751 *) Add command line password options to the remaining applications.
7754 *) Bug fix for BN_div_recp() for numerators with an even number of
7758 *) More tests in bntest.c, and changed test_bn output.
7761 *) ./config recognizes MacOS X now.
7764 *) Bug fix for BN_div() when the first words of num and divsor are
7765 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
7768 *) Add support for various broken PKCS#8 formats, and command line
7769 options to produce them.
7772 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
7773 get temporary BIGNUMs from a BN_CTX.
7776 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
7780 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
7781 include a #define from the old name to the new. The original intent
7782 was that statically linked binaries could for example just call
7783 SSLeay_add_all_ciphers() to just add ciphers to the table and not
7784 link with digests. This never worked becayse SSLeay_add_all_digests()
7785 and SSLeay_add_all_ciphers() were in the same source file so calling
7786 one would link with the other. They are now in separate source files.
7789 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
7792 *) Use a less unusual form of the Miller-Rabin primality test (it used
7793 a binary algorithm for exponentiation integrated into the Miller-Rabin
7794 loop, our standard modexp algorithms are faster).
7797 *) Support for the EBCDIC character set completed.
7798 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
7800 *) Source code cleanups: use const where appropriate, eliminate casts,
7801 use void * instead of char * in lhash.
7804 *) Bugfix: ssl3_send_server_key_exchange was not restartable
7805 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
7806 this the server could overwrite ephemeral keys that the client
7810 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
7811 using 50 iterations of the Rabin-Miller test.
7813 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
7814 iterations of the Rabin-Miller test as required by the appendix
7815 to FIPS PUB 186[-1]) instead of DSA_is_prime.
7816 As BN_is_prime_fasttest includes trial division, DSA parameter
7817 generation becomes much faster.
7819 This implies a change for the callback functions in DSA_is_prime
7820 and DSA_generate_parameters: The callback function is called once
7821 for each positive witness in the Rabin-Miller test, not just
7822 occasionally in the inner loop; and the parameters to the
7823 callback function now provide an iteration count for the outer
7824 loop rather than for the current invocation of the inner loop.
7825 DSA_generate_parameters additionally can call the callback
7826 function with an 'iteration count' of -1, meaning that a
7827 candidate has passed the trial division test (when q is generated
7828 from an application-provided seed, trial division is skipped).
7831 *) New function BN_is_prime_fasttest that optionally does trial
7832 division before starting the Rabin-Miller test and has
7833 an additional BN_CTX * argument (whereas BN_is_prime always
7834 has to allocate at least one BN_CTX).
7835 'callback(1, -1, cb_arg)' is called when a number has passed the
7836 trial division stage.
7839 *) Fix for bug in CRL encoding. The validity dates weren't being handled
7843 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
7846 *) New function BN_pseudo_rand().
7849 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
7850 bignum version of BN_from_montgomery() with the working code from
7851 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
7855 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
7856 made it impossible to use the same SSL_SESSION data structure in
7857 SSL2 clients in multiple threads.
7860 *) The return value of RAND_load_file() no longer counts bytes obtained
7861 by stat(). RAND_load_file(..., -1) is new and uses the complete file
7862 to seed the PRNG (previously an explicit byte count was required).
7863 [Ulf Möller, Bodo Möller]
7865 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
7866 used (char *) instead of (void *) and had casts all over the place.
7869 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
7872 *) Retain source code compatibility for BN_prime_checks macro:
7873 BN_is_prime(..., BN_prime_checks, ...) now uses
7874 BN_prime_checks_for_size to determine the appropriate number of
7875 Rabin-Miller iterations.
7878 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
7879 DH_CHECK_P_NOT_SAFE_PRIME.
7880 (Check if this is true? OpenPGP calls them "strong".)
7883 *) Merge the functionality of "dh" and "gendh" programs into a new program
7884 "dhparam". The old programs are retained for now but will handle DH keys
7885 (instead of parameters) in future.
7888 *) Make the ciphers, s_server and s_client programs check the return values
7889 when a new cipher list is set.
7892 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
7893 ciphers. Before when the 56bit ciphers were enabled the sorting was
7896 The syntax for the cipher sorting has been extended to support sorting by
7897 cipher-strength (using the strength_bits hard coded in the tables).
7898 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
7900 Fix a bug in the cipher-command parser: when supplying a cipher command
7901 string with an "undefined" symbol (neither command nor alphanumeric
7902 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
7903 an error is flagged.
7905 Due to the strength-sorting extension, the code of the
7906 ssl_create_cipher_list() function was completely rearranged. I hope that
7907 the readability was also increased :-)
7908 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7910 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
7911 for the first serial number and places 2 in the serial number file. This
7912 avoids problems when the root CA is created with serial number zero and
7913 the first user certificate has the same issuer name and serial number
7917 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
7918 the new code. Add documentation for this stuff.
7921 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
7922 X509_*() to X509at_*() on the grounds that they don't handle X509
7923 structures and behave in an analagous way to the X509v3 functions:
7924 they shouldn't be called directly but wrapper functions should be used
7927 So we also now have some wrapper functions that call the X509at functions
7928 when passed certificate requests. (TO DO: similar things can be done with
7929 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
7930 things. Some of these need some d2i or i2d and print functionality
7931 because they handle more complex structures.)
7934 *) Add missing #ifndefs that caused missing symbols when building libssl
7935 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
7936 NO_RSA in ssl/s2*.c.
7937 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
7939 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
7940 has a return value which indicates the quality of the random data
7941 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
7942 error queue. New function RAND_pseudo_bytes() generates output that is
7943 guaranteed to be unique but not unpredictable. RAND_add is like
7944 RAND_seed, but takes an extra argument for an entropy estimate
7945 (RAND_seed always assumes full entropy).
7948 *) Do more iterations of Rabin-Miller probable prime test (specifically,
7949 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
7950 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
7951 in crypto/bn/bn_prime.c for the complete table). This guarantees a
7952 false-positive rate of at most 2^-80 for random input.
7955 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
7958 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
7959 in the 0.9.5 release), this returns the chain
7960 from an X509_CTX structure with a dup of the stack and all
7961 the X509 reference counts upped: so the stack will exist
7962 after X509_CTX_cleanup() has been called. Modify pkcs12.c
7965 Also make SSL_SESSION_print() print out the verify return
7969 *) Add manpage for the pkcs12 command. Also change the default
7970 behaviour so MAC iteration counts are used unless the new
7971 -nomaciter option is used. This improves file security and
7972 only older versions of MSIE (4.0 for example) need it.
7975 *) Honor the no-xxx Configure options when creating .DEF files.
7978 *) Add PKCS#10 attributes to field table: challengePassword,
7979 unstructuredName and unstructuredAddress. These are taken from
7980 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
7981 international characters are used.
7983 More changes to X509_ATTRIBUTE code: allow the setting of types
7984 based on strings. Remove the 'loc' parameter when adding
7985 attributes because these will be a SET OF encoding which is sorted
7989 *) Initial changes to the 'req' utility to allow request generation
7990 automation. This will allow an application to just generate a template
7991 file containing all the field values and have req construct the
7994 Initial support for X509_ATTRIBUTE handling. Stacks of these are
7995 used all over the place including certificate requests and PKCS#7
7996 structures. They are currently handled manually where necessary with
7997 some primitive wrappers for PKCS#7. The new functions behave in a
7998 manner analogous to the X509 extension functions: they allow
7999 attributes to be looked up by NID and added.
8001 Later something similar to the X509V3 code would be desirable to
8002 automatically handle the encoding, decoding and printing of the
8003 more complex types. The string types like challengePassword can
8004 be handled by the string table functions.
8006 Also modified the multi byte string table handling. Now there is
8007 a 'global mask' which masks out certain types. The table itself
8008 can use the flag STABLE_NO_MASK to ignore the mask setting: this
8009 is useful when for example there is only one permissible type
8010 (as in countryName) and using the mask might result in no valid
8014 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
8015 SSL_get_peer_finished to allow applications to obtain the latest
8016 Finished messages sent to the peer or expected from the peer,
8017 respectively. (SSL_get_peer_finished is usually the Finished message
8018 actually received from the peer, otherwise the protocol will be aborted.)
8020 As the Finished message are message digests of the complete handshake
8021 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
8022 be used for external authentication procedures when the authentication
8023 provided by SSL/TLS is not desired or is not enough.
8026 *) Enhanced support for Alpha Linux is added. Now ./config checks if
8027 the host supports BWX extension and if Compaq C is present on the
8028 $PATH. Just exploiting of the BWX extension results in 20-30%
8029 performance kick for some algorithms, e.g. DES and RC4 to mention
8030 a couple. Compaq C in turn generates ~20% faster code for MD5 and
8034 *) Add support for MS "fast SGC". This is arguably a violation of the
8035 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
8036 weak crypto and after checking the certificate is SGC a second one
8037 with strong crypto. MS SGC stops the first handshake after receiving
8038 the server certificate message and sends a second client hello. Since
8039 a server will typically do all the time consuming operations before
8040 expecting any further messages from the client (server key exchange
8041 is the most expensive) there is little difference between the two.
8043 To get OpenSSL to support MS SGC we have to permit a second client
8044 hello message after we have sent server done. In addition we have to
8045 reset the MAC if we do get this second client hello.
8048 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
8049 if a DER encoded private key is RSA or DSA traditional format. Changed
8050 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
8051 format DER encoded private key. Newer code should use PKCS#8 format which
8052 has the key type encoded in the ASN1 structure. Added DER private key
8053 support to pkcs8 application.
8056 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
8057 ciphersuites has been selected (as required by the SSL 3/TLS 1
8058 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
8059 is set, we interpret this as a request to violate the specification
8060 (the worst that can happen is a handshake failure, and 'correct'
8061 behaviour would result in a handshake failure anyway).
8064 *) In SSL_CTX_add_session, take into account that there might be multiple
8065 SSL_SESSION structures with the same session ID (e.g. when two threads
8066 concurrently obtain them from an external cache).
8067 The internal cache can handle only one SSL_SESSION with a given ID,
8068 so if there's a conflict, we now throw out the old one to achieve
8072 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
8073 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
8074 some routines that use cipher OIDs: some ciphers do not have OIDs
8075 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
8079 *) Simplify the trust setting structure and code. Now we just have
8080 two sequences of OIDs for trusted and rejected settings. These will
8081 typically have values the same as the extended key usage extension
8082 and any application specific purposes.
8084 The trust checking code now has a default behaviour: it will just
8085 check for an object with the same NID as the passed id. Functions can
8086 be provided to override either the default behaviour or the behaviour
8087 for a given id. SSL client, server and email already have functions
8088 in place for compatibility: they check the NID and also return "trusted"
8089 if the certificate is self signed.
8092 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
8093 traditional format into an EVP_PKEY structure.
8096 *) Add a password callback function PEM_cb() which either prompts for
8097 a password if usr_data is NULL or otherwise assumes it is a null
8098 terminated password. Allow passwords to be passed on command line
8099 environment or config files in a few more utilities.
8102 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
8103 keys. Add some short names for PKCS#8 PBE algorithms and allow them
8104 to be specified on the command line for the pkcs8 and pkcs12 utilities.
8105 Update documentation.
8108 *) Support for ASN1 "NULL" type. This could be handled before by using
8109 ASN1_TYPE but there wasn't any function that would try to read a NULL
8110 and produce an error if it couldn't. For compatibility we also have
8111 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
8112 don't allocate anything because they don't need to.
8115 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
8117 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
8119 *) Rebuild of the memory allocation routines used by OpenSSL code and
8120 possibly others as well. The purpose is to make an interface that
8121 provide hooks so anyone can build a separate set of allocation and
8122 deallocation routines to be used by OpenSSL, for example memory
8123 pool implementations, or something else, which was previously hard
8124 since Malloc(), Realloc() and Free() were defined as macros having
8125 the values malloc, realloc and free, respectively (except for Win32
8126 compilations). The same is provided for memory debugging code.
8127 OpenSSL already comes with functionality to find memory leaks, but
8128 this gives people a chance to debug other memory problems.
8130 With these changes, a new set of functions and macros have appeared:
8132 CRYPTO_set_mem_debug_functions() [F]
8133 CRYPTO_get_mem_debug_functions() [F]
8134 CRYPTO_dbg_set_options() [F]
8135 CRYPTO_dbg_get_options() [F]
8136 CRYPTO_malloc_debug_init() [M]
8138 The memory debug functions are NULL by default, unless the library
8139 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
8140 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
8141 gives the standard debugging functions that come with OpenSSL) or
8142 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
8143 provided by the library user) must be used. When the standard
8144 debugging functions are used, CRYPTO_dbg_set_options can be used to
8145 request additional information:
8146 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
8147 the CRYPTO_MDEBUG_xxx macro when compiling the library.
8149 Also, things like CRYPTO_set_mem_functions will always give the
8150 expected result (the new set of functions is used for allocation
8151 and deallocation) at all times, regardless of platform and compiler
8154 To finish it up, some functions that were never use in any other
8155 way than through macros have a new API and new semantic:
8158 CRYPTO_dbg_realloc()
8161 All macros of value have retained their old syntax.
8162 [Richard Levitte and Bodo Moeller]
8164 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
8165 ordering of SMIMECapabilities wasn't in "strength order" and there
8166 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
8170 *) Some ASN1 types with illegal zero length encoding (INTEGER,
8171 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
8172 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
8174 *) Merge in my S/MIME library for OpenSSL. This provides a simple
8175 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
8176 functionality to handle multipart/signed properly) and a utility
8177 called 'smime' to call all this stuff. This is based on code I
8178 originally wrote for Celo who have kindly allowed it to be
8179 included in OpenSSL.
8182 *) Add variants des_set_key_checked and des_set_key_unchecked of
8183 des_set_key (aka des_key_sched). Global variable des_check_key
8184 decides which of these is called by des_set_key; this way
8185 des_check_key behaves as it always did, but applications and
8186 the library itself, which was buggy for des_check_key == 1,
8187 have a cleaner way to pick the version they need.
8190 *) New function PKCS12_newpass() which changes the password of a
8194 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
8195 dynamic mix. In both cases the ids can be used as an index into the
8196 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
8197 functions so they accept a list of the field values and the
8198 application doesn't need to directly manipulate the X509_TRUST
8202 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
8206 *) Modify the way the V3 extension code looks up extensions. This now
8207 works in a similar way to the object code: we have some "standard"
8208 extensions in a static table which is searched with OBJ_bsearch()
8209 and the application can add dynamic ones if needed. The file
8210 crypto/x509v3/ext_dat.h now has the info: this file needs to be
8211 updated whenever a new extension is added to the core code and kept
8212 in ext_nid order. There is a simple program 'tabtest.c' which checks
8213 this. New extensions are not added too often so this file can readily
8214 be maintained manually.
8216 There are two big advantages in doing things this way. The extensions
8217 can be looked up immediately and no longer need to be "added" using
8218 X509V3_add_standard_extensions(): this function now does nothing.
8219 [Side note: I get *lots* of email saying the extension code doesn't
8220 work because people forget to call this function]
8221 Also no dynamic allocation is done unless new extensions are added:
8222 so if we don't add custom extensions there is no need to call
8223 X509V3_EXT_cleanup().
8226 *) Modify enc utility's salting as follows: make salting the default. Add a
8227 magic header, so unsalted files fail gracefully instead of just decrypting
8228 to garbage. This is because not salting is a big security hole, so people
8229 should be discouraged from doing it.
8232 *) Fixes and enhancements to the 'x509' utility. It allowed a message
8233 digest to be passed on the command line but it only used this
8234 parameter when signing a certificate. Modified so all relevant
8235 operations are affected by the digest parameter including the
8236 -fingerprint and -x509toreq options. Also -x509toreq choked if a
8237 DSA key was used because it didn't fix the digest.
8240 *) Initial certificate chain verify code. Currently tests the untrusted
8241 certificates for consistency with the verify purpose (which is set
8242 when the X509_STORE_CTX structure is set up) and checks the pathlength.
8244 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
8245 this is because it will reject chains with invalid extensions whereas
8246 every previous version of OpenSSL and SSLeay made no checks at all.
8248 Trust code: checks the root CA for the relevant trust settings. Trust
8249 settings have an initial value consistent with the verify purpose: e.g.
8250 if the verify purpose is for SSL client use it expects the CA to be
8251 trusted for SSL client use. However the default value can be changed to
8252 permit custom trust settings: one example of this would be to only trust
8253 certificates from a specific "secure" set of CAs.
8255 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
8256 which should be used for version portability: especially since the
8257 verify structure is likely to change more often now.
8259 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
8260 to set them. If not set then assume SSL clients will verify SSL servers
8263 Two new options to the verify program: -untrusted allows a set of
8264 untrusted certificates to be passed in and -purpose which sets the
8265 intended purpose of the certificate. If a purpose is set then the
8266 new chain verify code is used to check extension consistency.
8269 *) Support for the authority information access extension.
8272 *) Modify RSA and DSA PEM read routines to transparently handle
8273 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
8274 public keys in a format compatible with certificate
8275 SubjectPublicKeyInfo structures. Unfortunately there were already
8276 functions called *_PublicKey_* which used various odd formats so
8277 these are retained for compatibility: however the DSA variants were
8278 never in a public release so they have been deleted. Changed dsa/rsa
8279 utilities to handle the new format: note no releases ever handled public
8280 keys so we should be OK.
8282 The primary motivation for this change is to avoid the same fiasco
8283 that dogs private keys: there are several incompatible private key
8284 formats some of which are standard and some OpenSSL specific and
8285 require various evil hacks to allow partial transparent handling and
8286 even then it doesn't work with DER formats. Given the option anything
8287 other than PKCS#8 should be dumped: but the other formats have to
8288 stay in the name of compatibility.
8290 With public keys and the benefit of hindsight one standard format
8291 is used which works with EVP_PKEY, RSA or DSA structures: though
8292 it clearly returns an error if you try to read the wrong kind of key.
8294 Added a -pubkey option to the 'x509' utility to output the public key.
8295 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
8296 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
8297 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
8298 that do the same as the EVP_PKEY_assign_*() except they up the
8299 reference count of the added key (they don't "swallow" the
8303 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
8304 CRLs would fail if the file contained no certificates or no CRLs:
8305 added a new function to read in both types and return the number
8306 read: this means that if none are read it will be an error. The
8307 DER versions of the certificate and CRL reader would always fail
8308 because it isn't possible to mix certificates and CRLs in DER format
8309 without choking one or the other routine. Changed this to just read
8310 a certificate: this is the best we can do. Also modified the code
8311 in apps/verify.c to take notice of return codes: it was previously
8312 attempting to read in certificates from NULL pointers and ignoring
8313 any errors: this is one reason why the cert and CRL reader seemed
8314 to work. It doesn't check return codes from the default certificate
8315 routines: these may well fail if the certificates aren't installed.
8318 *) Code to support otherName option in GeneralName.
8321 *) First update to verify code. Change the verify utility
8322 so it warns if it is passed a self signed certificate:
8323 for consistency with the normal behaviour. X509_verify
8324 has been modified to it will now verify a self signed
8325 certificate if *exactly* the same certificate appears
8326 in the store: it was previously impossible to trust a
8327 single self signed certificate. This means that:
8328 openssl verify ss.pem
8329 now gives a warning about a self signed certificate but
8330 openssl verify -CAfile ss.pem ss.pem
8334 *) For servers, store verify_result in SSL_SESSION data structure
8335 (and add it to external session representation).
8336 This is needed when client certificate verifications fails,
8337 but an application-provided verification callback (set by
8338 SSL_CTX_set_cert_verify_callback) allows accepting the session
8339 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
8340 but returns 1): When the session is reused, we have to set
8341 ssl->verify_result to the appropriate error code to avoid
8343 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
8345 *) Fix a bug in the new PKCS#7 code: it didn't consider the
8346 case in PKCS7_dataInit() where the signed PKCS7 structure
8347 didn't contain any existing data because it was being created.
8348 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
8350 *) Add a salt to the key derivation routines in enc.c. This
8351 forms the first 8 bytes of the encrypted file. Also add a
8352 -S option to allow a salt to be input on the command line.
8355 *) New function X509_cmp(). Oddly enough there wasn't a function
8356 to compare two certificates. We do this by working out the SHA1
8357 hash and comparing that. X509_cmp() will be needed by the trust
8361 *) SSL_get1_session() is like SSL_get_session(), but increments
8362 the reference count in the SSL_SESSION returned.
8363 [Geoff Thorpe <geoff@eu.c2.net>]
8365 *) Fix for 'req': it was adding a null to request attributes.
8366 Also change the X509_LOOKUP and X509_INFO code to handle
8367 certificate auxiliary information.
8370 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
8374 *) Add the possibility to add extra information to the memory leak
8375 detecting output, to form tracebacks, showing from where each
8376 allocation was originated: CRYPTO_push_info("constant string") adds
8377 the string plus current file name and line number to a per-thread
8378 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
8379 is like calling CYRPTO_pop_info() until the stack is empty.
8380 Also updated memory leak detection code to be multi-thread-safe.
8383 *) Add options -text and -noout to pkcs7 utility and delete the
8384 encryption options which never did anything. Update docs.
8387 *) Add options to some of the utilities to allow the pass phrase
8388 to be included on either the command line (not recommended on
8389 OSes like Unix) or read from the environment. Update the
8390 manpages and fix a few bugs.
8393 *) Add a few manpages for some of the openssl commands.
8396 *) Fix the -revoke option in ca. It was freeing up memory twice,
8397 leaking and not finding already revoked certificates.
8400 *) Extensive changes to support certificate auxiliary information.
8401 This involves the use of X509_CERT_AUX structure and X509_AUX
8402 functions. An X509_AUX function such as PEM_read_X509_AUX()
8403 can still read in a certificate file in the usual way but it
8404 will also read in any additional "auxiliary information". By
8405 doing things this way a fair degree of compatibility can be
8406 retained: existing certificates can have this information added
8407 using the new 'x509' options.
8409 Current auxiliary information includes an "alias" and some trust
8410 settings. The trust settings will ultimately be used in enhanced
8411 certificate chain verification routines: currently a certificate
8412 can only be trusted if it is self signed and then it is trusted
8416 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
8417 The problem was that one of the replacement routines had not been working
8418 since SSLeay releases. For now the offending routine has been replaced
8419 with non-optimised assembler. Even so, this now gives around 95%
8420 performance improvement for 1024 bit RSA signs.
8423 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
8424 handling. Most clients have the effective key size in bits equal to
8425 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
8426 A few however don't do this and instead use the size of the decrypted key
8427 to determine the RC2 key length and the AlgorithmIdentifier to determine
8428 the effective key length. In this case the effective key length can still
8429 be 40 bits but the key length can be 168 bits for example. This is fixed
8430 by manually forcing an RC2 key into the EVP_PKEY structure because the
8431 EVP code can't currently handle unusual RC2 key sizes: it always assumes
8432 the key length and effective key length are equal.
8435 *) Add a bunch of functions that should simplify the creation of
8436 X509_NAME structures. Now you should be able to do:
8437 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
8438 and have it automatically work out the correct field type and fill in
8439 the structures. The more adventurous can try:
8440 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
8441 and it will (hopefully) work out the correct multibyte encoding.
8444 *) Change the 'req' utility to use the new field handling and multibyte
8445 copy routines. Before the DN field creation was handled in an ad hoc
8446 way in req, ca, and x509 which was rather broken and didn't support
8447 BMPStrings or UTF8Strings. Since some software doesn't implement
8448 BMPStrings or UTF8Strings yet, they can be enabled using the config file
8449 using the dirstring_type option. See the new comment in the default
8450 openssl.cnf for more info.
8453 *) Make crypto/rand/md_rand.c more robust:
8454 - Assure unique random numbers after fork().
8455 - Make sure that concurrent threads access the global counter and
8456 md serializably so that we never lose entropy in them
8457 or use exactly the same state in multiple threads.
8458 Access to the large state is not always serializable because
8459 the additional locking could be a performance killer, and
8460 md should be large enough anyway.
8463 *) New file apps/app_rand.c with commonly needed functionality
8464 for handling the random seed file.
8466 Use the random seed file in some applications that previously did not:
8468 dsaparam -genkey (which also ignored its '-rand' option),
8471 x509 (when signing).
8472 Except on systems with /dev/urandom, it is crucial to have a random
8473 seed file at least for key creation, DSA signing, and for DH exchanges;
8474 for RSA signatures we could do without one.
8476 gendh and gendsa (unlike genrsa) used to read only the first byte
8477 of each file listed in the '-rand' option. The function as previously
8478 found in genrsa is now in app_rand.c and is used by all programs
8479 that support '-rand'.
8482 *) In RAND_write_file, use mode 0600 for creating files;
8483 don't just chmod when it may be too late.
8486 *) Report an error from X509_STORE_load_locations
8487 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
8490 *) New function ASN1_mbstring_copy() this copies a string in either
8491 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
8492 into an ASN1_STRING type. A mask of permissible types is passed
8493 and it chooses the "minimal" type to use or an error if not type
8497 *) Add function equivalents to the various macros in asn1.h. The old
8498 macros are retained with an M_ prefix. Code inside the library can
8499 use the M_ macros. External code (including the openssl utility)
8500 should *NOT* in order to be "shared library friendly".
8503 *) Add various functions that can check a certificate's extensions
8504 to see if it usable for various purposes such as SSL client,
8505 server or S/MIME and CAs of these types. This is currently
8506 VERY EXPERIMENTAL but will ultimately be used for certificate chain
8507 verification. Also added a -purpose flag to x509 utility to
8508 print out all the purposes.
8511 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
8515 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
8516 for, obtain and decode and extension and obtain its critical flag.
8517 This allows all the necessary extension code to be handled in a
8518 single function call.
8521 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
8522 platforms. See crypto/rc4/rc4_enc.c for further details.
8525 *) New -noout option to asn1parse. This causes no output to be produced
8526 its main use is when combined with -strparse and -out to extract data
8527 from a file (which may not be in ASN.1 format).
8530 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
8531 when producing the local key id.
8532 [Richard Levitte <levitte@stacken.kth.se>]
8534 *) New option -dhparam in s_server. This allows a DH parameter file to be
8535 stated explicitly. If it is not stated then it tries the first server
8536 certificate file. The previous behaviour hard coded the filename
8540 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
8541 a public key to be input or output. For example:
8542 openssl rsa -in key.pem -pubout -out pubkey.pem
8543 Also added necessary DSA public key functions to handle this.
8546 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
8547 in the message. This was handled by allowing
8548 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8549 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8551 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8552 to the end of the strings whereas this didn't. This would cause problems
8553 if strings read with d2i_ASN1_bytes() were later modified.
8554 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8556 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8557 data and it contains EOF it will end up returning an error. This is
8558 caused by input 46 bytes long. The cause is due to the way base64
8559 BIOs find the start of base64 encoded data. They do this by trying a
8560 trial decode on each line until they find one that works. When they
8561 do a flag is set and it starts again knowing it can pass all the
8562 data directly through the decoder. Unfortunately it doesn't reset
8563 the context it uses. This means that if EOF is reached an attempt
8564 is made to pass two EOFs through the context and this causes the
8565 resulting error. This can also cause other problems as well. As is
8566 usual with these problems it takes *ages* to find and the fix is
8567 trivial: move one line.
8568 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8570 *) Ugly workaround to get s_client and s_server working under Windows. The
8571 old code wouldn't work because it needed to select() on sockets and the
8572 tty (for keypresses and to see if data could be written). Win32 only
8573 supports select() on sockets so we select() with a 1s timeout on the
8574 sockets and then see if any characters are waiting to be read, if none
8575 are present then we retry, we also assume we can always write data to
8576 the tty. This isn't nice because the code then blocks until we've
8577 received a complete line of data and it is effectively polling the
8578 keyboard at 1s intervals: however it's quite a bit better than not
8579 working at all :-) A dedicated Windows application might handle this
8580 with an event loop for example.
8583 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
8584 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
8585 will be called when RSA_sign() and RSA_verify() are used. This is useful
8586 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
8587 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
8588 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
8589 This necessitated the support of an extra signature type NID_md5_sha1
8590 for SSL signatures and modifications to the SSL library to use it instead
8591 of calling RSA_public_decrypt() and RSA_private_encrypt().
8594 *) Add new -verify -CAfile and -CApath options to the crl program, these
8595 will lookup a CRL issuers certificate and verify the signature in a
8596 similar way to the verify program. Tidy up the crl program so it
8597 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
8598 less strict. It will now permit CRL extensions even if it is not
8599 a V2 CRL: this will allow it to tolerate some broken CRLs.
8602 *) Initialize all non-automatic variables each time one of the openssl
8603 sub-programs is started (this is necessary as they may be started
8604 multiple times from the "OpenSSL>" prompt).
8605 [Lennart Bang, Bodo Moeller]
8607 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
8608 removing all other RSA functionality (this is what NO_RSA does). This
8609 is so (for example) those in the US can disable those operations covered
8610 by the RSA patent while allowing storage and parsing of RSA keys and RSA
8614 *) Non-copying interface to BIO pairs.
8615 (still largely untested)
8618 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
8619 ASCII string. This was handled independently in various places before.
8622 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
8623 UTF8 strings a character at a time.
8626 *) Use client_version from client hello to select the protocol
8627 (s23_srvr.c) and for RSA client key exchange verification
8628 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
8631 *) Add various utility functions to handle SPKACs, these were previously
8632 handled by poking round in the structure internals. Added new function
8633 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
8634 print, verify and generate SPKACs. Based on an original idea from
8635 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
8638 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
8641 *) Allow the config file extension section to be overwritten on the
8642 command line. Based on an original idea from Massimiliano Pala
8643 <madwolf@comune.modena.it>. The new option is called -extensions
8644 and can be applied to ca, req and x509. Also -reqexts to override
8645 the request extensions in req and -crlexts to override the crl extensions
8649 *) Add new feature to the SPKAC handling in ca. Now you can include
8650 the same field multiple times by preceding it by "XXXX." for example:
8653 this is the same syntax as used in the req config file.
8656 *) Allow certificate extensions to be added to certificate requests. These
8657 are specified in a 'req_extensions' option of the req section of the
8658 config file. They can be printed out with the -text option to req but
8659 are otherwise ignored at present.
8662 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
8663 data read consists of only the final block it would not decrypted because
8664 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
8665 A misplaced 'break' also meant the decrypted final block might not be
8666 copied until the next read.
8669 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
8670 a few extra parameters to the DH structure: these will be useful if
8671 for example we want the value of 'q' or implement X9.42 DH.
8674 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
8675 provides hooks that allow the default DSA functions or functions on a
8676 "per key" basis to be replaced. This allows hardware acceleration and
8677 hardware key storage to be handled without major modification to the
8678 library. Also added low level modexp hooks and CRYPTO_EX structure and
8679 associated functions.
8682 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
8683 as "read only": it can't be written to and the buffer it points to will
8684 not be freed. Reading from a read only BIO is much more efficient than
8685 a normal memory BIO. This was added because there are several times when
8686 an area of memory needs to be read from a BIO. The previous method was
8687 to create a memory BIO and write the data to it, this results in two
8688 copies of the data and an O(n^2) reading algorithm. There is a new
8689 function BIO_new_mem_buf() which creates a read only memory BIO from
8690 an area of memory. Also modified the PKCS#7 routines to use read only
8694 *) Bugfix: ssl23_get_client_hello did not work properly when called in
8695 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
8696 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
8697 but a retry condition occured while trying to read the rest.
8700 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
8701 NID_pkcs7_encrypted by default: this was wrong since this should almost
8702 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
8703 the encrypted data type: this is a more sensible place to put it and it
8704 allows the PKCS#12 code to be tidied up that duplicated this
8708 *) Changed obj_dat.pl script so it takes its input and output files on
8709 the command line. This should avoid shell escape redirection problems
8713 *) Initial support for certificate extension requests, these are included
8714 in things like Xenroll certificate requests. Included functions to allow
8715 extensions to be obtained and added.
8718 *) -crlf option to s_client and s_server for sending newlines as
8719 CRLF (as required by many protocols).
8722 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
8724 *) Install libRSAglue.a when OpenSSL is built with RSAref.
8725 [Ralf S. Engelschall]
8727 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
8728 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
8730 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
8734 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
8735 DH parameters/keys (q is lost during that conversion, but the resulting
8736 DH parameters contain its length).
8738 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
8739 much faster than DH_generate_parameters (which creates parameters
8740 where p = 2*q + 1), and also the smaller q makes DH computations
8741 much more efficient (160-bit exponentiation instead of 1024-bit
8742 exponentiation); so this provides a convenient way to support DHE
8743 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
8744 utter importance to use
8745 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8747 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8748 when such DH parameters are used, because otherwise small subgroup
8749 attacks may become possible!
8752 *) Avoid memory leak in i2d_DHparams.
8755 *) Allow the -k option to be used more than once in the enc program:
8756 this allows the same encrypted message to be read by multiple recipients.
8759 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
8760 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
8761 it will always use the numerical form of the OID, even if it has a short
8765 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
8766 method only got called if p,q,dmp1,dmq1,iqmp components were present,
8767 otherwise bn_mod_exp was called. In the case of hardware keys for example
8768 no private key components need be present and it might store extra data
8769 in the RSA structure, which cannot be accessed from bn_mod_exp.
8770 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
8771 private key operations.
8774 *) Added support for SPARC Linux.
8777 *) pem_password_cb function type incompatibly changed from
8778 typedef int pem_password_cb(char *buf, int size, int rwflag);
8780 ....(char *buf, int size, int rwflag, void *userdata);
8781 so that applications can pass data to their callbacks:
8782 The PEM[_ASN1]_{read,write}... functions and macros now take an
8783 additional void * argument, which is just handed through whenever
8784 the password callback is called.
8785 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
8787 New function SSL_CTX_set_default_passwd_cb_userdata.
8789 Compatibility note: As many C implementations push function arguments
8790 onto the stack in reverse order, the new library version is likely to
8791 interoperate with programs that have been compiled with the old
8792 pem_password_cb definition (PEM_whatever takes some data that
8793 happens to be on the stack as its last argument, and the callback
8794 just ignores this garbage); but there is no guarantee whatsoever that
8797 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
8798 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
8799 problems not only on Windows, but also on some Unix platforms.
8800 To avoid problematic command lines, these definitions are now in an
8801 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
8802 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
8805 *) MIPS III/IV assembler module is reimplemented.
8808 *) More DES library cleanups: remove references to srand/rand and
8809 delete an unused file.
8812 *) Add support for the the free Netwide assembler (NASM) under Win32,
8813 since not many people have MASM (ml) and it can be hard to obtain.
8814 This is currently experimental but it seems to work OK and pass all
8815 the tests. Check out INSTALL.W32 for info.
8818 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
8819 without temporary keys kept an extra copy of the server key,
8820 and connections with temporary keys did not free everything in case
8824 *) New function RSA_check_key and new openssl rsa option -check
8825 for verifying the consistency of RSA keys.
8826 [Ulf Moeller, Bodo Moeller]
8828 *) Various changes to make Win32 compile work:
8829 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
8830 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
8831 comparison" warnings.
8832 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
8835 *) Add a debugging option to PKCS#5 v2 key generation function: when
8836 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
8837 derived keys are printed to stderr.
8840 *) Copy the flags in ASN1_STRING_dup().
8841 [Roman E. Pavlov <pre@mo.msk.ru>]
8843 *) The x509 application mishandled signing requests containing DSA
8844 keys when the signing key was also DSA and the parameters didn't match.
8846 It was supposed to omit the parameters when they matched the signing key:
8847 the verifying software was then supposed to automatically use the CA's
8848 parameters if they were absent from the end user certificate.
8850 Omitting parameters is no longer recommended. The test was also
8851 the wrong way round! This was probably due to unusual behaviour in
8852 EVP_cmp_parameters() which returns 1 if the parameters match.
8853 This meant that parameters were omitted when they *didn't* match and
8854 the certificate was useless. Certificates signed with 'ca' didn't have
8856 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
8858 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
8859 The interface is as follows:
8860 Applications can use
8861 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
8862 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
8863 "off" is now the default.
8864 The library internally uses
8865 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
8866 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
8867 to disable memory-checking temporarily.
8869 Some inconsistent states that previously were possible (and were
8870 even the default) are now avoided.
8872 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
8873 with each memory chunk allocated; this is occasionally more helpful
8874 than just having a counter.
8876 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
8878 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
8882 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
8883 which largely parallels "options", but is for changing API behaviour,
8884 whereas "options" are about protocol behaviour.
8885 Initial "mode" flags are:
8887 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
8888 a single record has been written.
8889 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
8890 retries use the same buffer location.
8891 (But all of the contents must be
8895 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
8898 *) Fix problems with no-hmac etc.
8899 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
8901 *) New functions RSA_get_default_method(), RSA_set_method() and
8902 RSA_get_method(). These allows replacement of RSA_METHODs without having
8903 to mess around with the internals of an RSA structure.
8906 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
8907 Also really enable memory leak checks in openssl.c and in some
8909 [Chad C. Mulligan, Bodo Moeller]
8911 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
8912 up the length of negative integers. This has now been simplified to just
8913 store the length when it is first determined and use it later, rather
8914 than trying to keep track of where data is copied and updating it to
8916 [Steve Henson, reported by Brien Wheeler
8917 <bwheeler@authentica-security.com>]
8919 *) Add a new function PKCS7_signatureVerify. This allows the verification
8920 of a PKCS#7 signature but with the signing certificate passed to the
8921 function itself. This contrasts with PKCS7_dataVerify which assumes the
8922 certificate is present in the PKCS#7 structure. This isn't always the
8923 case: certificates can be omitted from a PKCS#7 structure and be
8924 distributed by "out of band" means (such as a certificate database).
8927 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
8928 function prototypes in pem.h, also change util/mkdef.pl to add the
8929 necessary function names.
8932 *) mk1mf.pl (used by Windows builds) did not properly read the
8933 options set by Configure in the top level Makefile, and Configure
8934 was not even able to write more than one option correctly.
8935 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
8938 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
8939 file to be loaded from a BIO or FILE pointer. The BIO version will
8940 for example allow memory BIOs to contain config info.
8943 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
8944 Whoever hopes to achieve shared-library compatibility across versions
8945 must use this, not the compile-time macro.
8946 (Exercise 0.9.4: Which is the minimum library version required by
8948 Note: All this applies only to multi-threaded programs, others don't
8952 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
8953 through a BIO pair triggered the default case, i.e.
8954 SSLerr(...,SSL_R_UNKNOWN_STATE).
8957 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
8958 can use the SSL library even if none of the specific BIOs is
8962 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
8963 for the encoded length.
8964 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
8966 *) Add initial documentation of the X509V3 functions.
8969 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
8970 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
8971 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
8972 secure PKCS#8 private key format with a high iteration count.
8975 *) Fix determination of Perl interpreter: A perl or perl5
8976 _directory_ in $PATH was also accepted as the interpreter.
8977 [Ralf S. Engelschall]
8979 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
8980 wrong with it but it was very old and did things like calling
8981 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
8985 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
8986 to use the new extension code.
8989 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
8990 with macros. This should make it easier to change their form, add extra
8991 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
8995 *) Add to configuration table a new entry that can specify an alternative
8996 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
8997 according to Mark Crispin <MRC@Panda.COM>.
9001 *) DES CBC did not update the IV. Weird.
9004 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
9005 Changing the behaviour of the former might break existing programs --
9006 where IV updating is needed, des_ncbc_encrypt can be used.
9009 *) When bntest is run from "make test" it drives bc to check its
9010 calculations, as well as internally checking them. If an internal check
9011 fails, it needs to cause bc to give a non-zero result or make test carries
9012 on without noticing the failure. Fixed.
9015 *) DES library cleanups.
9018 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
9019 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
9020 ciphers. NOTE: although the key derivation function has been verified
9021 against some published test vectors it has not been extensively tested
9022 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
9026 *) Instead of "mkdir -p", which is not fully portable, use new
9027 Perl script "util/mkdir-p.pl".
9030 *) Rewrite the way password based encryption (PBE) is handled. It used to
9031 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
9032 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
9033 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
9034 the 'parameter' field of the AlgorithmIdentifier is passed to the
9035 underlying key generation function so it must do its own ASN1 parsing.
9036 This has also changed the EVP_PBE_CipherInit() function which now has a
9037 'parameter' argument instead of literal salt and iteration count values
9038 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
9041 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
9042 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
9043 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
9044 KEY" because this clashed with PKCS#8 unencrypted string. Since this
9045 value was just used as a "magic string" and not used directly its
9046 value doesn't matter.
9049 *) Introduce some semblance of const correctness to BN. Shame C doesn't
9053 *) "linux-sparc64" configuration (ultrapenguin).
9054 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
9055 "linux-sparc" configuration.
9056 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
9058 *) config now generates no-xxx options for missing ciphers.
9061 *) Support the EBCDIC character set (work in progress).
9062 File ebcdic.c not yet included because it has a different license.
9063 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9065 *) Support BS2000/OSD-POSIX.
9066 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9068 *) Make callbacks for key generation use void * instead of char *.
9071 *) Make S/MIME samples compile (not yet tested).
9074 *) Additional typesafe stacks.
9077 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
9081 Changes between 0.9.3 and 0.9.3a [29 May 1999]
9083 *) New configuration variant "sco5-gcc".
9085 *) Updated some demos.
9086 [Sean O Riordain, Wade Scholine]
9088 *) Add missing BIO_free at exit of pkcs12 application.
9091 *) Fix memory leak in conf.c.
9094 *) Updates for Win32 to assembler version of MD5.
9097 *) Set #! path to perl in apps/der_chop to where we found it
9098 instead of using a fixed path.
9101 *) SHA library changes for irix64-mips4-cc.
9104 *) Improvements for VMS support.
9108 Changes between 0.9.2b and 0.9.3 [24 May 1999]
9110 *) Bignum library bug fix. IRIX 6 passes "make test" now!
9111 This also avoids the problems with SC4.2 and unpatched SC5.
9112 [Andy Polyakov <appro@fy.chalmers.se>]
9114 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
9115 These are required because of the typesafe stack would otherwise break
9116 existing code. If old code used a structure member which used to be STACK
9117 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
9118 sk_num or sk_value it would produce an error because the num, data members
9119 are not present in STACK_OF. Now it just produces a warning. sk_set
9120 replaces the old method of assigning a value to sk_value
9121 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
9122 that does this will no longer work (and should use sk_set instead) but
9123 this could be regarded as a "questionable" behaviour anyway.
9126 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
9127 correctly handle encrypted S/MIME data.
9130 *) Change type of various DES function arguments from des_cblock
9131 (which means, in function argument declarations, pointer to char)
9132 to des_cblock * (meaning pointer to array with 8 char elements),
9133 which allows the compiler to do more typechecking; it was like
9134 that back in SSLeay, but with lots of ugly casts.
9136 Introduce new type const_des_cblock.
9139 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
9140 problems: find RecipientInfo structure that matches recipient certificate
9141 and initialise the ASN1 structures properly based on passed cipher.
9144 *) Belatedly make the BN tests actually check the results.
9147 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
9148 to and from BNs: it was completely broken. New compilation option
9149 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
9150 key elements as negative integers.
9153 *) Reorganize and speed up MD5.
9154 [Andy Polyakov <appro@fy.chalmers.se>]
9157 [Richard Levitte <richard@levitte.org>]
9159 *) New option -out to asn1parse to allow the parsed structure to be
9160 output to a file. This is most useful when combined with the -strparse
9161 option to examine the output of things like OCTET STRINGS.
9164 *) Make SSL library a little more fool-proof by not requiring any longer
9165 that SSL_set_{accept,connect}_state be called before
9166 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
9167 in many applications because usually everything *appeared* to work as
9168 intended anyway -- now it really works as intended).
9171 *) Move openssl.cnf out of lib/.
9174 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
9175 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
9176 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
9177 [Ralf S. Engelschall]
9179 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
9180 handle PKCS#7 enveloped data properly.
9181 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
9183 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
9184 copying pointers. The cert_st handling is changed by this in
9185 various ways (and thus what used to be known as ctx->default_cert
9186 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
9187 any longer when s->cert does not give us what we need).
9188 ssl_cert_instantiate becomes obsolete by this change.
9189 As soon as we've got the new code right (possibly it already is?),
9190 we have solved a couple of bugs of the earlier code where s->cert
9191 was used as if it could not have been shared with other SSL structures.
9193 Note that using the SSL API in certain dirty ways now will result
9194 in different behaviour than observed with earlier library versions:
9195 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
9196 does not influence s as it used to.
9198 In order to clean up things more thoroughly, inside SSL_SESSION
9199 we don't use CERT any longer, but a new structure SESS_CERT
9200 that holds per-session data (if available); currently, this is
9201 the peer's certificate chain and, for clients, the server's certificate
9202 and temporary key. CERT holds only those values that can have
9203 meaningful defaults in an SSL_CTX.
9206 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
9207 from the internal representation. Various PKCS#7 fixes: remove some
9208 evil casts and set the enc_dig_alg field properly based on the signing
9212 *) Allow PKCS#12 password to be set from the command line or the
9213 environment. Let 'ca' get its config file name from the environment
9214 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
9218 *) Allow certificate policies extension to use an IA5STRING for the
9219 organization field. This is contrary to the PKIX definition but
9220 VeriSign uses it and IE5 only recognises this form. Document 'x509'
9224 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
9225 without disallowing inline assembler and the like for non-pedantic builds.
9228 *) Support Borland C++ builder.
9229 [Janez Jere <jj@void.si>, modified by Ulf Möller]
9234 *) SHA-1 cleanups and performance enhancements.
9235 [Andy Polyakov <appro@fy.chalmers.se>]
9237 *) Sparc v8plus assembler for the bignum library.
9238 [Andy Polyakov <appro@fy.chalmers.se>]
9240 *) Accept any -xxx and +xxx compiler options in Configure.
9243 *) Update HPUX configuration.
9246 *) Add missing sk_<type>_unshift() function to safestack.h
9247 [Ralf S. Engelschall]
9249 *) New function SSL_CTX_use_certificate_chain_file that sets the
9250 "extra_cert"s in addition to the certificate. (This makes sense
9251 only for "PEM" format files, as chains as a whole are not
9255 *) Support verify_depth from the SSL API.
9256 x509_vfy.c had what can be considered an off-by-one-error:
9257 Its depth (which was not part of the external interface)
9258 was actually counting the number of certificates in a chain;
9259 now it really counts the depth.
9262 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
9263 instead of X509err, which often resulted in confusing error
9264 messages since the error codes are not globally unique
9265 (e.g. an alleged error in ssl3_accept when a certificate
9266 didn't match the private key).
9268 *) New function SSL_CTX_set_session_id_context that allows to set a default
9269 value (so that you don't need SSL_set_session_id_context for each
9270 connection using the SSL_CTX).
9273 *) OAEP decoding bug fix.
9276 *) Support INSTALL_PREFIX for package builders, as proposed by
9280 *) New Configure options "threads" and "no-threads". For systems
9281 where the proper compiler options are known (currently Solaris
9282 and Linux), "threads" is the default.
9285 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
9288 *) Install various scripts to $(OPENSSLDIR)/misc, not to
9289 $(INSTALLTOP)/bin -- they shouldn't clutter directories
9290 such as /usr/local/bin.
9293 *) "make linux-shared" to build shared libraries.
9294 [Niels Poppe <niels@netbox.org>]
9296 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
9299 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
9300 extension adding in x509 utility.
9303 *) Remove NOPROTO sections and error code comments.
9306 *) Partial rewrite of the DEF file generator to now parse the ANSI
9310 *) New Configure options --prefix=DIR and --openssldir=DIR.
9313 *) Complete rewrite of the error code script(s). It is all now handled
9314 by one script at the top level which handles error code gathering,
9315 header rewriting and C source file generation. It should be much better
9316 than the old method: it now uses a modified version of Ulf's parser to
9317 read the ANSI prototypes in all header files (thus the old K&R definitions
9318 aren't needed for error creation any more) and do a better job of
9319 translating function codes into names. The old 'ASN1 error code imbedded
9320 in a comment' is no longer necessary and it doesn't use .err files which
9321 have now been deleted. Also the error code call doesn't have to appear all
9322 on one line (which resulted in some large lines...).
9325 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
9328 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
9329 0 (which usually indicates a closed connection), but continue reading.
9332 *) Fix some race conditions.
9335 *) Add support for CRL distribution points extension. Add Certificate
9336 Policies and CRL distribution points documentation.
9339 *) Move the autogenerated header file parts to crypto/opensslconf.h.
9342 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
9343 8 of keying material. Merlin has also confirmed interop with this fix
9344 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
9345 [Merlin Hughes <merlin@baltimore.ie>]
9347 *) Fix lots of warnings.
9348 [Richard Levitte <levitte@stacken.kth.se>]
9350 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
9351 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
9352 [Richard Levitte <levitte@stacken.kth.se>]
9354 *) Fix problems with sizeof(long) == 8.
9355 [Andy Polyakov <appro@fy.chalmers.se>]
9357 *) Change functions to ANSI C.
9360 *) Fix typos in error codes.
9361 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
9363 *) Remove defunct assembler files from Configure.
9366 *) SPARC v8 assembler BIGNUM implementation.
9367 [Andy Polyakov <appro@fy.chalmers.se>]
9369 *) Support for Certificate Policies extension: both print and set.
9370 Various additions to support the r2i method this uses.
9373 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
9374 return a const string when you are expecting an allocated buffer.
9377 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
9378 types DirectoryString and DisplayText.
9381 *) Add code to allow r2i extensions to access the configuration database,
9382 add an LHASH database driver and add several ctx helper functions.
9385 *) Fix an evil bug in bn_expand2() which caused various BN functions to
9386 fail when they extended the size of a BIGNUM.
9389 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
9390 support typesafe stack.
9393 *) Fix typo in SSL_[gs]et_options().
9394 [Nils Frostberg <nils@medcom.se>]
9396 *) Delete various functions and files that belonged to the (now obsolete)
9397 old X509V3 handling code.
9400 *) New Configure option "rsaref".
9403 *) Don't auto-generate pem.h.
9406 *) Introduce type-safe ASN.1 SETs.
9409 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
9410 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
9412 *) Introduce type-safe STACKs. This will almost certainly break lots of code
9413 that links with OpenSSL (well at least cause lots of warnings), but fear
9414 not: the conversion is trivial, and it eliminates loads of evil casts. A
9415 few STACKed things have been converted already. Feel free to convert more.
9416 In the fullness of time, I'll do away with the STACK type altogether.
9419 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
9420 specified in <certfile> by updating the entry in the index.txt file.
9421 This way one no longer has to edit the index.txt file manually for
9422 revoking a certificate. The -revoke option does the gory details now.
9423 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
9425 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
9426 `-text' option at all and this way the `-noout -text' combination was
9427 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
9428 [Ralf S. Engelschall]
9430 *) Make sure a corresponding plain text error message exists for the
9431 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
9432 verify callback function determined that a certificate was revoked.
9433 [Ralf S. Engelschall]
9435 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
9436 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
9437 all available cipers including rc5, which was forgotten until now.
9438 In order to let the testing shell script know which algorithms
9439 are available, a new (up to now undocumented) command
9440 "openssl list-cipher-commands" is used.
9443 *) Bugfix: s_client occasionally would sleep in select() when
9444 it should have checked SSL_pending() first.
9447 *) New functions DSA_do_sign and DSA_do_verify to provide access to
9448 the raw DSA values prior to ASN.1 encoding.
9451 *) Tweaks to Configure
9452 [Niels Poppe <niels@netbox.org>]
9454 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
9458 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
9461 *) New config option to avoid instructions that are illegal on the 80386.
9462 The default code is faster, but requires at least a 486.
9465 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
9466 SSL2_SERVER_VERSION (not used at all) macros, which are now the
9467 same as SSL2_VERSION anyway.
9470 *) New "-showcerts" option for s_client.
9473 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
9474 application. Various cleanups and fixes.
9477 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
9478 modify error routines to work internally. Add error codes and PBE init
9479 to library startup routines.
9482 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
9483 packing functions to asn1 and evp. Changed function names and error
9484 codes along the way.
9487 *) PKCS12 integration: and so it begins... First of several patches to
9488 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
9489 objects to objects.h
9492 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
9493 and display support for Thawte strong extranet extension.
9496 *) Add LinuxPPC support.
9497 [Jeff Dubrule <igor@pobox.org>]
9499 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
9500 bn_div_words in alpha.s.
9501 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
9503 *) Make sure the RSA OAEP test is skipped under -DRSAref because
9504 OAEP isn't supported when OpenSSL is built with RSAref.
9505 [Ulf Moeller <ulf@fitug.de>]
9507 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
9508 so they no longer are missing under -DNOPROTO.
9509 [Soren S. Jorvang <soren@t.dk>]
9512 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
9514 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
9515 doesn't work when the session is reused. Coming soon!
9518 *) Fix a security hole, that allows sessions to be reused in the wrong
9519 context thus bypassing client cert protection! All software that uses
9520 client certs and session caches in multiple contexts NEEDS PATCHING to
9521 allow session reuse! A fuller solution is in the works.
9522 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
9524 *) Some more source tree cleanups (removed obsolete files
9525 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
9526 permission on "config" script to be executable) and a fix for the INSTALL
9528 [Ulf Moeller <ulf@fitug.de>]
9530 *) Remove some legacy and erroneous uses of malloc, free instead of
9532 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
9534 *) Make rsa_oaep_test return non-zero on error.
9535 [Ulf Moeller <ulf@fitug.de>]
9537 *) Add support for native Solaris shared libraries. Configure
9538 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
9539 if someone would make that last step automatic.
9540 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
9542 *) ctx_size was not built with the right compiler during "make links". Fixed.
9545 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
9546 except NULL ciphers". This means the default cipher list will no longer
9547 enable NULL ciphers. They need to be specifically enabled e.g. with
9548 the string "DEFAULT:eNULL".
9551 *) Fix to RSA private encryption routines: if p < q then it would
9552 occasionally produce an invalid result. This will only happen with
9553 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9556 *) Be less restrictive and allow also `perl util/perlpath.pl
9557 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9558 because this way one can also use an interpreter named `perl5' (which is
9559 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9560 installed as `perl').
9561 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9563 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9564 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9566 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9567 advapi32.lib to Win32 build and change the pem test comparision
9568 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9569 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9570 and crypto/des/ede_cbcm_enc.c.
9573 *) DES quad checksum was broken on big-endian architectures. Fixed.
9576 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9577 Win32 test batch file so it (might) work again. The Win32 test batch file
9578 is horrible: I feel ill....
9581 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
9582 in e_os.h. Audit of header files to check ANSI and non ANSI
9583 sections: 10 functions were absent from non ANSI section and not exported
9584 from Windows DLLs. Fixed up libeay.num for new functions.
9587 *) Make `openssl version' output lines consistent.
9588 [Ralf S. Engelschall]
9590 *) Fix Win32 symbol export lists for BIO functions: Added
9591 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
9592 to ms/libeay{16,32}.def.
9593 [Ralf S. Engelschall]
9595 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
9596 fine under Unix and passes some trivial tests I've now added. But the
9597 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
9598 added to make sure no one expects that this stuff really works in the
9599 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
9600 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
9602 [Ralf S. Engelschall]
9604 *) Fix the generation of two part addresses in perl.
9605 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
9607 *) Add config entry for Linux on MIPS.
9608 [John Tobey <jtobey@channel1.com>]
9610 *) Make links whenever Configure is run, unless we are on Windoze.
9613 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
9614 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
9618 *) Add a useful kludge to allow package maintainers to specify compiler and
9619 other platforms details on the command line without having to patch the
9620 Configure script everytime: One now can use ``perl Configure
9621 <id>:<details>'', i.e. platform ids are allowed to have details appended
9622 to them (seperated by colons). This is treated as there would be a static
9623 pre-configured entry in Configure's %table under key <id> with value
9624 <details> and ``perl Configure <id>'' is called. So, when you want to
9625 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
9626 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
9627 now, which overrides the FreeBSD-elf entry on-the-fly.
9628 [Ralf S. Engelschall]
9630 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
9633 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
9634 on the `perl Configure ...' command line. This way one can compile
9635 OpenSSL libraries with Position Independent Code (PIC) which is needed
9636 for linking it into DSOs.
9637 [Ralf S. Engelschall]
9639 *) Remarkably, export ciphers were totally broken and no-one had noticed!
9643 *) Cleaned up the LICENSE document: The official contact for any license
9644 questions now is the OpenSSL core team under openssl-core@openssl.org.
9645 And add a paragraph about the dual-license situation to make sure people
9646 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
9647 to the OpenSSL toolkit.
9648 [Ralf S. Engelschall]
9650 *) General source tree makefile cleanups: Made `making xxx in yyy...'
9651 display consistent in the source tree and replaced `/bin/rm' by `rm'.
9652 Additonally cleaned up the `make links' target: Remove unnecessary
9653 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
9654 to speed processing and no longer clutter the display with confusing
9655 stuff. Instead only the actually done links are displayed.
9656 [Ralf S. Engelschall]
9658 *) Permit null encryption ciphersuites, used for authentication only. It used
9659 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
9660 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
9664 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
9665 signed attributes when verifying signatures (this would break them),
9666 the detached data encoding was wrong and public keys obtained using
9667 X509_get_pubkey() weren't freed.
9670 *) Add text documentation for the BUFFER functions. Also added a work around
9671 to a Win95 console bug. This was triggered by the password read stuff: the
9672 last character typed gets carried over to the next fread(). If you were
9673 generating a new cert request using 'req' for example then the last
9674 character of the passphrase would be CR which would then enter the first
9678 *) Added the new `Includes OpenSSL Cryptography Software' button as
9679 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
9680 button and can be used by applications based on OpenSSL to show the
9681 relationship to the OpenSSL project.
9682 [Ralf S. Engelschall]
9684 *) Remove confusing variables in function signatures in files
9685 ssl/ssl_lib.c and ssl/ssl.h.
9686 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9688 *) Don't install bss_file.c under PREFIX/include/
9689 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9691 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
9692 functions that return function pointers and has support for NT specific
9693 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
9694 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
9695 unsigned to signed types: this was killing the Win32 compile.
9698 *) Add new certificate file to stack functions,
9699 SSL_add_dir_cert_subjects_to_stack() and
9700 SSL_add_file_cert_subjects_to_stack(). These largely supplant
9701 SSL_load_client_CA_file(), and can be used to add multiple certs easily
9702 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
9703 This means that Apache-SSL and similar packages don't have to mess around
9704 to add as many CAs as they want to the preferred list.
9707 *) Experiment with doxygen documentation. Currently only partially applied to
9709 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
9710 openssl.doxy as the configuration file.
9713 *) Get rid of remaining C++-style comments which strict C compilers hate.
9714 [Ralf S. Engelschall, pointed out by Carlos Amengual]
9716 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
9717 compiled in by default: it has problems with large keys.
9720 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
9721 DH private keys and/or callback functions which directly correspond to
9722 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
9723 is needed for applications which have to configure certificates on a
9724 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
9726 For the RSA certificate situation is makes no difference, but
9727 for the DSA certificate situation this fixes the "no shared cipher"
9728 problem where the OpenSSL cipher selection procedure failed because the
9729 temporary keys were not overtaken from the context and the API provided
9730 no way to reconfigure them.
9731 The new functions now let applications reconfigure the stuff and they
9732 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
9733 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
9734 non-public-API function ssl_cert_instantiate() is used as a helper
9735 function and also to reduce code redundancy inside ssl_rsa.c.
9736 [Ralf S. Engelschall]
9738 *) Move s_server -dcert and -dkey options out of the undocumented feature
9739 area because they are useful for the DSA situation and should be
9740 recognized by the users.
9741 [Ralf S. Engelschall]
9743 *) Fix the cipher decision scheme for export ciphers: the export bits are
9744 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
9745 SSL_EXP_MASK. So, the original variable has to be used instead of the
9746 already masked variable.
9747 [Richard Levitte <levitte@stacken.kth.se>]
9749 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
9750 [Richard Levitte <levitte@stacken.kth.se>]
9752 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
9753 from `int' to `unsigned int' because it's a length and initialized by
9754 EVP_DigestFinal() which expects an `unsigned int *'.
9755 [Richard Levitte <levitte@stacken.kth.se>]
9757 *) Don't hard-code path to Perl interpreter on shebang line of Configure
9758 script. Instead use the usual Shell->Perl transition trick.
9759 [Ralf S. Engelschall]
9761 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
9762 (in addition to RSA certificates) to match the behaviour of `openssl dsa
9763 -noout -modulus' as it's already the case for `openssl rsa -noout
9764 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
9765 currently the public key is printed (a decision which was already done by
9766 `openssl dsa -modulus' in the past) which serves a similar purpose.
9767 Additionally the NO_RSA no longer completely removes the whole -modulus
9768 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
9770 [Ralf S. Engelschall]
9772 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
9773 BIO. See the source (crypto/evp/bio_ok.c) for more info.
9774 [Arne Ansper <arne@ats.cyber.ee>]
9776 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
9777 to be added. Now both 'req' and 'ca' can use new objects defined in the
9781 *) Add cool BIO that does syslog (or event log on NT).
9782 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
9784 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
9785 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
9786 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
9787 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
9790 *) Add preliminary config info for new extension code.
9793 *) Make RSA_NO_PADDING really use no padding.
9794 [Ulf Moeller <ulf@fitug.de>]
9796 *) Generate errors when private/public key check is done.
9799 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
9800 for some CRL extensions and new objects added.
9803 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
9804 key usage extension and fuller support for authority key id.
9807 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
9808 padding method for RSA, which is recommended for new applications in PKCS
9809 #1 v2.0 (RFC 2437, October 1998).
9810 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
9811 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
9812 against Bleichbacher's attack on RSA.
9813 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
9816 *) Updates to the new SSL compression code
9817 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9819 *) Fix so that the version number in the master secret, when passed
9820 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
9821 (because the server will not accept higher), that the version number
9822 is 0x03,0x01, not 0x03,0x00
9823 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9825 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
9826 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
9827 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
9830 *) Support for RAW extensions where an arbitrary extension can be
9831 created by including its DER encoding. See apps/openssl.cnf for
9835 *) Make sure latest Perl versions don't interpret some generated C array
9836 code as Perl array code in the crypto/err/err_genc.pl script.
9837 [Lars Weber <3weber@informatik.uni-hamburg.de>]
9839 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
9840 not many people have the assembler. Various Win32 compilation fixes and
9841 update to the INSTALL.W32 file with (hopefully) more accurate Win32
9845 *) Modify configure script 'Configure' to automatically create crypto/date.h
9846 file under Win32 and also build pem.h from pem.org. New script
9847 util/mkfiles.pl to create the MINFO file on environments that can't do a
9848 'make files': perl util/mkfiles.pl >MINFO should work.
9851 *) Major rework of DES function declarations, in the pursuit of correctness
9852 and purity. As a result, many evil casts evaporated, and some weirdness,
9853 too. You may find this causes warnings in your code. Zapping your evil
9854 casts will probably fix them. Mostly.
9857 *) Fix for a typo in asn1.h. Bug fix to object creation script
9858 obj_dat.pl. It considered a zero in an object definition to mean
9859 "end of object": none of the objects in objects.h have any zeros
9860 so it wasn't spotted.
9861 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
9863 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
9864 Masking (CBCM). In the absence of test vectors, the best I have been able
9865 to do is check that the decrypt undoes the encrypt, so far. Send me test
9866 vectors if you have them.
9869 *) Correct calculation of key length for export ciphers (too much space was
9870 allocated for null ciphers). This has not been tested!
9873 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
9874 message is now correct (it understands "crypto" and "ssl" on its
9875 command line). There is also now an "update" option. This will update
9876 the util/ssleay.num and util/libeay.num files with any new functions.
9878 perl util/mkdef.pl crypto ssl update
9879 it will update them.
9882 *) Overhauled the Perl interface (perl/*):
9883 - ported BN stuff to OpenSSL's different BN library
9884 - made the perl/ source tree CVS-aware
9885 - renamed the package from SSLeay to OpenSSL (the files still contain
9886 their history because I've copied them in the repository)
9887 - removed obsolete files (the test scripts will be replaced
9888 by better Test::Harness variants in the future)
9889 [Ralf S. Engelschall]
9891 *) First cut for a very conservative source tree cleanup:
9892 1. merge various obsolete readme texts into doc/ssleay.txt
9893 where we collect the old documents and readme texts.
9894 2. remove the first part of files where I'm already sure that we no
9895 longer need them because of three reasons: either they are just temporary
9896 files which were left by Eric or they are preserved original files where
9897 I've verified that the diff is also available in the CVS via "cvs diff
9898 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
9899 the crypto/md/ stuff).
9900 [Ralf S. Engelschall]
9902 *) More extension code. Incomplete support for subject and issuer alt
9903 name, issuer and authority key id. Change the i2v function parameters
9904 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
9905 what that's for :-) Fix to ASN1 macro which messed up
9906 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
9909 *) Preliminary support for ENUMERATED type. This is largely copied from the
9913 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
9914 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9916 *) Make sure `make rehash' target really finds the `openssl' program.
9917 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9919 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
9920 like to hear about it if this slows down other processors.
9923 *) Add CygWin32 platform information to Configure script.
9924 [Alan Batie <batie@aahz.jf.intel.com>]
9926 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
9927 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
9929 *) New program nseq to manipulate netscape certificate sequences
9932 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
9936 *) Fixes to BN code. Previously the default was to define BN_RECURSION
9937 but the BN code had some problems that would cause failures when
9938 doing certificate verification and some other functions.
9939 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9941 *) Add ASN1 and PEM code to support netscape certificate sequences.
9944 *) Add ASN1 and PEM code to support netscape certificate sequences.
9947 *) Add several PKIX and private extended key usage OIDs.
9950 *) Modify the 'ca' program to handle the new extension code. Modify
9951 openssl.cnf for new extension format, add comments.
9954 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
9955 and add a sample to openssl.cnf so req -x509 now adds appropriate
9959 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
9960 error code, add initial support to X509_print() and x509 application.
9963 *) Takes a deep breath and start addding X509 V3 extension support code. Add
9964 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
9965 stuff is currently isolated and isn't even compiled yet.
9968 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
9969 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
9970 Removed the versions check from X509 routines when loading extensions:
9971 this allows certain broken certificates that don't set the version
9972 properly to be processed.
9975 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
9976 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
9977 can still be regenerated with "make depend".
9980 *) Spelling mistake in C version of CAST-128.
9981 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
9983 *) Changes to the error generation code. The perl script err-code.pl
9984 now reads in the old error codes and retains the old numbers, only
9985 adding new ones if necessary. It also only changes the .err files if new
9986 codes are added. The makefiles have been modified to only insert errors
9987 when needed (to avoid needlessly modifying header files). This is done
9988 by only inserting errors if the .err file is newer than the auto generated
9989 C file. To rebuild all the error codes from scratch (the old behaviour)
9990 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
9991 or delete all the .err files.
9994 *) CAST-128 was incorrectly implemented for short keys. The C version has
9995 been fixed, but is untested. The assembler versions are also fixed, but
9996 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
9997 to regenerate it if needed.
9998 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
9999 Hagino <itojun@kame.net>]
10001 *) File was opened incorrectly in randfile.c.
10002 [Ulf Möller <ulf@fitug.de>]
10004 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
10005 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
10006 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
10007 al: it's just almost always a UTCTime. Note this patch adds new error
10008 codes so do a "make errors" if there are problems.
10011 *) Correct Linux 1 recognition in config.
10012 [Ulf Möller <ulf@fitug.de>]
10014 *) Remove pointless MD5 hash when using DSA keys in ca.
10015 [Anonymous <nobody@replay.com>]
10017 *) Generate an error if given an empty string as a cert directory. Also
10018 generate an error if handed NULL (previously returned 0 to indicate an
10019 error, but didn't set one).
10020 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
10022 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
10025 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
10026 parameters. This was causing a warning which killed off the Win32 compile.
10029 *) Remove C++ style comments from crypto/bn/bn_local.h.
10030 [Neil Costigan <neil.costigan@celocom.com>]
10032 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
10033 based on a text string, looking up short and long names and finally
10034 "dot" format. The "dot" format stuff didn't work. Added new function
10035 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
10036 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
10037 OID is not part of the table.
10040 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
10041 X509_LOOKUP_by_alias().
10044 *) Sort openssl functions by name.
10047 *) Get the gendsa program working (hopefully) and add it to app list. Remove
10048 encryption from sample DSA keys (in case anyone is interested the password
10052 *) Make _all_ *_free functions accept a NULL pointer.
10053 [Frans Heymans <fheymans@isaserver.be>]
10055 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
10057 [Anonymous <nobody@replay.com>]
10059 *) s_server should send the CAfile as acceptable CAs, not its own cert.
10060 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10062 *) Don't blow it for numeric -newkey arguments to apps/req.
10063 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10065 *) Temp key "for export" tests were wrong in s3_srvr.c.
10066 [Anonymous <nobody@replay.com>]
10068 *) Add prototype for temp key callback functions
10069 SSL_CTX_set_tmp_{rsa,dh}_callback().
10072 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
10073 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
10076 *) X509_name_add_entry() freed the wrong thing after an error.
10077 [Arne Ansper <arne@ats.cyber.ee>]
10079 *) rsa_eay.c would attempt to free a NULL context.
10080 [Arne Ansper <arne@ats.cyber.ee>]
10082 *) BIO_s_socket() had a broken should_retry() on Windoze.
10083 [Arne Ansper <arne@ats.cyber.ee>]
10085 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
10086 [Arne Ansper <arne@ats.cyber.ee>]
10088 *) Make sure the already existing X509_STORE->depth variable is initialized
10089 in X509_STORE_new(), but document the fact that this variable is still
10090 unused in the certificate verification process.
10091 [Ralf S. Engelschall]
10093 *) Fix the various library and apps files to free up pkeys obtained from
10094 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
10097 *) Fix reference counting in X509_PUBKEY_get(). This makes
10098 demos/maurice/example2.c work, amongst others, probably.
10099 [Steve Henson and Ben Laurie]
10101 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
10102 `openssl' and second, the shortcut symlinks for the `openssl <command>'
10103 are no longer created. This way we have a single and consistent command
10104 line interface `openssl <command>', similar to `cvs <command>'.
10105 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
10107 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
10108 BIT STRING wrapper always have zero unused bits.
10111 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
10114 *) Make the top-level INSTALL documentation easier to understand.
10117 *) Makefiles updated to exit if an error occurs in a sub-directory
10118 make (including if user presses ^C) [Paul Sutton]
10120 *) Make Montgomery context stuff explicit in RSA data structure.
10123 *) Fix build order of pem and err to allow for generated pem.h.
10126 *) Fix renumbering bug in X509_NAME_delete_entry().
10129 *) Enhanced the err-ins.pl script so it makes the error library number
10130 global and can add a library name. This is needed for external ASN1 and
10131 other error libraries.
10134 *) Fixed sk_insert which never worked properly.
10137 *) Fix ASN1 macros so they can handle indefinite length construted
10138 EXPLICIT tags. Some non standard certificates use these: they can now
10142 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
10143 into a single doc/ssleay.txt bundle. This way the information is still
10144 preserved but no longer messes up this directory. Now it's new room for
10145 the new set of documenation files.
10146 [Ralf S. Engelschall]
10148 *) SETs were incorrectly DER encoded. This was a major pain, because they
10149 shared code with SEQUENCEs, which aren't coded the same. This means that
10150 almost everything to do with SETs or SEQUENCEs has either changed name or
10151 number of arguments.
10152 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
10154 *) Fix test data to work with the above.
10157 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
10158 was already fixed by Eric for 0.9.1 it seems.
10159 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
10161 *) Autodetect FreeBSD3.
10164 *) Fix various bugs in Configure. This affects the following platforms:
10168 unixware-2.0-pentium
10172 *) Eliminate generated files from CVS. Reorder tests to regenerate files
10173 before they are needed.
10176 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
10180 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
10182 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
10183 changed SSLeay to OpenSSL in version strings.
10184 [Ralf S. Engelschall]
10186 *) Some fixups to the top-level documents.
10189 *) Fixed the nasty bug where rsaref.h was not found under compile-time
10190 because the symlink to include/ was missing.
10191 [Ralf S. Engelschall]
10193 *) Incorporated the popular no-RSA/DSA-only patches
10194 which allow to compile a RSA-free SSLeay.
10195 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
10197 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
10198 when "ssleay" is still not found.
10199 [Ralf S. Engelschall]
10201 *) Added more platforms to Configure: Cray T3E, HPUX 11,
10202 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
10204 *) Updated the README file.
10205 [Ralf S. Engelschall]
10207 *) Added various .cvsignore files in the CVS repository subdirs
10208 to make a "cvs update" really silent.
10209 [Ralf S. Engelschall]
10211 *) Recompiled the error-definition header files and added
10212 missing symbols to the Win32 linker tables.
10213 [Ralf S. Engelschall]
10215 *) Cleaned up the top-level documents;
10216 o new files: CHANGES and LICENSE
10217 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
10218 o merged COPYRIGHT into LICENSE
10219 o removed obsolete TODO file
10220 o renamed MICROSOFT to INSTALL.W32
10221 [Ralf S. Engelschall]
10223 *) Removed dummy files from the 0.9.1b source tree:
10224 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
10225 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
10226 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
10227 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
10228 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
10229 [Ralf S. Engelschall]
10231 *) Added various platform portability fixes.
10234 *) The Genesis of the OpenSSL rpject:
10235 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
10236 Young and Tim J. Hudson created while they were working for C2Net until
10238 [The OpenSSL Project]
10241 Changes between 0.9.0b and 0.9.1b [not released]
10243 *) Updated a few CA certificates under certs/
10246 *) Changed some BIGNUM api stuff.
10249 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
10250 DGUX x86, Linux Alpha, etc.
10253 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
10254 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
10258 *) Add -strparse option to asn1pars program which parses nested
10260 [Dr Stephen Henson <shenson@bigfoot.com>]
10262 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
10265 *) DSA fix for "ca" program.
10268 *) Added "-genkey" option to "dsaparam" program.
10271 *) Added RIPE MD160 (rmd160) message digest.
10274 *) Added -a (all) option to "ssleay version" command.
10277 *) Added PLATFORM define which is the id given to Configure.
10280 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
10283 *) Extended the ASN.1 parser routines.
10286 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
10289 *) Added a BN_CTX to the BN library.
10292 *) Fixed the weak key values in DES library
10295 *) Changed API in EVP library for cipher aliases.
10298 *) Added support for RC2/64bit cipher.
10301 *) Converted the lhash library to the crypto/mem.c functions.
10304 *) Added more recognized ASN.1 object ids.
10307 *) Added more RSA padding checks for SSL/TLS.
10310 *) Added BIO proxy/filter functionality.
10313 *) Added extra_certs to SSL_CTX which can be used
10314 send extra CA certificates to the client in the CA cert chain sending
10315 process. It can be configured with SSL_CTX_add_extra_chain_cert().
10318 *) Now Fortezza is denied in the authentication phase because
10319 this is key exchange mechanism is not supported by SSLeay at all.
10322 *) Additional PKCS1 checks.
10325 *) Support the string "TLSv1" for all TLS v1 ciphers.
10328 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
10329 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
10332 *) Fixed a few memory leaks.
10335 *) Fixed various code and comment typos.
10338 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
10339 bytes sent in the client random.
10340 [Edward Bishop <ebishop@spyglass.com>]