5 Changes between 1.0.1m and 1.0.1n [xx XXX xxxx]
7 *) Reject DH handshakes with parameters shorter than 768 bits.
8 [Kurt Roeckx and Emilia Kasper]
10 Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
12 *) Segmentation fault in ASN1_TYPE_cmp fix
14 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
15 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
16 certificate signature algorithm consistency this can be used to crash any
17 certificate verification operation and exploited in a DoS attack. Any
18 application which performs certificate verification is vulnerable including
19 OpenSSL clients and servers which enable client authentication.
23 *) ASN.1 structure reuse memory corruption fix
25 Reusing a structure in ASN.1 parsing may allow an attacker to cause
26 memory corruption via an invalid write. Such reuse is and has been
27 strongly discouraged and is believed to be rare.
29 Applications that parse structures containing CHOICE or ANY DEFINED BY
30 components may be affected. Certificate parsing (d2i_X509 and related
31 functions) are however not affected. OpenSSL clients and servers are
36 *) PKCS7 NULL pointer dereferences fix
38 The PKCS#7 parsing code does not handle missing outer ContentInfo
39 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
40 missing content and trigger a NULL pointer dereference on parsing.
42 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
43 otherwise parse PKCS#7 structures from untrusted sources are
44 affected. OpenSSL clients and servers are not affected.
46 This issue was reported to OpenSSL by Michal Zalewski (Google).
50 *) DoS via reachable assert in SSLv2 servers fix
52 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
53 servers that both support SSLv2 and enable export cipher suites by sending
54 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
56 This issue was discovered by Sean Burford (Google) and Emilia Käsper
57 (OpenSSL development team).
61 *) Use After Free following d2i_ECPrivatekey error fix
63 A malformed EC private key file consumed via the d2i_ECPrivateKey function
64 could cause a use after free condition. This, in turn, could cause a double
65 free in several private key parsing functions (such as d2i_PrivateKey
66 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
67 for applications that receive EC private keys from untrusted
68 sources. This scenario is considered rare.
70 This issue was discovered by the BoringSSL project and fixed in their
75 *) X509_to_X509_REQ NULL pointer deref fix
77 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
78 the certificate key is invalid. This function is rarely used in practice.
80 This issue was discovered by Brian Carpenter.
84 *) Removed the export ciphers from the DEFAULT ciphers
87 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
89 *) Build fixes for the Windows and OpenVMS platforms
90 [Matt Caswell and Richard Levitte]
92 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
94 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
95 message can cause a segmentation fault in OpenSSL due to a NULL pointer
96 dereference. This could lead to a Denial Of Service attack. Thanks to
97 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
101 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
102 dtls1_buffer_record function under certain conditions. In particular this
103 could occur if an attacker sent repeated DTLS records with the same
104 sequence number but for the next epoch. The memory leak could be exploited
105 by an attacker in a Denial of Service attack through memory exhaustion.
106 Thanks to Chris Mueller for reporting this issue.
110 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
111 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
112 method would be set to NULL which could later result in a NULL pointer
113 dereference. Thanks to Frank Schmirler for reporting this issue.
117 *) Abort handshake if server key exchange message is omitted for ephemeral
120 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
121 reporting this issue.
125 *) Remove non-export ephemeral RSA code on client and server. This code
126 violated the TLS standard by allowing the use of temporary RSA keys in
127 non-export ciphersuites and could be used by a server to effectively
128 downgrade the RSA key length used to a value smaller than the server
129 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
130 INRIA or reporting this issue.
134 *) Fixed issue where DH client certificates are accepted without verification.
135 An OpenSSL server will accept a DH certificate for client authentication
136 without the certificate verify message. This effectively allows a client to
137 authenticate without the use of a private key. This only affects servers
138 which trust a client certificate authority which issues certificates
139 containing DH keys: these are extremely rare and hardly ever encountered.
140 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
145 *) Ensure that the session ID context of an SSL is updated when its
146 SSL_CTX is updated via SSL_set_SSL_CTX.
148 The session ID context is typically set from the parent SSL_CTX,
149 and can vary with the CTX.
152 *) Fix various certificate fingerprint issues.
154 By using non-DER or invalid encodings outside the signed portion of a
155 certificate the fingerprint can be changed without breaking the signature.
156 Although no details of the signed portion of the certificate can be changed
157 this can cause problems with some applications: e.g. those using the
158 certificate fingerprint for blacklists.
160 1. Reject signatures with non zero unused bits.
162 If the BIT STRING containing the signature has non zero unused bits reject
163 the signature. All current signature algorithms require zero unused bits.
165 2. Check certificate algorithm consistency.
167 Check the AlgorithmIdentifier inside TBS matches the one in the
168 certificate signature. NB: this will result in signature failure
169 errors for some broken certificates.
171 Thanks to Konrad Kraszewski from Google for reporting this issue.
173 3. Check DSA/ECDSA signatures use DER.
175 Reencode DSA/ECDSA signatures and compare with the original received
176 signature. Return an error if there is a mismatch.
178 This will reject various cases including garbage after signature
179 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
180 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
181 (negative or with leading zeroes).
183 Further analysis was conducted and fixes were developed by Stephen Henson
184 of the OpenSSL core team.
189 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
190 results on some platforms, including x86_64. This bug occurs at random
191 with a very low probability, and is not known to be exploitable in any
192 way, though its exact impact is difficult to determine. Thanks to Pieter
193 Wuille (Blockstream) who reported this issue and also suggested an initial
194 fix. Further analysis was conducted by the OpenSSL development team and
195 Adam Langley of Google. The final fix was developed by Andy Polyakov of
196 the OpenSSL core team.
200 *) Do not resume sessions on the server if the negotiated protocol
201 version does not match the session's version. Resuming with a different
202 version, while not strictly forbidden by the RFC, is of questionable
203 sanity and breaks all known clients.
204 [David Benjamin, Emilia Käsper]
206 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
207 early CCS messages during renegotiation. (Note that because
208 renegotiation is encrypted, this early CCS was not exploitable.)
211 *) Tighten client-side session ticket handling during renegotiation:
212 ensure that the client only accepts a session ticket if the server sends
213 the extension anew in the ServerHello. Previously, a TLS client would
214 reuse the old extension state and thus accept a session ticket if one was
215 announced in the initial ServerHello.
217 Similarly, ensure that the client requires a session ticket if one
218 was advertised in the ServerHello. Previously, a TLS client would
219 ignore a missing NewSessionTicket message.
222 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
226 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
227 sends a carefully crafted handshake message, to cause OpenSSL to fail
228 to free up to 64k of memory causing a memory leak. This could be
229 exploited in a Denial Of Service attack. This issue affects OpenSSL
230 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
231 whether SRTP is used or configured. Implementations of OpenSSL that
232 have been compiled with OPENSSL_NO_SRTP defined are not affected.
234 The fix was developed by the OpenSSL team.
238 *) Session Ticket Memory Leak.
240 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
241 integrity of that ticket is first verified. In the event of a session
242 ticket integrity check failing, OpenSSL will fail to free memory
243 causing a memory leak. By sending a large number of invalid session
244 tickets an attacker could exploit this issue in a Denial Of Service
249 *) Build option no-ssl3 is incomplete.
251 When OpenSSL is configured with "no-ssl3" as a build option, servers
252 could accept and complete a SSL 3.0 handshake, and clients could be
253 configured to send them.
255 [Akamai and the OpenSSL team]
257 *) Add support for TLS_FALLBACK_SCSV.
258 Client applications doing fallback retries should call
259 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
261 [Adam Langley, Bodo Moeller]
263 *) Add additional DigestInfo checks.
265 Reencode DigestInto in DER and check against the original when
266 verifying RSA signature: this will reject any improperly encoded
267 DigestInfo structures.
269 Note: this is a precautionary measure and no attacks are currently known.
273 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
275 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
276 SRP code can be overrun an internal buffer. Add sanity check that
277 g, A, B < N to SRP code.
279 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
280 Group for discovering this issue.
284 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
285 TLS 1.0 instead of higher protocol versions when the ClientHello message
286 is badly fragmented. This allows a man-in-the-middle attacker to force a
287 downgrade to TLS 1.0 even if both the server and the client support a
288 higher protocol version, by modifying the client's TLS records.
290 Thanks to David Benjamin and Adam Langley (Google) for discovering and
291 researching this issue.
295 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
296 to a denial of service attack. A malicious server can crash the client
297 with a null pointer dereference (read) by specifying an anonymous (EC)DH
298 ciphersuite and sending carefully crafted handshake messages.
300 Thanks to Felix Gröbert (Google) for discovering and researching this
305 *) By sending carefully crafted DTLS packets an attacker could cause openssl
306 to leak memory. This can be exploited through a Denial of Service attack.
307 Thanks to Adam Langley for discovering and researching this issue.
311 *) An attacker can force openssl to consume large amounts of memory whilst
312 processing DTLS handshake messages. This can be exploited through a
313 Denial of Service attack.
314 Thanks to Adam Langley for discovering and researching this issue.
318 *) An attacker can force an error condition which causes openssl to crash
319 whilst processing DTLS packets due to memory being freed twice. This
320 can be exploited through a Denial of Service attack.
321 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
326 *) If a multithreaded client connects to a malicious server using a resumed
327 session and the server sends an ec point format extension it could write
328 up to 255 bytes to freed memory.
330 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
335 *) A malicious server can crash an OpenSSL client with a null pointer
336 dereference (read) by specifying an SRP ciphersuite even though it was not
337 properly negotiated with the client. This can be exploited through a
338 Denial of Service attack.
340 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
341 discovering and researching this issue.
345 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
346 X509_name_oneline, X509_name_print_ex et al. to leak some information
347 from the stack. Applications may be affected if they echo pretty printing
348 output to the attacker.
350 Thanks to Ivan Fratric (Google) for discovering this issue.
352 [Emilia Käsper, and Steve Henson]
354 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
355 for corner cases. (Certain input points at infinity could lead to
356 bogus results, with non-infinity inputs mapped to infinity too.)
359 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
361 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
362 handshake can force the use of weak keying material in OpenSSL
363 SSL/TLS clients and servers.
365 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
366 researching this issue. (CVE-2014-0224)
367 [KIKUCHI Masashi, Steve Henson]
369 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
370 OpenSSL DTLS client the code can be made to recurse eventually crashing
373 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
375 [Imre Rad, Steve Henson]
377 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
378 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
379 client or server. This is potentially exploitable to run arbitrary
380 code on a vulnerable client or server.
382 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
383 [Jüri Aedla, Steve Henson]
385 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
386 are subject to a denial of service attack.
388 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
389 this issue. (CVE-2014-3470)
390 [Felix Gröbert, Ivan Fratric, Steve Henson]
392 *) Harmonize version and its documentation. -f flag is used to display
394 [mancha <mancha1@zoho.com>]
396 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
398 [mancha <mancha1@zoho.com>]
400 *) Fix some double frees. These are not thought to be exploitable.
401 [mancha <mancha1@zoho.com>]
403 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
405 *) A missing bounds check in the handling of the TLS heartbeat extension
406 can be used to reveal up to 64k of memory to a connected client or
409 Thanks for Neel Mehta of Google Security for discovering this bug and to
410 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
411 preparing the fix (CVE-2014-0160)
412 [Adam Langley, Bodo Moeller]
414 *) Fix for the attack described in the paper "Recovering OpenSSL
415 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
416 by Yuval Yarom and Naomi Benger. Details can be obtained from:
417 http://eprint.iacr.org/2014/140
419 Thanks to Yuval Yarom and Naomi Benger for discovering this
420 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
421 [Yuval Yarom and Naomi Benger]
423 *) TLS pad extension: draft-agl-tls-padding-03
425 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
426 TLS client Hello record length value would otherwise be > 255 and
427 less that 512 pad with a dummy extension containing zeroes so it
428 is at least 512 bytes long.
430 [Adam Langley, Steve Henson]
432 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
434 *) Fix for TLS record tampering bug. A carefully crafted invalid
435 handshake could crash OpenSSL with a NULL pointer exception.
436 Thanks to Anton Johansson for reporting this issues.
439 *) Keep original DTLS digest and encryption contexts in retransmission
440 structures so we can use the previous session parameters if they need
441 to be resent. (CVE-2013-6450)
444 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
445 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
446 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
447 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
448 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
449 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
450 [Rob Stradling, Adam Langley]
452 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
454 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
455 supporting platforms or when small records were transferred.
456 [Andy Polyakov, Steve Henson]
458 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
460 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
462 This addresses the flaw in CBC record processing discovered by
463 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
464 at: http://www.isg.rhul.ac.uk/tls/
466 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
467 Security Group at Royal Holloway, University of London
468 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
469 Emilia Käsper for the initial patch.
471 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
473 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
474 ciphersuites which can be exploited in a denial of service attack.
475 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
476 and detecting this bug and to Wolfgang Ettlinger
477 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
481 *) Return an error when checking OCSP signatures when key is NULL.
482 This fixes a DoS attack. (CVE-2013-0166)
485 *) Make openssl verify return errors.
486 [Chris Palmer <palmer@google.com> and Ben Laurie]
488 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
489 the right response is stapled. Also change SSL_get_certificate()
490 so it returns the certificate actually sent.
491 See http://rt.openssl.org/Ticket/Display.html?id=2836.
492 [Rob Stradling <rob.stradling@comodo.com>]
494 *) Fix possible deadlock when decoding public keys.
497 *) Don't use TLS 1.0 record version number in initial client hello
501 Changes between 1.0.1b and 1.0.1c [10 May 2012]
503 *) Sanity check record length before skipping explicit IV in TLS
504 1.2, 1.1 and DTLS to fix DoS attack.
506 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
507 fuzzing as a service testing platform.
511 *) Initialise tkeylen properly when encrypting CMS messages.
512 Thanks to Solar Designer of Openwall for reporting this issue.
515 *) In FIPS mode don't try to use composite ciphers as they are not
519 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
521 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
522 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
523 mean any application compiled against OpenSSL 1.0.0 headers setting
524 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
525 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
526 0x10000000L Any application which was previously compiled against
527 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
528 will need to be recompiled as a result. Letting be results in
529 inability to disable specifically TLS 1.1 and in client context,
530 in unlike event, limit maximum offered version to TLS 1.0 [see below].
533 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
534 disable just protocol X, but all protocols above X *if* there are
535 protocols *below* X still enabled. In more practical terms it means
536 that if application wants to disable TLS1.0 in favor of TLS1.1 and
537 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
538 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
542 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
544 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
545 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
546 in CRYPTO_realloc_clean.
548 Thanks to Tavis Ormandy, Google Security Team, for discovering this
549 issue and to Adam Langley <agl@chromium.org> for fixing it.
551 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
553 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
556 *) Workarounds for some broken servers that "hang" if a client hello
557 record length exceeds 255 bytes.
559 1. Do not use record version number > TLS 1.0 in initial client
560 hello: some (but not all) hanging servers will now work.
561 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
562 the number of ciphers sent in the client hello. This should be
563 set to an even number, such as 50, for example by passing:
564 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
565 Most broken servers should now work.
566 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
567 TLS 1.2 client support entirely.
570 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
573 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
575 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
576 STRING form instead of a DigestInfo.
579 *) The format used for MDC2 RSA signatures is inconsistent between EVP
580 and the RSA_sign/RSA_verify functions. This was made more apparent when
581 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
582 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
583 the correct format in RSA_verify so both forms transparently work.
586 *) Some servers which support TLS 1.0 can choke if we initially indicate
587 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
588 encrypted premaster secret. As a workaround use the maximum pemitted
589 client version in client hello, this should keep such servers happy
590 and still work with previous versions of OpenSSL.
593 *) Add support for TLS/DTLS heartbeats.
594 [Robin Seggelmann <seggelmann@fh-muenster.de>]
596 *) Add support for SCTP.
597 [Robin Seggelmann <seggelmann@fh-muenster.de>]
599 *) Improved PRNG seeding for VOS.
600 [Paul Green <Paul.Green@stratus.com>]
602 *) Extensive assembler packs updates, most notably:
604 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
605 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
606 - x86_64: bit-sliced AES implementation;
607 - ARM: NEON support, contemporary platforms optimizations;
608 - s390x: z196 support;
609 - *: GHASH and GF(2^m) multiplication implementations;
613 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
614 (removal of unnecessary code)
615 [Peter Sylvester <peter.sylvester@edelweb.fr>]
617 *) Add TLS key material exporter from RFC 5705.
620 *) Add DTLS-SRTP negotiation from RFC 5764.
623 *) Add Next Protocol Negotiation,
624 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
625 disabled with a no-npn flag to config or Configure. Code donated
627 [Adam Langley <agl@google.com> and Ben Laurie]
629 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
630 NIST-P256, NIST-P521, with constant-time single point multiplication on
631 typical inputs. Compiler support for the nonstandard type __uint128_t is
632 required to use this (present in gcc 4.4 and later, for 64-bit builds).
633 Code made available under Apache License version 2.0.
635 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
636 line to include this in your build of OpenSSL, and run "make depend" (or
637 "make update"). This enables the following EC_METHODs:
639 EC_GFp_nistp224_method()
640 EC_GFp_nistp256_method()
641 EC_GFp_nistp521_method()
643 EC_GROUP_new_by_curve_name() will automatically use these (while
644 EC_GROUP_new_curve_GFp() currently prefers the more flexible
646 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
648 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
649 all platforms. Move ssize_t definition from e_os.h to the public
650 header file e_os2.h as it now appears in public header file cms.h
653 *) New -sigopt option to the ca, req and x509 utilities. Additional
654 signature parameters can be passed using this option and in
658 *) Add RSA PSS signing function. This will generate and set the
659 appropriate AlgorithmIdentifiers for PSS based on those in the
660 corresponding EVP_MD_CTX structure. No application support yet.
663 *) Support for companion algorithm specific ASN1 signing routines.
664 New function ASN1_item_sign_ctx() signs a pre-initialised
665 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
666 the appropriate parameters.
669 *) Add new algorithm specific ASN1 verification initialisation function
670 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
671 handling will be the same no matter what EVP_PKEY_METHOD is used.
672 Add a PSS handler to support verification of PSS signatures: checked
673 against a number of sample certificates.
676 *) Add signature printing for PSS. Add PSS OIDs.
677 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
679 *) Add algorithm specific signature printing. An individual ASN1 method
680 can now print out signatures instead of the standard hex dump.
682 More complex signatures (e.g. PSS) can print out more meaningful
683 information. Include DSA version that prints out the signature
687 *) Password based recipient info support for CMS library: implementing
691 *) Split password based encryption into PBES2 and PBKDF2 functions. This
692 neatly separates the code into cipher and PBE sections and is required
693 for some algorithms that split PBES2 into separate pieces (such as
697 *) Session-handling fixes:
698 - Fix handling of connections that are resuming with a session ID,
699 but also support Session Tickets.
700 - Fix a bug that suppressed issuing of a new ticket if the client
701 presented a ticket with an expired session.
702 - Try to set the ticket lifetime hint to something reasonable.
703 - Make tickets shorter by excluding irrelevant information.
704 - On the client side, don't ignore renewed tickets.
705 [Adam Langley, Bodo Moeller (Google)]
707 *) Fix PSK session representation.
710 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
712 This work was sponsored by Intel.
715 *) Add GCM support to TLS library. Some custom code is needed to split
716 the IV between the fixed (from PRF) and explicit (from TLS record)
717 portions. This adds all GCM ciphersuites supported by RFC5288 and
718 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
719 add a special AESGCM string for GCM only.
722 *) Expand range of ctrls for AES GCM. Permit setting invocation
723 field on decrypt and retrieval of invocation field only on encrypt.
726 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
727 As required by RFC5289 these ciphersuites cannot be used if for
728 versions of TLS earlier than 1.2.
731 *) For FIPS capable OpenSSL interpret a NULL default public key method
732 as unset and return the appopriate default but do *not* set the default.
733 This means we can return the appopriate method in applications that
734 swicth between FIPS and non-FIPS modes.
737 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
738 ENGINE is used then we cannot handle that in the FIPS module so we
739 keep original code iff non-FIPS operations are allowed.
742 *) Add -attime option to openssl utilities.
743 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
745 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
748 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
749 FIPS EC methods unconditionally for now.
752 *) New build option no-ec2m to disable characteristic 2 code.
755 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
756 all cases can be covered as some introduce binary incompatibilities.
759 *) Redirect RSA operations to FIPS module including keygen,
760 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
763 *) Add similar low level API blocking to ciphers.
766 *) Low level digest APIs are not approved in FIPS mode: any attempt
767 to use these will cause a fatal error. Applications that *really* want
768 to use them can use the private_* version instead.
771 *) Redirect cipher operations to FIPS module for FIPS builds.
774 *) Redirect digest operations to FIPS module for FIPS builds.
777 *) Update build system to add "fips" flag which will link in fipscanister.o
778 for static and shared library builds embedding a signature if needed.
781 *) Output TLS supported curves in preference order instead of numerical
782 order. This is currently hardcoded for the highest order curves first.
783 This should be configurable so applications can judge speed vs strength.
786 *) Add TLS v1.2 server support for client authentication.
789 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
793 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
794 FIPS modules versions.
797 *) Add TLS v1.2 client side support for client authentication. Keep cache
798 of handshake records longer as we don't know the hash algorithm to use
799 until after the certificate request message is received.
802 *) Initial TLS v1.2 client support. Add a default signature algorithms
803 extension including all the algorithms we support. Parse new signature
804 format in client key exchange. Relax some ECC signing restrictions for
805 TLS v1.2 as indicated in RFC5246.
808 *) Add server support for TLS v1.2 signature algorithms extension. Switch
809 to new signature format when needed using client digest preference.
810 All server ciphersuites should now work correctly in TLS v1.2. No client
811 support yet and no support for client certificates.
814 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
815 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
816 ciphersuites. At present only RSA key exchange ciphersuites work with
817 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
818 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
819 and version checking.
822 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
823 with this defined it will not be affected by any changes to ssl internal
824 structures. Add several utility functions to allow openssl application
825 to work with OPENSSL_NO_SSL_INTERN defined.
829 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
831 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
834 *) Permit abbreviated handshakes when renegotiating using the function
835 SSL_renegotiate_abbreviated().
836 [Robin Seggelmann <seggelmann@fh-muenster.de>]
838 *) Add call to ENGINE_register_all_complete() to
839 ENGINE_load_builtin_engines(), so some implementations get used
840 automatically instead of needing explicit application support.
843 *) Add support for TLS key exporter as described in RFC5705.
844 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
846 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
847 a few changes are required:
849 Add SSL_OP_NO_TLSv1_1 flag.
851 Update version checking logic to handle version 1.1.
852 Add explicit IV handling (ported from DTLS code).
853 Add command line options to s_client/s_server.
856 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
858 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
859 in CMS and PKCS7 code. When RSA decryption fails use a random key for
860 content decryption and always return the same error. Note: this attack
861 needs on average 2^20 messages so it only affects automated senders. The
862 old behaviour can be reenabled in the CMS code by setting the
863 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
864 an MMA defence is not necessary.
865 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
866 this issue. (CVE-2012-0884)
869 *) Fix CVE-2011-4619: make sure we really are receiving a
870 client hello before rejecting multiple SGC restarts. Thanks to
871 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
874 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
876 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
877 Thanks to Antonio Martin, Enterprise Secure Access Research and
878 Development, Cisco Systems, Inc. for discovering this bug and
879 preparing a fix. (CVE-2012-0050)
882 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
884 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
885 of the Vaudenay padding oracle attack on CBC mode encryption
886 which enables an efficient plaintext recovery attack against
887 the OpenSSL implementation of DTLS. Their attack exploits timing
888 differences arising during decryption processing. A research
889 paper describing this attack can be found at:
890 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
891 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
892 Security Group at Royal Holloway, University of London
893 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
894 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
895 for preparing the fix. (CVE-2011-4108)
896 [Robin Seggelmann, Michael Tuexen]
898 *) Clear bytes used for block padding of SSL 3.0 records.
900 [Adam Langley (Google)]
902 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
903 Kadianakis <desnacked@gmail.com> for discovering this issue and
904 Adam Langley for preparing the fix. (CVE-2011-4619)
905 [Adam Langley (Google)]
907 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
908 [Andrey Kulikov <amdeich@gmail.com>]
910 *) Prevent malformed RFC3779 data triggering an assertion failure.
911 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
912 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
913 [Rob Austein <sra@hactrn.net>]
915 *) Improved PRNG seeding for VOS.
916 [Paul Green <Paul.Green@stratus.com>]
918 *) Fix ssl_ciph.c set-up race.
919 [Adam Langley (Google)]
921 *) Fix spurious failures in ecdsatest.c.
922 [Emilia Käsper (Google)]
924 *) Fix the BIO_f_buffer() implementation (which was mixing different
925 interpretations of the '..._len' fields).
926 [Adam Langley (Google)]
928 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
929 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
930 threads won't reuse the same blinding coefficients.
932 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
933 lock to call BN_BLINDING_invert_ex, and avoids one use of
934 BN_BLINDING_update for each BN_BLINDING structure (previously,
935 the last update always remained unused).
936 [Emilia Käsper (Google)]
938 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
939 [Bob Buckholz (Google)]
941 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
943 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
944 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
945 [Kaspar Brand <ossl@velox.ch>]
947 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
948 for multi-threaded use of ECDH. (CVE-2011-3210)
949 [Adam Langley (Google)]
951 *) Fix x509_name_ex_d2i memory leak on bad inputs.
954 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
955 signature public key algorithm by using OID xref utilities instead.
956 Before this you could only use some ECC ciphersuites with SHA1 only.
959 *) Add protection against ECDSA timing attacks as mentioned in the paper
960 by Billy Bob Brumley and Nicola Tuveri, see:
962 http://eprint.iacr.org/2011/232.pdf
964 [Billy Bob Brumley and Nicola Tuveri]
966 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
968 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
969 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
971 *) Fix bug in string printing code: if *any* escaping is enabled we must
972 escape the escape character (backslash) or the resulting string is
976 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
978 *) Disable code workaround for ancient and obsolete Netscape browsers
979 and servers: an attacker can use it in a ciphersuite downgrade attack.
980 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
983 *) Fixed J-PAKE implementation error, originally discovered by
984 Sebastien Martini, further info and confirmation from Stefan
985 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
988 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
990 *) Fix extension code to avoid race conditions which can result in a buffer
991 overrun vulnerability: resumed sessions must not be modified as they can
992 be shared by multiple threads. CVE-2010-3864
995 *) Fix WIN32 build system to correctly link an ENGINE directory into
999 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
1001 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
1003 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
1005 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
1007 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
1008 context. The operation can be customised via the ctrl mechanism in
1009 case ENGINEs want to include additional functionality.
1012 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
1015 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
1016 output hashes compatible with older versions of OpenSSL.
1017 [Willy Weisz <weisz@vcpc.univie.ac.at>]
1019 *) Fix compression algorithm handling: if resuming a session use the
1020 compression algorithm of the resumed session instead of determining
1021 it from client hello again. Don't allow server to change algorithm.
1024 *) Add load_crls() function to apps tidying load_certs() too. Add option
1025 to verify utility to allow additional CRLs to be included.
1028 *) Update OCSP request code to permit adding custom headers to the request:
1029 some responders need this.
1032 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
1034 [Julia Lawall <julia@diku.dk>]
1036 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
1037 needlessly dereferenced structures, used obsolete functions and
1038 didn't handle all updated verify codes correctly.
1041 *) Disable MD2 in the default configuration.
1044 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
1045 indicate the initial BIO being pushed or popped. This makes it possible
1046 to determine whether the BIO is the one explicitly called or as a result
1047 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
1048 it handles reference counts correctly and doesn't zero out the I/O bio
1049 when it is not being explicitly popped. WARNING: applications which
1050 included workarounds for the old buggy behaviour will need to be modified
1051 or they could free up already freed BIOs.
1054 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
1055 renaming to all platforms (within the 0.9.8 branch, this was
1056 done conditionally on Netware platforms to avoid a name clash).
1057 [Guenter <lists@gknw.net>]
1059 *) Add ECDHE and PSK support to DTLS.
1060 [Michael Tuexen <tuexen@fh-muenster.de>]
1062 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
1066 *) Add "missing" function EVP_MD_flags() (without this the only way to
1067 retrieve a digest flags is by accessing the structure directly. Update
1068 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
1069 or cipher is registered as in the "from" argument. Print out all
1070 registered digests in the dgst usage message instead of manually
1071 attempting to work them out.
1074 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
1075 this allows the use of compression and extensions. Change default cipher
1076 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
1077 by default unless an application cipher string requests it.
1080 *) Alter match criteria in PKCS12_parse(). It used to try to use local
1081 key ids to find matching certificates and keys but some PKCS#12 files
1082 don't follow the (somewhat unwritten) rules and this strategy fails.
1083 Now just gather all certificates together and the first private key
1084 then look for the first certificate that matches the key.
1087 *) Support use of registered digest and cipher names for dgst and cipher
1088 commands instead of having to add each one as a special case. So now
1095 openssl dgst -sha256 foo
1097 and this works for ENGINE based algorithms too.
1101 *) Update Gost ENGINE to support parameter files.
1102 [Victor B. Wagner <vitus@cryptocom.ru>]
1104 *) Support GeneralizedTime in ca utility.
1105 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
1107 *) Enhance the hash format used for certificate directory links. The new
1108 form uses the canonical encoding (meaning equivalent names will work
1109 even if they aren't identical) and uses SHA1 instead of MD5. This form
1110 is incompatible with the older format and as a result c_rehash should
1111 be used to rebuild symbolic links.
1114 *) Make PKCS#8 the default write format for private keys, replacing the
1115 traditional format. This form is standardised, more secure and doesn't
1116 include an implicit MD5 dependency.
1119 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
1120 committed to OpenSSL should pass this lot as a minimum.
1123 *) Add session ticket override functionality for use by EAP-FAST.
1124 [Jouni Malinen <j@w1.fi>]
1126 *) Modify HMAC functions to return a value. Since these can be implemented
1127 in an ENGINE errors can occur.
1130 *) Type-checked OBJ_bsearch_ex.
1133 *) Type-checked OBJ_bsearch. Also some constification necessitated
1134 by type-checking. Still to come: TXT_DB, bsearch(?),
1135 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
1139 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
1140 seconds to a tm structure directly, instead of going through OS
1141 specific date routines. This avoids any issues with OS routines such
1142 as the year 2038 bug. New *_adj() functions for ASN1 time structures
1143 and X509_time_adj_ex() to cover the extended range. The existing
1144 X509_time_adj() is still usable and will no longer have any date issues.
1147 *) Delta CRL support. New use deltas option which will attempt to locate
1148 and search any appropriate delta CRLs available.
1150 This work was sponsored by Google.
1153 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
1154 code and add additional score elements. Validate alternate CRL paths
1155 as part of the CRL checking and indicate a new error "CRL path validation
1156 error" in this case. Applications wanting additional details can use
1157 the verify callback and check the new "parent" field. If this is not
1158 NULL CRL path validation is taking place. Existing applications wont
1159 see this because it requires extended CRL support which is off by
1162 This work was sponsored by Google.
1165 *) Support for freshest CRL extension.
1167 This work was sponsored by Google.
1170 *) Initial indirect CRL support. Currently only supported in the CRLs
1171 passed directly and not via lookup. Process certificate issuer
1172 CRL entry extension and lookup CRL entries by bother issuer name
1173 and serial number. Check and process CRL issuer entry in IDP extension.
1175 This work was sponsored by Google.
1178 *) Add support for distinct certificate and CRL paths. The CRL issuer
1179 certificate is validated separately in this case. Only enabled if
1180 an extended CRL support flag is set: this flag will enable additional
1181 CRL functionality in future.
1183 This work was sponsored by Google.
1186 *) Add support for policy mappings extension.
1188 This work was sponsored by Google.
1191 *) Fixes to pathlength constraint, self issued certificate handling,
1192 policy processing to align with RFC3280 and PKITS tests.
1194 This work was sponsored by Google.
1197 *) Support for name constraints certificate extension. DN, email, DNS
1198 and URI types are currently supported.
1200 This work was sponsored by Google.
1203 *) To cater for systems that provide a pointer-based thread ID rather
1204 than numeric, deprecate the current numeric thread ID mechanism and
1205 replace it with a structure and associated callback type. This
1206 mechanism allows a numeric "hash" to be extracted from a thread ID in
1207 either case, and on platforms where pointers are larger than 'long',
1208 mixing is done to help ensure the numeric 'hash' is usable even if it
1209 can't be guaranteed unique. The default mechanism is to use "&errno"
1210 as a pointer-based thread ID to distinguish between threads.
1212 Applications that want to provide their own thread IDs should now use
1213 CRYPTO_THREADID_set_callback() to register a callback that will call
1214 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
1216 Note that ERR_remove_state() is now deprecated, because it is tied
1217 to the assumption that thread IDs are numeric. ERR_remove_state(0)
1218 to free the current thread's error state should be replaced by
1219 ERR_remove_thread_state(NULL).
1221 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
1222 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
1223 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
1224 application was previously providing a numeric thread callback that
1225 was inappropriate for distinguishing threads, then uniqueness might
1226 have been obtained with &errno that happened immediately in the
1227 intermediate development versions of OpenSSL; this is no longer the
1228 case, the numeric thread callback will now override the automatic use
1230 [Geoff Thorpe, with help from Bodo Moeller]
1232 *) Initial support for different CRL issuing certificates. This covers a
1233 simple case where the self issued certificates in the chain exist and
1234 the real CRL issuer is higher in the existing chain.
1236 This work was sponsored by Google.
1239 *) Removed effectively defunct crypto/store from the build.
1242 *) Revamp of STACK to provide stronger type-checking. Still to come:
1243 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
1244 ASN1_STRING, CONF_VALUE.
1247 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
1248 RAM on SSL connections. This option can save about 34k per idle SSL.
1251 *) Revamp of LHASH to provide stronger type-checking. Still to come:
1252 STACK, TXT_DB, bsearch, qsort.
1255 *) Initial support for Cryptographic Message Syntax (aka CMS) based
1256 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
1257 support for data, signedData, compressedData, digestedData and
1258 encryptedData, envelopedData types included. Scripts to check against
1259 RFC4134 examples draft and interop and consistency checks of many
1260 content types and variants.
1263 *) Add options to enc utility to support use of zlib compression BIO.
1266 *) Extend mk1mf to support importing of options and assembly language
1267 files from Configure script, currently only included in VC-WIN32.
1268 The assembly language rules can now optionally generate the source
1269 files from the associated perl scripts.
1272 *) Implement remaining functionality needed to support GOST ciphersuites.
1273 Interop testing has been performed using CryptoPro implementations.
1274 [Victor B. Wagner <vitus@cryptocom.ru>]
1276 *) s390x assembler pack.
1279 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
1283 *) Implement Opaque PRF Input TLS extension as specified in
1284 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
1285 official specification yet and no extension type assignment by
1286 IANA exists, this extension (for now) will have to be explicitly
1287 enabled when building OpenSSL by providing the extension number
1288 to use. For example, specify an option
1290 -DTLSEXT_TYPE_opaque_prf_input=0x9527
1292 to the "config" or "Configure" script to enable the extension,
1293 assuming extension number 0x9527 (which is a completely arbitrary
1294 and unofficial assignment based on the MD5 hash of the Internet
1295 Draft). Note that by doing so, you potentially lose
1296 interoperability with other TLS implementations since these might
1297 be using the same extension number for other purposes.
1299 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
1300 opaque PRF input value to use in the handshake. This will create
1301 an interal copy of the length-'len' string at 'src', and will
1302 return non-zero for success.
1304 To get more control and flexibility, provide a callback function
1307 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
1308 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
1312 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
1315 Callback function 'cb' will be called in handshakes, and is
1316 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
1317 Argument 'arg' is for application purposes (the value as given to
1318 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
1319 be provided to the callback function). The callback function
1320 has to return non-zero to report success: usually 1 to use opaque
1321 PRF input just if possible, or 2 to enforce use of the opaque PRF
1322 input. In the latter case, the library will abort the handshake
1323 if opaque PRF input is not successfully negotiated.
1325 Arguments 'peerinput' and 'len' given to the callback function
1326 will always be NULL and 0 in the case of a client. A server will
1327 see the client's opaque PRF input through these variables if
1328 available (NULL and 0 otherwise). Note that if the server
1329 provides an opaque PRF input, the length must be the same as the
1330 length of the client's opaque PRF input.
1332 Note that the callback function will only be called when creating
1333 a new session (session resumption can resume whatever was
1334 previously negotiated), and will not be called in SSL 2.0
1335 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
1336 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
1337 for applications that need to enforce opaque PRF input.
1341 *) Update ssl code to support digests other than SHA1+MD5 for handshake
1344 [Victor B. Wagner <vitus@cryptocom.ru>]
1346 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1347 RFC4507bis. The encrypted ticket format is an encrypted encoded
1348 SSL_SESSION structure, that way new session features are automatically
1351 If a client application caches session in an SSL_SESSION structure
1352 support is transparent because tickets are now stored in the encoded
1355 The SSL_CTX structure automatically generates keys for ticket
1356 protection in servers so again support should be possible
1357 with no application modification.
1359 If a client or server wishes to disable RFC4507 support then the option
1360 SSL_OP_NO_TICKET can be set.
1362 Add a TLS extension debugging callback to allow the contents of any client
1363 or server extensions to be examined.
1365 This work was sponsored by Google.
1368 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
1369 OpenSSL should now compile cleanly on gcc 4.2
1370 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
1372 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
1373 support including streaming MAC support: this is required for GOST
1374 ciphersuite support.
1375 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
1377 *) Add option -stream to use PKCS#7 streaming in smime utility. New
1378 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
1379 to output in BER and PEM format.
1382 *) Experimental support for use of HMAC via EVP_PKEY interface. This
1383 allows HMAC to be handled via the EVP_DigestSign*() interface. The
1384 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
1385 ENGINE support for HMAC keys which are unextractable. New -mac and
1386 -macopt options to dgst utility.
1389 *) New option -sigopt to dgst utility. Update dgst to use
1390 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
1391 alternative signing paramaters such as X9.31 or PSS in the dgst
1395 *) Change ssl_cipher_apply_rule(), the internal function that does
1396 the work each time a ciphersuite string requests enabling
1397 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
1398 removing ("!foo+bar") a class of ciphersuites: Now it maintains
1399 the order of disabled ciphersuites such that those ciphersuites
1400 that most recently went from enabled to disabled not only stay
1401 in order with respect to each other, but also have higher priority
1402 than other disabled ciphersuites the next time ciphersuites are
1405 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
1406 the same ciphersuites as with "HIGH" alone, but in a specific
1407 order where the PSK ciphersuites come first (since they are the
1408 most recently disabled ciphersuites when "HIGH" is parsed).
1410 Also, change ssl_create_cipher_list() (using this new
1411 funcionality) such that between otherwise identical
1412 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
1416 *) Change ssl_create_cipher_list() so that it automatically
1417 arranges the ciphersuites in reasonable order before starting
1418 to process the rule string. Thus, the definition for "DEFAULT"
1419 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
1420 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
1421 This makes it much easier to arrive at a reasonable default order
1422 in applications for which anonymous ciphers are OK (meaning
1423 that you can't actually use DEFAULT).
1424 [Bodo Moeller; suggested by Victor Duchovni]
1426 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
1427 processing) into multiple integers instead of setting
1428 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
1429 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
1430 (These masks as well as the individual bit definitions are hidden
1431 away into the non-exported interface ssl/ssl_locl.h, so this
1432 change to the definition of the SSL_CIPHER structure shouldn't
1433 affect applications.) This give us more bits for each of these
1434 categories, so there is no longer a need to coagulate AES128 and
1435 AES256 into a single algorithm bit, and to coagulate Camellia128
1436 and Camellia256 into a single algorithm bit, which has led to all
1439 Thus, among other things, the kludge introduced in 0.9.7m and
1440 0.9.8e for masking out AES256 independently of AES128 or masking
1441 out Camellia256 independently of AES256 is not needed here in 0.9.9.
1443 With the change, we also introduce new ciphersuite aliases that
1444 so far were missing: "AES128", "AES256", "CAMELLIA128", and
1448 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
1449 Use the leftmost N bytes of the signature input if the input is
1450 larger than the prime q (with N being the size in bytes of q).
1453 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
1454 it yet and it is largely untested.
1457 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
1460 *) Initial incomplete changes to avoid need for function casts in OpenSSL
1461 some compilers (gcc 4.2 and later) reject their use. Safestack is
1462 reimplemented. Update ASN1 to avoid use of legacy functions.
1465 *) Win32/64 targets are linked with Winsock2.
1468 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
1469 to external functions. This can be used to increase CRL handling
1470 efficiency especially when CRLs are very large by (for example) storing
1471 the CRL revoked certificates in a database.
1474 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
1475 new CRLs added to a directory can be used. New command line option
1476 -verify_return_error to s_client and s_server. This causes real errors
1477 to be returned by the verify callback instead of carrying on no matter
1478 what. This reflects the way a "real world" verify callback would behave.
1481 *) GOST engine, supporting several GOST algorithms and public key formats.
1482 Kindly donated by Cryptocom.
1485 *) Partial support for Issuing Distribution Point CRL extension. CRLs
1486 partitioned by DP are handled but no indirect CRL or reason partitioning
1487 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
1488 selected via a scoring technique which handles IDP and AKID in CRLs.
1491 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
1492 will ultimately be used for all verify operations: this will remove the
1493 X509_STORE dependency on certificate verification and allow alternative
1494 lookup methods. X509_STORE based implementations of these two callbacks.
1497 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
1498 Modify get_crl() to find a valid (unexpired) CRL if possible.
1501 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
1502 this would be called X509_CRL_cmp() but that name is already used by
1503 a function that just compares CRL issuer names. Cache several CRL
1504 extensions in X509_CRL structure and cache CRLDP in X509.
1507 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
1508 this maps equivalent X509_NAME structures into a consistent structure.
1509 Name comparison can then be performed rapidly using memcmp().
1512 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
1516 *) Allow digests to supply their own micalg string for S/MIME type using
1517 the ctrl EVP_MD_CTRL_MICALG.
1520 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
1521 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
1522 ctrl. It can then customise the structure before and/or after signing
1526 *) New function OBJ_add_sigid() to allow application defined signature OIDs
1527 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
1528 to free up any added signature OIDs.
1531 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
1532 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
1533 digest and cipher tables. New options added to openssl utility:
1534 list-message-digest-algorithms and list-cipher-algorithms.
1537 *) Change the array representation of binary polynomials: the list
1538 of degrees of non-zero coefficients is now terminated with -1.
1539 Previously it was terminated with 0, which was also part of the
1540 value; thus, the array representation was not applicable to
1541 polynomials where t^0 has coefficient zero. This change makes
1542 the array representation useful in a more general context.
1545 *) Various modifications and fixes to SSL/TLS cipher string
1546 handling. For ECC, the code now distinguishes between fixed ECDH
1547 with RSA certificates on the one hand and with ECDSA certificates
1548 on the other hand, since these are separate ciphersuites. The
1549 unused code for Fortezza ciphersuites has been removed.
1551 For consistency with EDH, ephemeral ECDH is now called "EECDH"
1552 (not "ECDHE"). For consistency with the code for DH
1553 certificates, use of ECDH certificates is now considered ECDH
1554 authentication, not RSA or ECDSA authentication (the latter is
1555 merely the CA's signing algorithm and not actively used in the
1558 The temporary ciphersuite alias "ECCdraft" is no longer
1559 available, and ECC ciphersuites are no longer excluded from "ALL"
1560 and "DEFAULT". The following aliases now exist for RFC 4492
1561 ciphersuites, most of these by analogy with the DH case:
1563 kECDHr - ECDH cert, signed with RSA
1564 kECDHe - ECDH cert, signed with ECDSA
1565 kECDH - ECDH cert (signed with either RSA or ECDSA)
1566 kEECDH - ephemeral ECDH
1567 ECDH - ECDH cert or ephemeral ECDH
1573 AECDH - anonymous ECDH
1574 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
1578 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
1579 Use correct micalg parameters depending on digest(s) in signed message.
1582 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
1583 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
1586 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
1587 an engine to register a method. Add ENGINE lookups for methods and
1588 functional reference processing.
1591 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
1592 EVP_{Sign,Verify}* which allow an application to customise the signature
1596 *) New -resign option to smime utility. This adds one or more signers
1597 to an existing PKCS#7 signedData structure. Also -md option to use an
1598 alternative message digest algorithm for signing.
1601 *) Tidy up PKCS#7 routines and add new functions to make it easier to
1602 create PKCS7 structures containing multiple signers. Update smime
1603 application to support multiple signers.
1606 *) New -macalg option to pkcs12 utility to allow setting of an alternative
1610 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
1611 Reorganize PBE internals to lookup from a static table using NIDs,
1612 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
1613 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
1614 PRF which will be automatically used with PBES2.
1617 *) Replace the algorithm specific calls to generate keys in "req" with the
1621 *) Update PKCS#7 enveloped data routines to use new API. This is now
1622 supported by any public key method supporting the encrypt operation. A
1623 ctrl is added to allow the public key algorithm to examine or modify
1624 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
1628 *) Add a ctrl to asn1 method to allow a public key algorithm to express
1629 a default digest type to use. In most cases this will be SHA1 but some
1630 algorithms (such as GOST) need to specify an alternative digest. The
1631 return value indicates how strong the prefernce is 1 means optional and
1632 2 is mandatory (that is it is the only supported type). Modify
1633 ASN1_item_sign() to accept a NULL digest argument to indicate it should
1634 use the default md. Update openssl utilities to use the default digest
1635 type for signing if it is not explicitly indicated.
1638 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
1639 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
1640 signing method from the key type. This effectively removes the link
1641 between digests and public key types.
1644 *) Add an OID cross reference table and utility functions. Its purpose is to
1645 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
1646 rsaEncryption. This will allow some of the algorithm specific hackery
1647 needed to use the correct OID to be removed.
1650 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
1651 structures for PKCS7_sign(). They are now set up by the relevant public
1655 *) Add provisional EC pkey method with support for ECDSA and ECDH.
1658 *) Add support for key derivation (agreement) in the API, DH method and
1662 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
1663 public and private key formats. As a side effect these add additional
1664 command line functionality not previously available: DSA signatures can be
1665 generated and verified using pkeyutl and DH key support and generation in
1670 [Oliver Tappe <zooey@hirschkaefer.de>]
1672 *) New make target "install_html_docs" installs HTML renditions of the
1674 [Oliver Tappe <zooey@hirschkaefer.de>]
1676 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
1677 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
1678 support key and parameter generation and add initial key generation
1679 functionality for RSA.
1682 *) Add functions for main EVP_PKEY_method operations. The undocumented
1683 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
1684 EVP_PKEY_{encrypt,decrypt}_old.
1687 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
1688 key API, doesn't do much yet.
1691 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
1692 public key algorithms. New option to openssl utility:
1693 "list-public-key-algorithms" to print out info.
1696 *) Implement the Supported Elliptic Curves Extension for
1697 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1700 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
1701 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
1704 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
1705 utilities such as rsa, dsa, dsaparam etc except they process any key
1709 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1710 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1711 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
1715 *) Initial support for pluggable public key ASN1.
1716 De-spaghettify the public key ASN1 handling. Move public and private
1717 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
1718 algorithm specific handling to a single module within the relevant
1719 algorithm directory. Add functions to allow (near) opaque processing
1720 of public and private key structures.
1723 *) Implement the Supported Point Formats Extension for
1724 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1727 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
1728 for the psk identity [hint] and the psk callback functions to the
1729 SSL_SESSION, SSL and SSL_CTX structure.
1732 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
1736 SSL_CTX_use_psk_identity_hint
1737 SSL_get_psk_identity_hint
1738 SSL_get_psk_identity
1739 SSL_use_psk_identity_hint
1741 [Mika Kousa and Pasi Eronen of Nokia Corporation]
1743 *) Add RFC 3161 compliant time stamp request creation, response generation
1744 and response verification functionality.
1745 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
1747 *) Add initial support for TLS extensions, specifically for the server_name
1748 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1749 have new members for a host name. The SSL data structure has an
1750 additional member SSL_CTX *initial_ctx so that new sessions can be
1751 stored in that context to allow for session resumption, even after the
1752 SSL has been switched to a new SSL_CTX in reaction to a client's
1753 server_name extension.
1755 New functions (subject to change):
1757 SSL_get_servername()
1758 SSL_get_servername_type()
1761 New CTRL codes and macros (subject to change):
1763 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1764 - SSL_CTX_set_tlsext_servername_callback()
1765 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1766 - SSL_CTX_set_tlsext_servername_arg()
1767 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1769 openssl s_client has a new '-servername ...' option.
1771 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1772 '-key2 ...', '-servername_fatal' (subject to change). This allows
1773 testing the HostName extension for a specific single host name ('-cert'
1774 and '-key' remain fallbacks for handshakes without HostName
1775 negotiation). If the unrecogninzed_name alert has to be sent, this by
1776 default is a warning; it becomes fatal with the '-servername_fatal'
1779 [Peter Sylvester, Remy Allais, Christophe Renou]
1781 *) Whirlpool hash implementation is added.
1784 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1785 bn(64,32). Because of instruction set limitations it doesn't have
1786 any negative impact on performance. This was done mostly in order
1787 to make it possible to share assembler modules, such as bn_mul_mont
1788 implementations, between 32- and 64-bit builds without hassle.
1791 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1792 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1796 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1797 dedicated Montgomery multiplication procedure, is introduced.
1798 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1799 "64-bit" performance on certain 32-bit targets.
1802 *) New option SSL_OP_NO_COMP to disable use of compression selectively
1803 in SSL structures. New SSL ctrl to set maximum send fragment size.
1804 Save memory by seeting the I/O buffer sizes dynamically instead of
1805 using the maximum available value.
1808 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
1809 in addition to the text details.
1812 *) Very, very preliminary EXPERIMENTAL support for printing of general
1813 ASN1 structures. This currently produces rather ugly output and doesn't
1814 handle several customised structures at all.
1817 *) Integrated support for PVK file format and some related formats such
1818 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
1819 these in the 'rsa' and 'dsa' utilities.
1822 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
1825 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
1826 place for the (very old) "NETSCAPE" format certificates which are now
1827 handled using new ASN1 code equivalents.
1830 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
1831 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
1832 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
1835 *) Modify CRL distribution points extension code to print out previously
1836 unsupported fields. Enhance extension setting code to allow setting of
1840 *) Add print and set support for Issuing Distribution Point CRL extension.
1843 *) Change 'Configure' script to enable Camellia by default.
1846 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
1848 *) When rejecting SSL/TLS records due to an incorrect version number, never
1849 update s->server with a new major version number. As of
1850 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
1851 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
1852 the previous behavior could result in a read attempt at NULL when
1853 receiving specific incorrect SSL/TLS records once record payload
1854 protection is active. (CVE-2010-0740)
1855 [Bodo Moeller, Adam Langley <agl@chromium.org>]
1857 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
1858 could be crashed if the relevant tables were not present (e.g. chrooted).
1859 [Tomas Hoger <thoger@redhat.com>]
1861 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
1863 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
1864 [Martin Olsson, Neel Mehta]
1866 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
1867 accommodate for stack sorting, always a write lock!).
1870 *) On some versions of WIN32 Heap32Next is very slow. This can cause
1871 excessive delays in the RAND_poll(): over a minute. As a workaround
1872 include a time check in the inner Heap32Next loop too.
1875 *) The code that handled flushing of data in SSL/TLS originally used the
1876 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
1877 the problem outlined in PR#1949. The fix suggested there however can
1878 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
1879 of Apache). So instead simplify the code to flush unconditionally.
1880 This should be fine since flushing with no data to flush is a no op.
1883 *) Handle TLS versions 2.0 and later properly and correctly use the
1884 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
1885 off ancient servers have a habit of sticking around for a while...
1888 *) Modify compression code so it frees up structures without using the
1889 ex_data callbacks. This works around a problem where some applications
1890 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
1891 restarting) then use compression (e.g. SSL with compression) later.
1892 This results in significant per-connection memory leaks and
1893 has caused some security issues including CVE-2008-1678 and
1897 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
1898 change when encrypting or decrypting.
1901 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
1902 connect and renegotiate with servers which do not support RI.
1903 Until RI is more widely deployed this option is enabled by default.
1906 *) Add "missing" ssl ctrls to clear options and mode.
1909 *) If client attempts to renegotiate and doesn't support RI respond with
1910 a no_renegotiation alert as required by RFC5746. Some renegotiating
1911 TLS clients will continue a connection gracefully when they receive
1912 the alert. Unfortunately OpenSSL mishandled this alert and would hang
1913 waiting for a server hello which it will never receive. Now we treat a
1914 received no_renegotiation alert as a fatal error. This is because
1915 applications requesting a renegotiation might well expect it to succeed
1916 and would have no code in place to handle the server denying it so the
1917 only safe thing to do is to terminate the connection.
1920 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
1921 peer supports secure renegotiation and 0 otherwise. Print out peer
1922 renegotiation support in s_client/s_server.
1925 *) Replace the highly broken and deprecated SPKAC certification method with
1926 the updated NID creation version. This should correctly handle UTF8.
1929 *) Implement RFC5746. Re-enable renegotiation but require the extension
1930 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
1931 turns out to be a bad idea. It has been replaced by
1932 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
1933 SSL_CTX_set_options(). This is really not recommended unless you
1934 know what you are doing.
1935 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
1937 *) Fixes to stateless session resumption handling. Use initial_ctx when
1938 issuing and attempting to decrypt tickets in case it has changed during
1939 servername handling. Use a non-zero length session ID when attempting
1940 stateless session resumption: this makes it possible to determine if
1941 a resumption has occurred immediately after receiving server hello
1942 (several places in OpenSSL subtly assume this) instead of later in
1946 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
1947 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
1948 fixes for a few places where the return code is not checked
1950 [Julia Lawall <julia@diku.dk>]
1952 *) Add --strict-warnings option to Configure script to include devteam
1953 warnings in other configurations.
1956 *) Add support for --libdir option and LIBDIR variable in makefiles. This
1957 makes it possible to install openssl libraries in locations which
1958 have names other than "lib", for example "/usr/lib64" which some
1960 [Steve Henson, based on patch from Jeremy Utley]
1962 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
1963 X690 8.9.12 and can produce some misleading textual output of OIDs.
1964 [Steve Henson, reported by Dan Kaminsky]
1966 *) Delete MD2 from algorithm tables. This follows the recommendation in
1967 several standards that it is not used in new applications due to
1968 several cryptographic weaknesses. For binary compatibility reasons
1969 the MD2 API is still compiled in by default.
1972 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
1976 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
1977 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
1979 [Guenter <lists@gknw.net>]
1981 *) Fix the server certificate chain building code to use X509_verify_cert(),
1982 it used to have an ad-hoc builder which was unable to cope with anything
1983 other than a simple chain.
1984 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
1986 *) Don't check self signed certificate signatures in X509_verify_cert()
1987 by default (a flag can override this): it just wastes time without
1988 adding any security. As a useful side effect self signed root CAs
1989 with non-FIPS digests are now usable in FIPS mode.
1992 *) In dtls1_process_out_of_seq_message() the check if the current message
1993 is already buffered was missing. For every new message was memory
1994 allocated, allowing an attacker to perform an denial of service attack
1995 with sending out of seq handshake messages until there is no memory
1996 left. Additionally every future messege was buffered, even if the
1997 sequence number made no sense and would be part of another handshake.
1998 So only messages with sequence numbers less than 10 in advance will be
1999 buffered. (CVE-2009-1378)
2000 [Robin Seggelmann, discovered by Daniel Mentz]
2002 *) Records are buffered if they arrive with a future epoch to be
2003 processed after finishing the corresponding handshake. There is
2004 currently no limitation to this buffer allowing an attacker to perform
2005 a DOS attack with sending records with future epochs until there is no
2006 memory left. This patch adds the pqueue_size() function to detemine
2007 the size of a buffer and limits the record buffer to 100 entries.
2009 [Robin Seggelmann, discovered by Daniel Mentz]
2011 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
2012 parent structure is freed. (CVE-2009-1379)
2015 *) Handle non-blocking I/O properly in SSL_shutdown() call.
2016 [Darryl Miles <darryl-mailinglists@netbauds.net>]
2019 [Ilya O. <vrghost@gmail.com>]
2021 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
2023 *) Disable renegotiation completely - this fixes a severe security
2024 problem (CVE-2009-3555) at the cost of breaking all
2025 renegotiation. Renegotiation can be re-enabled by setting
2026 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
2027 run-time. This is really not recommended unless you know what
2031 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
2033 *) Don't set val to NULL when freeing up structures, it is freed up by
2034 underlying code. If sizeof(void *) > sizeof(long) this can result in
2035 zeroing past the valid field. (CVE-2009-0789)
2036 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
2038 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
2039 checked correctly. This would allow some invalid signed attributes to
2040 appear to verify correctly. (CVE-2009-0591)
2041 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2043 *) Reject UniversalString and BMPString types with invalid lengths. This
2044 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
2045 a legal length. (CVE-2009-0590)
2048 *) Set S/MIME signing as the default purpose rather than setting it
2049 unconditionally. This allows applications to override it at the store
2053 *) Permit restricted recursion of ASN1 strings. This is needed in practice
2054 to handle some structures.
2057 *) Improve efficiency of mem_gets: don't search whole buffer each time
2059 [Jeremy Shapiro <jnshapir@us.ibm.com>]
2061 *) New -hex option for openssl rand.
2064 *) Print out UTF8String and NumericString when parsing ASN1.
2067 *) Support NumericString type for name components.
2070 *) Allow CC in the environment to override the automatically chosen
2071 compiler. Note that nothing is done to ensure flags work with the
2075 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
2077 *) Properly check EVP_VerifyFinal() and similar return values
2079 [Ben Laurie, Bodo Moeller, Google Security Team]
2081 *) Enable TLS extensions by default.
2084 *) Allow the CHIL engine to be loaded, whether the application is
2085 multithreaded or not. (This does not release the developer from the
2086 obligation to set up the dynamic locking callbacks.)
2087 [Sander Temme <sander@temme.net>]
2089 *) Use correct exit code if there is an error in dgst command.
2090 [Steve Henson; problem pointed out by Roland Dirlewanger]
2092 *) Tweak Configure so that you need to say "experimental-jpake" to enable
2093 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
2096 *) Add experimental JPAKE support, including demo authentication in
2097 s_client and s_server.
2100 *) Set the comparison function in v3_addr_canonize().
2101 [Rob Austein <sra@hactrn.net>]
2103 *) Add support for XMPP STARTTLS in s_client.
2104 [Philip Paeps <philip@freebsd.org>]
2106 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
2107 to ensure that even with this option, only ciphersuites in the
2108 server's preference list will be accepted. (Note that the option
2109 applies only when resuming a session, so the earlier behavior was
2110 just about the algorithm choice for symmetric cryptography.)
2113 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
2115 *) Fix NULL pointer dereference if a DTLS server received
2116 ChangeCipherSpec as first record (CVE-2009-1386).
2119 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
2120 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
2123 *) The fix in 0.9.8c that supposedly got rid of unsafe
2124 double-checked locking was incomplete for RSA blinding,
2125 addressing just one layer of what turns out to have been
2126 doubly unsafe triple-checked locking.
2128 So now fix this for real by retiring the MONT_HELPER macro
2129 in crypto/rsa/rsa_eay.c.
2131 [Bodo Moeller; problem pointed out by Marius Schilder]
2133 *) Various precautionary measures:
2135 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
2137 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
2138 (NB: This would require knowledge of the secret session ticket key
2139 to exploit, in which case you'd be SOL either way.)
2141 - Change bn_nist.c so that it will properly handle input BIGNUMs
2142 outside the expected range.
2144 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
2147 [Neel Mehta, Bodo Moeller]
2149 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
2150 the load fails. Useful for distros.
2151 [Ben Laurie and the FreeBSD team]
2153 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
2156 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
2159 *) Expand ENGINE to support engine supplied SSL client certificate functions.
2161 This work was sponsored by Logica.
2164 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
2165 keystores. Support for SSL/TLS client authentication too.
2166 Not compiled unless enable-capieng specified to Configure.
2168 This work was sponsored by Logica.
2171 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
2172 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
2173 attribute creation routines such as certifcate requests and PKCS#12
2177 Changes between 0.9.8g and 0.9.8h [28 May 2008]
2179 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
2180 handshake which could lead to a cilent crash as found using the
2181 Codenomicon TLS test suite (CVE-2008-1672)
2182 [Steve Henson, Mark Cox]
2184 *) Fix double free in TLS server name extensions which could lead to
2185 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
2188 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
2190 Clear the error queue to ensure that error entries left from
2191 older function calls do not interfere with the correct operation.
2192 [Lutz Jaenicke, Erik de Castro Lopo]
2194 *) Remove root CA certificates of commercial CAs:
2196 The OpenSSL project does not recommend any specific CA and does not
2197 have any policy with respect to including or excluding any CA.
2198 Therefore it does not make any sense to ship an arbitrary selection
2199 of root CA certificates with the OpenSSL software.
2202 *) RSA OAEP patches to fix two separate invalid memory reads.
2203 The first one involves inputs when 'lzero' is greater than
2204 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
2205 before the beginning of from). The second one involves inputs where
2206 the 'db' section contains nothing but zeroes (there is a one-byte
2207 invalid read after the end of 'db').
2208 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2210 *) Partial backport from 0.9.9-dev:
2212 Introduce bn_mul_mont (dedicated Montgomery multiplication
2213 procedure) as a candidate for BIGNUM assembler implementation.
2214 While 0.9.9-dev uses assembler for various architectures, only
2215 x86_64 is available by default here in the 0.9.8 branch, and
2216 32-bit x86 is available through a compile-time setting.
2218 To try the 32-bit x86 assembler implementation, use Configure
2219 option "enable-montasm" (which exists only for this backport).
2221 As "enable-montasm" for 32-bit x86 disclaims code stability
2222 anyway, in this constellation we activate additional code
2223 backported from 0.9.9-dev for further performance improvements,
2224 namely BN_from_montgomery_word. (To enable this otherwise,
2225 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
2227 [Andy Polyakov (backport partially by Bodo Moeller)]
2229 *) Add TLS session ticket callback. This allows an application to set
2230 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
2231 values. This is useful for key rollover for example where several key
2232 sets may exist with different names.
2235 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
2236 This was broken until now in 0.9.8 releases, such that the only way
2237 a registered ENGINE could be used (assuming it initialises
2238 successfully on the host) was to explicitly set it as the default
2239 for the relevant algorithms. This is in contradiction with 0.9.7
2240 behaviour and the documentation. With this fix, when an ENGINE is
2241 registered into a given algorithm's table of implementations, the
2242 'uptodate' flag is reset so that auto-discovery will be used next
2243 time a new context for that algorithm attempts to select an
2245 [Ian Lister (tweaked by Geoff Thorpe)]
2247 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
2248 implemention in the following ways:
2250 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
2253 Lack of BER streaming support means one pass streaming processing is
2254 only supported if data is detached: setting the streaming flag is
2255 ignored for embedded content.
2257 CMS support is disabled by default and must be explicitly enabled
2258 with the enable-cms configuration option.
2261 *) Update the GMP engine glue to do direct copies between BIGNUM and
2262 mpz_t when openssl and GMP use the same limb size. Otherwise the
2263 existing "conversion via a text string export" trick is still used.
2264 [Paul Sheer <paulsheer@gmail.com>]
2266 *) Zlib compression BIO. This is a filter BIO which compressed and
2267 uncompresses any data passed through it.
2270 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
2271 RFC3394 compatible AES key wrapping.
2274 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
2275 sets string data without copying. X509_ALGOR_set0() and
2276 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
2277 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
2278 from an X509_ATTRIBUTE structure optionally checking it occurs only
2279 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
2283 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
2284 to get the expected BN_FLG_CONSTTIME behavior.
2285 [Bodo Moeller (Google)]
2289 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
2290 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
2291 - added some more tests to do_tests.pl
2292 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
2293 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
2294 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
2295 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
2296 - various changes to netware.pl to enable gcc-cross builds on Win32
2298 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
2299 - various changes to fix missing prototype warnings
2300 - fixed x86nasm.pl to create correct asm files for NASM COFF output
2301 - added AES, WHIRLPOOL and CPUID assembler code to build files
2302 - added missing AES assembler make rules to mk1mf.pl
2303 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
2304 [Guenter Knauf <eflash@gmx.net>]
2306 *) Implement certificate status request TLS extension defined in RFC3546.
2307 A client can set the appropriate parameters and receive the encoded
2308 OCSP response via a callback. A server can query the supplied parameters
2309 and set the encoded OCSP response in the callback. Add simplified examples
2310 to s_client and s_server.
2313 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
2315 *) Fix various bugs:
2316 + Binary incompatibility of ssl_ctx_st structure
2317 + DTLS interoperation with non-compliant servers
2318 + Don't call get_session_cb() without proposed session
2319 + Fix ia64 assembler code
2320 [Andy Polyakov, Steve Henson]
2322 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
2324 *) DTLS Handshake overhaul. There were longstanding issues with
2325 OpenSSL DTLS implementation, which were making it impossible for
2326 RFC 4347 compliant client to communicate with OpenSSL server.
2327 Unfortunately just fixing these incompatibilities would "cut off"
2328 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
2329 server keeps tolerating non RFC compliant syntax. The opposite is
2330 not true, 0.9.8f client can not communicate with earlier server.
2331 This update even addresses CVE-2007-4995.
2334 *) Changes to avoid need for function casts in OpenSSL: some compilers
2335 (gcc 4.2 and later) reject their use.
2336 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
2339 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2340 RFC4507bis. The encrypted ticket format is an encrypted encoded
2341 SSL_SESSION structure, that way new session features are automatically
2344 If a client application caches session in an SSL_SESSION structure
2345 support is transparent because tickets are now stored in the encoded
2348 The SSL_CTX structure automatically generates keys for ticket
2349 protection in servers so again support should be possible
2350 with no application modification.
2352 If a client or server wishes to disable RFC4507 support then the option
2353 SSL_OP_NO_TICKET can be set.
2355 Add a TLS extension debugging callback to allow the contents of any client
2356 or server extensions to be examined.
2358 This work was sponsored by Google.
2361 *) Add initial support for TLS extensions, specifically for the server_name
2362 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2363 have new members for a host name. The SSL data structure has an
2364 additional member SSL_CTX *initial_ctx so that new sessions can be
2365 stored in that context to allow for session resumption, even after the
2366 SSL has been switched to a new SSL_CTX in reaction to a client's
2367 server_name extension.
2369 New functions (subject to change):
2371 SSL_get_servername()
2372 SSL_get_servername_type()
2375 New CTRL codes and macros (subject to change):
2377 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2378 - SSL_CTX_set_tlsext_servername_callback()
2379 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2380 - SSL_CTX_set_tlsext_servername_arg()
2381 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2383 openssl s_client has a new '-servername ...' option.
2385 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2386 '-key2 ...', '-servername_fatal' (subject to change). This allows
2387 testing the HostName extension for a specific single host name ('-cert'
2388 and '-key' remain fallbacks for handshakes without HostName
2389 negotiation). If the unrecogninzed_name alert has to be sent, this by
2390 default is a warning; it becomes fatal with the '-servername_fatal'
2393 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
2395 *) Add AES and SSE2 assembly language support to VC++ build.
2398 *) Mitigate attack on final subtraction in Montgomery reduction.
2401 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
2402 (which previously caused an internal error).
2405 *) Squeeze another 10% out of IGE mode when in != out.
2408 *) AES IGE mode speedup.
2409 [Dean Gaudet (Google)]
2411 *) Add the Korean symmetric 128-bit cipher SEED (see
2412 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
2413 add SEED ciphersuites from RFC 4162:
2415 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
2416 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
2417 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
2418 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
2420 To minimize changes between patchlevels in the OpenSSL 0.9.8
2421 series, SEED remains excluded from compilation unless OpenSSL
2422 is configured with 'enable-seed'.
2423 [KISA, Bodo Moeller]
2425 *) Mitigate branch prediction attacks, which can be practical if a
2426 single processor is shared, allowing a spy process to extract
2427 information. For detailed background information, see
2428 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
2429 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
2430 and Necessary Software Countermeasures"). The core of the change
2431 are new versions BN_div_no_branch() and
2432 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
2433 respectively, which are slower, but avoid the security-relevant
2434 conditional branches. These are automatically called by BN_div()
2435 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
2436 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
2437 remove a conditional branch.
2439 BN_FLG_CONSTTIME is the new name for the previous
2440 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
2441 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
2442 in the exponent causes BN_mod_exp_mont() to use the alternative
2443 implementation in BN_mod_exp_mont_consttime().) The old name
2444 remains as a deprecated alias.
2446 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
2447 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
2448 constant-time implementations for more than just exponentiation.
2449 Here too the old name is kept as a deprecated alias.
2451 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2452 the BN_BLINDING structure gets an independent copy of the
2453 modulus. This means that the previous "BIGNUM *m" argument to
2454 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2455 essentially becomes "const BIGNUM *m", although we can't actually
2456 change this in the header file before 0.9.9. It allows
2457 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2458 enable BN_FLG_CONSTTIME.
2460 [Matthew D Wood (Intel Corp)]
2462 *) In the SSL/TLS server implementation, be strict about session ID
2463 context matching (which matters if an application uses a single
2464 external cache for different purposes). Previously,
2465 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2466 set. This did ensure strict client verification, but meant that,
2467 with applications using a single external cache for quite
2468 different requirements, clients could circumvent ciphersuite
2469 restrictions for a given session ID context by starting a session
2470 in a different context.
2473 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2474 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2475 authentication-only ciphersuites.
2478 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
2479 not complete and could lead to a possible single byte overflow
2480 (CVE-2007-5135) [Ben Laurie]
2482 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
2484 *) Since AES128 and AES256 (and similarly Camellia128 and
2485 Camellia256) share a single mask bit in the logic of
2486 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2487 kludge to work properly if AES128 is available and AES256 isn't
2488 (or if Camellia128 is available and Camellia256 isn't).
2491 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
2492 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
2493 When a point or a seed is encoded in a BIT STRING, we need to
2494 prevent the removal of trailing zero bits to get the proper DER
2495 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
2496 of a NamedBitList, for which trailing 0 bits need to be removed.)
2499 *) Have SSL/TLS server implementation tolerate "mismatched" record
2500 protocol version while receiving ClientHello even if the
2501 ClientHello is fragmented. (The server can't insist on the
2502 particular protocol version it has chosen before the ServerHello
2503 message has informed the client about his choice.)
2506 *) Add RFC 3779 support.
2507 [Rob Austein for ARIN, Ben Laurie]
2509 *) Load error codes if they are not already present instead of using a
2510 static variable. This allows them to be cleanly unloaded and reloaded.
2511 Improve header file function name parsing.
2514 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
2515 or CAPABILITY handshake as required by RFCs.
2518 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
2520 *) Introduce limits to prevent malicious keys being able to
2521 cause a denial of service. (CVE-2006-2940)
2522 [Steve Henson, Bodo Moeller]
2524 *) Fix ASN.1 parsing of certain invalid structures that can result
2525 in a denial of service. (CVE-2006-2937) [Steve Henson]
2527 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2528 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2530 *) Fix SSL client code which could crash if connecting to a
2531 malicious SSLv2 server. (CVE-2006-4343)
2532 [Tavis Ormandy and Will Drewry, Google Security Team]
2534 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
2535 match only those. Before that, "AES256-SHA" would be interpreted
2536 as a pattern and match "AES128-SHA" too (since AES128-SHA got
2537 the same strength classification in 0.9.7h) as we currently only
2538 have a single AES bit in the ciphersuite description bitmap.
2539 That change, however, also applied to ciphersuite strings such as
2540 "RC4-MD5" that intentionally matched multiple ciphersuites --
2541 namely, SSL 2.0 ciphersuites in addition to the more common ones
2542 from SSL 3.0/TLS 1.0.
2544 So we change the selection algorithm again: Naming an explicit
2545 ciphersuite selects this one ciphersuite, and any other similar
2546 ciphersuite (same bitmap) from *other* protocol versions.
2547 Thus, "RC4-MD5" again will properly select both the SSL 2.0
2548 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
2550 Since SSL 2.0 does not have any ciphersuites for which the
2551 128/256 bit distinction would be relevant, this works for now.
2552 The proper fix will be to use different bits for AES128 and
2553 AES256, which would have avoided the problems from the beginning;
2554 however, bits are scarce, so we can only do this in a new release
2555 (not just a patchlevel) when we can change the SSL_CIPHER
2556 definition to split the single 'unsigned long mask' bitmap into
2557 multiple values to extend the available space.
2561 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2563 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2564 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2566 *) Add AES IGE and biIGE modes.
2569 *) Change the Unix randomness entropy gathering to use poll() when
2570 possible instead of select(), since the latter has some
2571 undesirable limitations.
2572 [Darryl Miles via Richard Levitte and Bodo Moeller]
2574 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2575 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2576 cannot be implicitly activated as part of, e.g., the "AES" alias.
2577 However, please upgrade to OpenSSL 0.9.9[-dev] for
2578 non-experimental use of the ECC ciphersuites to get TLS extension
2579 support, which is required for curve and point format negotiation
2580 to avoid potential handshake problems.
2583 *) Disable rogue ciphersuites:
2585 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2586 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2587 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2589 The latter two were purportedly from
2590 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2593 Also deactivate the remaining ciphersuites from
2594 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2595 unofficial, and the ID has long expired.
2598 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2599 dual-core machines) and other potential thread-safety issues.
2602 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
2603 versions), which is now available for royalty-free use
2604 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
2605 Also, add Camellia TLS ciphersuites from RFC 4132.
2607 To minimize changes between patchlevels in the OpenSSL 0.9.8
2608 series, Camellia remains excluded from compilation unless OpenSSL
2609 is configured with 'enable-camellia'.
2612 *) Disable the padding bug check when compression is in use. The padding
2613 bug check assumes the first packet is of even length, this is not
2614 necessarily true if compresssion is enabled and can result in false
2615 positives causing handshake failure. The actual bug test is ancient
2616 code so it is hoped that implementations will either have fixed it by
2617 now or any which still have the bug do not support compression.
2620 Changes between 0.9.8a and 0.9.8b [04 May 2006]
2622 *) When applying a cipher rule check to see if string match is an explicit
2623 cipher suite and only match that one cipher suite if it is.
2626 *) Link in manifests for VC++ if needed.
2627 [Austin Ziegler <halostatue@gmail.com>]
2629 *) Update support for ECC-based TLS ciphersuites according to
2630 draft-ietf-tls-ecc-12.txt with proposed changes (but without
2631 TLS extensions, which are supported starting with the 0.9.9
2632 branch, not in the OpenSSL 0.9.8 branch).
2635 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
2636 opaque EVP_CIPHER_CTX handling.
2639 *) Fixes and enhancements to zlib compression code. We now only use
2640 "zlib1.dll" and use the default __cdecl calling convention on Win32
2641 to conform with the standards mentioned here:
2642 http://www.zlib.net/DLL_FAQ.txt
2643 Static zlib linking now works on Windows and the new --with-zlib-include
2644 --with-zlib-lib options to Configure can be used to supply the location
2645 of the headers and library. Gracefully handle case where zlib library
2649 *) Several fixes and enhancements to the OID generation code. The old code
2650 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
2651 handle numbers larger than ULONG_MAX, truncated printing and had a
2652 non standard OBJ_obj2txt() behaviour.
2655 *) Add support for building of engines under engine/ as shared libraries
2656 under VC++ build system.
2659 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
2660 Hopefully, we will not see any false combination of paths any more.
2663 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
2665 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2666 (part of SSL_OP_ALL). This option used to disable the
2667 countermeasure against man-in-the-middle protocol-version
2668 rollback in the SSL 2.0 server implementation, which is a bad
2669 idea. (CVE-2005-2969)
2671 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2672 for Information Security, National Institute of Advanced Industrial
2673 Science and Technology [AIST], Japan)]
2675 *) Add two function to clear and return the verify parameter flags.
2678 *) Keep cipherlists sorted in the source instead of sorting them at
2679 runtime, thus removing the need for a lock.
2682 *) Avoid some small subgroup attacks in Diffie-Hellman.
2683 [Nick Mathewson and Ben Laurie]
2685 *) Add functions for well-known primes.
2688 *) Extended Windows CE support.
2689 [Satoshi Nakamura and Andy Polyakov]
2691 *) Initialize SSL_METHOD structures at compile time instead of during
2692 runtime, thus removing the need for a lock.
2695 *) Make PKCS7_decrypt() work even if no certificate is supplied by
2696 attempting to decrypt each encrypted key in turn. Add support to
2700 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
2702 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
2705 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
2708 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
2709 key into the same file any more.
2712 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
2715 *) Add -utf8 command line and config file option to 'ca'.
2716 [Stefan <stf@udoma.org]
2718 *) Removed the macro des_crypt(), as it seems to conflict with some
2719 libraries. Use DES_crypt().
2722 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
2723 involves renaming the source and generated shared-libs for
2724 both. The engines will accept the corrected or legacy ids
2725 ('ncipher' and '4758_cca' respectively) when binding. NB,
2726 this only applies when building 'shared'.
2727 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
2729 *) Add attribute functions to EVP_PKEY structure. Modify
2730 PKCS12_create() to recognize a CSP name attribute and
2731 use it. Make -CSP option work again in pkcs12 utility.
2734 *) Add new functionality to the bn blinding code:
2735 - automatic re-creation of the BN_BLINDING parameters after
2736 a fixed number of uses (currently 32)
2737 - add new function for parameter creation
2738 - introduce flags to control the update behaviour of the
2739 BN_BLINDING parameters
2740 - hide BN_BLINDING structure
2741 Add a second BN_BLINDING slot to the RSA structure to improve
2742 performance when a single RSA object is shared among several
2746 *) Add support for DTLS.
2747 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
2749 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
2750 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
2753 *) Remove buggy and incompletet DH cert support from
2754 ssl/ssl_rsa.c and ssl/s3_both.c
2757 *) Use SHA-1 instead of MD5 as the default digest algorithm for
2758 the apps/openssl applications.
2761 *) Compile clean with "-Wall -Wmissing-prototypes
2762 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
2763 DEBUG_SAFESTACK must also be set.
2766 *) Change ./Configure so that certain algorithms can be disabled by default.
2767 The new counterpiece to "no-xxx" is "enable-xxx".
2769 The patented RC5 and MDC2 algorithms will now be disabled unless
2770 "enable-rc5" and "enable-mdc2", respectively, are specified.
2772 (IDEA remains enabled despite being patented. This is because IDEA
2773 is frequently required for interoperability, and there is no license
2774 fee for non-commercial use. As before, "no-idea" can be used to
2775 avoid this algorithm.)
2779 *) Add processing of proxy certificates (see RFC 3820). This work was
2780 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
2781 EGEE (Enabling Grids for E-science in Europe).
2784 *) RC4 performance overhaul on modern architectures/implementations, such
2785 as Intel P4, IA-64 and AMD64.
2788 *) New utility extract-section.pl. This can be used specify an alternative
2789 section number in a pod file instead of having to treat each file as
2790 a separate case in Makefile. This can be done by adding two lines to the
2793 =for comment openssl_section:XXX
2795 The blank line is mandatory.
2799 *) New arguments -certform, -keyform and -pass for s_client and s_server
2800 to allow alternative format key and certificate files and passphrase
2804 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
2805 update associated structures and add various utility functions.
2807 Add new policy related verify parameters, include policy checking in
2808 standard verify code. Enhance 'smime' application with extra parameters
2809 to support policy checking and print out.
2812 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
2813 Nehemiah processors. These extensions support AES encryption in hardware
2814 as well as RNG (though RNG support is currently disabled).
2815 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
2817 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
2820 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
2821 [Andy Polyakov and a number of other people]
2823 *) Improved PowerPC platform support. Most notably BIGNUM assembler
2824 implementation contributed by IBM.
2825 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
2827 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
2828 exponent rather than 'unsigned long'. There is a corresponding change to
2829 the new 'rsa_keygen' element of the RSA_METHOD structure.
2830 [Jelte Jansen, Geoff Thorpe]
2832 *) Functionality for creating the initial serial number file is now
2833 moved from CA.pl to the 'ca' utility with a new option -create_serial.
2835 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
2836 number file to 1, which is bound to cause problems. To avoid
2837 the problems while respecting compatibility between different 0.9.7
2838 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
2839 CA.pl for serial number initialization. With the new release 0.9.8,
2840 we can fix the problem directly in the 'ca' utility.)
2843 *) Reduced header interdepencies by declaring more opaque objects in
2844 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
2845 give fewer recursive includes, which could break lazy source code - so
2846 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
2847 developers should define this symbol when building and using openssl to
2848 ensure they track the recommended behaviour, interfaces, [etc], but
2849 backwards-compatible behaviour prevails when this isn't defined.
2852 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
2855 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
2856 This will generate a random key of the appropriate length based on the
2857 cipher context. The EVP_CIPHER can provide its own random key generation
2858 routine to support keys of a specific form. This is used in the des and
2859 3des routines to generate a key of the correct parity. Update S/MIME
2860 code to use new functions and hence generate correct parity DES keys.
2861 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
2862 valid (weak or incorrect parity).
2865 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
2866 as looking them up. This is useful when the verified structure may contain
2867 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
2868 present unless the new PKCS7_NO_CRL flag is asserted.
2871 *) Extend ASN1 oid configuration module. It now additionally accepts the
2874 shortName = some long name, 1.2.3.4
2877 *) Reimplemented the BN_CTX implementation. There is now no more static
2878 limitation on the number of variables it can handle nor the depth of the
2879 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
2880 information can now expand as required, and rather than having a single
2881 static array of bignums, BN_CTX now uses a linked-list of such arrays
2882 allowing it to expand on demand whilst maintaining the usefulness of
2883 BN_CTX's "bundling".
2886 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
2887 to allow all RSA operations to function using a single BN_CTX.
2890 *) Preliminary support for certificate policy evaluation and checking. This
2891 is initially intended to pass the tests outlined in "Conformance Testing
2892 of Relying Party Client Certificate Path Processing Logic" v1.07.
2895 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
2896 remained unused and not that useful. A variety of other little bignum
2897 tweaks and fixes have also been made continuing on from the audit (see
2901 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
2902 associated ASN1, EVP and SSL functions and old ASN1 macros.
2905 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
2906 and this should never fail. So the return value from the use of
2907 BN_set_word() (which can fail due to needless expansion) is now deprecated;
2908 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
2911 *) BN_CTX_get() should return zero-valued bignums, providing the same
2912 initialised value as BN_new().
2913 [Geoff Thorpe, suggested by Ulf Möller]
2915 *) Support for inhibitAnyPolicy certificate extension.
2918 *) An audit of the BIGNUM code is underway, for which debugging code is
2919 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
2920 is considered valid when processing BIGNUMs, and causes execution to
2921 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
2922 further steps are taken to deliberately pollute unused data in BIGNUM
2923 structures to try and expose faulty code further on. For now, openssl will
2924 (in its default mode of operation) continue to tolerate the inconsistent
2925 forms that it has tolerated in the past, but authors and packagers should
2926 consider trying openssl and their own applications when compiled with
2927 these debugging symbols defined. It will help highlight potential bugs in
2928 their own code, and will improve the test coverage for OpenSSL itself. At
2929 some point, these tighter rules will become openssl's default to improve
2930 maintainability, though the assert()s and other overheads will remain only
2931 in debugging configurations. See bn.h for more details.
2932 [Geoff Thorpe, Nils Larsch, Ulf Möller]
2934 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
2935 that can only be obtained through BN_CTX_new() (which implicitly
2936 initialises it). The presence of this function only made it possible
2937 to overwrite an existing structure (and cause memory leaks).
2940 *) Because of the callback-based approach for implementing LHASH as a
2941 template type, lh_insert() adds opaque objects to hash-tables and
2942 lh_doall() or lh_doall_arg() are typically used with a destructor callback
2943 to clean up those corresponding objects before destroying the hash table
2944 (and losing the object pointers). So some over-zealous constifications in
2945 LHASH have been relaxed so that lh_insert() does not take (nor store) the
2946 objects as "const" and the lh_doall[_arg] callback wrappers are not
2947 prototyped to have "const" restrictions on the object pointers they are
2948 given (and so aren't required to cast them away any more).
2951 *) The tmdiff.h API was so ugly and minimal that our own timing utility
2952 (speed) prefers to use its own implementation. The two implementations
2953 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
2954 its object type properly exposed (MS_TM) instead of casting to/from "char
2955 *". This may still change yet if someone realises MS_TM and "ms_time_***"
2956 aren't necessarily the greatest nomenclatures - but this is what was used
2957 internally to the implementation so I've used that for now.
2960 *) Ensure that deprecated functions do not get compiled when
2961 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
2962 the self-tests were still using deprecated key-generation functions so
2963 these have been updated also.
2966 *) Reorganise PKCS#7 code to separate the digest location functionality
2967 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
2968 New function PKCS7_set_digest() to set the digest type for PKCS#7
2969 digestedData type. Add additional code to correctly generate the
2970 digestedData type and add support for this type in PKCS7 initialization
2974 *) New function PKCS7_set0_type_other() this initializes a PKCS7
2975 structure of type "other".
2978 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
2979 sure the loop does correctly stop and breaking ("division by zero")
2980 modulus operations are not performed. The (pre-generated) prime
2981 table crypto/bn/bn_prime.h was already correct, but it could not be
2982 re-generated on some platforms because of the "division by zero"
2983 situation in the script.
2984 [Ralf S. Engelschall]
2986 *) Update support for ECC-based TLS ciphersuites according to
2987 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
2988 SHA-1 now is only used for "small" curves (where the
2989 representation of a field element takes up to 24 bytes); for
2990 larger curves, the field element resulting from ECDH is directly
2991 used as premaster secret.
2992 [Douglas Stebila (Sun Microsystems Laboratories)]
2994 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
2995 curve secp160r1 to the tests.
2996 [Douglas Stebila (Sun Microsystems Laboratories)]
2998 *) Add the possibility to load symbols globally with DSO.
2999 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
3001 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
3002 control of the error stack.
3005 *) Add support for STORE in ENGINE.
3008 *) Add the STORE type. The intention is to provide a common interface
3009 to certificate and key stores, be they simple file-based stores, or
3010 HSM-type store, or LDAP stores, or...
3011 NOTE: The code is currently UNTESTED and isn't really used anywhere.
3014 *) Add a generic structure called OPENSSL_ITEM. This can be used to
3015 pass a list of arguments to any function as well as provide a way
3016 for a function to pass data back to the caller.
3019 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
3020 works like BUF_strdup() but can be used to duplicate a portion of
3021 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
3025 *) Add the function sk_find_ex() which works like sk_find(), but will
3026 return an index to an element even if an exact match couldn't be
3027 found. The index is guaranteed to point at the element where the
3028 searched-for key would be inserted to preserve sorting order.
3031 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
3032 takes an extra flags argument for optional functionality. Currently,
3033 the following flags are defined:
3035 OBJ_BSEARCH_VALUE_ON_NOMATCH
3036 This one gets OBJ_bsearch_ex() to return a pointer to the first
3037 element where the comparing function returns a negative or zero
3040 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
3041 This one gets OBJ_bsearch_ex() to return a pointer to the first
3042 element where the comparing function returns zero. This is useful
3043 if there are more than one element where the comparing function
3047 *) Make it possible to create self-signed certificates with 'openssl ca'
3048 in such a way that the self-signed certificate becomes part of the
3049 CA database and uses the same mechanisms for serial number generation
3050 as all other certificate signing. The new flag '-selfsign' enables
3051 this functionality. Adapt CA.sh and CA.pl.in.
3054 *) Add functionality to check the public key of a certificate request
3055 against a given private. This is useful to check that a certificate
3056 request can be signed by that key (self-signing).
3059 *) Make it possible to have multiple active certificates with the same
3060 subject in the CA index file. This is done only if the keyword
3061 'unique_subject' is set to 'no' in the main CA section (default
3062 if 'CA_default') of the configuration file. The value is saved
3063 with the database itself in a separate index attribute file,
3064 named like the index file with '.attr' appended to the name.
3067 *) Generate muti valued AVAs using '+' notation in config files for
3071 *) Support for nameConstraints certificate extension.
3074 *) Support for policyConstraints certificate extension.
3077 *) Support for policyMappings certificate extension.
3080 *) Make sure the default DSA_METHOD implementation only uses its
3081 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
3082 and change its own handlers to be NULL so as to remove unnecessary
3083 indirection. This lets alternative implementations fallback to the
3084 default implementation more easily.
3087 *) Support for directoryName in GeneralName related extensions
3091 *) Make it possible to link applications using Makefile.shared.
3092 Make that possible even when linking against static libraries!
3095 *) Support for single pass processing for S/MIME signing. This now
3096 means that S/MIME signing can be done from a pipe, in addition
3097 cleartext signing (multipart/signed type) is effectively streaming
3098 and the signed data does not need to be all held in memory.
3100 This is done with a new flag PKCS7_STREAM. When this flag is set
3101 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
3102 is done after the data is output (and digests calculated) in
3103 SMIME_write_PKCS7().
3106 *) Add full support for -rpath/-R, both in shared libraries and
3107 applications, at least on the platforms where it's known how
3111 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
3112 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
3113 will now compute a table of multiples of the generator that
3114 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
3115 faster (notably in the case of a single point multiplication,
3116 scalar * generator).
3117 [Nils Larsch, Bodo Moeller]
3119 *) IPv6 support for certificate extensions. The various extensions
3120 which use the IP:a.b.c.d can now take IPv6 addresses using the
3121 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
3125 *) Added an ENGINE that implements RSA by performing private key
3126 exponentiations with the GMP library. The conversions to and from
3127 GMP's mpz_t format aren't optimised nor are any montgomery forms
3128 cached, and on x86 it appears OpenSSL's own performance has caught up.
3129 However there are likely to be other architectures where GMP could
3130 provide a boost. This ENGINE is not built in by default, but it can be
3131 specified at Configure time and should be accompanied by the necessary
3132 linker additions, eg;
3133 ./config -DOPENSSL_USE_GMP -lgmp
3136 *) "openssl engine" will not display ENGINE/DSO load failure errors when
3137 testing availability of engines with "-t" - the old behaviour is
3138 produced by increasing the feature's verbosity with "-tt".
3141 *) ECDSA routines: under certain error conditions uninitialized BN objects
3142 could be freed. Solution: make sure initialization is performed early
3143 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
3147 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
3148 and DH_METHOD (eg. by ENGINE implementations) to override the normal
3149 software implementations. For DSA and DH, parameter generation can
3150 also be overriden by providing the appropriate method callbacks.
3153 *) Change the "progress" mechanism used in key-generation and
3154 primality testing to functions that take a new BN_GENCB pointer in
3155 place of callback/argument pairs. The new API functions have "_ex"
3156 postfixes and the older functions are reimplemented as wrappers for
3157 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
3158 declarations of the old functions to help (graceful) attempts to
3159 migrate to the new functions. Also, the new key-generation API
3160 functions operate on a caller-supplied key-structure and return
3161 success/failure rather than returning a key or NULL - this is to
3162 help make "keygen" another member function of RSA_METHOD etc.
3164 Example for using the new callback interface:
3166 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
3170 BN_GENCB_set(&my_cb, my_callback, my_arg);
3172 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
3173 /* For the meaning of a, b in calls to my_callback(), see the
3174 * documentation of the function that calls the callback.
3175 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
3176 * my_callback should return 1 if it wants BN_is_prime_ex()
3177 * to continue, or 0 to stop.
3182 *) Change the ZLIB compression method to be stateful, and make it
3183 available to TLS with the number defined in
3184 draft-ietf-tls-compression-04.txt.
3187 *) Add the ASN.1 structures and functions for CertificatePair, which
3188 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
3190 CertificatePair ::= SEQUENCE {
3191 forward [0] Certificate OPTIONAL,
3192 reverse [1] Certificate OPTIONAL,
3193 -- at least one of the pair shall be present -- }
3195 Also implement the PEM functions to read and write certificate
3196 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
3198 This needed to be defined, mostly for the sake of the LDAP
3199 attribute crossCertificatePair, but may prove useful elsewhere as
3203 *) Make it possible to inhibit symlinking of shared libraries in
3204 Makefile.shared, for Cygwin's sake.
3207 *) Extend the BIGNUM API by creating a function
3208 void BN_set_negative(BIGNUM *a, int neg);
3209 and a macro that behave like
3210 int BN_is_negative(const BIGNUM *a);
3212 to avoid the need to access 'a->neg' directly in applications.
3215 *) Implement fast modular reduction for pseudo-Mersenne primes
3216 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
3217 EC_GROUP_new_curve_GFp() will now automatically use this
3219 [Nils Larsch <nla@trustcenter.de>]
3221 *) Add new lock type (CRYPTO_LOCK_BN).
3224 *) Change the ENGINE framework to automatically load engines
3225 dynamically from specific directories unless they could be
3226 found to already be built in or loaded. Move all the
3227 current engines except for the cryptodev one to a new
3229 The engines in engines/ are built as shared libraries if
3230 the "shared" options was given to ./Configure or ./config.
3231 Otherwise, they are inserted in libcrypto.a.
3232 /usr/local/ssl/engines is the default directory for dynamic
3233 engines, but that can be overriden at configure time through
3234 the usual use of --prefix and/or --openssldir, and at run
3235 time with the environment variable OPENSSL_ENGINES.
3236 [Geoff Thorpe and Richard Levitte]
3238 *) Add Makefile.shared, a helper makefile to build shared
3239 libraries. Addapt Makefile.org.
3242 *) Add version info to Win32 DLLs.
3243 [Peter 'Luna' Runestig" <peter@runestig.com>]
3245 *) Add new 'medium level' PKCS#12 API. Certificates and keys
3246 can be added using this API to created arbitrary PKCS#12
3247 files while avoiding the low level API.
3249 New options to PKCS12_create(), key or cert can be NULL and
3250 will then be omitted from the output file. The encryption
3251 algorithm NIDs can be set to -1 for no encryption, the mac
3252 iteration count can be set to 0 to omit the mac.
3254 Enhance pkcs12 utility by making the -nokeys and -nocerts
3255 options work when creating a PKCS#12 file. New option -nomac
3256 to omit the mac, NONE can be set for an encryption algorithm.
3257 New code is modified to use the enhanced PKCS12_create()
3258 instead of the low level API.
3261 *) Extend ASN1 encoder to support indefinite length constructed
3262 encoding. This can output sequences tags and octet strings in
3263 this form. Modify pk7_asn1.c to support indefinite length
3264 encoding. This is experimental and needs additional code to
3265 be useful, such as an ASN1 bio and some enhanced streaming
3268 Extend template encode functionality so that tagging is passed
3269 down to the template encoder.
3272 *) Let 'openssl req' fail if an argument to '-newkey' is not
3273 recognized instead of using RSA as a default.
3276 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
3277 As these are not official, they are not included in "ALL";
3278 the "ECCdraft" ciphersuite group alias can be used to select them.
3279 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
3281 *) Add ECDH engine support.
3282 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
3284 *) Add ECDH in new directory crypto/ecdh/.
3285 [Douglas Stebila (Sun Microsystems Laboratories)]
3287 *) Let BN_rand_range() abort with an error after 100 iterations
3288 without success (which indicates a broken PRNG).
3291 *) Change BN_mod_sqrt() so that it verifies that the input value
3292 is really the square of the return value. (Previously,
3293 BN_mod_sqrt would show GIGO behaviour.)
3296 *) Add named elliptic curves over binary fields from X9.62, SECG,
3297 and WAP/WTLS; add OIDs that were still missing.
3299 [Sheueling Chang Shantz and Douglas Stebila
3300 (Sun Microsystems Laboratories)]
3302 *) Extend the EC library for elliptic curves over binary fields
3303 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
3306 EC_GF2m_simple_method
3310 EC_GROUP_new_curve_GF2m
3311 EC_GROUP_set_curve_GF2m
3312 EC_GROUP_get_curve_GF2m
3313 EC_POINT_set_affine_coordinates_GF2m
3314 EC_POINT_get_affine_coordinates_GF2m
3315 EC_POINT_set_compressed_coordinates_GF2m
3317 Point compression for binary fields is disabled by default for
3318 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
3321 As binary polynomials are represented as BIGNUMs, various members
3322 of the EC_GROUP and EC_POINT data structures can be shared
3323 between the implementations for prime fields and binary fields;
3324 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
3325 are essentially identical to their ..._GFp counterparts.
3326 (For simplicity, the '..._GFp' prefix has been dropped from
3327 various internal method names.)
3329 An internal 'field_div' method (similar to 'field_mul' and
3330 'field_sqr') has been added; this is used only for binary fields.
3332 [Sheueling Chang Shantz and Douglas Stebila
3333 (Sun Microsystems Laboratories)]
3335 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
3336 through methods ('mul', 'precompute_mult').
3338 The generic implementations (now internally called 'ec_wNAF_mul'
3339 and 'ec_wNAF_precomputed_mult') remain the default if these
3340 methods are undefined.
3342 [Sheueling Chang Shantz and Douglas Stebila
3343 (Sun Microsystems Laboratories)]
3345 *) New function EC_GROUP_get_degree, which is defined through
3346 EC_METHOD. For curves over prime fields, this returns the bit
3347 length of the modulus.
3349 [Sheueling Chang Shantz and Douglas Stebila
3350 (Sun Microsystems Laboratories)]
3352 *) New functions EC_GROUP_dup, EC_POINT_dup.
3353 (These simply call ..._new and ..._copy).
3355 [Sheueling Chang Shantz and Douglas Stebila
3356 (Sun Microsystems Laboratories)]
3358 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
3359 Polynomials are represented as BIGNUMs (where the sign bit is not
3360 used) in the following functions [macros]:
3363 BN_GF2m_sub [= BN_GF2m_add]
3364 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
3365 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
3366 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
3368 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
3369 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
3370 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
3371 BN_GF2m_cmp [= BN_ucmp]
3373 (Note that only the 'mod' functions are actually for fields GF(2^m).
3374 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
3376 For some functions, an the irreducible polynomial defining a
3377 field can be given as an 'unsigned int[]' with strictly
3378 decreasing elements giving the indices of those bits that are set;
3379 i.e., p[] represents the polynomial
3380 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
3382 p[0] > p[1] > ... > p[k] = 0.
3383 This applies to the following functions:
3388 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
3389 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
3391 BN_GF2m_mod_sqrt_arr
3392 BN_GF2m_mod_solve_quad_arr
3396 Conversion can be performed by the following functions:
3401 bntest.c has additional tests for binary polynomial arithmetic.
3403 Two implementations for BN_GF2m_mod_div() are available.
3404 The default algorithm simply uses BN_GF2m_mod_inv() and
3405 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
3406 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
3407 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
3409 [Sheueling Chang Shantz and Douglas Stebila
3410 (Sun Microsystems Laboratories)]
3412 *) Add new error code 'ERR_R_DISABLED' that can be used when some
3413 functionality is disabled at compile-time.
3414 [Douglas Stebila <douglas.stebila@sun.com>]
3416 *) Change default behaviour of 'openssl asn1parse' so that more
3417 information is visible when viewing, e.g., a certificate:
3419 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
3420 mode the content of non-printable OCTET STRINGs is output in a
3421 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
3422 avoid the appearance of a printable string.
3423 [Nils Larsch <nla@trustcenter.de>]
3425 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
3427 EC_GROUP_set_asn1_flag()
3428 EC_GROUP_get_asn1_flag()
3429 EC_GROUP_set_point_conversion_form()
3430 EC_GROUP_get_point_conversion_form()
3431 These control ASN1 encoding details:
3432 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
3433 has been set to OPENSSL_EC_NAMED_CURVE.
3434 - Points are encoded in uncompressed form by default; options for
3435 asn1_for are as for point2oct, namely
3436 POINT_CONVERSION_COMPRESSED
3437 POINT_CONVERSION_UNCOMPRESSED
3438 POINT_CONVERSION_HYBRID
3440 Also add 'seed' and 'seed_len' members to EC_GROUP with access
3443 EC_GROUP_get0_seed()
3444 EC_GROUP_get_seed_len()
3445 This is used only for ASN1 purposes (so far).
3446 [Nils Larsch <nla@trustcenter.de>]
3448 *) Add 'field_type' member to EC_METHOD, which holds the NID
3449 of the appropriate field type OID. The new function
3450 EC_METHOD_get_field_type() returns this value.
3451 [Nils Larsch <nla@trustcenter.de>]
3456 EC_POINT_point2hex()
3457 EC_POINT_hex2point()
3458 providing useful interfaces to EC_POINT_point2oct() and
3459 EC_POINT_oct2point().
3460 [Nils Larsch <nla@trustcenter.de>]
3462 *) Change internals of the EC library so that the functions
3463 EC_GROUP_set_generator()
3464 EC_GROUP_get_generator()
3465 EC_GROUP_get_order()
3466 EC_GROUP_get_cofactor()
3467 are implemented directly in crypto/ec/ec_lib.c and not dispatched
3468 to methods, which would lead to unnecessary code duplication when
3469 adding different types of curves.
3470 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
3472 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
3473 arithmetic, and such that modified wNAFs are generated
3474 (which avoid length expansion in many cases).
3477 *) Add a function EC_GROUP_check_discriminant() (defined via
3478 EC_METHOD) that verifies that the curve discriminant is non-zero.
3480 Add a function EC_GROUP_check() that makes some sanity tests
3481 on a EC_GROUP, its generator and order. This includes
3482 EC_GROUP_check_discriminant().
3483 [Nils Larsch <nla@trustcenter.de>]
3485 *) Add ECDSA in new directory crypto/ecdsa/.
3487 Add applications 'openssl ecparam' and 'openssl ecdsa'
3488 (these are based on 'openssl dsaparam' and 'openssl dsa').
3490 ECDSA support is also included in various other files across the
3491 library. Most notably,
3492 - 'openssl req' now has a '-newkey ecdsa:file' option;
3493 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
3494 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
3495 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
3496 them suitable for ECDSA where domain parameters must be
3497 extracted before the specific public key;
3498 - ECDSA engine support has been added.
3499 [Nils Larsch <nla@trustcenter.de>]
3501 *) Include some named elliptic curves, and add OIDs from X9.62,
3502 SECG, and WAP/WTLS. Each curve can be obtained from the new
3504 EC_GROUP_new_by_curve_name(),
3505 and the list of available named curves can be obtained with
3506 EC_get_builtin_curves().
3507 Also add a 'curve_name' member to EC_GROUP objects, which can be
3509 EC_GROUP_set_curve_name()
3510 EC_GROUP_get_curve_name()
3511 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
3513 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3514 was actually never needed) and in BN_mul(). The removal in BN_mul()
3515 required a small change in bn_mul_part_recursive() and the addition
3516 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3517 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3518 bn_sub_words() and bn_add_words() except they take arrays with
3522 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
3524 *) Cleanse PEM buffers before freeing them since they may contain
3526 [Benjamin Bennett <ben@psc.edu>]
3528 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3529 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3530 authentication-only ciphersuites.
3533 *) Since AES128 and AES256 share a single mask bit in the logic of
3534 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3535 kludge to work properly if AES128 is available and AES256 isn't.
3538 *) Expand security boundary to match 1.1.1 module.
3541 *) Remove redundant features: hash file source, editing of test vectors
3542 modify fipsld to use external fips_premain.c signature.
3545 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
3546 run algorithm test programs.
3549 *) Make algorithm test programs more tolerant of whitespace.
3552 *) Have SSL/TLS server implementation tolerate "mismatched" record
3553 protocol version while receiving ClientHello even if the
3554 ClientHello is fragmented. (The server can't insist on the
3555 particular protocol version it has chosen before the ServerHello
3556 message has informed the client about his choice.)
3559 *) Load error codes if they are not already present instead of using a
3560 static variable. This allows them to be cleanly unloaded and reloaded.
3563 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3565 *) Introduce limits to prevent malicious keys being able to
3566 cause a denial of service. (CVE-2006-2940)
3567 [Steve Henson, Bodo Moeller]
3569 *) Fix ASN.1 parsing of certain invalid structures that can result
3570 in a denial of service. (CVE-2006-2937) [Steve Henson]
3572 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3573 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3575 *) Fix SSL client code which could crash if connecting to a
3576 malicious SSLv2 server. (CVE-2006-4343)
3577 [Tavis Ormandy and Will Drewry, Google Security Team]
3579 *) Change ciphersuite string processing so that an explicit
3580 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3581 will no longer include "AES128-SHA"), and any other similar
3582 ciphersuite (same bitmap) from *other* protocol versions (so that
3583 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3584 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3585 changes from 0.9.8b and 0.9.8d.
3588 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
3590 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3591 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3593 *) Change the Unix randomness entropy gathering to use poll() when
3594 possible instead of select(), since the latter has some
3595 undesirable limitations.
3596 [Darryl Miles via Richard Levitte and Bodo Moeller]
3598 *) Disable rogue ciphersuites:
3600 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3601 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3602 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3604 The latter two were purportedly from
3605 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3608 Also deactive the remaining ciphersuites from
3609 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3610 unofficial, and the ID has long expired.
3613 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3614 dual-core machines) and other potential thread-safety issues.
3617 Changes between 0.9.7i and 0.9.7j [04 May 2006]
3619 *) Adapt fipsld and the build system to link against the validated FIPS
3620 module in FIPS mode.
3623 *) Fixes for VC++ 2005 build under Windows.
3626 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
3627 from a Windows bash shell such as MSYS. It is autodetected from the
3628 "config" script when run from a VC++ environment. Modify standard VC++
3629 build to use fipscanister.o from the GNU make build.
3632 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
3634 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
3635 The value now differs depending on if you build for FIPS or not.
3636 BEWARE! A program linked with a shared FIPSed libcrypto can't be
3637 safely run with a non-FIPSed libcrypto, as it may crash because of
3638 the difference induced by this change.
3641 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
3643 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3644 (part of SSL_OP_ALL). This option used to disable the
3645 countermeasure against man-in-the-middle protocol-version
3646 rollback in the SSL 2.0 server implementation, which is a bad
3647 idea. (CVE-2005-2969)
3649 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3650 for Information Security, National Institute of Advanced Industrial
3651 Science and Technology [AIST], Japan)]
3653 *) Minimal support for X9.31 signatures and PSS padding modes. This is
3654 mainly for FIPS compliance and not fully integrated at this stage.
3657 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
3658 the exponentiation using a fixed-length exponent. (Otherwise,
3659 the information leaked through timing could expose the secret key
3660 after many signatures; cf. Bleichenbacher's attack on DSA with
3664 *) Make a new fixed-window mod_exp implementation the default for
3665 RSA, DSA, and DH private-key operations so that the sequence of
3666 squares and multiplies and the memory access pattern are
3667 independent of the particular secret key. This will mitigate
3668 cache-timing and potential related attacks.
3670 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
3671 and this is automatically used by BN_mod_exp_mont() if the new flag
3672 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
3673 will use this BN flag for private exponents unless the flag
3674 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
3675 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
3677 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
3679 *) Change the client implementation for SSLv23_method() and
3680 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
3681 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
3682 (Previously, the SSL 2.0 backwards compatible Client Hello
3683 message format would be used even with SSL_OP_NO_SSLv2.)
3686 *) Add support for smime-type MIME parameter in S/MIME messages which some
3690 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
3691 a threadsafe manner. Modify rsa code to use new function and add calls
3692 to dsa and dh code (which had race conditions before).
3695 *) Include the fixed error library code in the C error file definitions
3696 instead of fixing them up at runtime. This keeps the error code
3697 structures constant.
3700 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
3702 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
3705 *) Fixes for newer kerberos headers. NB: the casts are needed because
3706 the 'length' field is signed on one version and unsigned on another
3707 with no (?) obvious way to tell the difference, without these VC++
3708 complains. Also the "definition" of FAR (blank) is no longer included
3709 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
3710 some needed definitions.
3713 *) Undo Cygwin change.
3716 *) Added support for proxy certificates according to RFC 3820.
3717 Because they may be a security thread to unaware applications,
3718 they must be explicitely allowed in run-time. See
3719 docs/HOWTO/proxy_certificates.txt for further information.
3722 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
3724 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
3725 server and client random values. Previously
3726 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
3727 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
3729 This change has negligible security impact because:
3731 1. Server and client random values still have 24 bytes of pseudo random
3734 2. Server and client random values are sent in the clear in the initial
3737 3. The master secret is derived using the premaster secret (48 bytes in
3738 size for static RSA ciphersuites) as well as client server and random
3741 The OpenSSL team would like to thank the UK NISCC for bringing this issue
3744 [Stephen Henson, reported by UK NISCC]
3746 *) Use Windows randomness collection on Cygwin.
3749 *) Fix hang in EGD/PRNGD query when communication socket is closed
3750 prematurely by EGD/PRNGD.
3751 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
3753 *) Prompt for pass phrases when appropriate for PKCS12 input format.
3756 *) Back-port of selected performance improvements from development
3757 branch, as well as improved support for PowerPC platforms.
3760 *) Add lots of checks for memory allocation failure, error codes to indicate
3761 failure and freeing up memory if a failure occurs.
3762 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
3764 *) Add new -passin argument to dgst.
3767 *) Perform some character comparisons of different types in X509_NAME_cmp:
3768 this is needed for some certificates that reencode DNs into UTF8Strings
3769 (in violation of RFC3280) and can't or wont issue name rollover
3773 *) Make an explicit check during certificate validation to see that
3774 the CA setting in each certificate on the chain is correct. As a
3775 side effect always do the following basic checks on extensions,
3776 not just when there's an associated purpose to the check:
3778 - if there is an unhandled critical extension (unless the user
3779 has chosen to ignore this fault)
3780 - if the path length has been exceeded (if one is set at all)
3781 - that certain extensions fit the associated purpose (if one has
3785 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
3787 *) Avoid a race condition when CRLs are checked in a multi threaded
3788 environment. This would happen due to the reordering of the revoked
3789 entries during signature checking and serial number lookup. Now the
3790 encoding is cached and the serial number sort performed under a lock.
3791 Add new STACK function sk_is_sorted().
3794 *) Add Delta CRL to the extension code.
3797 *) Various fixes to s3_pkt.c so alerts are sent properly.
3798 [David Holmes <d.holmes@f5.com>]
3800 *) Reduce the chances of duplicate issuer name and serial numbers (in
3801 violation of RFC3280) using the OpenSSL certificate creation utilities.
3802 This is done by creating a random 64 bit value for the initial serial
3803 number when a serial number file is created or when a self signed
3804 certificate is created using 'openssl req -x509'. The initial serial
3805 number file is created using 'openssl x509 -next_serial' in CA.pl
3806 rather than being initialized to 1.
3809 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
3811 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3812 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3813 [Joe Orton, Steve Henson]
3815 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
3817 [Joe Orton, Steve Henson]
3819 *) Make it possible to have multiple active certificates with the same
3820 subject in the CA index file. This is done only if the keyword
3821 'unique_subject' is set to 'no' in the main CA section (default
3822 if 'CA_default') of the configuration file. The value is saved
3823 with the database itself in a separate index attribute file,
3824 named like the index file with '.attr' appended to the name.
3827 *) X509 verify fixes. Disable broken certificate workarounds when
3828 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
3829 keyUsage extension present. Don't accept CRLs with unhandled critical
3830 extensions: since verify currently doesn't process CRL extensions this
3831 rejects a CRL with *any* critical extensions. Add new verify error codes
3835 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
3836 A clarification of RFC2560 will require the use of OCTET STRINGs and
3837 some implementations cannot handle the current raw format. Since OpenSSL
3838 copies and compares OCSP nonces as opaque blobs without any attempt at
3839 parsing them this should not create any compatibility issues.
3842 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
3843 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
3844 this HMAC (and other) operations are several times slower than OpenSSL
3848 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
3849 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3851 *) Use the correct content when signing type "other".
3854 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
3856 *) Fix various bugs revealed by running the NISCC test suite:
3858 Stop out of bounds reads in the ASN1 code when presented with
3859 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3861 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
3863 If verify callback ignores invalid public key errors don't try to check
3864 certificate signature with the NULL public key.
3868 *) New -ignore_err option in ocsp application to stop the server
3869 exiting on the first error in a request.
3872 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3873 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3877 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3878 extra data after the compression methods not only for TLS 1.0
3879 but also for SSL 3.0 (as required by the specification).
3880 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3882 *) Change X509_certificate_type() to mark the key as exported/exportable
3883 when it's 512 *bits* long, not 512 bytes.
3886 *) Change AES_cbc_encrypt() so it outputs exact multiple of
3887 blocks during encryption.
3890 *) Various fixes to base64 BIO and non blocking I/O. On write
3891 flushes were not handled properly if the BIO retried. On read
3892 data was not being buffered properly and had various logic bugs.
3893 This also affects blocking I/O when the data being decoded is a
3897 *) Various S/MIME bugfixes and compatibility changes:
3898 output correct application/pkcs7 MIME type if
3899 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
3900 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
3901 of files as .eml work). Correctly handle very long lines in MIME
3905 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
3907 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3908 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3909 a protocol version number mismatch like a decryption error
3910 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3913 *) Turn on RSA blinding by default in the default implementation
3914 to avoid a timing attack. Applications that don't want it can call
3915 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3916 They would be ill-advised to do so in most cases.
3917 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3919 *) Change RSA blinding code so that it works when the PRNG is not
3920 seeded (in this case, the secret RSA exponent is abused as
3921 an unpredictable seed -- if it is not unpredictable, there
3922 is no point in blinding anyway). Make RSA blinding thread-safe
3923 by remembering the creator's thread ID in rsa->blinding and
3924 having all other threads use local one-time blinding factors
3925 (this requires more computation than sharing rsa->blinding, but
3926 avoids excessive locking; and if an RSA object is not shared
3927 between threads, blinding will still be very fast).
3930 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
3931 ENGINE as defaults for all supported algorithms irrespective of
3932 the 'flags' parameter. 'flags' is now honoured, so applications
3933 should make sure they are passing it correctly.
3936 *) Target "mingw" now allows native Windows code to be generated in
3937 the Cygwin environment as well as with the MinGW compiler.
3940 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
3942 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3943 via timing by performing a MAC computation even if incorrrect
3944 block cipher padding has been found. This is a countermeasure
3945 against active attacks where the attacker has to distinguish
3946 between bad padding and a MAC verification error. (CVE-2003-0078)
3948 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3949 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3950 Martin Vuagnoux (EPFL, Ilion)]
3952 *) Make the no-err option work as intended. The intention with no-err
3953 is not to have the whole error stack handling routines removed from
3954 libcrypto, it's only intended to remove all the function name and
3955 reason texts, thereby removing some of the footprint that may not
3956 be interesting if those errors aren't displayed anyway.
3958 NOTE: it's still possible for any application or module to have it's
3959 own set of error texts inserted. The routines are there, just not
3960 used by default when no-err is given.
3963 *) Add support for FreeBSD on IA64.
3964 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
3966 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
3967 Kerberos function mit_des_cbc_cksum(). Before this change,
3968 the value returned by DES_cbc_cksum() was like the one from
3969 mit_des_cbc_cksum(), except the bytes were swapped.
3970 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
3972 *) Allow an application to disable the automatic SSL chain building.
3973 Before this a rather primitive chain build was always performed in
3974 ssl3_output_cert_chain(): an application had no way to send the
3975 correct chain if the automatic operation produced an incorrect result.
3977 Now the chain builder is disabled if either:
3979 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
3981 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
3983 The reasoning behind this is that an application would not want the
3984 auto chain building to take place if extra chain certificates are
3985 present and it might also want a means of sending no additional
3986 certificates (for example the chain has two certificates and the
3990 *) Add the possibility to build without the ENGINE framework.
3991 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3993 *) Under Win32 gmtime() can return NULL: check return value in
3994 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
3997 *) DSA routines: under certain error conditions uninitialized BN objects
3998 could be freed. Solution: make sure initialization is performed early
3999 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
4000 Nils Larsch <nla@trustcenter.de> via PR#459)
4003 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
4004 checked on reconnect on the client side, therefore session resumption
4005 could still fail with a "ssl session id is different" error. This
4006 behaviour is masked when SSL_OP_ALL is used due to
4007 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
4008 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4009 followup to PR #377.
4012 *) IA-32 assembler support enhancements: unified ELF targets, support
4013 for SCO/Caldera platforms, fix for Cygwin shared build.
4016 *) Add support for FreeBSD on sparc64. As a consequence, support for
4017 FreeBSD on non-x86 processors is separate from x86 processors on
4018 the config script, much like the NetBSD support.
4019 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
4021 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
4023 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
4026 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
4027 code (06) was taken as the first octet of the session ID and the last
4028 octet was ignored consequently. As a result SSLv2 client side session
4029 caching could not have worked due to the session ID mismatch between
4031 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4035 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
4036 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
4040 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
4041 seems that in spite of existing for more than a year, many application
4042 author have done nothing to provide the necessary callbacks, which
4043 means that this particular engine will not work properly anywhere.
4044 This is a very unfortunate situation which forces us, in the name
4045 of usability, to give the hw_ncipher.c a static lock, which is part
4047 NOTE: This is for the 0.9.7 series ONLY. This hack will never
4048 appear in 0.9.8 or later. We EXPECT application authors to have
4049 dealt properly with this when 0.9.8 is released (unless we actually
4050 make such changes in the libcrypto locking code that changes will
4051 have to be made anyway).
4054 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
4055 octets have been read, EOF or an error occurs. Without this change
4056 some truncated ASN1 structures will not produce an error.
4059 *) Disable Heimdal support, since it hasn't been fully implemented.
4060 Still give the possibility to force the use of Heimdal, but with
4061 warnings and a request that patches get sent to openssl-dev.
4064 *) Add the VC-CE target, introduce the WINCE sysname, and add
4065 INSTALL.WCE and appropriate conditionals to make it build.
4066 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4068 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
4069 cygssl-x.y.z.dll, where x, y and z are the major, minor and
4070 edit numbers of the version.
4071 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4073 *) Introduce safe string copy and catenation functions
4074 (BUF_strlcpy() and BUF_strlcat()).
4075 [Ben Laurie (CHATS) and Richard Levitte]
4077 *) Avoid using fixed-size buffers for one-line DNs.
4078 [Ben Laurie (CHATS)]
4080 *) Add BUF_MEM_grow_clean() to avoid information leakage when
4081 resizing buffers containing secrets, and use where appropriate.
4082 [Ben Laurie (CHATS)]
4084 *) Avoid using fixed size buffers for configuration file location.
4085 [Ben Laurie (CHATS)]
4087 *) Avoid filename truncation for various CA files.
4088 [Ben Laurie (CHATS)]
4090 *) Use sizeof in preference to magic numbers.
4091 [Ben Laurie (CHATS)]
4093 *) Avoid filename truncation in cert requests.
4094 [Ben Laurie (CHATS)]
4096 *) Add assertions to check for (supposedly impossible) buffer
4098 [Ben Laurie (CHATS)]
4100 *) Don't cache truncated DNS entries in the local cache (this could
4101 potentially lead to a spoofing attack).
4102 [Ben Laurie (CHATS)]
4104 *) Fix various buffers to be large enough for hex/decimal
4105 representations in a platform independent manner.
4106 [Ben Laurie (CHATS)]
4108 *) Add CRYPTO_realloc_clean() to avoid information leakage when
4109 resizing buffers containing secrets, and use where appropriate.
4110 [Ben Laurie (CHATS)]
4112 *) Add BIO_indent() to avoid much slightly worrying code to do
4114 [Ben Laurie (CHATS)]
4116 *) Convert sprintf()/BIO_puts() to BIO_printf().
4117 [Ben Laurie (CHATS)]
4119 *) buffer_gets() could terminate with the buffer only half
4121 [Ben Laurie (CHATS)]
4123 *) Add assertions to prevent user-supplied crypto functions from
4124 overflowing internal buffers by having large block sizes, etc.
4125 [Ben Laurie (CHATS)]
4127 *) New OPENSSL_assert() macro (similar to assert(), but enabled
4129 [Ben Laurie (CHATS)]
4131 *) Eliminate unused copy of key in RC4.
4132 [Ben Laurie (CHATS)]
4134 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
4135 [Ben Laurie (CHATS)]
4137 *) Fix off-by-one error in EGD path.
4138 [Ben Laurie (CHATS)]
4140 *) If RANDFILE path is too long, ignore instead of truncating.
4141 [Ben Laurie (CHATS)]
4143 *) Eliminate unused and incorrectly sized X.509 structure
4145 [Ben Laurie (CHATS)]
4147 *) Eliminate unused and dangerous function knumber().
4148 [Ben Laurie (CHATS)]
4150 *) Eliminate unused and dangerous structure, KSSL_ERR.
4151 [Ben Laurie (CHATS)]
4153 *) Protect against overlong session ID context length in an encoded
4154 session object. Since these are local, this does not appear to be
4156 [Ben Laurie (CHATS)]
4158 *) Change from security patch (see 0.9.6e below) that did not affect
4159 the 0.9.6 release series:
4161 Remote buffer overflow in SSL3 protocol - an attacker could
4162 supply an oversized master key in Kerberos-enabled versions.
4164 [Ben Laurie (CHATS)]
4166 *) Change the SSL kerb5 codes to match RFC 2712.
4169 *) Make -nameopt work fully for req and add -reqopt switch.
4170 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
4172 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
4173 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
4175 *) Make sure tests can be performed even if the corresponding algorithms
4176 have been removed entirely. This was also the last step to make
4177 OpenSSL compilable with DJGPP under all reasonable conditions.
4178 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
4180 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
4181 to allow version independent disabling of normally unselected ciphers,
4182 which may be activated as a side-effect of selecting a single cipher.
4184 (E.g., cipher list string "RSA" enables ciphersuites that are left
4185 out of "ALL" because they do not provide symmetric encryption.
4186 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
4187 [Lutz Jaenicke, Bodo Moeller]
4189 *) Add appropriate support for separate platform-dependent build
4190 directories. The recommended way to make a platform-dependent
4191 build directory is the following (tested on Linux), maybe with
4194 # Place yourself outside of the OpenSSL source tree. In
4195 # this example, the environment variable OPENSSL_SOURCE
4196 # is assumed to contain the absolute OpenSSL source directory.
4197 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
4198 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
4199 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
4200 mkdir -p `dirname $F`
4201 ln -s $OPENSSL_SOURCE/$F $F
4204 To be absolutely sure not to disturb the source tree, a "make clean"
4205 is a good thing. If it isn't successfull, don't worry about it,
4206 it probably means the source directory is very clean.
4209 *) Make sure any ENGINE control commands make local copies of string
4210 pointers passed to them whenever necessary. Otherwise it is possible
4211 the caller may have overwritten (or deallocated) the original string
4212 data when a later ENGINE operation tries to use the stored values.
4213 [Götz Babin-Ebell <babinebell@trustcenter.de>]
4215 *) Improve diagnostics in file reading and command-line digests.
4216 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
4218 *) Add AES modes CFB and OFB to the object database. Correct an
4219 error in AES-CFB decryption.
4222 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
4223 allows existing EVP_CIPHER_CTX structures to be reused after
4224 calling EVP_*Final(). This behaviour is used by encryption
4225 BIOs and some applications. This has the side effect that
4226 applications must explicitly clean up cipher contexts with
4227 EVP_CIPHER_CTX_cleanup() or they will leak memory.
4230 *) Check the values of dna and dnb in bn_mul_recursive before calling
4231 bn_mul_comba (a non zero value means the a or b arrays do not contain
4232 n2 elements) and fallback to bn_mul_normal if either is not zero.
4235 *) Fix escaping of non-ASCII characters when using the -subj option
4236 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
4239 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
4240 form for "surname", serialNumber has no short form.
4241 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
4242 therefore remove "mail" short name for "internet 7".
4243 The OID for unique identifiers in X509 certificates is
4244 x500UniqueIdentifier, not uniqueIdentifier.
4245 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
4248 *) Add an "init" command to the ENGINE config module and auto initialize
4249 ENGINEs. Without any "init" command the ENGINE will be initialized
4250 after all ctrl commands have been executed on it. If init=1 the
4251 ENGINE is initailized at that point (ctrls before that point are run
4252 on the uninitialized ENGINE and after on the initialized one). If
4253 init=0 then the ENGINE will not be iniatialized at all.
4256 *) Fix the 'app_verify_callback' interface so that the user-defined
4257 argument is actually passed to the callback: In the
4258 SSL_CTX_set_cert_verify_callback() prototype, the callback
4259 declaration has been changed from
4262 int (*cb)(X509_STORE_CTX *,void *);
4263 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
4264 i=s->ctx->app_verify_callback(&ctx)
4265 has been changed into
4266 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
4268 To update applications using SSL_CTX_set_cert_verify_callback(),
4269 a dummy argument can be added to their callback functions.
4270 [D. K. Smetters <smetters@parc.xerox.com>]
4272 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
4273 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
4275 *) Add and OPENSSL_LOAD_CONF define which will cause
4276 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
4277 This allows older applications to transparently support certain
4278 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
4279 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
4280 load the config file and OPENSSL_add_all_algorithms_conf() which will
4281 always load it have also been added.
4284 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
4285 Adjust NIDs and EVP layer.
4286 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4288 *) Config modules support in openssl utility.
4290 Most commands now load modules from the config file,
4291 though in a few (such as version) this isn't done
4292 because it couldn't be used for anything.
4294 In the case of ca and req the config file used is
4295 the same as the utility itself: that is the -config
4296 command line option can be used to specify an
4300 *) Move default behaviour from OPENSSL_config(). If appname is NULL
4301 use "openssl_conf" if filename is NULL use default openssl config file.
4304 *) Add an argument to OPENSSL_config() to allow the use of an alternative
4305 config section name. Add a new flag to tolerate a missing config file
4306 and move code to CONF_modules_load_file().
4309 *) Support for crypto accelerator cards from Accelerated Encryption
4310 Processing, www.aep.ie. (Use engine 'aep')
4311 The support was copied from 0.9.6c [engine] and adapted/corrected
4312 to work with the new engine framework.
4313 [AEP Inc. and Richard Levitte]
4315 *) Support for SureWare crypto accelerator cards from Baltimore
4316 Technologies. (Use engine 'sureware')
4317 The support was copied from 0.9.6c [engine] and adapted
4318 to work with the new engine framework.
4321 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
4322 make the newer ENGINE framework commands for the CHIL engine work.
4323 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
4325 *) Make it possible to produce shared libraries on ReliantUNIX.
4326 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
4328 *) Add the configuration target debug-linux-ppro.
4329 Make 'openssl rsa' use the general key loading routines
4330 implemented in apps.c, and make those routines able to
4331 handle the key format FORMAT_NETSCAPE and the variant
4333 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4335 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4336 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4338 *) Add -keyform to rsautl, and document -engine.
4339 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
4341 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
4342 BIO_R_NO_SUCH_FILE error code rather than the generic
4343 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
4346 *) Add new functions
4348 ERR_peek_last_error_line
4349 ERR_peek_last_error_line_data.
4350 These are similar to
4353 ERR_peek_error_line_data,
4354 but report on the latest error recorded rather than the first one
4355 still in the error queue.
4356 [Ben Laurie, Bodo Moeller]
4358 *) default_algorithms option in ENGINE config module. This allows things
4360 default_algorithms = ALL
4361 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
4364 *) Prelminary ENGINE config module.
4367 *) New experimental application configuration code.
4370 *) Change the AES code to follow the same name structure as all other
4371 symmetric ciphers, and behave the same way. Move everything to
4372 the directory crypto/aes, thereby obsoleting crypto/rijndael.
4373 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4375 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
4376 [Ben Laurie and Theo de Raadt]
4378 *) Add option to output public keys in req command.
4379 [Massimiliano Pala madwolf@openca.org]
4381 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
4382 (up to about 10% better than before for P-192 and P-224).
4385 *) New functions/macros
4387 SSL_CTX_set_msg_callback(ctx, cb)
4388 SSL_CTX_set_msg_callback_arg(ctx, arg)
4389 SSL_set_msg_callback(ssl, cb)
4390 SSL_set_msg_callback_arg(ssl, arg)
4392 to request calling a callback function
4394 void cb(int write_p, int version, int content_type,
4395 const void *buf, size_t len, SSL *ssl, void *arg)
4397 whenever a protocol message has been completely received
4398 (write_p == 0) or sent (write_p == 1). Here 'version' is the
4399 protocol version according to which the SSL library interprets
4400 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
4401 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
4402 the content type as defined in the SSL 3.0/TLS 1.0 protocol
4403 specification (change_cipher_spec(20), alert(21), handshake(22)).
4404 'buf' and 'len' point to the actual message, 'ssl' to the
4405 SSL object, and 'arg' is the application-defined value set by
4406 SSL[_CTX]_set_msg_callback_arg().
4408 'openssl s_client' and 'openssl s_server' have new '-msg' options
4409 to enable a callback that displays all protocol messages.
4412 *) Change the shared library support so shared libraries are built as
4413 soon as the corresponding static library is finished, and thereby get
4414 openssl and the test programs linked against the shared library.
4415 This still only happens when the keyword "shard" has been given to
4416 the configuration scripts.
4418 NOTE: shared library support is still an experimental thing, and
4419 backward binary compatibility is still not guaranteed.
4420 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
4422 *) Add support for Subject Information Access extension.
4423 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4425 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
4426 additional bytes when new memory had to be allocated, not just
4427 when reusing an existing buffer.
4430 *) New command line and configuration option 'utf8' for the req command.
4431 This allows field values to be specified as UTF8 strings.
4434 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
4435 runs for the former and machine-readable output for the latter.
4438 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
4439 of the e-mail address in the DN (i.e., it will go into a certificate
4440 extension only). The new configuration file option 'email_in_dn = no'
4441 has the same effect.
4442 [Massimiliano Pala madwolf@openca.org]
4444 *) Change all functions with names starting with des_ to be starting
4445 with DES_ instead. Add wrappers that are compatible with libdes,
4446 but are named _ossl_old_des_*. Finally, add macros that map the
4447 des_* symbols to the corresponding _ossl_old_des_* if libdes
4448 compatibility is desired. If OpenSSL 0.9.6c compatibility is
4449 desired, the des_* symbols will be mapped to DES_*, with one
4452 Since we provide two compatibility mappings, the user needs to
4453 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
4454 compatibility is desired. The default (i.e., when that macro
4455 isn't defined) is OpenSSL 0.9.6c compatibility.
4457 There are also macros that enable and disable the support of old
4458 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
4459 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
4460 are defined, the default will apply: to support the old des routines.
4462 In either case, one must include openssl/des.h to get the correct
4463 definitions. Do not try to just include openssl/des_old.h, that
4466 NOTE: This is a major break of an old API into a new one. Software
4467 authors are encouraged to switch to the DES_ style functions. Some
4468 time in the future, des_old.h and the libdes compatibility functions
4469 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
4470 default), and then completely removed.
4473 *) Test for certificates which contain unsupported critical extensions.
4474 If such a certificate is found during a verify operation it is
4475 rejected by default: this behaviour can be overridden by either
4476 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
4477 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
4478 X509_supported_extension() has also been added which returns 1 if a
4479 particular extension is supported.
4482 *) Modify the behaviour of EVP cipher functions in similar way to digests
4483 to retain compatibility with existing code.
4486 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
4487 compatibility with existing code. In particular the 'ctx' parameter does
4488 not have to be to be initialized before the call to EVP_DigestInit() and
4489 it is tidied up after a call to EVP_DigestFinal(). New function
4490 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
4491 EVP_MD_CTX_copy() changed to not require the destination to be
4492 initialized valid and new function EVP_MD_CTX_copy_ex() added which
4493 requires the destination to be valid.
4495 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
4496 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
4499 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
4500 so that complete 'Handshake' protocol structures are kept in memory
4501 instead of overwriting 'msg_type' and 'length' with 'body' data.
4504 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
4505 [Massimo Santin via Richard Levitte]
4507 *) Major restructuring to the underlying ENGINE code. This includes
4508 reduction of linker bloat, separation of pure "ENGINE" manipulation
4509 (initialisation, etc) from functionality dealing with implementations
4510 of specific crypto iterfaces. This change also introduces integrated
4511 support for symmetric ciphers and digest implementations - so ENGINEs
4512 can now accelerate these by providing EVP_CIPHER and EVP_MD
4513 implementations of their own. This is detailed in crypto/engine/README
4514 as it couldn't be adequately described here. However, there are a few
4515 API changes worth noting - some RSA, DSA, DH, and RAND functions that
4516 were changed in the original introduction of ENGINE code have now
4517 reverted back - the hooking from this code to ENGINE is now a good
4518 deal more passive and at run-time, operations deal directly with
4519 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
4520 dereferencing through an ENGINE pointer any more. Also, the ENGINE
4521 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
4522 they were not being used by the framework as there is no concept of a
4523 BIGNUM_METHOD and they could not be generalised to the new
4524 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
4525 ENGINE_cpy() has been removed as it cannot be consistently defined in
4529 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
4532 *) Change mkdef.pl to sort symbols that get the same entry number,
4533 and make sure the automatically generated functions ERR_load_*
4534 become part of libeay.num as well.
4537 *) New function SSL_renegotiate_pending(). This returns true once
4538 renegotiation has been requested (either SSL_renegotiate() call
4539 or HelloRequest/ClientHello receveived from the peer) and becomes
4540 false once a handshake has been completed.
4541 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
4542 sends a HelloRequest, but does not ensure that a handshake takes
4543 place. SSL_renegotiate_pending() is useful for checking if the
4544 client has followed the request.)
4547 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
4548 By default, clients may request session resumption even during
4549 renegotiation (if session ID contexts permit); with this option,
4550 session resumption is possible only in the first handshake.
4552 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4553 more bits available for options that should not be part of
4554 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4557 *) Add some demos for certificate and certificate request creation.
4560 *) Make maximum certificate chain size accepted from the peer application
4561 settable (SSL*_get/set_max_cert_list()), as proposed by
4562 "Douglas E. Engert" <deengert@anl.gov>.
4565 *) Add support for shared libraries for Unixware-7
4566 (Boyd Lynn Gerber <gerberb@zenez.com>).
4569 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4570 be done prior to destruction. Use this to unload error strings from
4571 ENGINEs that load their own error strings. NB: This adds two new API
4572 functions to "get" and "set" this destroy handler in an ENGINE.
4575 *) Alter all existing ENGINE implementations (except "openssl" and
4576 "openbsd") to dynamically instantiate their own error strings. This
4577 makes them more flexible to be built both as statically-linked ENGINEs
4578 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4579 Also, add stub code to each that makes building them as self-contained
4580 shared-libraries easier (see README.ENGINE).
4583 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4584 implementations into applications that are completely implemented in
4585 self-contained shared-libraries. The "dynamic" ENGINE exposes control
4586 commands that can be used to configure what shared-library to load and
4587 to control aspects of the way it is handled. Also, made an update to
4588 the README.ENGINE file that brings its information up-to-date and
4589 provides some information and instructions on the "dynamic" ENGINE
4590 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
4593 *) Make it possible to unload ranges of ERR strings with a new
4594 "ERR_unload_strings" function.
4597 *) Add a copy() function to EVP_MD.
4600 *) Make EVP_MD routines take a context pointer instead of just the
4601 md_data void pointer.
4604 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
4605 that the digest can only process a single chunk of data
4606 (typically because it is provided by a piece of
4607 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
4608 is only going to provide a single chunk of data, and hence the
4609 framework needn't accumulate the data for oneshot drivers.
4612 *) As with "ERR", make it possible to replace the underlying "ex_data"
4613 functions. This change also alters the storage and management of global
4614 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
4615 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
4616 index counters. The API functions that use this state have been changed
4617 to take a "class_index" rather than pointers to the class's local STACK
4618 and counter, and there is now an API function to dynamically create new
4619 classes. This centralisation allows us to (a) plug a lot of the
4620 thread-safety problems that existed, and (b) makes it possible to clean
4621 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
4622 such data would previously have always leaked in application code and
4623 workarounds were in place to make the memory debugging turn a blind eye
4624 to it. Application code that doesn't use this new function will still
4625 leak as before, but their memory debugging output will announce it now
4626 rather than letting it slide.
4628 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
4629 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
4630 has a return value to indicate success or failure.
4633 *) Make it possible to replace the underlying "ERR" functions such that the
4634 global state (2 LHASH tables and 2 locks) is only used by the "default"
4635 implementation. This change also adds two functions to "get" and "set"
4636 the implementation prior to it being automatically set the first time
4637 any other ERR function takes place. Ie. an application can call "get",
4638 pass the return value to a module it has just loaded, and that module
4639 can call its own "set" function using that value. This means the
4640 module's "ERR" operations will use (and modify) the error state in the
4641 application and not in its own statically linked copy of OpenSSL code.
4644 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
4645 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
4646 the operation, and provides a more encapsulated way for external code
4647 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
4648 to use these functions rather than manually incrementing the counts.
4650 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
4653 *) Add EVP test program.
4656 *) Add symmetric cipher support to ENGINE. Expect the API to change!
4659 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
4660 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
4661 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
4662 These allow a CRL to be built without having to access X509_CRL fields
4663 directly. Modify 'ca' application to use new functions.
4666 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
4667 bug workarounds. Rollback attack detection is a security feature.
4668 The problem will only arise on OpenSSL servers when TLSv1 is not
4669 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
4670 Software authors not wanting to support TLSv1 will have special reasons
4671 for their choice and can explicitly enable this option.
4672 [Bodo Moeller, Lutz Jaenicke]
4674 *) Rationalise EVP so it can be extended: don't include a union of
4675 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
4676 (similar to those existing for EVP_CIPHER_CTX).
4681 EVP_MD_CTX_init(&md); /* new function call */
4682 EVP_DigestInit(&md, EVP_sha1());
4683 EVP_DigestUpdate(&md, in, len);
4684 EVP_DigestFinal(&md, out, NULL);
4685 EVP_MD_CTX_cleanup(&md); /* new function call */
4689 *) Make DES key schedule conform to the usual scheme, as well as
4690 correcting its structure. This means that calls to DES functions
4691 now have to pass a pointer to a des_key_schedule instead of a
4692 plain des_key_schedule (which was actually always a pointer
4695 des_key_schedule ks;
4697 des_set_key_checked(..., &ks);
4698 des_ncbc_encrypt(..., &ks, ...);
4700 (Note that a later change renames 'des_...' into 'DES_...'.)
4703 *) Initial reduction of linker bloat: the use of some functions, such as
4704 PEM causes large amounts of unused functions to be linked in due to
4705 poor organisation. For example pem_all.c contains every PEM function
4706 which has a knock on effect of linking in large amounts of (unused)
4707 ASN1 code. Grouping together similar functions and splitting unrelated
4708 functions prevents this.
4711 *) Cleanup of EVP macros.
4714 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
4715 correct _ecb suffix.
4718 *) Add initial OCSP responder support to ocsp application. The
4719 revocation information is handled using the text based index
4720 use by the ca application. The responder can either handle
4721 requests generated internally, supplied in files (for example
4722 via a CGI script) or using an internal minimal server.
4725 *) Add configuration choices to get zlib compression for TLS.
4728 *) Changes to Kerberos SSL for RFC 2712 compliance:
4729 1. Implemented real KerberosWrapper, instead of just using
4730 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
4731 2. Implemented optional authenticator field of KerberosWrapper.
4733 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
4734 and authenticator structs; see crypto/krb5/.
4736 Generalized Kerberos calls to support multiple Kerberos libraries.
4737 [Vern Staats <staatsvr@asc.hpc.mil>,
4738 Jeffrey Altman <jaltman@columbia.edu>
4739 via Richard Levitte]
4741 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
4742 already does with RSA. testdsa.h now has 'priv_key/pub_key'
4743 values for each of the key sizes rather than having just
4744 parameters (and 'speed' generating keys each time).
4747 *) Speed up EVP routines.
4750 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
4751 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
4752 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
4753 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
4755 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
4756 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
4757 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
4760 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
4762 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
4765 *) Added the OS2-EMX target.
4766 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
4768 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
4769 to support NCONF routines in extension code. New function CONF_set_nconf()
4770 to allow functions which take an NCONF to also handle the old LHASH
4771 structure: this means that the old CONF compatible routines can be
4772 retained (in particular wrt extensions) without having to duplicate the
4773 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
4776 *) Enhance the general user interface with mechanisms for inner control
4777 and with possibilities to have yes/no kind of prompts.
4780 *) Change all calls to low level digest routines in the library and
4781 applications to use EVP. Add missing calls to HMAC_cleanup() and
4782 don't assume HMAC_CTX can be copied using memcpy().
4783 [Verdon Walker <VWalker@novell.com>, Steve Henson]
4785 *) Add the possibility to control engines through control names but with
4786 arbitrary arguments instead of just a string.
4787 Change the key loaders to take a UI_METHOD instead of a callback
4788 function pointer. NOTE: this breaks binary compatibility with earlier
4789 versions of OpenSSL [engine].
4790 Adapt the nCipher code for these new conditions and add a card insertion
4794 *) Enhance the general user interface with mechanisms to better support
4795 dialog box interfaces, application-defined prompts, the possibility
4796 to use defaults (for example default passwords from somewhere else)
4797 and interrupts/cancellations.
4800 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
4801 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
4804 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
4805 tidy up some unnecessarily weird code in 'sk_new()').
4806 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
4808 *) Change the key loading routines for ENGINEs to use the same kind
4809 callback (pem_password_cb) as all other routines that need this
4813 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
4814 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
4815 than this minimum value is recommended.
4818 *) New random seeder for OpenVMS, using the system process statistics
4819 that are easily reachable.
4822 *) Windows apparently can't transparently handle global
4823 variables defined in DLLs. Initialisations such as:
4825 const ASN1_ITEM *it = &ASN1_INTEGER_it;
4827 wont compile. This is used by the any applications that need to
4828 declare their own ASN1 modules. This was fixed by adding the option
4829 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
4830 needed for static libraries under Win32.
4833 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
4834 setting of purpose and trust fields. New X509_STORE trust and
4835 purpose functions and tidy up setting in other SSL functions.
4838 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
4839 structure. These are inherited by X509_STORE_CTX when it is
4840 initialised. This allows various defaults to be set in the
4841 X509_STORE structure (such as flags for CRL checking and custom
4842 purpose or trust settings) for functions which only use X509_STORE_CTX
4843 internally such as S/MIME.
4845 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
4846 trust settings if they are not set in X509_STORE. This allows X509_STORE
4847 purposes and trust (in S/MIME for example) to override any set by default.
4849 Add command line options for CRL checking to smime, s_client and s_server
4853 *) Initial CRL based revocation checking. If the CRL checking flag(s)
4854 are set then the CRL is looked up in the X509_STORE structure and
4855 its validity and signature checked, then if the certificate is found
4856 in the CRL the verify fails with a revoked error.
4858 Various new CRL related callbacks added to X509_STORE_CTX structure.
4860 Command line options added to 'verify' application to support this.
4862 This needs some additional work, such as being able to handle multiple
4863 CRLs with different times, extension based lookup (rather than just
4864 by subject name) and ultimately more complete V2 CRL extension
4868 *) Add a general user interface API (crypto/ui/). This is designed
4869 to replace things like des_read_password and friends (backward
4870 compatibility functions using this new API are provided).
4871 The purpose is to remove prompting functions from the DES code
4872 section as well as provide for prompting through dialog boxes in
4873 a window system and the like.
4876 *) Add "ex_data" support to ENGINE so implementations can add state at a
4877 per-structure level rather than having to store it globally.
4880 *) Make it possible for ENGINE structures to be copied when retrieved by
4881 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
4882 This causes the "original" ENGINE structure to act like a template,
4883 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
4884 operational state can be localised to each ENGINE structure, despite the
4885 fact they all share the same "methods". New ENGINE structures returned in
4886 this case have no functional references and the return value is the single
4887 structural reference. This matches the single structural reference returned
4888 by ENGINE_by_id() normally, when it is incremented on the pre-existing
4892 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
4893 needs to match any other type at all we need to manually clear the
4897 *) Changes to the "openssl engine" utility to include;
4898 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
4899 about an ENGINE's available control commands.
4900 - executing control commands from command line arguments using the
4901 '-pre' and '-post' switches. '-post' is only used if '-t' is
4902 specified and the ENGINE is successfully initialised. The syntax for
4903 the individual commands are colon-separated, for example;
4904 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
4907 *) New dynamic control command support for ENGINEs. ENGINEs can now
4908 declare their own commands (numbers), names (strings), descriptions,
4909 and input types for run-time discovery by calling applications. A
4910 subset of these commands are implicitly classed as "executable"
4911 depending on their input type, and only these can be invoked through
4912 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
4913 can be based on user input, config files, etc). The distinction is
4914 that "executable" commands cannot return anything other than a boolean
4915 result and can only support numeric or string input, whereas some
4916 discoverable commands may only be for direct use through
4917 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
4918 pointers, or other custom uses. The "executable" commands are to
4919 support parameterisations of ENGINE behaviour that can be
4920 unambiguously defined by ENGINEs and used consistently across any
4921 OpenSSL-based application. Commands have been added to all the
4922 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
4923 control over shared-library paths without source code alterations.
4926 *) Changed all ENGINE implementations to dynamically allocate their
4927 ENGINEs rather than declaring them statically. Apart from this being
4928 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
4929 this also allows the implementations to compile without using the
4930 internal engine_int.h header.
4933 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
4934 'const' value. Any code that should be able to modify a RAND_METHOD
4935 should already have non-const pointers to it (ie. they should only
4936 modify their own ones).
4939 *) Made a variety of little tweaks to the ENGINE code.
4940 - "atalla" and "ubsec" string definitions were moved from header files
4941 to C code. "nuron" string definitions were placed in variables
4942 rather than hard-coded - allowing parameterisation of these values
4943 later on via ctrl() commands.
4944 - Removed unused "#if 0"'d code.
4945 - Fixed engine list iteration code so it uses ENGINE_free() to release
4946 structural references.
4947 - Constified the RAND_METHOD element of ENGINE structures.
4948 - Constified various get/set functions as appropriate and added
4949 missing functions (including a catch-all ENGINE_cpy that duplicates
4950 all ENGINE values onto a new ENGINE except reference counts/state).
4951 - Removed NULL parameter checks in get/set functions. Setting a method
4952 or function to NULL is a way of cancelling out a previously set
4953 value. Passing a NULL ENGINE parameter is just plain stupid anyway
4954 and doesn't justify the extra error symbols and code.
4955 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
4956 flags from engine_int.h to engine.h.
4957 - Changed prototypes for ENGINE handler functions (init(), finish(),
4958 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
4961 *) Implement binary inversion algorithm for BN_mod_inverse in addition
4962 to the algorithm using long division. The binary algorithm can be
4963 used only if the modulus is odd. On 32-bit systems, it is faster
4964 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
4965 roughly 5-15% for 256-bit moduli), so we use it only for moduli
4966 up to 450 bits. In 64-bit environments, the binary algorithm
4967 appears to be advantageous for much longer moduli; here we use it
4968 for moduli up to 2048 bits.
4971 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
4972 could not support the combine flag in choice fields.
4975 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
4976 extensions from a certificate request to the certificate.
4979 *) Allow multiple 'certopt' and 'nameopt' options to be separated
4980 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
4981 file: this allows the display of the certificate about to be
4982 signed to be customised, to allow certain fields to be included
4983 or excluded and extension details. The old system didn't display
4984 multicharacter strings properly, omitted fields not in the policy
4985 and couldn't display additional details such as extensions.
4988 *) Function EC_POINTs_mul for multiple scalar multiplication
4989 of an arbitrary number of elliptic curve points
4990 \sum scalars[i]*points[i],
4991 optionally including the generator defined for the EC_GROUP:
4992 scalar*generator + \sum scalars[i]*points[i].
4994 EC_POINT_mul is a simple wrapper function for the typical case
4995 that the point list has just one item (besides the optional
4999 *) First EC_METHODs for curves over GF(p):
5001 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
5002 operations and provides various method functions that can also
5003 operate with faster implementations of modular arithmetic.
5005 EC_GFp_mont_method() reuses most functions that are part of
5006 EC_GFp_simple_method, but uses Montgomery arithmetic.
5008 [Bodo Moeller; point addition and point doubling
5009 implementation directly derived from source code provided by
5010 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
5012 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
5013 crypto/ec/ec_lib.c):
5015 Curves are EC_GROUP objects (with an optional group generator)
5016 based on EC_METHODs that are built into the library.
5018 Points are EC_POINT objects based on EC_GROUP objects.
5020 Most of the framework would be able to handle curves over arbitrary
5021 finite fields, but as there are no obvious types for fields other
5022 than GF(p), some functions are limited to that for now.
5025 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
5026 that the file contains a complete HTTP response.
5029 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
5030 change the def and num file printf format specifier from "%-40sXXX"
5031 to "%-39s XXX". The latter will always guarantee a space after the
5032 field while the former will cause them to run together if the field
5033 is 40 of more characters long.
5036 *) Constify the cipher and digest 'method' functions and structures
5037 and modify related functions to take constant EVP_MD and EVP_CIPHER
5041 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
5042 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
5045 *) Modify EVP_Digest*() routines so they now return values. Although the
5046 internal software routines can never fail additional hardware versions
5050 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
5052 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
5053 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
5056 ERR_R_NESTED_ASN1_ERROR
5058 ERR_R_MISSING_ASN1_EOS
5059 were 4 .. 9, conflicting with
5060 ERR_LIB_RSA (= ERR_R_RSA_LIB)
5062 ERR_LIB_PEM (= ERR_R_PEM_LIB).
5063 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
5065 Add new error code 'ERR_R_INTERNAL_ERROR'.
5068 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
5072 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
5073 sets the subject name for a new request or supersedes the
5074 subject name in a given request. Formats that can be parsed are
5075 'CN=Some Name, OU=myOU, C=IT'
5077 'CN=Some Name/OU=myOU/C=IT'.
5079 Add options '-batch' and '-verbose' to 'openssl req'.
5080 [Massimiliano Pala <madwolf@hackmasters.net>]
5082 *) Introduce the possibility to access global variables through
5083 functions on platform were that's the best way to handle exporting
5084 global variables in shared libraries. To enable this functionality,
5085 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
5086 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
5087 is normally done by Configure or something similar).
5089 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
5090 in the source file (foo.c) like this:
5092 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
5093 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
5095 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
5096 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
5098 OPENSSL_DECLARE_GLOBAL(int,foo);
5099 #define foo OPENSSL_GLOBAL_REF(foo)
5100 OPENSSL_DECLARE_GLOBAL(double,bar);
5101 #define bar OPENSSL_GLOBAL_REF(bar)
5103 The #defines are very important, and therefore so is including the
5104 header file everywhere where the defined globals are used.
5106 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
5107 of ASN.1 items, but that structure is a bit different.
5109 The largest change is in util/mkdef.pl which has been enhanced with
5110 better and easier to understand logic to choose which symbols should
5111 go into the Windows .def files as well as a number of fixes and code
5112 cleanup (among others, algorithm keywords are now sorted
5113 lexicographically to avoid constant rewrites).
5116 *) In BN_div() keep a copy of the sign of 'num' before writing the
5117 result to 'rm' because if rm==num the value will be overwritten
5118 and produce the wrong result if 'num' is negative: this caused
5119 problems with BN_mod() and BN_nnmod().
5122 *) Function OCSP_request_verify(). This checks the signature on an
5123 OCSP request and verifies the signer certificate. The signer
5124 certificate is just checked for a generic purpose and OCSP request
5128 *) Add OCSP_check_validity() function to check the validity of OCSP
5129 responses. OCSP responses are prepared in real time and may only
5130 be a few seconds old. Simply checking that the current time lies
5131 between thisUpdate and nextUpdate max reject otherwise valid responses
5132 caused by either OCSP responder or client clock inaccuracy. Instead
5133 we allow thisUpdate and nextUpdate to fall within a certain period of
5134 the current time. The age of the response can also optionally be
5135 checked. Two new options -validity_period and -status_age added to
5139 *) If signature or public key algorithm is unrecognized print out its
5140 OID rather that just UNKNOWN.
5143 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
5144 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
5145 ID to be generated from the issuer certificate alone which can then be
5146 passed to OCSP_id_issuer_cmp().
5149 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
5150 ASN1 modules to export functions returning ASN1_ITEM pointers
5151 instead of the ASN1_ITEM structures themselves. This adds several
5152 new macros which allow the underlying ASN1 function/structure to
5153 be accessed transparently. As a result code should not use ASN1_ITEM
5154 references directly (such as &X509_it) but instead use the relevant
5155 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
5156 use of the new ASN1 code on platforms where exporting structures
5157 is problematical (for example in shared libraries) but exporting
5158 functions returning pointers to structures is not.
5161 *) Add support for overriding the generation of SSL/TLS session IDs.
5162 These callbacks can be registered either in an SSL_CTX or per SSL.
5163 The purpose of this is to allow applications to control, if they wish,
5164 the arbitrary values chosen for use as session IDs, particularly as it
5165 can be useful for session caching in multiple-server environments. A
5166 command-line switch for testing this (and any client code that wishes
5167 to use such a feature) has been added to "s_server".
5168 [Geoff Thorpe, Lutz Jaenicke]
5170 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
5171 of the form '#if defined(...) || defined(...) || ...' and
5172 '#if !defined(...) && !defined(...) && ...'. This also avoids
5173 the growing number of special cases it was previously handling.
5176 *) Make all configuration macros available for application by making
5177 sure they are available in opensslconf.h, by giving them names starting
5178 with "OPENSSL_" to avoid conflicts with other packages and by making
5179 sure e_os2.h will cover all platform-specific cases together with
5181 Additionally, it is now possible to define configuration/platform-
5182 specific names (called "system identities"). In the C code, these
5183 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
5184 macro with the name beginning with "OPENSSL_SYS_", which is determined
5185 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
5189 *) New option -set_serial to 'req' and 'x509' this allows the serial
5190 number to use to be specified on the command line. Previously self
5191 signed certificates were hard coded with serial number 0 and the
5192 CA options of 'x509' had to use a serial number in a file which was
5196 *) New options to 'ca' utility to support V2 CRL entry extensions.
5197 Currently CRL reason, invalidity date and hold instruction are
5198 supported. Add new CRL extensions to V3 code and some new objects.
5201 *) New function EVP_CIPHER_CTX_set_padding() this is used to
5202 disable standard block padding (aka PKCS#5 padding) in the EVP
5203 API, which was previously mandatory. This means that the data is
5204 not padded in any way and so the total length much be a multiple
5205 of the block size, otherwise an error occurs.
5208 *) Initial (incomplete) OCSP SSL support.
5211 *) New function OCSP_parse_url(). This splits up a URL into its host,
5212 port and path components: primarily to parse OCSP URLs. New -url
5213 option to ocsp utility.
5216 *) New nonce behavior. The return value of OCSP_check_nonce() now
5217 reflects the various checks performed. Applications can decide
5218 whether to tolerate certain situations such as an absent nonce
5219 in a response when one was present in a request: the ocsp application
5220 just prints out a warning. New function OCSP_add1_basic_nonce()
5221 this is to allow responders to include a nonce in a response even if
5222 the request is nonce-less.
5225 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
5226 skipped when using openssl x509 multiple times on a single input file,
5227 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
5230 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
5231 set string type: to handle setting ASN1_TIME structures. Fix ca
5232 utility to correctly initialize revocation date of CRLs.
5235 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
5236 the clients preferred ciphersuites and rather use its own preferences.
5237 Should help to work around M$ SGC (Server Gated Cryptography) bug in
5238 Internet Explorer by ensuring unchanged hash method during stepup.
5239 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
5242 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
5243 to aes and add a new 'exist' option to print out symbols that don't
5247 *) Additional options to ocsp utility to allow flags to be set and
5248 additional certificates supplied.
5251 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
5252 OCSP client a number of certificate to only verify the response
5256 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
5257 handle the new API. Currently only ECB, CBC modes supported. Add new
5260 Add TLS AES ciphersuites as described in RFC3268, "Advanced
5261 Encryption Standard (AES) Ciphersuites for Transport Layer
5262 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
5263 not enabled by default and were not part of the "ALL" ciphersuite
5264 alias because they were not yet official; they could be
5265 explicitly requested by specifying the "AESdraft" ciphersuite
5266 group alias. In the final release of OpenSSL 0.9.7, the group
5267 alias is called "AES" and is part of "ALL".)
5268 [Ben Laurie, Steve Henson, Bodo Moeller]
5270 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
5271 request to response.
5274 *) Functions for OCSP responders. OCSP_request_onereq_count(),
5275 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
5276 extract information from a certificate request. OCSP_response_create()
5277 creates a response and optionally adds a basic response structure.
5278 OCSP_basic_add1_status() adds a complete single response to a basic
5279 response and returns the OCSP_SINGLERESP structure just added (to allow
5280 extensions to be included for example). OCSP_basic_add1_cert() adds a
5281 certificate to a basic response and OCSP_basic_sign() signs a basic
5282 response with various flags. New helper functions ASN1_TIME_check()
5283 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
5284 (converts ASN1_TIME to GeneralizedTime).
5287 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
5288 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
5289 structure from a certificate. X509_pubkey_digest() digests the public_key
5290 contents: this is used in various key identifiers.
5293 *) Make sk_sort() tolerate a NULL argument.
5294 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
5296 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
5297 passed by the function are trusted implicitly. If any of them signed the
5298 response then it is assumed to be valid and is not verified.
5301 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
5302 to data. This was previously part of the PKCS7 ASN1 code. This
5303 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
5304 [Steve Henson, reported by Kenneth R. Robinette
5305 <support@securenetterm.com>]
5307 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
5308 routines: without these tracing memory leaks is very painful.
5309 Fix leaks in PKCS12 and PKCS7 routines.
5312 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
5313 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
5314 effectively meant GeneralizedTime would never be used. Now it
5315 is initialised to -1 but X509_time_adj() now has to check the value
5316 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
5317 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
5318 [Steve Henson, reported by Kenneth R. Robinette
5319 <support@securenetterm.com>]
5321 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
5322 result in a zero length in the ASN1_INTEGER structure which was
5323 not consistent with the structure when d2i_ASN1_INTEGER() was used
5324 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
5325 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
5326 where it did not print out a minus for negative ASN1_INTEGER.
5329 *) Add summary printout to ocsp utility. The various functions which
5330 convert status values to strings have been renamed to:
5331 OCSP_response_status_str(), OCSP_cert_status_str() and
5332 OCSP_crl_reason_str() and are no longer static. New options
5333 to verify nonce values and to disable verification. OCSP response
5334 printout format cleaned up.
5337 *) Add additional OCSP certificate checks. These are those specified
5338 in RFC2560. This consists of two separate checks: the CA of the
5339 certificate being checked must either be the OCSP signer certificate
5340 or the issuer of the OCSP signer certificate. In the latter case the
5341 OCSP signer certificate must contain the OCSP signing extended key
5342 usage. This check is performed by attempting to match the OCSP
5343 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
5344 in the OCSP_CERTID structures of the response.
5347 *) Initial OCSP certificate verification added to OCSP_basic_verify()
5348 and related routines. This uses the standard OpenSSL certificate
5349 verify routines to perform initial checks (just CA validity) and
5350 to obtain the certificate chain. Then additional checks will be
5351 performed on the chain. Currently the root CA is checked to see
5352 if it is explicitly trusted for OCSP signing. This is used to set
5353 a root CA as a global signing root: that is any certificate that
5354 chains to that CA is an acceptable OCSP signing certificate.
5357 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
5358 extensions from a separate configuration file.
5359 As when reading extensions from the main configuration file,
5360 the '-extensions ...' option may be used for specifying the
5362 [Massimiliano Pala <madwolf@comune.modena.it>]
5364 *) New OCSP utility. Allows OCSP requests to be generated or
5365 read. The request can be sent to a responder and the output
5366 parsed, outputed or printed in text form. Not complete yet:
5367 still needs to check the OCSP response validity.
5370 *) New subcommands for 'openssl ca':
5371 'openssl ca -status <serial>' prints the status of the cert with
5372 the given serial number (according to the index file).
5373 'openssl ca -updatedb' updates the expiry status of certificates
5375 [Massimiliano Pala <madwolf@comune.modena.it>]
5377 *) New '-newreq-nodes' command option to CA.pl. This is like
5378 '-newreq', but calls 'openssl req' with the '-nodes' option
5379 so that the resulting key is not encrypted.
5380 [Damien Miller <djm@mindrot.org>]
5382 *) New configuration for the GNU Hurd.
5383 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
5385 *) Initial code to implement OCSP basic response verify. This
5386 is currently incomplete. Currently just finds the signer's
5387 certificate and verifies the signature on the response.
5390 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
5391 value of OPENSSLDIR. This is available via the new '-d' option
5392 to 'openssl version', and is also included in 'openssl version -a'.
5395 *) Allowing defining memory allocation callbacks that will be given
5396 file name and line number information in additional arguments
5397 (a const char* and an int). The basic functionality remains, as
5398 well as the original possibility to just replace malloc(),
5399 realloc() and free() by functions that do not know about these
5400 additional arguments. To register and find out the current
5401 settings for extended allocation functions, the following
5402 functions are provided:
5404 CRYPTO_set_mem_ex_functions
5405 CRYPTO_set_locked_mem_ex_functions
5406 CRYPTO_get_mem_ex_functions
5407 CRYPTO_get_locked_mem_ex_functions
5409 These work the same way as CRYPTO_set_mem_functions and friends.
5410 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
5411 extended allocation function is enabled.
5412 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
5413 a conventional allocation function is enabled.
5414 [Richard Levitte, Bodo Moeller]
5416 *) Finish off removing the remaining LHASH function pointer casts.
5417 There should no longer be any prototype-casting required when using
5418 the LHASH abstraction, and any casts that remain are "bugs". See
5419 the callback types and macros at the head of lhash.h for details
5420 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
5423 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
5424 If /dev/[u]random devices are not available or do not return enough
5425 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
5427 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
5428 /etc/entropy will be queried once each in this sequence, quering stops
5429 when enough entropy was collected without querying more sockets.
5432 *) Change the Unix RAND_poll() variant to be able to poll several
5433 random devices, as specified by DEVRANDOM, until a sufficient amount
5434 of data has been collected. We spend at most 10 ms on each file
5435 (select timeout) and read in non-blocking mode. DEVRANDOM now
5436 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
5437 (previously it was just the string "/dev/urandom"), so on typical
5438 platforms the 10 ms delay will never occur.
5439 Also separate out the Unix variant to its own file, rand_unix.c.
5440 For VMS, there's a currently-empty rand_vms.c.
5443 *) Move OCSP client related routines to ocsp_cl.c. These
5444 provide utility functions which an application needing
5445 to issue a request to an OCSP responder and analyse the
5446 response will typically need: as opposed to those which an
5447 OCSP responder itself would need which will be added later.
5449 OCSP_request_sign() signs an OCSP request with an API similar
5450 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
5451 response. OCSP_response_get1_basic() extracts basic response
5452 from response. OCSP_resp_find_status(): finds and extracts status
5453 information from an OCSP_CERTID structure (which will be created
5454 when the request structure is built). These are built from lower
5455 level functions which work on OCSP_SINGLERESP structures but
5456 wont normally be used unless the application wishes to examine
5457 extensions in the OCSP response for example.
5459 Replace nonce routines with a pair of functions.
5460 OCSP_request_add1_nonce() adds a nonce value and optionally
5461 generates a random value. OCSP_check_nonce() checks the
5462 validity of the nonce in an OCSP response.
5465 *) Change function OCSP_request_add() to OCSP_request_add0_id().
5466 This doesn't copy the supplied OCSP_CERTID and avoids the
5467 need to free up the newly created id. Change return type
5468 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
5469 This can then be used to add extensions to the request.
5470 Deleted OCSP_request_new(), since most of its functionality
5471 is now in OCSP_REQUEST_new() (and the case insensitive name
5472 clash) apart from the ability to set the request name which
5473 will be added elsewhere.
5476 *) Update OCSP API. Remove obsolete extensions argument from
5477 various functions. Extensions are now handled using the new
5478 OCSP extension code. New simple OCSP HTTP function which
5479 can be used to send requests and parse the response.
5482 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
5483 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
5484 uses the special reorder version of SET OF to sort the attributes
5485 and reorder them to match the encoded order. This resolves a long
5486 standing problem: a verify on a PKCS7 structure just after signing
5487 it used to fail because the attribute order did not match the
5488 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
5489 it uses the received order. This is necessary to tolerate some broken
5490 software that does not order SET OF. This is handled by encoding
5491 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
5492 to produce the required SET OF.
5495 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
5496 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
5497 files to get correct declarations of the ASN.1 item variables.
5500 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
5501 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
5502 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
5503 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
5504 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
5505 ASN1_ITEM and no wrapper functions.
5508 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
5509 replace the old function pointer based I/O routines. Change most of
5510 the *_d2i_bio() and *_d2i_fp() functions to use these.
5513 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
5514 lines, recognice more "algorithms" that can be deselected, and make
5515 it complain about algorithm deselection that isn't recognised.
5518 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
5519 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
5520 to use new functions. Add NO_ASN1_OLD which can be set to remove
5521 some old style ASN1 functions: this can be used to determine if old
5522 code will still work when these eventually go away.
5525 *) New extension functions for OCSP structures, these follow the
5526 same conventions as certificates and CRLs.
5529 *) New function X509V3_add1_i2d(). This automatically encodes and
5530 adds an extension. Its behaviour can be customised with various
5531 flags to append, replace or delete. Various wrappers added for
5532 certifcates and CRLs.
5535 *) Fix to avoid calling the underlying ASN1 print routine when
5536 an extension cannot be parsed. Correct a typo in the
5537 OCSP_SERVICELOC extension. Tidy up print OCSP format.
5540 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
5541 entries for variables.
5544 *) Add functionality to apps/openssl.c for detecting locking
5545 problems: As the program is single-threaded, all we have
5546 to do is register a locking callback using an array for
5547 storing which locks are currently held by the program.
5550 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
5551 SSL_get_ex_data_X509_STORE_idx(), which is used in
5552 ssl_verify_cert_chain() and thus can be called at any time
5553 during TLS/SSL handshakes so that thread-safety is essential.
5554 Unfortunately, the ex_data design is not at all suited
5555 for multi-threaded use, so it probably should be abolished.
5558 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5559 [Broadcom, tweaked and integrated by Geoff Thorpe]
5561 *) Move common extension printing code to new function
5562 X509V3_print_extensions(). Reorganise OCSP print routines and
5563 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5566 *) New function X509_signature_print() to remove duplication in some
5570 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5571 set (this was treated exactly the same as SET OF previously). This
5572 is used to reorder the STACK representing the structure to match the
5573 encoding. This will be used to get round a problem where a PKCS7
5574 structure which was signed could not be verified because the STACK
5575 order did not reflect the encoded order.
5578 *) Reimplement the OCSP ASN1 module using the new code.
5581 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5582 for its ASN1 operations. The old style function pointers still exist
5583 for now but they will eventually go away.
5586 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
5587 completely replaces the old ASN1 functionality with a table driven
5588 encoder and decoder which interprets an ASN1_ITEM structure describing
5589 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
5590 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
5591 has also been converted to the new form.
5594 *) Change BN_mod_exp_recp so that negative moduli are tolerated
5595 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
5596 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
5597 for negative moduli.
5600 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
5601 of not touching the result's sign bit.
5604 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
5608 *) Changed the LHASH code to use prototypes for callbacks, and created
5609 macros to declare and implement thin (optionally static) functions
5610 that provide type-safety and avoid function pointer casting for the
5611 type-specific callbacks.
5614 *) Added Kerberos Cipher Suites to be used with TLS, as written in
5616 [Veers Staats <staatsvr@asc.hpc.mil>,
5617 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
5619 *) Reformat the FAQ so the different questions and answers can be divided
5620 in sections depending on the subject.
5623 *) Have the zlib compression code load ZLIB.DLL dynamically under
5627 *) New function BN_mod_sqrt for computing square roots modulo a prime
5628 (using the probabilistic Tonelli-Shanks algorithm unless
5629 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
5630 be handled deterministically).
5631 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5633 *) Make BN_mod_inverse faster by explicitly handling small quotients
5634 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
5635 512 bits], about 30% for larger ones [1024 or 2048 bits].)
5638 *) New function BN_kronecker.
5641 *) Fix BN_gcd so that it works on negative inputs; the result is
5642 positive unless both parameters are zero.
5643 Previously something reasonably close to an infinite loop was
5644 possible because numbers could be growing instead of shrinking
5645 in the implementation of Euclid's algorithm.
5648 *) Fix BN_is_word() and BN_is_one() macros to take into account the
5649 sign of the number in question.
5651 Fix BN_is_word(a,w) to work correctly for w == 0.
5653 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
5654 because its test if the absolute value of 'a' equals 'w'.
5655 Note that BN_abs_is_word does *not* handle w == 0 reliably;
5656 it exists mostly for use in the implementations of BN_is_zero(),
5657 BN_is_one(), and BN_is_word().
5660 *) New function BN_swap.
5663 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
5664 the exponentiation functions are more likely to produce reasonable
5665 results on negative inputs.
5668 *) Change BN_mod_mul so that the result is always non-negative.
5669 Previously, it could be negative if one of the factors was negative;
5670 I don't think anyone really wanted that behaviour.
5673 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
5674 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
5675 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
5676 and add new functions:
5685 BN_mod_lshift1_quick
5689 These functions always generate non-negative results.
5691 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
5692 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
5694 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
5695 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
5696 be reduced modulo m.
5697 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5700 The following entry accidentily appeared in the CHANGES file
5701 distributed with OpenSSL 0.9.7. The modifications described in
5702 it do *not* apply to OpenSSL 0.9.7.
5704 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5705 was actually never needed) and in BN_mul(). The removal in BN_mul()
5706 required a small change in bn_mul_part_recursive() and the addition
5707 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5708 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5709 bn_sub_words() and bn_add_words() except they take arrays with
5714 *) In 'openssl passwd', verify passwords read from the terminal
5715 unless the '-salt' option is used (which usually means that
5716 verification would just waste user's time since the resulting
5717 hash is going to be compared with some given password hash)
5718 or the new '-noverify' option is used.
5720 This is an incompatible change, but it does not affect
5721 non-interactive use of 'openssl passwd' (passwords on the command
5722 line, '-stdin' option, '-in ...' option) and thus should not
5726 *) Remove all references to RSAref, since there's no more need for it.
5729 *) Make DSO load along a path given through an environment variable
5730 (SHLIB_PATH) with shl_load().
5733 *) Constify the ENGINE code as a result of BIGNUM constification.
5734 Also constify the RSA code and most things related to it. In a
5735 few places, most notable in the depth of the ASN.1 code, ugly
5736 casts back to non-const were required (to be solved at a later
5740 *) Make it so the openssl application has all engines loaded by default.
5743 *) Constify the BIGNUM routines a little more.
5746 *) Add the following functions:
5748 ENGINE_load_cswift()
5750 ENGINE_load_atalla()
5752 ENGINE_load_builtin_engines()
5754 That way, an application can itself choose if external engines that
5755 are built-in in OpenSSL shall ever be used or not. The benefit is
5756 that applications won't have to be linked with libdl or other dso
5757 libraries unless it's really needed.
5759 Changed 'openssl engine' to load all engines on demand.
5760 Changed the engine header files to avoid the duplication of some
5761 declarations (they differed!).
5764 *) 'openssl engine' can now list capabilities.
5767 *) Better error reporting in 'openssl engine'.
5770 *) Never call load_dh_param(NULL) in s_server.
5773 *) Add engine application. It can currently list engines by name and
5774 identity, and test if they are actually available.
5777 *) Improve RPM specification file by forcing symbolic linking and making
5778 sure the installed documentation is also owned by root.root.
5779 [Damien Miller <djm@mindrot.org>]
5781 *) Give the OpenSSL applications more possibilities to make use of
5782 keys (public as well as private) handled by engines.
5785 *) Add OCSP code that comes from CertCo.
5788 *) Add VMS support for the Rijndael code.
5791 *) Added untested support for Nuron crypto accelerator.
5794 *) Add support for external cryptographic devices. This code was
5795 previously distributed separately as the "engine" branch.
5796 [Geoff Thorpe, Richard Levitte]
5798 *) Rework the filename-translation in the DSO code. It is now possible to
5799 have far greater control over how a "name" is turned into a filename
5800 depending on the operating environment and any oddities about the
5801 different shared library filenames on each system.
5804 *) Support threads on FreeBSD-elf in Configure.
5807 *) Fix for SHA1 assembly problem with MASM: it produces
5808 warnings about corrupt line number information when assembling
5809 with debugging information. This is caused by the overlapping
5811 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
5814 NCONF_get_number() has no error checking at all. As a replacement,
5815 NCONF_get_number_e() is defined (_e for "error checking") and is
5816 promoted strongly. The old NCONF_get_number is kept around for
5817 binary backward compatibility.
5818 Make it possible for methods to load from something other than a BIO,
5819 by providing a function pointer that is given a name instead of a BIO.
5820 For example, this could be used to load configuration data from an
5824 *) Fix for non blocking accept BIOs. Added new I/O special reason
5825 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
5826 with non blocking I/O was not possible because no retry code was
5827 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
5831 *) Added the beginnings of Rijndael support.
5834 *) Fix for bug in DirectoryString mask setting. Add support for
5835 X509_NAME_print_ex() in 'req' and X509_print_ex() function
5836 to allow certificate printing to more controllable, additional
5837 'certopt' option to 'x509' to allow new printing options to be
5841 *) Clean old EAY MD5 hack from e_os.h.
5844 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
5846 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5847 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5848 [Joe Orton, Steve Henson]
5850 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
5852 *) Fix additional bug revealed by the NISCC test suite:
5854 Stop bug triggering large recursion when presented with
5855 certain ASN.1 tags (CVE-2003-0851)
5858 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
5860 *) Fix various bugs revealed by running the NISCC test suite:
5862 Stop out of bounds reads in the ASN1 code when presented with
5863 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5865 If verify callback ignores invalid public key errors don't try to check
5866 certificate signature with the NULL public key.
5870 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5871 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5875 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5876 extra data after the compression methods not only for TLS 1.0
5877 but also for SSL 3.0 (as required by the specification).
5878 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5880 *) Change X509_certificate_type() to mark the key as exported/exportable
5881 when it's 512 *bits* long, not 512 bytes.
5884 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
5886 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5887 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5888 a protocol version number mismatch like a decryption error
5889 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5892 *) Turn on RSA blinding by default in the default implementation
5893 to avoid a timing attack. Applications that don't want it can call
5894 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5895 They would be ill-advised to do so in most cases.
5896 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5898 *) Change RSA blinding code so that it works when the PRNG is not
5899 seeded (in this case, the secret RSA exponent is abused as
5900 an unpredictable seed -- if it is not unpredictable, there
5901 is no point in blinding anyway). Make RSA blinding thread-safe
5902 by remembering the creator's thread ID in rsa->blinding and
5903 having all other threads use local one-time blinding factors
5904 (this requires more computation than sharing rsa->blinding, but
5905 avoids excessive locking; and if an RSA object is not shared
5906 between threads, blinding will still be very fast).
5909 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
5911 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5912 via timing by performing a MAC computation even if incorrrect
5913 block cipher padding has been found. This is a countermeasure
5914 against active attacks where the attacker has to distinguish
5915 between bad padding and a MAC verification error. (CVE-2003-0078)
5917 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5918 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5919 Martin Vuagnoux (EPFL, Ilion)]
5921 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
5923 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
5924 memory from it's contents. This is done with a counter that will
5925 place alternating values in each byte. This can be used to solve
5926 two issues: 1) the removal of calls to memset() by highly optimizing
5927 compilers, and 2) cleansing with other values than 0, since those can
5928 be read through on certain media, for example a swap space on disk.
5931 *) Bugfix: client side session caching did not work with external caching,
5932 because the session->cipher setting was not restored when reloading
5933 from the external cache. This problem was masked, when
5934 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
5935 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
5938 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
5939 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
5940 [Zeev Lieber <zeev-l@yahoo.com>]
5942 *) Undo an undocumented change introduced in 0.9.6e which caused
5943 repeated calls to OpenSSL_add_all_ciphers() and
5944 OpenSSL_add_all_digests() to be ignored, even after calling
5948 *) Change the default configuration reader to deal with last line not
5949 being properly terminated.
5952 *) Change X509_NAME_cmp() so it applies the special rules on handling
5953 DN values that are of type PrintableString, as well as RDNs of type
5954 emailAddress where the value has the type ia5String.
5955 [stefank@valicert.com via Richard Levitte]
5957 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
5958 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
5959 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
5960 the bitwise-OR of the two for use by the majority of applications
5961 wanting this behaviour, and update the docs. The documented
5962 behaviour and actual behaviour were inconsistent and had been
5963 changing anyway, so this is more a bug-fix than a behavioural
5965 [Geoff Thorpe, diagnosed by Nadav Har'El]
5967 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
5968 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
5971 *) Fix initialization code race conditions in
5972 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
5973 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
5974 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
5975 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
5976 ssl2_get_cipher_by_char(),
5977 ssl3_get_cipher_by_char().
5978 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
5980 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
5981 the cached sessions are flushed, as the remove_cb() might use ex_data
5982 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
5983 (see [openssl.org #212]).
5984 [Geoff Thorpe, Lutz Jaenicke]
5986 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
5987 length, instead of the encoding length to d2i_ASN1_OBJECT.
5990 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
5992 *) [In 0.9.6g-engine release:]
5993 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
5994 [Lynn Gazis <lgazis@rainbow.com>]
5996 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
5998 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
5999 and get fix the header length calculation.
6000 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
6001 Alon Kantor <alonk@checkpoint.com> (and others),
6004 *) Use proper error handling instead of 'assertions' in buffer
6005 overflow checks added in 0.9.6e. This prevents DoS (the
6006 assertions could call abort()).
6007 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
6009 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
6011 *) Add various sanity checks to asn1_get_length() to reject
6012 the ASN1 length bytes if they exceed sizeof(long), will appear
6013 negative or the content length exceeds the length of the
6015 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6017 *) Fix cipher selection routines: ciphers without encryption had no flags
6018 for the cipher strength set and where therefore not handled correctly
6019 by the selection routines (PR #130).
6022 *) Fix EVP_dsa_sha macro.
6026 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
6027 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
6028 that was added in OpenSSL 0.9.6d.
6030 As the countermeasure turned out to be incompatible with some
6031 broken SSL implementations, the new option is part of SSL_OP_ALL.
6032 SSL_OP_ALL is usually employed when compatibility with weird SSL
6033 implementations is desired (e.g. '-bugs' option to 's_client' and
6034 's_server'), so the new option is automatically set in many
6038 *) Changes in security patch:
6040 Changes marked "(CHATS)" were sponsored by the Defense Advanced
6041 Research Projects Agency (DARPA) and Air Force Research Laboratory,
6042 Air Force Materiel Command, USAF, under agreement number
6045 *) Add various sanity checks to asn1_get_length() to reject
6046 the ASN1 length bytes if they exceed sizeof(long), will appear
6047 negative or the content length exceeds the length of the
6048 supplied buffer. (CVE-2002-0659)
6049 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6051 *) Assertions for various potential buffer overflows, not known to
6053 [Ben Laurie (CHATS)]
6055 *) Various temporary buffers to hold ASCII versions of integers were
6056 too small for 64 bit platforms. (CVE-2002-0655)
6057 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
6059 *) Remote buffer overflow in SSL3 protocol - an attacker could
6060 supply an oversized session ID to a client. (CVE-2002-0656)
6061 [Ben Laurie (CHATS)]
6063 *) Remote buffer overflow in SSL2 protocol - an attacker could
6064 supply an oversized client master key. (CVE-2002-0656)
6065 [Ben Laurie (CHATS)]
6067 Changes between 0.9.6c and 0.9.6d [9 May 2002]
6069 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
6070 encoded as NULL) with id-dsa-with-sha1.
6071 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
6073 *) Check various X509_...() return values in apps/req.c.
6074 [Nils Larsch <nla@trustcenter.de>]
6076 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
6077 an end-of-file condition would erronously be flagged, when the CRLF
6078 was just at the end of a processed block. The bug was discovered when
6079 processing data through a buffering memory BIO handing the data to a
6080 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
6081 <ptsekov@syntrex.com> and Nedelcho Stanev.
6084 *) Implement a countermeasure against a vulnerability recently found
6085 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
6086 before application data chunks to avoid the use of known IVs
6087 with data potentially chosen by the attacker.
6090 *) Fix length checks in ssl3_get_client_hello().
6093 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
6094 to prevent ssl3_read_internal() from incorrectly assuming that
6095 ssl3_read_bytes() found application data while handshake
6096 processing was enabled when in fact s->s3->in_read_app_data was
6097 merely automatically cleared during the initial handshake.
6098 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
6100 *) Fix object definitions for Private and Enterprise: they were not
6101 recognized in their shortname (=lowercase) representation. Extend
6102 obj_dat.pl to issue an error when using undefined keywords instead
6103 of silently ignoring the problem (Svenning Sorensen
6104 <sss@sss.dnsalias.net>).
6107 *) Fix DH_generate_parameters() so that it works for 'non-standard'
6108 generators, i.e. generators other than 2 and 5. (Previously, the
6109 code did not properly initialise the 'add' and 'rem' values to
6110 BN_generate_prime().)
6112 In the new general case, we do not insist that 'generator' is
6113 actually a primitive root: This requirement is rather pointless;
6114 a generator of the order-q subgroup is just as good, if not
6118 *) Map new X509 verification errors to alerts. Discovered and submitted by
6119 Tom Wu <tom@arcot.com>.
6122 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
6123 returning non-zero before the data has been completely received
6124 when using non-blocking I/O.
6125 [Bodo Moeller; problem pointed out by John Hughes]
6127 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
6128 [Ben Laurie, Lutz Jaenicke]
6130 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
6131 Yoram Zahavi <YoramZ@gilian.com>).
6134 *) Add information about CygWin 1.3 and on, and preserve proper
6135 configuration for the versions before that.
6136 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6138 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
6139 check whether we deal with a copy of a session and do not delete from
6140 the cache in this case. Problem reported by "Izhar Shoshani Levi"
6141 <izhar@checkpoint.com>.
6144 *) Do not store session data into the internal session cache, if it
6145 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
6146 flag is set). Proposed by Aslam <aslam@funk.com>.
6149 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
6153 *) [In 0.9.6d-engine release:]
6154 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6155 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6157 *) Add the configuration target linux-s390x.
6158 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
6160 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
6161 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
6162 variable as an indication that a ClientHello message has been
6163 received. As the flag value will be lost between multiple
6164 invocations of ssl3_accept when using non-blocking I/O, the
6165 function may not be aware that a handshake has actually taken
6166 place, thus preventing a new session from being added to the
6169 To avoid this problem, we now set s->new_session to 2 instead of
6170 using a local variable.
6171 [Lutz Jaenicke, Bodo Moeller]
6173 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
6174 if the SSL_R_LENGTH_MISMATCH error is detected.
6175 [Geoff Thorpe, Bodo Moeller]
6177 *) New 'shared_ldflag' column in Configure platform table.
6180 *) Fix EVP_CIPHER_mode macro.
6181 ["Dan S. Camper" <dan@bti.net>]
6183 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
6184 type, we must throw them away by setting rr->length to 0.
6185 [D P Chang <dpc@qualys.com>]
6187 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
6189 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
6190 <Dominikus.Scherkl@biodata.com>. (The previous implementation
6191 worked incorrectly for those cases where range = 10..._2 and
6192 3*range is two bits longer than range.)
6195 *) Only add signing time to PKCS7 structures if it is not already
6199 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
6200 OBJ_ld_ce should be OBJ_id_ce.
6201 Also some ip-pda OIDs in crypto/objects/objects.txt were
6202 incorrect (cf. RFC 3039).
6203 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
6205 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
6206 returns early because it has nothing to do.
6207 [Andy Schneider <andy.schneider@bjss.co.uk>]
6209 *) [In 0.9.6c-engine release:]
6210 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
6211 [Andy Schneider <andy.schneider@bjss.co.uk>]
6213 *) [In 0.9.6c-engine release:]
6214 Add support for Cryptographic Appliance's keyserver technology.
6215 (Use engine 'keyclient')
6216 [Cryptographic Appliances and Geoff Thorpe]
6218 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
6219 is called via tools/c89.sh because arguments have to be
6220 rearranged (all '-L' options must appear before the first object
6222 [Richard Shapiro <rshapiro@abinitio.com>]
6224 *) [In 0.9.6c-engine release:]
6225 Add support for Broadcom crypto accelerator cards, backported
6227 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
6229 *) [In 0.9.6c-engine release:]
6230 Add support for SureWare crypto accelerator cards from
6231 Baltimore Technologies. (Use engine 'sureware')
6232 [Baltimore Technologies and Mark Cox]
6234 *) [In 0.9.6c-engine release:]
6235 Add support for crypto accelerator cards from Accelerated
6236 Encryption Processing, www.aep.ie. (Use engine 'aep')
6237 [AEP Inc. and Mark Cox]
6239 *) Add a configuration entry for gcc on UnixWare.
6240 [Gary Benson <gbenson@redhat.com>]
6242 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
6243 messages are stored in a single piece (fixed-length part and
6244 variable-length part combined) and fix various bugs found on the way.
6247 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
6248 instead. BIO_gethostbyname() does not know what timeouts are
6249 appropriate, so entries would stay in cache even when they have
6251 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
6253 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
6254 faced with a pathologically small ClientHello fragment that does
6255 not contain client_version: Instead of aborting with an error,
6256 simply choose the highest available protocol version (i.e.,
6257 TLS 1.0 unless it is disabled). In practice, ClientHello
6258 messages are never sent like this, but this change gives us
6259 strictly correct behaviour at least for TLS.
6262 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
6263 never resets s->method to s->ctx->method when called from within
6264 one of the SSL handshake functions.
6265 [Bodo Moeller; problem pointed out by Niko Baric]
6267 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
6268 (sent using the client's version number) if client_version is
6269 smaller than the protocol version in use. Also change
6270 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
6271 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
6272 the client will at least see that alert.
6275 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
6279 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
6280 client receives HelloRequest while in a handshake.
6281 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
6283 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
6284 should end in 'break', not 'goto end' which circuments various
6285 cleanups done in state SSL_ST_OK. But session related stuff
6286 must be disabled for SSL_ST_OK in the case that we just sent a
6289 Also avoid some overhead by not calling ssl_init_wbio_buffer()
6290 before just sending a HelloRequest.
6291 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
6293 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
6294 reveal whether illegal block cipher padding was found or a MAC
6295 verification error occured. (Neither SSLerr() codes nor alerts
6296 are directly visible to potential attackers, but the information
6297 may leak via logfiles.)
6299 Similar changes are not required for the SSL 2.0 implementation
6300 because the number of padding bytes is sent in clear for SSL 2.0,
6301 and the extra bytes are just ignored. However ssl/s2_pkt.c
6302 failed to verify that the purported number of padding bytes is in
6306 *) Add OpenUNIX-8 support including shared libraries
6307 (Boyd Lynn Gerber <gerberb@zenez.com>).
6310 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
6311 'wristwatch attack' using huge encoding parameters (cf.
6312 James H. Manger's CRYPTO 2001 paper). Note that the
6313 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
6314 encoding parameters and hence was not vulnerable.
6317 *) BN_sqr() bug fix.
6318 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
6320 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
6321 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
6322 followed by modular reduction.
6323 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
6325 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
6326 equivalent based on BN_pseudo_rand() instead of BN_rand().
6329 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
6330 This function was broken, as the check for a new client hello message
6331 to handle SGC did not allow these large messages.
6332 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
6335 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
6338 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
6339 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
6342 *) Rework the configuration and shared library support for Tru64 Unix.
6343 The configuration part makes use of modern compiler features and
6344 still retains old compiler behavior for those that run older versions
6345 of the OS. The shared library support part includes a variant that
6346 uses the RPATH feature, and is available through the special
6347 configuration target "alpha-cc-rpath", which will never be selected
6349 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
6351 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
6352 with the same message size as in ssl3_get_certificate_request().
6353 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
6354 messages might inadvertently be reject as too long.
6355 [Petr Lampa <lampa@fee.vutbr.cz>]
6357 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
6360 *) Modified SSL library such that the verify_callback that has been set
6361 specificly for an SSL object with SSL_set_verify() is actually being
6362 used. Before the change, a verify_callback set with this function was
6363 ignored and the verify_callback() set in the SSL_CTX at the time of
6364 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
6365 to allow the necessary settings.
6368 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
6369 explicitly to NULL, as at least on Solaris 8 this seems not always to be
6370 done automatically (in contradiction to the requirements of the C
6371 standard). This made problems when used from OpenSSH.
6374 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
6375 dh->length and always used
6377 BN_rand_range(priv_key, dh->p).
6379 BN_rand_range() is not necessary for Diffie-Hellman, and this
6380 specific range makes Diffie-Hellman unnecessarily inefficient if
6381 dh->length (recommended exponent length) is much smaller than the
6382 length of dh->p. We could use BN_rand_range() if the order of
6383 the subgroup was stored in the DH structure, but we only have
6388 BN_rand(priv_key, l, ...)
6390 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
6396 RSA_eay_public_encrypt
6397 RSA_eay_private_decrypt
6398 RSA_eay_private_encrypt (signing)
6399 RSA_eay_public_decrypt (signature verification)
6401 (default implementations for RSA_public_encrypt,
6402 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
6403 always reject numbers >= n.
6406 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
6407 to synchronize access to 'locking_thread'. This is necessary on
6408 systems where access to 'locking_thread' (an 'unsigned long'
6409 variable) is not atomic.
6412 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
6413 *before* setting the 'crypto_lock_rand' flag. The previous code had
6414 a race condition if 0 is a valid thread ID.
6415 [Travis Vitek <vitek@roguewave.com>]
6417 *) Add support for shared libraries under Irix.
6418 [Albert Chin-A-Young <china@thewrittenword.com>]
6420 *) Add configuration option to build on Linux on both big-endian and
6422 [Ralf Baechle <ralf@uni-koblenz.de>]
6424 *) Add the possibility to create shared libraries on HP-UX.
6427 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
6429 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
6430 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
6431 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
6432 PRNG state recovery was possible based on the output of
6433 one PRNG request appropriately sized to gain knowledge on
6434 'md' followed by enough consecutive 1-byte PRNG requests
6435 to traverse all of 'state'.
6437 1. When updating 'md_local' (the current thread's copy of 'md')
6438 during PRNG output generation, hash all of the previous
6439 'md_local' value, not just the half used for PRNG output.
6441 2. Make the number of bytes from 'state' included into the hash
6442 independent from the number of PRNG bytes requested.
6444 The first measure alone would be sufficient to avoid
6445 Markku-Juhani's attack. (Actually it had never occurred
6446 to me that the half of 'md_local' used for chaining was the
6447 half from which PRNG output bytes were taken -- I had always
6448 assumed that the secret half would be used.) The second
6449 measure makes sure that additional data from 'state' is never
6450 mixed into 'md_local' in small portions; this heuristically
6451 further strengthens the PRNG.
6454 *) Fix crypto/bn/asm/mips3.s.
6457 *) When only the key is given to "enc", the IV is undefined. Print out
6458 an error message in this case.
6461 *) Handle special case when X509_NAME is empty in X509 printing routines.
6464 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
6465 positive and less than q.
6468 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
6469 used: it isn't thread safe and the add_lock_callback should handle
6471 [Paul Rose <Paul.Rose@bridge.com>]
6473 *) Verify that incoming data obeys the block size in
6474 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
6478 [Ulf Möller, Bodo Möller]
6480 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
6481 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
6482 when fixing the server behaviour for backwards-compatible 'client
6483 hello' messages. (Note that the attack is impractical against
6484 SSL 3.0 and TLS 1.0 anyway because length and version checking
6485 means that the probability of guessing a valid ciphertext is
6486 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
6489 Before 0.9.5, the countermeasure (hide the error by generating a
6490 random 'decryption result') did not work properly because
6491 ERR_clear_error() was missing, meaning that SSL_get_error() would
6492 detect the supposedly ignored error.
6494 Both problems are now fixed.
6497 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
6498 (previously it was 1024).
6501 *) Fix for compatibility mode trust settings: ignore trust settings
6502 unless some valid trust or reject settings are present.
6505 *) Fix for blowfish EVP: its a variable length cipher.
6508 *) Fix various bugs related to DSA S/MIME verification. Handle missing
6509 parameters in DSA public key structures and return an error in the
6510 DSA routines if parameters are absent.
6513 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
6514 in the current directory if neither $RANDFILE nor $HOME was set.
6515 RAND_file_name() in 0.9.6a returned NULL in this case. This has
6516 caused some confusion to Windows users who haven't defined $HOME.
6517 Thus RAND_file_name() is changed again: e_os.h can define a
6518 DEFAULT_HOME, which will be used if $HOME is not set.
6519 For Windows, we use "C:"; on other platforms, we still require
6520 environment variables.
6522 *) Move 'if (!initialized) RAND_poll()' into regions protected by
6523 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
6524 having multiple threads call RAND_poll() concurrently.
6527 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
6528 combination of a flag and a thread ID variable.
6529 Otherwise while one thread is in ssleay_rand_bytes (which sets the
6530 flag), *other* threads can enter ssleay_add_bytes without obeying
6531 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
6532 that they do not hold after the first thread unsets add_do_not_lock).
6535 *) Change bctest again: '-x' expressions are not available in all
6539 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
6541 *) Fix a couple of memory leaks in PKCS7_dataDecode()
6542 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
6544 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
6545 the default extension for executables, if any. Also, make the perl
6546 scripts that use symlink() to test if it really exists and use "cp"
6547 if it doesn't. All this made OpenSSL compilable and installable in
6551 *) Fix for asn1_GetSequence() for indefinite length constructed data.
6552 If SEQUENCE is length is indefinite just set c->slen to the total
6553 amount of data available.
6554 [Steve Henson, reported by shige@FreeBSD.org]
6555 [This change does not apply to 0.9.7.]
6557 *) Change bctest to avoid here-documents inside command substitution
6558 (workaround for FreeBSD /bin/sh bug).
6559 For compatibility with Ultrix, avoid shell functions (introduced
6560 in the bctest version that searches along $PATH).
6563 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6564 with des_encrypt() defined on some operating systems, like Solaris
6568 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6569 On the Importance of Eliminating Errors in Cryptographic
6570 Computations, J. Cryptology 14 (2001) 2, 101-119,
6571 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6574 *) MIPS assembler BIGNUM division bug fix.
6577 *) Disabled incorrect Alpha assembler code.
6580 *) Fix PKCS#7 decode routines so they correctly update the length
6581 after reading an EOC for the EXPLICIT tag.
6583 [This change does not apply to 0.9.7.]
6585 *) Fix bug in PKCS#12 key generation routines. This was triggered
6586 if a 3DES key was generated with a 0 initial byte. Include
6587 PKCS12_BROKEN_KEYGEN compilation option to retain the old
6588 (but broken) behaviour.
6591 *) Enhance bctest to search for a working bc along $PATH and print
6593 [Tim Rice <tim@multitalents.net> via Richard Levitte]
6595 *) Fix memory leaks in err.c: free err_data string if necessary;
6596 don't write to the wrong index in ERR_set_error_data.
6599 *) Implement ssl23_peek (analogous to ssl23_read), which previously
6603 *) Replace rdtsc with _emit statements for VC++ version 5.
6604 [Jeremy Cooper <jeremy@baymoo.org>]
6606 *) Make it possible to reuse SSLv2 sessions.
6609 *) In copy_email() check for >= 0 as a return value for
6610 X509_NAME_get_index_by_NID() since 0 is a valid index.
6611 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
6613 *) Avoid coredump with unsupported or invalid public keys by checking if
6614 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
6615 PKCS7_verify() fails with non detached data.
6618 *) Don't use getenv in library functions when run as setuid/setgid.
6619 New function OPENSSL_issetugid().
6622 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
6623 due to incorrect handling of multi-threading:
6625 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
6627 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
6629 3. Count how many times MemCheck_off() has been called so that
6630 nested use can be treated correctly. This also avoids
6631 inband-signalling in the previous code (which relied on the
6632 assumption that thread ID 0 is impossible).
6635 *) Add "-rand" option also to s_client and s_server.
6638 *) Fix CPU detection on Irix 6.x.
6639 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
6640 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6642 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
6645 [This change does not apply to 0.9.7.]
6647 *) Use the cached encoding of an X509_NAME structure rather than
6648 copying it. This is apparently the reason for the libsafe "errors"
6649 but the code is actually correct.
6652 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
6653 Bleichenbacher's DSA attack.
6654 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
6655 to be set and top=0 forces the highest bit to be set; top=-1 is new
6656 and leaves the highest bit random.
6657 [Ulf Moeller, Bodo Moeller]
6659 *) In the NCONF_...-based implementations for CONF_... queries
6660 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
6661 a temporary CONF structure with the data component set to NULL
6662 (which gives segmentation faults in lh_retrieve).
6663 Instead, use NULL for the CONF pointer in CONF_get_string and
6664 CONF_get_number (which may use environment variables) and directly
6665 return NULL from CONF_get_section.
6668 *) Fix potential buffer overrun for EBCDIC.
6671 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
6672 keyUsage if basicConstraints absent for a CA.
6675 *) Make SMIME_write_PKCS7() write mail header values with a format that
6676 is more generally accepted (no spaces before the semicolon), since
6677 some programs can't parse those values properly otherwise. Also make
6678 sure BIO's that break lines after each write do not create invalid
6682 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
6683 macros previously used would not encode an empty SEQUENCE OF
6684 and break the signature.
6686 [This change does not apply to 0.9.7.]
6688 *) Zero the premaster secret after deriving the master secret in
6692 *) Add some EVP_add_digest_alias registrations (as found in
6693 OpenSSL_add_all_digests()) to SSL_library_init()
6694 aka OpenSSL_add_ssl_algorithms(). This provides improved
6695 compatibility with peers using X.509 certificates
6696 with unconventional AlgorithmIdentifier OIDs.
6699 *) Fix for Irix with NO_ASM.
6700 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6702 *) ./config script fixes.
6703 [Ulf Moeller, Richard Levitte]
6705 *) Fix 'openssl passwd -1'.
6708 *) Change PKCS12_key_gen_asc() so it can cope with non null
6709 terminated strings whose length is passed in the passlen
6710 parameter, for example from PEM callbacks. This was done
6711 by adding an extra length parameter to asc2uni().
6712 [Steve Henson, reported by <oddissey@samsung.co.kr>]
6714 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
6715 call failed, free the DSA structure.
6718 *) Fix to uni2asc() to cope with zero length Unicode strings.
6719 These are present in some PKCS#12 files.
6722 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
6723 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
6724 when writing a 32767 byte record.
6725 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
6727 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
6728 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
6730 (RSA objects have a reference count access to which is protected
6731 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
6732 so they are meant to be shared between threads.)
6733 [Bodo Moeller, Geoff Thorpe; original patch submitted by
6734 "Reddie, Steven" <Steven.Reddie@ca.com>]
6736 *) Fix a deadlock in CRYPTO_mem_leaks().
6739 *) Use better test patterns in bntest.
6742 *) rand_win.c fix for Borland C.
6745 *) BN_rshift bugfix for n == 0.
6748 *) Add a 'bctest' script that checks for some known 'bc' bugs
6749 so that 'make test' does not abort just because 'bc' is broken.
6752 *) Store verify_result within SSL_SESSION also for client side to
6753 avoid potential security hole. (Re-used sessions on the client side
6754 always resulted in verify_result==X509_V_OK, not using the original
6755 result of the server certificate verification.)
6758 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
6759 SSL3_RT_APPLICATION_DATA, return 0.
6760 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
6764 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
6765 releases, have been re-implemented by renaming the previous
6766 implementations of ssl2_read and ssl3_read to ssl2_read_internal
6767 and ssl3_read_internal, respectively, and adding 'peek' parameters
6768 to them. The new ssl[23]_{read,peek} functions are calls to
6769 ssl[23]_read_internal with the 'peek' flag set appropriately.
6770 A 'peek' parameter has also been added to ssl3_read_bytes, which
6771 does the actual work for ssl3_read_internal.
6774 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
6775 the method-specific "init()" handler. Also clean up ex_data after
6776 calling the method-specific "finish()" handler. Previously, this was
6777 happening the other way round.
6780 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
6781 The previous value, 12, was not always sufficient for BN_mod_exp().
6784 *) Make sure that shared libraries get the internal name engine with
6785 the full version number and not just 0. This should mark the
6786 shared libraries as not backward compatible. Of course, this should
6787 be changed again when we can guarantee backward binary compatibility.
6790 *) Fix typo in get_cert_by_subject() in by_dir.c
6791 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
6793 *) Rework the system to generate shared libraries:
6795 - Make note of the expected extension for the shared libraries and
6796 if there is a need for symbolic links from for example libcrypto.so.0
6797 to libcrypto.so.0.9.7. There is extended info in Configure for
6800 - Make as few rebuilds of the shared libraries as possible.
6802 - Still avoid linking the OpenSSL programs with the shared libraries.
6804 - When installing, install the shared libraries separately from the
6808 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
6810 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
6811 and not in SSL_clear because the latter is also used by the
6812 accept/connect functions; previously, the settings made by
6813 SSL_set_read_ahead would be lost during the handshake.
6814 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
6816 *) Correct util/mkdef.pl to be selective about disabled algorithms.
6817 Previously, it would create entries for disableed algorithms no
6821 *) Added several new manual pages for SSL_* function.
6824 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
6826 *) In ssl23_get_client_hello, generate an error message when faced
6827 with an initial SSL 3.0/TLS record that is too small to contain the
6828 first two bytes of the ClientHello message, i.e. client_version.
6829 (Note that this is a pathologic case that probably has never happened
6830 in real life.) The previous approach was to use the version number
6831 from the record header as a substitute; but our protocol choice
6832 should not depend on that one because it is not authenticated
6833 by the Finished messages.
6836 *) More robust randomness gathering functions for Windows.
6837 [Jeffrey Altman <jaltman@columbia.edu>]
6839 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
6840 not set then we don't setup the error code for issuer check errors
6841 to avoid possibly overwriting other errors which the callback does
6842 handle. If an application does set the flag then we assume it knows
6843 what it is doing and can handle the new informational codes
6847 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
6848 a general "ANY" type, as such it should be able to decode anything
6849 including tagged types. However it didn't check the class so it would
6850 wrongly interpret tagged types in the same way as their universal
6851 counterpart and unknown types were just rejected. Changed so that the
6852 tagged and unknown types are handled in the same way as a SEQUENCE:
6853 that is the encoding is stored intact. There is also a new type
6854 "V_ASN1_OTHER" which is used when the class is not universal, in this
6855 case we have no idea what the actual type is so we just lump them all
6859 *) On VMS, stdout may very well lead to a file that is written to
6860 in a record-oriented fashion. That means that every write() will
6861 write a separate record, which will be read separately by the
6862 programs trying to read from it. This can be very confusing.
6864 The solution is to put a BIO filter in the way that will buffer
6865 text until a linefeed is reached, and then write everything a
6866 line at a time, so every record written will be an actual line,
6867 not chunks of lines and not (usually doesn't happen, but I've
6868 seen it once) several lines in one record. BIO_f_linebuffer() is
6871 Currently, it's a VMS-only method, because that's where it has
6872 been tested well enough.
6875 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
6876 it can return incorrect results.
6877 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
6878 but it was in 0.9.6-beta[12].)
6881 *) Disable the check for content being present when verifying detached
6882 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
6883 include zero length content when signing messages.
6886 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
6887 BIO_ctrl (for BIO pairs).
6890 *) Add DSO method for VMS.
6893 *) Bug fix: Montgomery multiplication could produce results with the
6897 *) Add RPM specification openssl.spec and modify it to build three
6898 packages. The default package contains applications, application
6899 documentation and run-time libraries. The devel package contains
6900 include files, static libraries and function documentation. The
6901 doc package contains the contents of the doc directory. The original
6902 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
6905 *) Add a large number of documentation files for many SSL routines.
6906 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6908 *) Add a configuration entry for Sony News 4.
6909 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
6911 *) Don't set the two most significant bits to one when generating a
6912 random number < q in the DSA library.
6915 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
6916 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
6917 the underlying transport is blocking) if a handshake took place.
6918 (The default behaviour is needed by applications such as s_client
6919 and s_server that use select() to determine when to use SSL_read;
6920 but for applications that know in advance when to expect data, it
6921 just makes things more complicated.)
6924 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
6928 *) Add a few more EBCDIC conditionals that make `req' and `x509'
6929 work better on such systems.
6930 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6932 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
6933 Update PKCS12_parse() so it copies the friendlyName and the
6934 keyid to the certificates aux info.
6937 *) Fix bug in PKCS7_verify() which caused an infinite loop
6938 if there was more than one signature.
6939 [Sven Uszpelkat <su@celocom.de>]
6941 *) Major change in util/mkdef.pl to include extra information
6942 about each symbol, as well as presentig variables as well
6943 as functions. This change means that there's n more need
6944 to rebuild the .num files when some algorithms are excluded.
6947 *) Allow the verify time to be set by an application,
6948 rather than always using the current time.
6951 *) Phase 2 verify code reorganisation. The certificate
6952 verify code now looks up an issuer certificate by a
6953 number of criteria: subject name, authority key id
6954 and key usage. It also verifies self signed certificates
6955 by the same criteria. The main comparison function is
6956 X509_check_issued() which performs these checks.
6958 Lot of changes were necessary in order to support this
6959 without completely rewriting the lookup code.
6961 Authority and subject key identifier are now cached.
6963 The LHASH 'certs' is X509_STORE has now been replaced
6964 by a STACK_OF(X509_OBJECT). This is mainly because an
6965 LHASH can't store or retrieve multiple objects with
6966 the same hash value.
6968 As a result various functions (which were all internal
6969 use only) have changed to handle the new X509_STORE
6970 structure. This will break anything that messed round
6971 with X509_STORE internally.
6973 The functions X509_STORE_add_cert() now checks for an
6974 exact match, rather than just subject name.
6976 The X509_STORE API doesn't directly support the retrieval
6977 of multiple certificates matching a given criteria, however
6978 this can be worked round by performing a lookup first
6979 (which will fill the cache with candidate certificates)
6980 and then examining the cache for matches. This is probably
6981 the best we can do without throwing out X509_LOOKUP
6982 entirely (maybe later...).
6984 The X509_VERIFY_CTX structure has been enhanced considerably.
6986 All certificate lookup operations now go via a get_issuer()
6987 callback. Although this currently uses an X509_STORE it
6988 can be replaced by custom lookups. This is a simple way
6989 to bypass the X509_STORE hackery necessary to make this
6990 work and makes it possible to use more efficient techniques
6991 in future. A very simple version which uses a simple
6992 STACK for its trusted certificate store is also provided
6993 using X509_STORE_CTX_trusted_stack().
6995 The verify_cb() and verify() callbacks now have equivalents
6996 in the X509_STORE_CTX structure.
6998 X509_STORE_CTX also has a 'flags' field which can be used
6999 to customise the verify behaviour.
7002 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
7003 excludes S/MIME capabilities.
7006 *) When a certificate request is read in keep a copy of the
7007 original encoding of the signed data and use it when outputing
7008 again. Signatures then use the original encoding rather than
7009 a decoded, encoded version which may cause problems if the
7010 request is improperly encoded.
7013 *) For consistency with other BIO_puts implementations, call
7014 buffer_write(b, ...) directly in buffer_puts instead of calling
7017 In BIO_puts, increment b->num_write as in BIO_write.
7018 [Peter.Sylvester@EdelWeb.fr]
7020 *) Fix BN_mul_word for the case where the word is 0. (We have to use
7021 BN_zero, we may not return a BIGNUM with an array consisting of
7025 *) Avoid calling abort() from within the library when problems are
7026 detected, except if preprocessor symbols have been defined
7027 (such as REF_CHECK, BN_DEBUG etc.).
7030 *) New openssl application 'rsautl'. This utility can be
7031 used for low level RSA operations. DER public key
7032 BIO/fp routines also added.
7035 *) New Configure entry and patches for compiling on QNX 4.
7036 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
7038 *) A demo state-machine implementation was sponsored by
7039 Nuron (http://www.nuron.com/) and is now available in
7040 demos/state_machine.
7043 *) New options added to the 'dgst' utility for signature
7044 generation and verification.
7047 *) Unrecognized PKCS#7 content types are now handled via a
7048 catch all ASN1_TYPE structure. This allows unsupported
7049 types to be stored as a "blob" and an application can
7050 encode and decode it manually.
7053 *) Fix various signed/unsigned issues to make a_strex.c
7055 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
7057 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
7058 length if passed a buffer. ASN1_INTEGER_to_BN failed
7059 if passed a NULL BN and its argument was negative.
7060 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
7062 *) Modification to PKCS#7 encoding routines to output definite
7063 length encoding. Since currently the whole structures are in
7064 memory there's not real point in using indefinite length
7065 constructed encoding. However if OpenSSL is compiled with
7066 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
7069 *) Added BIO_vprintf() and BIO_vsnprintf().
7072 *) Added more prefixes to parse for in the the strings written
7073 through a logging bio, to cover all the levels that are available
7074 through syslog. The prefixes are now:
7076 PANIC, EMERG, EMR => LOG_EMERG
7077 ALERT, ALR => LOG_ALERT
7078 CRIT, CRI => LOG_CRIT
7079 ERROR, ERR => LOG_ERR
7080 WARNING, WARN, WAR => LOG_WARNING
7081 NOTICE, NOTE, NOT => LOG_NOTICE
7082 INFO, INF => LOG_INFO
7083 DEBUG, DBG => LOG_DEBUG
7085 and as before, if none of those prefixes are present at the
7086 beginning of the string, LOG_ERR is chosen.
7088 On Win32, the LOG_* levels are mapped according to this:
7090 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
7091 LOG_WARNING => EVENTLOG_WARNING_TYPE
7092 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
7096 *) Made it possible to reconfigure with just the configuration
7097 argument "reconf" or "reconfigure". The command line arguments
7098 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
7099 and are retrieved from there when reconfiguring.
7103 [Assar Westerlund <assar@sics.se>, Richard Levitte]
7105 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
7108 *) The obj_dat.pl script was messing up the sorting of object
7109 names. The reason was that it compared the quoted version
7110 of strings as a result "OCSP" > "OCSP Signing" because
7111 " > SPACE. Changed script to store unquoted versions of
7112 names and add quotes on output. It was also omitting some
7113 names from the lookup table if they were given a default
7114 value (that is if SN is missing it is given the same
7115 value as LN and vice versa), these are now added on the
7116 grounds that if an object has a name we should be able to
7117 look it up. Finally added warning output when duplicate
7118 short or long names are found.
7121 *) Changes needed for Tandem NSK.
7122 [Scott Uroff <scott@xypro.com>]
7124 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
7125 RSA_padding_check_SSLv23(), special padding was never detected
7126 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
7127 version rollback attacks was not effective.
7129 In s23_clnt.c, don't use special rollback-attack detection padding
7130 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
7131 client; similarly, in s23_srvr.c, don't do the rollback check if
7132 SSL 2.0 is the only protocol enabled in the server.
7135 *) Make it possible to get hexdumps of unprintable data with 'openssl
7136 asn1parse'. By implication, the functions ASN1_parse_dump() and
7137 BIO_dump_indent() are added.
7140 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
7141 these print out strings and name structures based on various
7142 flags including RFC2253 support and proper handling of
7143 multibyte characters. Added options to the 'x509' utility
7144 to allow the various flags to be set.
7147 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
7148 Also change the functions X509_cmp_current_time() and
7149 X509_gmtime_adj() work with an ASN1_TIME structure,
7150 this will enable certificates using GeneralizedTime in validity
7151 dates to be checked.
7154 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
7155 negative public key encodings) on by default,
7156 NO_NEG_PUBKEY_BUG can be set to disable it.
7159 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
7160 content octets. An i2c_ASN1_OBJECT is unnecessary because
7161 the encoding can be trivially obtained from the structure.
7164 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
7165 not read locks (CRYPTO_r_[un]lock).
7168 *) A first attempt at creating official support for shared
7169 libraries through configuration. I've kept it so the
7170 default is static libraries only, and the OpenSSL programs
7171 are always statically linked for now, but there are
7172 preparations for dynamic linking in place.
7173 This has been tested on Linux and Tru64.
7176 *) Randomness polling function for Win9x, as described in:
7177 Peter Gutmann, Software Generation of Practically Strong
7181 *) Fix so PRNG is seeded in req if using an already existing
7185 *) New options to smime application. -inform and -outform
7186 allow alternative formats for the S/MIME message including
7187 PEM and DER. The -content option allows the content to be
7188 specified separately. This should allow things like Netscape
7189 form signing output easier to verify.
7192 *) Fix the ASN1 encoding of tags using the 'long form'.
7195 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
7196 STRING types. These convert content octets to and from the
7197 underlying type. The actual tag and length octets are
7198 already assumed to have been read in and checked. These
7199 are needed because all other string types have virtually
7200 identical handling apart from the tag. By having versions
7201 of the ASN1 functions that just operate on content octets
7202 IMPLICIT tagging can be handled properly. It also allows
7203 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
7204 and ASN1_INTEGER are identical apart from the tag.
7207 *) Change the handling of OID objects as follows:
7209 - New object identifiers are inserted in objects.txt, following
7210 the syntax given in objects.README.
7211 - objects.pl is used to process obj_mac.num and create a new
7213 - obj_dat.pl is used to create a new obj_dat.h, using the data in
7216 This is currently kind of a hack, and the perl code in objects.pl
7217 isn't very elegant, but it works as I intended. The simplest way
7218 to check that it worked correctly is to look in obj_dat.h and
7219 check the array nid_objs and make sure the objects haven't moved
7220 around (this is important!). Additions are OK, as well as
7221 consistent name changes.
7224 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
7227 *) Addition of the command line parameter '-rand file' to 'openssl req'.
7228 The given file adds to whatever has already been seeded into the
7229 random pool through the RANDFILE configuration file option or
7230 environment variable, or the default random state file.
7233 *) mkstack.pl now sorts each macro group into lexical order.
7234 Previously the output order depended on the order the files
7235 appeared in the directory, resulting in needless rewriting
7239 *) Patches to make OpenSSL compile under Win32 again. Mostly
7240 work arounds for the VC++ problem that it treats func() as
7241 func(void). Also stripped out the parts of mkdef.pl that
7242 added extra typesafe functions: these no longer exist.
7245 *) Reorganisation of the stack code. The macros are now all
7246 collected in safestack.h . Each macro is defined in terms of
7247 a "stack macro" of the form SKM_<name>(type, a, b). The
7248 DEBUG_SAFESTACK is now handled in terms of function casts,
7249 this has the advantage of retaining type safety without the
7250 use of additional functions. If DEBUG_SAFESTACK is not defined
7251 then the non typesafe macros are used instead. Also modified the
7252 mkstack.pl script to handle the new form. Needs testing to see
7253 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
7254 the default if no major problems. Similar behaviour for ASN1_SET_OF
7255 and PKCS12_STACK_OF.
7258 *) When some versions of IIS use the 'NET' form of private key the
7259 key derivation algorithm is different. Normally MD5(password) is
7260 used as a 128 bit RC4 key. In the modified case
7261 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
7262 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
7263 as the old Netscape_RSA functions except they have an additional
7264 'sgckey' parameter which uses the modified algorithm. Also added
7265 an -sgckey command line option to the rsa utility. Thanks to
7266 Adrian Peck <bertie@ncipher.com> for posting details of the modified
7267 algorithm to openssl-dev.
7270 *) The evp_local.h macros were using 'c.##kname' which resulted in
7271 invalid expansion on some systems (SCO 5.0.5 for example).
7272 Corrected to 'c.kname'.
7273 [Phillip Porch <root@theporch.com>]
7275 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
7276 a STACK of email addresses from a certificate or request, these look
7277 in the subject name and the subject alternative name extensions and
7278 omit any duplicate addresses.
7281 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
7282 This makes DSA verification about 2 % faster.
7285 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
7286 (meaning that now 2^5 values will be precomputed, which is only 4 KB
7287 plus overhead for 1024 bit moduli).
7288 This makes exponentiations about 0.5 % faster for 1024 bit
7289 exponents (as measured by "openssl speed rsa2048").
7292 *) Rename memory handling macros to avoid conflicts with other
7294 Malloc => OPENSSL_malloc
7295 Malloc_locked => OPENSSL_malloc_locked
7296 Realloc => OPENSSL_realloc
7297 Free => OPENSSL_free
7300 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
7301 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
7304 *) CygWin32 support.
7305 [John Jarvie <jjarvie@newsguy.com>]
7307 *) The type-safe stack code has been rejigged. It is now only compiled
7308 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
7309 by default all type-specific stack functions are "#define"d back to
7310 standard stack functions. This results in more streamlined output
7311 but retains the type-safety checking possibilities of the original
7315 *) The STACK code has been cleaned up, and certain type declarations
7316 that didn't make a lot of sense have been brought in line. This has
7317 also involved a cleanup of sorts in safestack.h to more correctly
7318 map type-safe stack functions onto their plain stack counterparts.
7319 This work has also resulted in a variety of "const"ifications of
7320 lots of the code, especially "_cmp" operations which should normally
7321 be prototyped with "const" parameters anyway.
7324 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
7325 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
7326 (The PRNG state consists of two parts, the large pool 'state' and 'md',
7327 where all of 'md' is used each time the PRNG is used, but 'state'
7328 is used only indexed by a cyclic counter. As entropy may not be
7329 well distributed from the beginning, 'md' is important as a
7330 chaining variable. However, the output function chains only half
7331 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
7332 all of 'md', and seeding with STATE_SIZE dummy bytes will result
7333 in all of 'state' being rewritten, with the new values depending
7334 on virtually all of 'md'. This overcomes the 80 bit limitation.)
7337 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
7338 the handshake is continued after ssl_verify_cert_chain();
7339 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
7340 can lead to 'unexplainable' connection aborts later.
7341 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
7343 *) Major EVP API cipher revision.
7344 Add hooks for extra EVP features. This allows various cipher
7345 parameters to be set in the EVP interface. Support added for variable
7346 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
7347 setting of RC2 and RC5 parameters.
7349 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
7352 Remove lots of duplicated code from the EVP library. For example *every*
7353 cipher init() function handles the 'iv' in the same way according to the
7354 cipher mode. They also all do nothing if the 'key' parameter is NULL and
7355 for CFB and OFB modes they zero ctx->num.
7357 New functionality allows removal of S/MIME code RC2 hack.
7359 Most of the routines have the same form and so can be declared in terms
7362 By shifting this to the top level EVP_CipherInit() it can be removed from
7363 all individual ciphers. If the cipher wants to handle IVs or keys
7364 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
7367 Change lots of functions like EVP_EncryptUpdate() to now return a
7368 value: although software versions of the algorithms cannot fail
7369 any installed hardware versions can.
7372 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
7373 this option is set, tolerate broken clients that send the negotiated
7374 protocol version number instead of the requested protocol version
7378 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
7379 i.e. non-zero for export ciphersuites, zero otherwise.
7380 Previous versions had this flag inverted, inconsistent with
7381 rsa_tmp_cb (..._TMP_RSA_CB).
7382 [Bodo Moeller; problem reported by Amit Chopra]
7384 *) Add missing DSA library text string. Work around for some IIS
7385 key files with invalid SEQUENCE encoding.
7388 *) Add a document (doc/standards.txt) that list all kinds of standards
7389 and so on that are implemented in OpenSSL.
7392 *) Enhance c_rehash script. Old version would mishandle certificates
7393 with the same subject name hash and wouldn't handle CRLs at all.
7394 Added -fingerprint option to crl utility, to support new c_rehash
7398 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
7401 *) Fix for SSL server purpose checking. Server checking was
7402 rejecting certificates which had extended key usage present
7403 but no ssl client purpose.
7404 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
7406 *) Make PKCS#12 code work with no password. The PKCS#12 spec
7407 is a little unclear about how a blank password is handled.
7408 Since the password in encoded as a BMPString with terminating
7409 double NULL a zero length password would end up as just the
7410 double NULL. However no password at all is different and is
7411 handled differently in the PKCS#12 key generation code. NS
7412 treats a blank password as zero length. MSIE treats it as no
7413 password on export: but it will try both on import. We now do
7414 the same: PKCS12_parse() tries zero length and no password if
7415 the password is set to "" or NULL (NULL is now a valid password:
7416 it wasn't before) as does the pkcs12 application.
7419 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
7420 perror when PEM_read_bio_X509_REQ fails, the error message must
7421 be obtained from the error queue.
7424 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
7425 it in ERR_remove_state if appropriate, and change ERR_get_state
7426 accordingly to avoid race conditions (this is necessary because
7427 thread_hash is no longer constant once set).
7430 *) Bugfix for linux-elf makefile.one.
7433 *) RSA_get_default_method() will now cause a default
7434 RSA_METHOD to be chosen if one doesn't exist already.
7435 Previously this was only set during a call to RSA_new()
7436 or RSA_new_method(NULL) meaning it was possible for
7437 RSA_get_default_method() to return NULL.
7440 *) Added native name translation to the existing DSO code
7441 that will convert (if the flag to do so is set) filenames
7442 that are sufficiently small and have no path information
7443 into a canonical native form. Eg. "blah" converted to
7444 "libblah.so" or "blah.dll" etc.
7447 *) New function ERR_error_string_n(e, buf, len) which is like
7448 ERR_error_string(e, buf), but writes at most 'len' bytes
7449 including the 0 terminator. For ERR_error_string_n, 'buf'
7451 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
7453 *) CONF library reworked to become more general. A new CONF
7454 configuration file reader "class" is implemented as well as a
7455 new functions (NCONF_*, for "New CONF") to handle it. The now
7456 old CONF_* functions are still there, but are reimplemented to
7457 work in terms of the new functions. Also, a set of functions
7458 to handle the internal storage of the configuration data is
7459 provided to make it easier to write new configuration file
7460 reader "classes" (I can definitely see something reading a
7461 configuration file in XML format, for example), called _CONF_*,
7462 or "the configuration storage API"...
7464 The new configuration file reading functions are:
7466 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
7467 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
7469 NCONF_default, NCONF_WIN32
7471 NCONF_dump_fp, NCONF_dump_bio
7473 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
7474 NCONF_new creates a new CONF object. This works in the same way
7475 as other interfaces in OpenSSL, like the BIO interface.
7476 NCONF_dump_* dump the internal storage of the configuration file,
7477 which is useful for debugging. All other functions take the same
7478 arguments as the old CONF_* functions wth the exception of the
7479 first that must be a `CONF *' instead of a `LHASH *'.
7481 To make it easer to use the new classes with the old CONF_* functions,
7482 the function CONF_set_default_method is provided.
7485 *) Add '-tls1' option to 'openssl ciphers', which was already
7486 mentioned in the documentation but had not been implemented.
7487 (This option is not yet really useful because even the additional
7488 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
7491 *) Initial DSO code added into libcrypto for letting OpenSSL (and
7492 OpenSSL-based applications) load shared libraries and bind to
7493 them in a portable way.
7494 [Geoff Thorpe, with contributions from Richard Levitte]
7496 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
7498 *) Make sure _lrotl and _lrotr are only used with MSVC.
7500 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
7501 (the default implementation of RAND_status).
7503 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
7504 to '-clrext' (= clear extensions), as intended and documented.
7505 [Bodo Moeller; inconsistency pointed out by Michael Attili
7508 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
7509 was larger than the MD block size.
7510 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
7512 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
7513 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
7514 using the passed key: if the passed key was a private key the result
7515 of X509_print(), for example, would be to print out all the private key
7519 *) des_quad_cksum() byte order bug fix.
7520 [Ulf Möller, using the problem description in krb4-0.9.7, where
7521 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
7523 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
7525 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
7527 *) For easily testing in shell scripts whether some command
7528 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
7529 returns with exit code 0 iff no command of the given name is available.
7530 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
7531 the output goes to stdout and nothing is printed to stderr.
7532 Additional arguments are always ignored.
7534 Since for each cipher there is a command of the same name,
7535 the 'no-cipher' compilation switches can be tested this way.
7537 ('openssl no-XXX' is not able to detect pseudo-commands such
7538 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
7541 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
7544 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
7545 is set; it will be thrown away anyway because each handshake creates
7547 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
7548 to parameters -- in previous versions (since OpenSSL 0.9.3) the
7549 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
7550 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7553 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7554 'Q' and 'R' lose their special meanings (quit/renegotiate).
7555 This is part of what -quiet does; unlike -quiet, -ign_eof
7556 does not suppress any output.
7559 *) Add compatibility options to the purpose and trust code. The
7560 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7561 accepts a certificate or CA, this was the previous behaviour,
7562 with all the associated security issues.
7564 X509_TRUST_COMPAT is the old trust behaviour: only and
7565 automatically trust self signed roots in certificate store. A
7566 new trust setting X509_TRUST_DEFAULT is used to specify that
7567 a purpose has no associated trust setting and it should instead
7568 use the value in the default purpose.
7571 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7572 and fix a memory leak.
7575 *) In util/mkerr.pl (which implements 'make errors'), preserve
7576 reason strings from the previous version of the .c file, as
7577 the default to have only downcase letters (and digits) in
7578 automatically generated reasons codes is not always appropriate.
7581 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7582 using strerror. Previously, ERR_reason_error_string() returned
7583 library names as reason strings for SYSerr; but SYSerr is a special
7584 case where small numbers are errno values, not library numbers.
7587 *) Add '-dsaparam' option to 'openssl dhparam' application. This
7588 converts DSA parameters into DH parameters. (When creating parameters,
7589 DSA_generate_parameters is used.)
7592 *) Include 'length' (recommended exponent length) in C code generated
7593 by 'openssl dhparam -C'.
7596 *) The second argument to set_label in perlasm was already being used
7597 so couldn't be used as a "file scope" flag. Moved to third argument
7601 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
7602 instead of RAND_bytes for encryption IVs and salts.
7605 *) Include RAND_status() into RAND_METHOD instead of implementing
7606 it only for md_rand.c Otherwise replacing the PRNG by calling
7607 RAND_set_rand_method would be impossible.
7610 *) Don't let DSA_generate_key() enter an infinite loop if the random
7611 number generation fails.
7614 *) New 'rand' application for creating pseudo-random output.
7617 *) Added configuration support for Linux/IA64
7618 [Rolf Haberrecker <rolf@suse.de>]
7620 *) Assembler module support for Mingw32.
7623 *) Shared library support for HPUX (in shlib/).
7624 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
7626 *) Shared library support for Solaris gcc.
7627 [Lutz Behnke <behnke@trustcenter.de>]
7629 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
7631 *) PKCS7_encrypt() was adding text MIME headers twice because they
7632 were added manually and by SMIME_crlf_copy().
7635 *) In bntest.c don't call BN_rand with zero bits argument.
7636 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
7638 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
7639 case was implemented. This caused BN_div_recp() to fail occasionally.
7642 *) Add an optional second argument to the set_label() in the perl
7643 assembly language builder. If this argument exists and is set
7644 to 1 it signals that the assembler should use a symbol whose
7645 scope is the entire file, not just the current function. This
7646 is needed with MASM which uses the format label:: for this scope.
7647 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
7649 *) Change the ASN1 types so they are typedefs by default. Before
7650 almost all types were #define'd to ASN1_STRING which was causing
7651 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
7655 *) Change names of new functions to the new get1/get0 naming
7656 convention: After 'get1', the caller owns a reference count
7657 and has to call ..._free; 'get0' returns a pointer to some
7658 data structure without incrementing reference counters.
7659 (Some of the existing 'get' functions increment a reference
7660 counter, some don't.)
7661 Similarly, 'set1' and 'add1' functions increase reference
7662 counters or duplicate objects.
7665 *) Allow for the possibility of temp RSA key generation failure:
7666 the code used to assume it always worked and crashed on failure.
7669 *) Fix potential buffer overrun problem in BIO_printf().
7670 [Ulf Möller, using public domain code by Patrick Powell; problem
7671 pointed out by David Sacerdote <das33@cornell.edu>]
7673 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
7674 RAND_egd() and RAND_status(). In the command line application,
7675 the EGD socket can be specified like a seed file using RANDFILE
7679 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
7680 Some CAs (e.g. Verisign) distribute certificates in this form.
7683 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
7684 list to exclude them. This means that no special compilation option
7685 is needed to use anonymous DH: it just needs to be included in the
7689 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
7690 EVP_MD_type. The old functionality is available in a new macro called
7691 EVP_MD_md(). Change code that uses it and update docs.
7694 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
7695 where the 'void *' argument is replaced by a function pointer argument.
7696 Previously 'void *' was abused to point to functions, which works on
7697 many platforms, but is not correct. As these functions are usually
7698 called by macros defined in OpenSSL header files, most source code
7699 should work without changes.
7702 *) <openssl/opensslconf.h> (which is created by Configure) now contains
7703 sections with information on -D... compiler switches used for
7704 compiling the library so that applications can see them. To enable
7705 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
7706 must be defined. E.g.,
7707 #define OPENSSL_ALGORITHM_DEFINES
7708 #include <openssl/opensslconf.h>
7709 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
7710 [Richard Levitte, Ulf and Bodo Möller]
7712 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
7716 *) Change the 'other' type in certificate aux info to a STACK_OF
7717 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
7718 the required ASN1 format: arbitrary types determined by an OID.
7721 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
7722 argument to 'req'. This is not because the function is newer or
7723 better than others it just uses the work 'NEW' in the certificate
7724 request header lines. Some software needs this.
7727 *) Reorganise password command line arguments: now passwords can be
7728 obtained from various sources. Delete the PEM_cb function and make
7729 it the default behaviour: i.e. if the callback is NULL and the
7730 usrdata argument is not NULL interpret it as a null terminated pass
7731 phrase. If usrdata and the callback are NULL then the pass phrase
7732 is prompted for as usual.
7735 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
7736 the support is automatically enabled. The resulting binaries will
7737 autodetect the card and use it if present.
7738 [Ben Laurie and Compaq Inc.]
7740 *) Work around for Netscape hang bug. This sends certificate request
7741 and server done in one record. Since this is perfectly legal in the
7742 SSL/TLS protocol it isn't a "bug" option and is on by default. See
7743 the bugs/SSLv3 entry for more info.
7746 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
7749 *) Add -rand argument to smime and pkcs12 applications and read/write
7753 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
7756 *) Add command line password options to the remaining applications.
7759 *) Bug fix for BN_div_recp() for numerators with an even number of
7763 *) More tests in bntest.c, and changed test_bn output.
7766 *) ./config recognizes MacOS X now.
7769 *) Bug fix for BN_div() when the first words of num and divsor are
7770 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
7773 *) Add support for various broken PKCS#8 formats, and command line
7774 options to produce them.
7777 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
7778 get temporary BIGNUMs from a BN_CTX.
7781 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
7785 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
7786 include a #define from the old name to the new. The original intent
7787 was that statically linked binaries could for example just call
7788 SSLeay_add_all_ciphers() to just add ciphers to the table and not
7789 link with digests. This never worked becayse SSLeay_add_all_digests()
7790 and SSLeay_add_all_ciphers() were in the same source file so calling
7791 one would link with the other. They are now in separate source files.
7794 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
7797 *) Use a less unusual form of the Miller-Rabin primality test (it used
7798 a binary algorithm for exponentiation integrated into the Miller-Rabin
7799 loop, our standard modexp algorithms are faster).
7802 *) Support for the EBCDIC character set completed.
7803 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
7805 *) Source code cleanups: use const where appropriate, eliminate casts,
7806 use void * instead of char * in lhash.
7809 *) Bugfix: ssl3_send_server_key_exchange was not restartable
7810 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
7811 this the server could overwrite ephemeral keys that the client
7815 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
7816 using 50 iterations of the Rabin-Miller test.
7818 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
7819 iterations of the Rabin-Miller test as required by the appendix
7820 to FIPS PUB 186[-1]) instead of DSA_is_prime.
7821 As BN_is_prime_fasttest includes trial division, DSA parameter
7822 generation becomes much faster.
7824 This implies a change for the callback functions in DSA_is_prime
7825 and DSA_generate_parameters: The callback function is called once
7826 for each positive witness in the Rabin-Miller test, not just
7827 occasionally in the inner loop; and the parameters to the
7828 callback function now provide an iteration count for the outer
7829 loop rather than for the current invocation of the inner loop.
7830 DSA_generate_parameters additionally can call the callback
7831 function with an 'iteration count' of -1, meaning that a
7832 candidate has passed the trial division test (when q is generated
7833 from an application-provided seed, trial division is skipped).
7836 *) New function BN_is_prime_fasttest that optionally does trial
7837 division before starting the Rabin-Miller test and has
7838 an additional BN_CTX * argument (whereas BN_is_prime always
7839 has to allocate at least one BN_CTX).
7840 'callback(1, -1, cb_arg)' is called when a number has passed the
7841 trial division stage.
7844 *) Fix for bug in CRL encoding. The validity dates weren't being handled
7848 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
7851 *) New function BN_pseudo_rand().
7854 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
7855 bignum version of BN_from_montgomery() with the working code from
7856 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
7860 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
7861 made it impossible to use the same SSL_SESSION data structure in
7862 SSL2 clients in multiple threads.
7865 *) The return value of RAND_load_file() no longer counts bytes obtained
7866 by stat(). RAND_load_file(..., -1) is new and uses the complete file
7867 to seed the PRNG (previously an explicit byte count was required).
7868 [Ulf Möller, Bodo Möller]
7870 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
7871 used (char *) instead of (void *) and had casts all over the place.
7874 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
7877 *) Retain source code compatibility for BN_prime_checks macro:
7878 BN_is_prime(..., BN_prime_checks, ...) now uses
7879 BN_prime_checks_for_size to determine the appropriate number of
7880 Rabin-Miller iterations.
7883 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
7884 DH_CHECK_P_NOT_SAFE_PRIME.
7885 (Check if this is true? OpenPGP calls them "strong".)
7888 *) Merge the functionality of "dh" and "gendh" programs into a new program
7889 "dhparam". The old programs are retained for now but will handle DH keys
7890 (instead of parameters) in future.
7893 *) Make the ciphers, s_server and s_client programs check the return values
7894 when a new cipher list is set.
7897 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
7898 ciphers. Before when the 56bit ciphers were enabled the sorting was
7901 The syntax for the cipher sorting has been extended to support sorting by
7902 cipher-strength (using the strength_bits hard coded in the tables).
7903 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
7905 Fix a bug in the cipher-command parser: when supplying a cipher command
7906 string with an "undefined" symbol (neither command nor alphanumeric
7907 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
7908 an error is flagged.
7910 Due to the strength-sorting extension, the code of the
7911 ssl_create_cipher_list() function was completely rearranged. I hope that
7912 the readability was also increased :-)
7913 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7915 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
7916 for the first serial number and places 2 in the serial number file. This
7917 avoids problems when the root CA is created with serial number zero and
7918 the first user certificate has the same issuer name and serial number
7922 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
7923 the new code. Add documentation for this stuff.
7926 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
7927 X509_*() to X509at_*() on the grounds that they don't handle X509
7928 structures and behave in an analagous way to the X509v3 functions:
7929 they shouldn't be called directly but wrapper functions should be used
7932 So we also now have some wrapper functions that call the X509at functions
7933 when passed certificate requests. (TO DO: similar things can be done with
7934 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
7935 things. Some of these need some d2i or i2d and print functionality
7936 because they handle more complex structures.)
7939 *) Add missing #ifndefs that caused missing symbols when building libssl
7940 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
7941 NO_RSA in ssl/s2*.c.
7942 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
7944 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
7945 has a return value which indicates the quality of the random data
7946 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
7947 error queue. New function RAND_pseudo_bytes() generates output that is
7948 guaranteed to be unique but not unpredictable. RAND_add is like
7949 RAND_seed, but takes an extra argument for an entropy estimate
7950 (RAND_seed always assumes full entropy).
7953 *) Do more iterations of Rabin-Miller probable prime test (specifically,
7954 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
7955 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
7956 in crypto/bn/bn_prime.c for the complete table). This guarantees a
7957 false-positive rate of at most 2^-80 for random input.
7960 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
7963 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
7964 in the 0.9.5 release), this returns the chain
7965 from an X509_CTX structure with a dup of the stack and all
7966 the X509 reference counts upped: so the stack will exist
7967 after X509_CTX_cleanup() has been called. Modify pkcs12.c
7970 Also make SSL_SESSION_print() print out the verify return
7974 *) Add manpage for the pkcs12 command. Also change the default
7975 behaviour so MAC iteration counts are used unless the new
7976 -nomaciter option is used. This improves file security and
7977 only older versions of MSIE (4.0 for example) need it.
7980 *) Honor the no-xxx Configure options when creating .DEF files.
7983 *) Add PKCS#10 attributes to field table: challengePassword,
7984 unstructuredName and unstructuredAddress. These are taken from
7985 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
7986 international characters are used.
7988 More changes to X509_ATTRIBUTE code: allow the setting of types
7989 based on strings. Remove the 'loc' parameter when adding
7990 attributes because these will be a SET OF encoding which is sorted
7994 *) Initial changes to the 'req' utility to allow request generation
7995 automation. This will allow an application to just generate a template
7996 file containing all the field values and have req construct the
7999 Initial support for X509_ATTRIBUTE handling. Stacks of these are
8000 used all over the place including certificate requests and PKCS#7
8001 structures. They are currently handled manually where necessary with
8002 some primitive wrappers for PKCS#7. The new functions behave in a
8003 manner analogous to the X509 extension functions: they allow
8004 attributes to be looked up by NID and added.
8006 Later something similar to the X509V3 code would be desirable to
8007 automatically handle the encoding, decoding and printing of the
8008 more complex types. The string types like challengePassword can
8009 be handled by the string table functions.
8011 Also modified the multi byte string table handling. Now there is
8012 a 'global mask' which masks out certain types. The table itself
8013 can use the flag STABLE_NO_MASK to ignore the mask setting: this
8014 is useful when for example there is only one permissible type
8015 (as in countryName) and using the mask might result in no valid
8019 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
8020 SSL_get_peer_finished to allow applications to obtain the latest
8021 Finished messages sent to the peer or expected from the peer,
8022 respectively. (SSL_get_peer_finished is usually the Finished message
8023 actually received from the peer, otherwise the protocol will be aborted.)
8025 As the Finished message are message digests of the complete handshake
8026 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
8027 be used for external authentication procedures when the authentication
8028 provided by SSL/TLS is not desired or is not enough.
8031 *) Enhanced support for Alpha Linux is added. Now ./config checks if
8032 the host supports BWX extension and if Compaq C is present on the
8033 $PATH. Just exploiting of the BWX extension results in 20-30%
8034 performance kick for some algorithms, e.g. DES and RC4 to mention
8035 a couple. Compaq C in turn generates ~20% faster code for MD5 and
8039 *) Add support for MS "fast SGC". This is arguably a violation of the
8040 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
8041 weak crypto and after checking the certificate is SGC a second one
8042 with strong crypto. MS SGC stops the first handshake after receiving
8043 the server certificate message and sends a second client hello. Since
8044 a server will typically do all the time consuming operations before
8045 expecting any further messages from the client (server key exchange
8046 is the most expensive) there is little difference between the two.
8048 To get OpenSSL to support MS SGC we have to permit a second client
8049 hello message after we have sent server done. In addition we have to
8050 reset the MAC if we do get this second client hello.
8053 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
8054 if a DER encoded private key is RSA or DSA traditional format. Changed
8055 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
8056 format DER encoded private key. Newer code should use PKCS#8 format which
8057 has the key type encoded in the ASN1 structure. Added DER private key
8058 support to pkcs8 application.
8061 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
8062 ciphersuites has been selected (as required by the SSL 3/TLS 1
8063 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
8064 is set, we interpret this as a request to violate the specification
8065 (the worst that can happen is a handshake failure, and 'correct'
8066 behaviour would result in a handshake failure anyway).
8069 *) In SSL_CTX_add_session, take into account that there might be multiple
8070 SSL_SESSION structures with the same session ID (e.g. when two threads
8071 concurrently obtain them from an external cache).
8072 The internal cache can handle only one SSL_SESSION with a given ID,
8073 so if there's a conflict, we now throw out the old one to achieve
8077 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
8078 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
8079 some routines that use cipher OIDs: some ciphers do not have OIDs
8080 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
8084 *) Simplify the trust setting structure and code. Now we just have
8085 two sequences of OIDs for trusted and rejected settings. These will
8086 typically have values the same as the extended key usage extension
8087 and any application specific purposes.
8089 The trust checking code now has a default behaviour: it will just
8090 check for an object with the same NID as the passed id. Functions can
8091 be provided to override either the default behaviour or the behaviour
8092 for a given id. SSL client, server and email already have functions
8093 in place for compatibility: they check the NID and also return "trusted"
8094 if the certificate is self signed.
8097 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
8098 traditional format into an EVP_PKEY structure.
8101 *) Add a password callback function PEM_cb() which either prompts for
8102 a password if usr_data is NULL or otherwise assumes it is a null
8103 terminated password. Allow passwords to be passed on command line
8104 environment or config files in a few more utilities.
8107 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
8108 keys. Add some short names for PKCS#8 PBE algorithms and allow them
8109 to be specified on the command line for the pkcs8 and pkcs12 utilities.
8110 Update documentation.
8113 *) Support for ASN1 "NULL" type. This could be handled before by using
8114 ASN1_TYPE but there wasn't any function that would try to read a NULL
8115 and produce an error if it couldn't. For compatibility we also have
8116 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
8117 don't allocate anything because they don't need to.
8120 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
8122 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
8124 *) Rebuild of the memory allocation routines used by OpenSSL code and
8125 possibly others as well. The purpose is to make an interface that
8126 provide hooks so anyone can build a separate set of allocation and
8127 deallocation routines to be used by OpenSSL, for example memory
8128 pool implementations, or something else, which was previously hard
8129 since Malloc(), Realloc() and Free() were defined as macros having
8130 the values malloc, realloc and free, respectively (except for Win32
8131 compilations). The same is provided for memory debugging code.
8132 OpenSSL already comes with functionality to find memory leaks, but
8133 this gives people a chance to debug other memory problems.
8135 With these changes, a new set of functions and macros have appeared:
8137 CRYPTO_set_mem_debug_functions() [F]
8138 CRYPTO_get_mem_debug_functions() [F]
8139 CRYPTO_dbg_set_options() [F]
8140 CRYPTO_dbg_get_options() [F]
8141 CRYPTO_malloc_debug_init() [M]
8143 The memory debug functions are NULL by default, unless the library
8144 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
8145 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
8146 gives the standard debugging functions that come with OpenSSL) or
8147 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
8148 provided by the library user) must be used. When the standard
8149 debugging functions are used, CRYPTO_dbg_set_options can be used to
8150 request additional information:
8151 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
8152 the CRYPTO_MDEBUG_xxx macro when compiling the library.
8154 Also, things like CRYPTO_set_mem_functions will always give the
8155 expected result (the new set of functions is used for allocation
8156 and deallocation) at all times, regardless of platform and compiler
8159 To finish it up, some functions that were never use in any other
8160 way than through macros have a new API and new semantic:
8163 CRYPTO_dbg_realloc()
8166 All macros of value have retained their old syntax.
8167 [Richard Levitte and Bodo Moeller]
8169 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
8170 ordering of SMIMECapabilities wasn't in "strength order" and there
8171 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
8175 *) Some ASN1 types with illegal zero length encoding (INTEGER,
8176 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
8177 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
8179 *) Merge in my S/MIME library for OpenSSL. This provides a simple
8180 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
8181 functionality to handle multipart/signed properly) and a utility
8182 called 'smime' to call all this stuff. This is based on code I
8183 originally wrote for Celo who have kindly allowed it to be
8184 included in OpenSSL.
8187 *) Add variants des_set_key_checked and des_set_key_unchecked of
8188 des_set_key (aka des_key_sched). Global variable des_check_key
8189 decides which of these is called by des_set_key; this way
8190 des_check_key behaves as it always did, but applications and
8191 the library itself, which was buggy for des_check_key == 1,
8192 have a cleaner way to pick the version they need.
8195 *) New function PKCS12_newpass() which changes the password of a
8199 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
8200 dynamic mix. In both cases the ids can be used as an index into the
8201 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
8202 functions so they accept a list of the field values and the
8203 application doesn't need to directly manipulate the X509_TRUST
8207 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
8211 *) Modify the way the V3 extension code looks up extensions. This now
8212 works in a similar way to the object code: we have some "standard"
8213 extensions in a static table which is searched with OBJ_bsearch()
8214 and the application can add dynamic ones if needed. The file
8215 crypto/x509v3/ext_dat.h now has the info: this file needs to be
8216 updated whenever a new extension is added to the core code and kept
8217 in ext_nid order. There is a simple program 'tabtest.c' which checks
8218 this. New extensions are not added too often so this file can readily
8219 be maintained manually.
8221 There are two big advantages in doing things this way. The extensions
8222 can be looked up immediately and no longer need to be "added" using
8223 X509V3_add_standard_extensions(): this function now does nothing.
8224 [Side note: I get *lots* of email saying the extension code doesn't
8225 work because people forget to call this function]
8226 Also no dynamic allocation is done unless new extensions are added:
8227 so if we don't add custom extensions there is no need to call
8228 X509V3_EXT_cleanup().
8231 *) Modify enc utility's salting as follows: make salting the default. Add a
8232 magic header, so unsalted files fail gracefully instead of just decrypting
8233 to garbage. This is because not salting is a big security hole, so people
8234 should be discouraged from doing it.
8237 *) Fixes and enhancements to the 'x509' utility. It allowed a message
8238 digest to be passed on the command line but it only used this
8239 parameter when signing a certificate. Modified so all relevant
8240 operations are affected by the digest parameter including the
8241 -fingerprint and -x509toreq options. Also -x509toreq choked if a
8242 DSA key was used because it didn't fix the digest.
8245 *) Initial certificate chain verify code. Currently tests the untrusted
8246 certificates for consistency with the verify purpose (which is set
8247 when the X509_STORE_CTX structure is set up) and checks the pathlength.
8249 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
8250 this is because it will reject chains with invalid extensions whereas
8251 every previous version of OpenSSL and SSLeay made no checks at all.
8253 Trust code: checks the root CA for the relevant trust settings. Trust
8254 settings have an initial value consistent with the verify purpose: e.g.
8255 if the verify purpose is for SSL client use it expects the CA to be
8256 trusted for SSL client use. However the default value can be changed to
8257 permit custom trust settings: one example of this would be to only trust
8258 certificates from a specific "secure" set of CAs.
8260 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
8261 which should be used for version portability: especially since the
8262 verify structure is likely to change more often now.
8264 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
8265 to set them. If not set then assume SSL clients will verify SSL servers
8268 Two new options to the verify program: -untrusted allows a set of
8269 untrusted certificates to be passed in and -purpose which sets the
8270 intended purpose of the certificate. If a purpose is set then the
8271 new chain verify code is used to check extension consistency.
8274 *) Support for the authority information access extension.
8277 *) Modify RSA and DSA PEM read routines to transparently handle
8278 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
8279 public keys in a format compatible with certificate
8280 SubjectPublicKeyInfo structures. Unfortunately there were already
8281 functions called *_PublicKey_* which used various odd formats so
8282 these are retained for compatibility: however the DSA variants were
8283 never in a public release so they have been deleted. Changed dsa/rsa
8284 utilities to handle the new format: note no releases ever handled public
8285 keys so we should be OK.
8287 The primary motivation for this change is to avoid the same fiasco
8288 that dogs private keys: there are several incompatible private key
8289 formats some of which are standard and some OpenSSL specific and
8290 require various evil hacks to allow partial transparent handling and
8291 even then it doesn't work with DER formats. Given the option anything
8292 other than PKCS#8 should be dumped: but the other formats have to
8293 stay in the name of compatibility.
8295 With public keys and the benefit of hindsight one standard format
8296 is used which works with EVP_PKEY, RSA or DSA structures: though
8297 it clearly returns an error if you try to read the wrong kind of key.
8299 Added a -pubkey option to the 'x509' utility to output the public key.
8300 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
8301 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
8302 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
8303 that do the same as the EVP_PKEY_assign_*() except they up the
8304 reference count of the added key (they don't "swallow" the
8308 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
8309 CRLs would fail if the file contained no certificates or no CRLs:
8310 added a new function to read in both types and return the number
8311 read: this means that if none are read it will be an error. The
8312 DER versions of the certificate and CRL reader would always fail
8313 because it isn't possible to mix certificates and CRLs in DER format
8314 without choking one or the other routine. Changed this to just read
8315 a certificate: this is the best we can do. Also modified the code
8316 in apps/verify.c to take notice of return codes: it was previously
8317 attempting to read in certificates from NULL pointers and ignoring
8318 any errors: this is one reason why the cert and CRL reader seemed
8319 to work. It doesn't check return codes from the default certificate
8320 routines: these may well fail if the certificates aren't installed.
8323 *) Code to support otherName option in GeneralName.
8326 *) First update to verify code. Change the verify utility
8327 so it warns if it is passed a self signed certificate:
8328 for consistency with the normal behaviour. X509_verify
8329 has been modified to it will now verify a self signed
8330 certificate if *exactly* the same certificate appears
8331 in the store: it was previously impossible to trust a
8332 single self signed certificate. This means that:
8333 openssl verify ss.pem
8334 now gives a warning about a self signed certificate but
8335 openssl verify -CAfile ss.pem ss.pem
8339 *) For servers, store verify_result in SSL_SESSION data structure
8340 (and add it to external session representation).
8341 This is needed when client certificate verifications fails,
8342 but an application-provided verification callback (set by
8343 SSL_CTX_set_cert_verify_callback) allows accepting the session
8344 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
8345 but returns 1): When the session is reused, we have to set
8346 ssl->verify_result to the appropriate error code to avoid
8348 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
8350 *) Fix a bug in the new PKCS#7 code: it didn't consider the
8351 case in PKCS7_dataInit() where the signed PKCS7 structure
8352 didn't contain any existing data because it was being created.
8353 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
8355 *) Add a salt to the key derivation routines in enc.c. This
8356 forms the first 8 bytes of the encrypted file. Also add a
8357 -S option to allow a salt to be input on the command line.
8360 *) New function X509_cmp(). Oddly enough there wasn't a function
8361 to compare two certificates. We do this by working out the SHA1
8362 hash and comparing that. X509_cmp() will be needed by the trust
8366 *) SSL_get1_session() is like SSL_get_session(), but increments
8367 the reference count in the SSL_SESSION returned.
8368 [Geoff Thorpe <geoff@eu.c2.net>]
8370 *) Fix for 'req': it was adding a null to request attributes.
8371 Also change the X509_LOOKUP and X509_INFO code to handle
8372 certificate auxiliary information.
8375 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
8379 *) Add the possibility to add extra information to the memory leak
8380 detecting output, to form tracebacks, showing from where each
8381 allocation was originated: CRYPTO_push_info("constant string") adds
8382 the string plus current file name and line number to a per-thread
8383 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
8384 is like calling CYRPTO_pop_info() until the stack is empty.
8385 Also updated memory leak detection code to be multi-thread-safe.
8388 *) Add options -text and -noout to pkcs7 utility and delete the
8389 encryption options which never did anything. Update docs.
8392 *) Add options to some of the utilities to allow the pass phrase
8393 to be included on either the command line (not recommended on
8394 OSes like Unix) or read from the environment. Update the
8395 manpages and fix a few bugs.
8398 *) Add a few manpages for some of the openssl commands.
8401 *) Fix the -revoke option in ca. It was freeing up memory twice,
8402 leaking and not finding already revoked certificates.
8405 *) Extensive changes to support certificate auxiliary information.
8406 This involves the use of X509_CERT_AUX structure and X509_AUX
8407 functions. An X509_AUX function such as PEM_read_X509_AUX()
8408 can still read in a certificate file in the usual way but it
8409 will also read in any additional "auxiliary information". By
8410 doing things this way a fair degree of compatibility can be
8411 retained: existing certificates can have this information added
8412 using the new 'x509' options.
8414 Current auxiliary information includes an "alias" and some trust
8415 settings. The trust settings will ultimately be used in enhanced
8416 certificate chain verification routines: currently a certificate
8417 can only be trusted if it is self signed and then it is trusted
8421 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
8422 The problem was that one of the replacement routines had not been working
8423 since SSLeay releases. For now the offending routine has been replaced
8424 with non-optimised assembler. Even so, this now gives around 95%
8425 performance improvement for 1024 bit RSA signs.
8428 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
8429 handling. Most clients have the effective key size in bits equal to
8430 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
8431 A few however don't do this and instead use the size of the decrypted key
8432 to determine the RC2 key length and the AlgorithmIdentifier to determine
8433 the effective key length. In this case the effective key length can still
8434 be 40 bits but the key length can be 168 bits for example. This is fixed
8435 by manually forcing an RC2 key into the EVP_PKEY structure because the
8436 EVP code can't currently handle unusual RC2 key sizes: it always assumes
8437 the key length and effective key length are equal.
8440 *) Add a bunch of functions that should simplify the creation of
8441 X509_NAME structures. Now you should be able to do:
8442 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
8443 and have it automatically work out the correct field type and fill in
8444 the structures. The more adventurous can try:
8445 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
8446 and it will (hopefully) work out the correct multibyte encoding.
8449 *) Change the 'req' utility to use the new field handling and multibyte
8450 copy routines. Before the DN field creation was handled in an ad hoc
8451 way in req, ca, and x509 which was rather broken and didn't support
8452 BMPStrings or UTF8Strings. Since some software doesn't implement
8453 BMPStrings or UTF8Strings yet, they can be enabled using the config file
8454 using the dirstring_type option. See the new comment in the default
8455 openssl.cnf for more info.
8458 *) Make crypto/rand/md_rand.c more robust:
8459 - Assure unique random numbers after fork().
8460 - Make sure that concurrent threads access the global counter and
8461 md serializably so that we never lose entropy in them
8462 or use exactly the same state in multiple threads.
8463 Access to the large state is not always serializable because
8464 the additional locking could be a performance killer, and
8465 md should be large enough anyway.
8468 *) New file apps/app_rand.c with commonly needed functionality
8469 for handling the random seed file.
8471 Use the random seed file in some applications that previously did not:
8473 dsaparam -genkey (which also ignored its '-rand' option),
8476 x509 (when signing).
8477 Except on systems with /dev/urandom, it is crucial to have a random
8478 seed file at least for key creation, DSA signing, and for DH exchanges;
8479 for RSA signatures we could do without one.
8481 gendh and gendsa (unlike genrsa) used to read only the first byte
8482 of each file listed in the '-rand' option. The function as previously
8483 found in genrsa is now in app_rand.c and is used by all programs
8484 that support '-rand'.
8487 *) In RAND_write_file, use mode 0600 for creating files;
8488 don't just chmod when it may be too late.
8491 *) Report an error from X509_STORE_load_locations
8492 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
8495 *) New function ASN1_mbstring_copy() this copies a string in either
8496 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
8497 into an ASN1_STRING type. A mask of permissible types is passed
8498 and it chooses the "minimal" type to use or an error if not type
8502 *) Add function equivalents to the various macros in asn1.h. The old
8503 macros are retained with an M_ prefix. Code inside the library can
8504 use the M_ macros. External code (including the openssl utility)
8505 should *NOT* in order to be "shared library friendly".
8508 *) Add various functions that can check a certificate's extensions
8509 to see if it usable for various purposes such as SSL client,
8510 server or S/MIME and CAs of these types. This is currently
8511 VERY EXPERIMENTAL but will ultimately be used for certificate chain
8512 verification. Also added a -purpose flag to x509 utility to
8513 print out all the purposes.
8516 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
8520 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
8521 for, obtain and decode and extension and obtain its critical flag.
8522 This allows all the necessary extension code to be handled in a
8523 single function call.
8526 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
8527 platforms. See crypto/rc4/rc4_enc.c for further details.
8530 *) New -noout option to asn1parse. This causes no output to be produced
8531 its main use is when combined with -strparse and -out to extract data
8532 from a file (which may not be in ASN.1 format).
8535 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
8536 when producing the local key id.
8537 [Richard Levitte <levitte@stacken.kth.se>]
8539 *) New option -dhparam in s_server. This allows a DH parameter file to be
8540 stated explicitly. If it is not stated then it tries the first server
8541 certificate file. The previous behaviour hard coded the filename
8545 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
8546 a public key to be input or output. For example:
8547 openssl rsa -in key.pem -pubout -out pubkey.pem
8548 Also added necessary DSA public key functions to handle this.
8551 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
8552 in the message. This was handled by allowing
8553 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8554 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8556 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8557 to the end of the strings whereas this didn't. This would cause problems
8558 if strings read with d2i_ASN1_bytes() were later modified.
8559 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8561 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8562 data and it contains EOF it will end up returning an error. This is
8563 caused by input 46 bytes long. The cause is due to the way base64
8564 BIOs find the start of base64 encoded data. They do this by trying a
8565 trial decode on each line until they find one that works. When they
8566 do a flag is set and it starts again knowing it can pass all the
8567 data directly through the decoder. Unfortunately it doesn't reset
8568 the context it uses. This means that if EOF is reached an attempt
8569 is made to pass two EOFs through the context and this causes the
8570 resulting error. This can also cause other problems as well. As is
8571 usual with these problems it takes *ages* to find and the fix is
8572 trivial: move one line.
8573 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8575 *) Ugly workaround to get s_client and s_server working under Windows. The
8576 old code wouldn't work because it needed to select() on sockets and the
8577 tty (for keypresses and to see if data could be written). Win32 only
8578 supports select() on sockets so we select() with a 1s timeout on the
8579 sockets and then see if any characters are waiting to be read, if none
8580 are present then we retry, we also assume we can always write data to
8581 the tty. This isn't nice because the code then blocks until we've
8582 received a complete line of data and it is effectively polling the
8583 keyboard at 1s intervals: however it's quite a bit better than not
8584 working at all :-) A dedicated Windows application might handle this
8585 with an event loop for example.
8588 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
8589 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
8590 will be called when RSA_sign() and RSA_verify() are used. This is useful
8591 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
8592 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
8593 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
8594 This necessitated the support of an extra signature type NID_md5_sha1
8595 for SSL signatures and modifications to the SSL library to use it instead
8596 of calling RSA_public_decrypt() and RSA_private_encrypt().
8599 *) Add new -verify -CAfile and -CApath options to the crl program, these
8600 will lookup a CRL issuers certificate and verify the signature in a
8601 similar way to the verify program. Tidy up the crl program so it
8602 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
8603 less strict. It will now permit CRL extensions even if it is not
8604 a V2 CRL: this will allow it to tolerate some broken CRLs.
8607 *) Initialize all non-automatic variables each time one of the openssl
8608 sub-programs is started (this is necessary as they may be started
8609 multiple times from the "OpenSSL>" prompt).
8610 [Lennart Bang, Bodo Moeller]
8612 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
8613 removing all other RSA functionality (this is what NO_RSA does). This
8614 is so (for example) those in the US can disable those operations covered
8615 by the RSA patent while allowing storage and parsing of RSA keys and RSA
8619 *) Non-copying interface to BIO pairs.
8620 (still largely untested)
8623 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
8624 ASCII string. This was handled independently in various places before.
8627 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
8628 UTF8 strings a character at a time.
8631 *) Use client_version from client hello to select the protocol
8632 (s23_srvr.c) and for RSA client key exchange verification
8633 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
8636 *) Add various utility functions to handle SPKACs, these were previously
8637 handled by poking round in the structure internals. Added new function
8638 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
8639 print, verify and generate SPKACs. Based on an original idea from
8640 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
8643 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
8646 *) Allow the config file extension section to be overwritten on the
8647 command line. Based on an original idea from Massimiliano Pala
8648 <madwolf@comune.modena.it>. The new option is called -extensions
8649 and can be applied to ca, req and x509. Also -reqexts to override
8650 the request extensions in req and -crlexts to override the crl extensions
8654 *) Add new feature to the SPKAC handling in ca. Now you can include
8655 the same field multiple times by preceding it by "XXXX." for example:
8658 this is the same syntax as used in the req config file.
8661 *) Allow certificate extensions to be added to certificate requests. These
8662 are specified in a 'req_extensions' option of the req section of the
8663 config file. They can be printed out with the -text option to req but
8664 are otherwise ignored at present.
8667 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
8668 data read consists of only the final block it would not decrypted because
8669 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
8670 A misplaced 'break' also meant the decrypted final block might not be
8671 copied until the next read.
8674 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
8675 a few extra parameters to the DH structure: these will be useful if
8676 for example we want the value of 'q' or implement X9.42 DH.
8679 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
8680 provides hooks that allow the default DSA functions or functions on a
8681 "per key" basis to be replaced. This allows hardware acceleration and
8682 hardware key storage to be handled without major modification to the
8683 library. Also added low level modexp hooks and CRYPTO_EX structure and
8684 associated functions.
8687 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
8688 as "read only": it can't be written to and the buffer it points to will
8689 not be freed. Reading from a read only BIO is much more efficient than
8690 a normal memory BIO. This was added because there are several times when
8691 an area of memory needs to be read from a BIO. The previous method was
8692 to create a memory BIO and write the data to it, this results in two
8693 copies of the data and an O(n^2) reading algorithm. There is a new
8694 function BIO_new_mem_buf() which creates a read only memory BIO from
8695 an area of memory. Also modified the PKCS#7 routines to use read only
8699 *) Bugfix: ssl23_get_client_hello did not work properly when called in
8700 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
8701 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
8702 but a retry condition occured while trying to read the rest.
8705 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
8706 NID_pkcs7_encrypted by default: this was wrong since this should almost
8707 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
8708 the encrypted data type: this is a more sensible place to put it and it
8709 allows the PKCS#12 code to be tidied up that duplicated this
8713 *) Changed obj_dat.pl script so it takes its input and output files on
8714 the command line. This should avoid shell escape redirection problems
8718 *) Initial support for certificate extension requests, these are included
8719 in things like Xenroll certificate requests. Included functions to allow
8720 extensions to be obtained and added.
8723 *) -crlf option to s_client and s_server for sending newlines as
8724 CRLF (as required by many protocols).
8727 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
8729 *) Install libRSAglue.a when OpenSSL is built with RSAref.
8730 [Ralf S. Engelschall]
8732 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
8733 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
8735 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
8739 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
8740 DH parameters/keys (q is lost during that conversion, but the resulting
8741 DH parameters contain its length).
8743 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
8744 much faster than DH_generate_parameters (which creates parameters
8745 where p = 2*q + 1), and also the smaller q makes DH computations
8746 much more efficient (160-bit exponentiation instead of 1024-bit
8747 exponentiation); so this provides a convenient way to support DHE
8748 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
8749 utter importance to use
8750 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8752 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8753 when such DH parameters are used, because otherwise small subgroup
8754 attacks may become possible!
8757 *) Avoid memory leak in i2d_DHparams.
8760 *) Allow the -k option to be used more than once in the enc program:
8761 this allows the same encrypted message to be read by multiple recipients.
8764 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
8765 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
8766 it will always use the numerical form of the OID, even if it has a short
8770 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
8771 method only got called if p,q,dmp1,dmq1,iqmp components were present,
8772 otherwise bn_mod_exp was called. In the case of hardware keys for example
8773 no private key components need be present and it might store extra data
8774 in the RSA structure, which cannot be accessed from bn_mod_exp.
8775 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
8776 private key operations.
8779 *) Added support for SPARC Linux.
8782 *) pem_password_cb function type incompatibly changed from
8783 typedef int pem_password_cb(char *buf, int size, int rwflag);
8785 ....(char *buf, int size, int rwflag, void *userdata);
8786 so that applications can pass data to their callbacks:
8787 The PEM[_ASN1]_{read,write}... functions and macros now take an
8788 additional void * argument, which is just handed through whenever
8789 the password callback is called.
8790 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
8792 New function SSL_CTX_set_default_passwd_cb_userdata.
8794 Compatibility note: As many C implementations push function arguments
8795 onto the stack in reverse order, the new library version is likely to
8796 interoperate with programs that have been compiled with the old
8797 pem_password_cb definition (PEM_whatever takes some data that
8798 happens to be on the stack as its last argument, and the callback
8799 just ignores this garbage); but there is no guarantee whatsoever that
8802 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
8803 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
8804 problems not only on Windows, but also on some Unix platforms.
8805 To avoid problematic command lines, these definitions are now in an
8806 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
8807 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
8810 *) MIPS III/IV assembler module is reimplemented.
8813 *) More DES library cleanups: remove references to srand/rand and
8814 delete an unused file.
8817 *) Add support for the the free Netwide assembler (NASM) under Win32,
8818 since not many people have MASM (ml) and it can be hard to obtain.
8819 This is currently experimental but it seems to work OK and pass all
8820 the tests. Check out INSTALL.W32 for info.
8823 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
8824 without temporary keys kept an extra copy of the server key,
8825 and connections with temporary keys did not free everything in case
8829 *) New function RSA_check_key and new openssl rsa option -check
8830 for verifying the consistency of RSA keys.
8831 [Ulf Moeller, Bodo Moeller]
8833 *) Various changes to make Win32 compile work:
8834 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
8835 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
8836 comparison" warnings.
8837 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
8840 *) Add a debugging option to PKCS#5 v2 key generation function: when
8841 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
8842 derived keys are printed to stderr.
8845 *) Copy the flags in ASN1_STRING_dup().
8846 [Roman E. Pavlov <pre@mo.msk.ru>]
8848 *) The x509 application mishandled signing requests containing DSA
8849 keys when the signing key was also DSA and the parameters didn't match.
8851 It was supposed to omit the parameters when they matched the signing key:
8852 the verifying software was then supposed to automatically use the CA's
8853 parameters if they were absent from the end user certificate.
8855 Omitting parameters is no longer recommended. The test was also
8856 the wrong way round! This was probably due to unusual behaviour in
8857 EVP_cmp_parameters() which returns 1 if the parameters match.
8858 This meant that parameters were omitted when they *didn't* match and
8859 the certificate was useless. Certificates signed with 'ca' didn't have
8861 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
8863 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
8864 The interface is as follows:
8865 Applications can use
8866 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
8867 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
8868 "off" is now the default.
8869 The library internally uses
8870 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
8871 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
8872 to disable memory-checking temporarily.
8874 Some inconsistent states that previously were possible (and were
8875 even the default) are now avoided.
8877 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
8878 with each memory chunk allocated; this is occasionally more helpful
8879 than just having a counter.
8881 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
8883 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
8887 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
8888 which largely parallels "options", but is for changing API behaviour,
8889 whereas "options" are about protocol behaviour.
8890 Initial "mode" flags are:
8892 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
8893 a single record has been written.
8894 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
8895 retries use the same buffer location.
8896 (But all of the contents must be
8900 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
8903 *) Fix problems with no-hmac etc.
8904 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
8906 *) New functions RSA_get_default_method(), RSA_set_method() and
8907 RSA_get_method(). These allows replacement of RSA_METHODs without having
8908 to mess around with the internals of an RSA structure.
8911 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
8912 Also really enable memory leak checks in openssl.c and in some
8914 [Chad C. Mulligan, Bodo Moeller]
8916 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
8917 up the length of negative integers. This has now been simplified to just
8918 store the length when it is first determined and use it later, rather
8919 than trying to keep track of where data is copied and updating it to
8921 [Steve Henson, reported by Brien Wheeler
8922 <bwheeler@authentica-security.com>]
8924 *) Add a new function PKCS7_signatureVerify. This allows the verification
8925 of a PKCS#7 signature but with the signing certificate passed to the
8926 function itself. This contrasts with PKCS7_dataVerify which assumes the
8927 certificate is present in the PKCS#7 structure. This isn't always the
8928 case: certificates can be omitted from a PKCS#7 structure and be
8929 distributed by "out of band" means (such as a certificate database).
8932 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
8933 function prototypes in pem.h, also change util/mkdef.pl to add the
8934 necessary function names.
8937 *) mk1mf.pl (used by Windows builds) did not properly read the
8938 options set by Configure in the top level Makefile, and Configure
8939 was not even able to write more than one option correctly.
8940 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
8943 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
8944 file to be loaded from a BIO or FILE pointer. The BIO version will
8945 for example allow memory BIOs to contain config info.
8948 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
8949 Whoever hopes to achieve shared-library compatibility across versions
8950 must use this, not the compile-time macro.
8951 (Exercise 0.9.4: Which is the minimum library version required by
8953 Note: All this applies only to multi-threaded programs, others don't
8957 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
8958 through a BIO pair triggered the default case, i.e.
8959 SSLerr(...,SSL_R_UNKNOWN_STATE).
8962 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
8963 can use the SSL library even if none of the specific BIOs is
8967 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
8968 for the encoded length.
8969 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
8971 *) Add initial documentation of the X509V3 functions.
8974 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
8975 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
8976 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
8977 secure PKCS#8 private key format with a high iteration count.
8980 *) Fix determination of Perl interpreter: A perl or perl5
8981 _directory_ in $PATH was also accepted as the interpreter.
8982 [Ralf S. Engelschall]
8984 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
8985 wrong with it but it was very old and did things like calling
8986 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
8990 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
8991 to use the new extension code.
8994 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
8995 with macros. This should make it easier to change their form, add extra
8996 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
9000 *) Add to configuration table a new entry that can specify an alternative
9001 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
9002 according to Mark Crispin <MRC@Panda.COM>.
9006 *) DES CBC did not update the IV. Weird.
9009 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
9010 Changing the behaviour of the former might break existing programs --
9011 where IV updating is needed, des_ncbc_encrypt can be used.
9014 *) When bntest is run from "make test" it drives bc to check its
9015 calculations, as well as internally checking them. If an internal check
9016 fails, it needs to cause bc to give a non-zero result or make test carries
9017 on without noticing the failure. Fixed.
9020 *) DES library cleanups.
9023 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
9024 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
9025 ciphers. NOTE: although the key derivation function has been verified
9026 against some published test vectors it has not been extensively tested
9027 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
9031 *) Instead of "mkdir -p", which is not fully portable, use new
9032 Perl script "util/mkdir-p.pl".
9035 *) Rewrite the way password based encryption (PBE) is handled. It used to
9036 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
9037 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
9038 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
9039 the 'parameter' field of the AlgorithmIdentifier is passed to the
9040 underlying key generation function so it must do its own ASN1 parsing.
9041 This has also changed the EVP_PBE_CipherInit() function which now has a
9042 'parameter' argument instead of literal salt and iteration count values
9043 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
9046 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
9047 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
9048 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
9049 KEY" because this clashed with PKCS#8 unencrypted string. Since this
9050 value was just used as a "magic string" and not used directly its
9051 value doesn't matter.
9054 *) Introduce some semblance of const correctness to BN. Shame C doesn't
9058 *) "linux-sparc64" configuration (ultrapenguin).
9059 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
9060 "linux-sparc" configuration.
9061 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
9063 *) config now generates no-xxx options for missing ciphers.
9066 *) Support the EBCDIC character set (work in progress).
9067 File ebcdic.c not yet included because it has a different license.
9068 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9070 *) Support BS2000/OSD-POSIX.
9071 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9073 *) Make callbacks for key generation use void * instead of char *.
9076 *) Make S/MIME samples compile (not yet tested).
9079 *) Additional typesafe stacks.
9082 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
9086 Changes between 0.9.3 and 0.9.3a [29 May 1999]
9088 *) New configuration variant "sco5-gcc".
9090 *) Updated some demos.
9091 [Sean O Riordain, Wade Scholine]
9093 *) Add missing BIO_free at exit of pkcs12 application.
9096 *) Fix memory leak in conf.c.
9099 *) Updates for Win32 to assembler version of MD5.
9102 *) Set #! path to perl in apps/der_chop to where we found it
9103 instead of using a fixed path.
9106 *) SHA library changes for irix64-mips4-cc.
9109 *) Improvements for VMS support.
9113 Changes between 0.9.2b and 0.9.3 [24 May 1999]
9115 *) Bignum library bug fix. IRIX 6 passes "make test" now!
9116 This also avoids the problems with SC4.2 and unpatched SC5.
9117 [Andy Polyakov <appro@fy.chalmers.se>]
9119 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
9120 These are required because of the typesafe stack would otherwise break
9121 existing code. If old code used a structure member which used to be STACK
9122 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
9123 sk_num or sk_value it would produce an error because the num, data members
9124 are not present in STACK_OF. Now it just produces a warning. sk_set
9125 replaces the old method of assigning a value to sk_value
9126 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
9127 that does this will no longer work (and should use sk_set instead) but
9128 this could be regarded as a "questionable" behaviour anyway.
9131 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
9132 correctly handle encrypted S/MIME data.
9135 *) Change type of various DES function arguments from des_cblock
9136 (which means, in function argument declarations, pointer to char)
9137 to des_cblock * (meaning pointer to array with 8 char elements),
9138 which allows the compiler to do more typechecking; it was like
9139 that back in SSLeay, but with lots of ugly casts.
9141 Introduce new type const_des_cblock.
9144 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
9145 problems: find RecipientInfo structure that matches recipient certificate
9146 and initialise the ASN1 structures properly based on passed cipher.
9149 *) Belatedly make the BN tests actually check the results.
9152 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
9153 to and from BNs: it was completely broken. New compilation option
9154 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
9155 key elements as negative integers.
9158 *) Reorganize and speed up MD5.
9159 [Andy Polyakov <appro@fy.chalmers.se>]
9162 [Richard Levitte <richard@levitte.org>]
9164 *) New option -out to asn1parse to allow the parsed structure to be
9165 output to a file. This is most useful when combined with the -strparse
9166 option to examine the output of things like OCTET STRINGS.
9169 *) Make SSL library a little more fool-proof by not requiring any longer
9170 that SSL_set_{accept,connect}_state be called before
9171 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
9172 in many applications because usually everything *appeared* to work as
9173 intended anyway -- now it really works as intended).
9176 *) Move openssl.cnf out of lib/.
9179 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
9180 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
9181 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
9182 [Ralf S. Engelschall]
9184 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
9185 handle PKCS#7 enveloped data properly.
9186 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
9188 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
9189 copying pointers. The cert_st handling is changed by this in
9190 various ways (and thus what used to be known as ctx->default_cert
9191 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
9192 any longer when s->cert does not give us what we need).
9193 ssl_cert_instantiate becomes obsolete by this change.
9194 As soon as we've got the new code right (possibly it already is?),
9195 we have solved a couple of bugs of the earlier code where s->cert
9196 was used as if it could not have been shared with other SSL structures.
9198 Note that using the SSL API in certain dirty ways now will result
9199 in different behaviour than observed with earlier library versions:
9200 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
9201 does not influence s as it used to.
9203 In order to clean up things more thoroughly, inside SSL_SESSION
9204 we don't use CERT any longer, but a new structure SESS_CERT
9205 that holds per-session data (if available); currently, this is
9206 the peer's certificate chain and, for clients, the server's certificate
9207 and temporary key. CERT holds only those values that can have
9208 meaningful defaults in an SSL_CTX.
9211 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
9212 from the internal representation. Various PKCS#7 fixes: remove some
9213 evil casts and set the enc_dig_alg field properly based on the signing
9217 *) Allow PKCS#12 password to be set from the command line or the
9218 environment. Let 'ca' get its config file name from the environment
9219 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
9223 *) Allow certificate policies extension to use an IA5STRING for the
9224 organization field. This is contrary to the PKIX definition but
9225 VeriSign uses it and IE5 only recognises this form. Document 'x509'
9229 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
9230 without disallowing inline assembler and the like for non-pedantic builds.
9233 *) Support Borland C++ builder.
9234 [Janez Jere <jj@void.si>, modified by Ulf Möller]
9239 *) SHA-1 cleanups and performance enhancements.
9240 [Andy Polyakov <appro@fy.chalmers.se>]
9242 *) Sparc v8plus assembler for the bignum library.
9243 [Andy Polyakov <appro@fy.chalmers.se>]
9245 *) Accept any -xxx and +xxx compiler options in Configure.
9248 *) Update HPUX configuration.
9251 *) Add missing sk_<type>_unshift() function to safestack.h
9252 [Ralf S. Engelschall]
9254 *) New function SSL_CTX_use_certificate_chain_file that sets the
9255 "extra_cert"s in addition to the certificate. (This makes sense
9256 only for "PEM" format files, as chains as a whole are not
9260 *) Support verify_depth from the SSL API.
9261 x509_vfy.c had what can be considered an off-by-one-error:
9262 Its depth (which was not part of the external interface)
9263 was actually counting the number of certificates in a chain;
9264 now it really counts the depth.
9267 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
9268 instead of X509err, which often resulted in confusing error
9269 messages since the error codes are not globally unique
9270 (e.g. an alleged error in ssl3_accept when a certificate
9271 didn't match the private key).
9273 *) New function SSL_CTX_set_session_id_context that allows to set a default
9274 value (so that you don't need SSL_set_session_id_context for each
9275 connection using the SSL_CTX).
9278 *) OAEP decoding bug fix.
9281 *) Support INSTALL_PREFIX for package builders, as proposed by
9285 *) New Configure options "threads" and "no-threads". For systems
9286 where the proper compiler options are known (currently Solaris
9287 and Linux), "threads" is the default.
9290 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
9293 *) Install various scripts to $(OPENSSLDIR)/misc, not to
9294 $(INSTALLTOP)/bin -- they shouldn't clutter directories
9295 such as /usr/local/bin.
9298 *) "make linux-shared" to build shared libraries.
9299 [Niels Poppe <niels@netbox.org>]
9301 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
9304 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
9305 extension adding in x509 utility.
9308 *) Remove NOPROTO sections and error code comments.
9311 *) Partial rewrite of the DEF file generator to now parse the ANSI
9315 *) New Configure options --prefix=DIR and --openssldir=DIR.
9318 *) Complete rewrite of the error code script(s). It is all now handled
9319 by one script at the top level which handles error code gathering,
9320 header rewriting and C source file generation. It should be much better
9321 than the old method: it now uses a modified version of Ulf's parser to
9322 read the ANSI prototypes in all header files (thus the old K&R definitions
9323 aren't needed for error creation any more) and do a better job of
9324 translating function codes into names. The old 'ASN1 error code imbedded
9325 in a comment' is no longer necessary and it doesn't use .err files which
9326 have now been deleted. Also the error code call doesn't have to appear all
9327 on one line (which resulted in some large lines...).
9330 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
9333 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
9334 0 (which usually indicates a closed connection), but continue reading.
9337 *) Fix some race conditions.
9340 *) Add support for CRL distribution points extension. Add Certificate
9341 Policies and CRL distribution points documentation.
9344 *) Move the autogenerated header file parts to crypto/opensslconf.h.
9347 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
9348 8 of keying material. Merlin has also confirmed interop with this fix
9349 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
9350 [Merlin Hughes <merlin@baltimore.ie>]
9352 *) Fix lots of warnings.
9353 [Richard Levitte <levitte@stacken.kth.se>]
9355 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
9356 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
9357 [Richard Levitte <levitte@stacken.kth.se>]
9359 *) Fix problems with sizeof(long) == 8.
9360 [Andy Polyakov <appro@fy.chalmers.se>]
9362 *) Change functions to ANSI C.
9365 *) Fix typos in error codes.
9366 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
9368 *) Remove defunct assembler files from Configure.
9371 *) SPARC v8 assembler BIGNUM implementation.
9372 [Andy Polyakov <appro@fy.chalmers.se>]
9374 *) Support for Certificate Policies extension: both print and set.
9375 Various additions to support the r2i method this uses.
9378 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
9379 return a const string when you are expecting an allocated buffer.
9382 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
9383 types DirectoryString and DisplayText.
9386 *) Add code to allow r2i extensions to access the configuration database,
9387 add an LHASH database driver and add several ctx helper functions.
9390 *) Fix an evil bug in bn_expand2() which caused various BN functions to
9391 fail when they extended the size of a BIGNUM.
9394 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
9395 support typesafe stack.
9398 *) Fix typo in SSL_[gs]et_options().
9399 [Nils Frostberg <nils@medcom.se>]
9401 *) Delete various functions and files that belonged to the (now obsolete)
9402 old X509V3 handling code.
9405 *) New Configure option "rsaref".
9408 *) Don't auto-generate pem.h.
9411 *) Introduce type-safe ASN.1 SETs.
9414 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
9415 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
9417 *) Introduce type-safe STACKs. This will almost certainly break lots of code
9418 that links with OpenSSL (well at least cause lots of warnings), but fear
9419 not: the conversion is trivial, and it eliminates loads of evil casts. A
9420 few STACKed things have been converted already. Feel free to convert more.
9421 In the fullness of time, I'll do away with the STACK type altogether.
9424 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
9425 specified in <certfile> by updating the entry in the index.txt file.
9426 This way one no longer has to edit the index.txt file manually for
9427 revoking a certificate. The -revoke option does the gory details now.
9428 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
9430 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
9431 `-text' option at all and this way the `-noout -text' combination was
9432 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
9433 [Ralf S. Engelschall]
9435 *) Make sure a corresponding plain text error message exists for the
9436 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
9437 verify callback function determined that a certificate was revoked.
9438 [Ralf S. Engelschall]
9440 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
9441 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
9442 all available cipers including rc5, which was forgotten until now.
9443 In order to let the testing shell script know which algorithms
9444 are available, a new (up to now undocumented) command
9445 "openssl list-cipher-commands" is used.
9448 *) Bugfix: s_client occasionally would sleep in select() when
9449 it should have checked SSL_pending() first.
9452 *) New functions DSA_do_sign and DSA_do_verify to provide access to
9453 the raw DSA values prior to ASN.1 encoding.
9456 *) Tweaks to Configure
9457 [Niels Poppe <niels@netbox.org>]
9459 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
9463 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
9466 *) New config option to avoid instructions that are illegal on the 80386.
9467 The default code is faster, but requires at least a 486.
9470 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
9471 SSL2_SERVER_VERSION (not used at all) macros, which are now the
9472 same as SSL2_VERSION anyway.
9475 *) New "-showcerts" option for s_client.
9478 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
9479 application. Various cleanups and fixes.
9482 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
9483 modify error routines to work internally. Add error codes and PBE init
9484 to library startup routines.
9487 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
9488 packing functions to asn1 and evp. Changed function names and error
9489 codes along the way.
9492 *) PKCS12 integration: and so it begins... First of several patches to
9493 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
9494 objects to objects.h
9497 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
9498 and display support for Thawte strong extranet extension.
9501 *) Add LinuxPPC support.
9502 [Jeff Dubrule <igor@pobox.org>]
9504 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
9505 bn_div_words in alpha.s.
9506 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
9508 *) Make sure the RSA OAEP test is skipped under -DRSAref because
9509 OAEP isn't supported when OpenSSL is built with RSAref.
9510 [Ulf Moeller <ulf@fitug.de>]
9512 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
9513 so they no longer are missing under -DNOPROTO.
9514 [Soren S. Jorvang <soren@t.dk>]
9517 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
9519 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
9520 doesn't work when the session is reused. Coming soon!
9523 *) Fix a security hole, that allows sessions to be reused in the wrong
9524 context thus bypassing client cert protection! All software that uses
9525 client certs and session caches in multiple contexts NEEDS PATCHING to
9526 allow session reuse! A fuller solution is in the works.
9527 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
9529 *) Some more source tree cleanups (removed obsolete files
9530 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
9531 permission on "config" script to be executable) and a fix for the INSTALL
9533 [Ulf Moeller <ulf@fitug.de>]
9535 *) Remove some legacy and erroneous uses of malloc, free instead of
9537 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
9539 *) Make rsa_oaep_test return non-zero on error.
9540 [Ulf Moeller <ulf@fitug.de>]
9542 *) Add support for native Solaris shared libraries. Configure
9543 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
9544 if someone would make that last step automatic.
9545 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
9547 *) ctx_size was not built with the right compiler during "make links". Fixed.
9550 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
9551 except NULL ciphers". This means the default cipher list will no longer
9552 enable NULL ciphers. They need to be specifically enabled e.g. with
9553 the string "DEFAULT:eNULL".
9556 *) Fix to RSA private encryption routines: if p < q then it would
9557 occasionally produce an invalid result. This will only happen with
9558 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9561 *) Be less restrictive and allow also `perl util/perlpath.pl
9562 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9563 because this way one can also use an interpreter named `perl5' (which is
9564 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9565 installed as `perl').
9566 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9568 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9569 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9571 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9572 advapi32.lib to Win32 build and change the pem test comparision
9573 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9574 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9575 and crypto/des/ede_cbcm_enc.c.
9578 *) DES quad checksum was broken on big-endian architectures. Fixed.
9581 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9582 Win32 test batch file so it (might) work again. The Win32 test batch file
9583 is horrible: I feel ill....
9586 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
9587 in e_os.h. Audit of header files to check ANSI and non ANSI
9588 sections: 10 functions were absent from non ANSI section and not exported
9589 from Windows DLLs. Fixed up libeay.num for new functions.
9592 *) Make `openssl version' output lines consistent.
9593 [Ralf S. Engelschall]
9595 *) Fix Win32 symbol export lists for BIO functions: Added
9596 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
9597 to ms/libeay{16,32}.def.
9598 [Ralf S. Engelschall]
9600 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
9601 fine under Unix and passes some trivial tests I've now added. But the
9602 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
9603 added to make sure no one expects that this stuff really works in the
9604 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
9605 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
9607 [Ralf S. Engelschall]
9609 *) Fix the generation of two part addresses in perl.
9610 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
9612 *) Add config entry for Linux on MIPS.
9613 [John Tobey <jtobey@channel1.com>]
9615 *) Make links whenever Configure is run, unless we are on Windoze.
9618 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
9619 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
9623 *) Add a useful kludge to allow package maintainers to specify compiler and
9624 other platforms details on the command line without having to patch the
9625 Configure script everytime: One now can use ``perl Configure
9626 <id>:<details>'', i.e. platform ids are allowed to have details appended
9627 to them (seperated by colons). This is treated as there would be a static
9628 pre-configured entry in Configure's %table under key <id> with value
9629 <details> and ``perl Configure <id>'' is called. So, when you want to
9630 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
9631 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
9632 now, which overrides the FreeBSD-elf entry on-the-fly.
9633 [Ralf S. Engelschall]
9635 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
9638 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
9639 on the `perl Configure ...' command line. This way one can compile
9640 OpenSSL libraries with Position Independent Code (PIC) which is needed
9641 for linking it into DSOs.
9642 [Ralf S. Engelschall]
9644 *) Remarkably, export ciphers were totally broken and no-one had noticed!
9648 *) Cleaned up the LICENSE document: The official contact for any license
9649 questions now is the OpenSSL core team under openssl-core@openssl.org.
9650 And add a paragraph about the dual-license situation to make sure people
9651 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
9652 to the OpenSSL toolkit.
9653 [Ralf S. Engelschall]
9655 *) General source tree makefile cleanups: Made `making xxx in yyy...'
9656 display consistent in the source tree and replaced `/bin/rm' by `rm'.
9657 Additonally cleaned up the `make links' target: Remove unnecessary
9658 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
9659 to speed processing and no longer clutter the display with confusing
9660 stuff. Instead only the actually done links are displayed.
9661 [Ralf S. Engelschall]
9663 *) Permit null encryption ciphersuites, used for authentication only. It used
9664 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
9665 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
9669 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
9670 signed attributes when verifying signatures (this would break them),
9671 the detached data encoding was wrong and public keys obtained using
9672 X509_get_pubkey() weren't freed.
9675 *) Add text documentation for the BUFFER functions. Also added a work around
9676 to a Win95 console bug. This was triggered by the password read stuff: the
9677 last character typed gets carried over to the next fread(). If you were
9678 generating a new cert request using 'req' for example then the last
9679 character of the passphrase would be CR which would then enter the first
9683 *) Added the new `Includes OpenSSL Cryptography Software' button as
9684 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
9685 button and can be used by applications based on OpenSSL to show the
9686 relationship to the OpenSSL project.
9687 [Ralf S. Engelschall]
9689 *) Remove confusing variables in function signatures in files
9690 ssl/ssl_lib.c and ssl/ssl.h.
9691 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9693 *) Don't install bss_file.c under PREFIX/include/
9694 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9696 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
9697 functions that return function pointers and has support for NT specific
9698 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
9699 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
9700 unsigned to signed types: this was killing the Win32 compile.
9703 *) Add new certificate file to stack functions,
9704 SSL_add_dir_cert_subjects_to_stack() and
9705 SSL_add_file_cert_subjects_to_stack(). These largely supplant
9706 SSL_load_client_CA_file(), and can be used to add multiple certs easily
9707 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
9708 This means that Apache-SSL and similar packages don't have to mess around
9709 to add as many CAs as they want to the preferred list.
9712 *) Experiment with doxygen documentation. Currently only partially applied to
9714 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
9715 openssl.doxy as the configuration file.
9718 *) Get rid of remaining C++-style comments which strict C compilers hate.
9719 [Ralf S. Engelschall, pointed out by Carlos Amengual]
9721 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
9722 compiled in by default: it has problems with large keys.
9725 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
9726 DH private keys and/or callback functions which directly correspond to
9727 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
9728 is needed for applications which have to configure certificates on a
9729 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
9731 For the RSA certificate situation is makes no difference, but
9732 for the DSA certificate situation this fixes the "no shared cipher"
9733 problem where the OpenSSL cipher selection procedure failed because the
9734 temporary keys were not overtaken from the context and the API provided
9735 no way to reconfigure them.
9736 The new functions now let applications reconfigure the stuff and they
9737 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
9738 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
9739 non-public-API function ssl_cert_instantiate() is used as a helper
9740 function and also to reduce code redundancy inside ssl_rsa.c.
9741 [Ralf S. Engelschall]
9743 *) Move s_server -dcert and -dkey options out of the undocumented feature
9744 area because they are useful for the DSA situation and should be
9745 recognized by the users.
9746 [Ralf S. Engelschall]
9748 *) Fix the cipher decision scheme for export ciphers: the export bits are
9749 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
9750 SSL_EXP_MASK. So, the original variable has to be used instead of the
9751 already masked variable.
9752 [Richard Levitte <levitte@stacken.kth.se>]
9754 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
9755 [Richard Levitte <levitte@stacken.kth.se>]
9757 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
9758 from `int' to `unsigned int' because it's a length and initialized by
9759 EVP_DigestFinal() which expects an `unsigned int *'.
9760 [Richard Levitte <levitte@stacken.kth.se>]
9762 *) Don't hard-code path to Perl interpreter on shebang line of Configure
9763 script. Instead use the usual Shell->Perl transition trick.
9764 [Ralf S. Engelschall]
9766 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
9767 (in addition to RSA certificates) to match the behaviour of `openssl dsa
9768 -noout -modulus' as it's already the case for `openssl rsa -noout
9769 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
9770 currently the public key is printed (a decision which was already done by
9771 `openssl dsa -modulus' in the past) which serves a similar purpose.
9772 Additionally the NO_RSA no longer completely removes the whole -modulus
9773 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
9775 [Ralf S. Engelschall]
9777 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
9778 BIO. See the source (crypto/evp/bio_ok.c) for more info.
9779 [Arne Ansper <arne@ats.cyber.ee>]
9781 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
9782 to be added. Now both 'req' and 'ca' can use new objects defined in the
9786 *) Add cool BIO that does syslog (or event log on NT).
9787 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
9789 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
9790 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
9791 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
9792 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
9795 *) Add preliminary config info for new extension code.
9798 *) Make RSA_NO_PADDING really use no padding.
9799 [Ulf Moeller <ulf@fitug.de>]
9801 *) Generate errors when private/public key check is done.
9804 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
9805 for some CRL extensions and new objects added.
9808 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
9809 key usage extension and fuller support for authority key id.
9812 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
9813 padding method for RSA, which is recommended for new applications in PKCS
9814 #1 v2.0 (RFC 2437, October 1998).
9815 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
9816 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
9817 against Bleichbacher's attack on RSA.
9818 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
9821 *) Updates to the new SSL compression code
9822 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9824 *) Fix so that the version number in the master secret, when passed
9825 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
9826 (because the server will not accept higher), that the version number
9827 is 0x03,0x01, not 0x03,0x00
9828 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9830 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
9831 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
9832 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
9835 *) Support for RAW extensions where an arbitrary extension can be
9836 created by including its DER encoding. See apps/openssl.cnf for
9840 *) Make sure latest Perl versions don't interpret some generated C array
9841 code as Perl array code in the crypto/err/err_genc.pl script.
9842 [Lars Weber <3weber@informatik.uni-hamburg.de>]
9844 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
9845 not many people have the assembler. Various Win32 compilation fixes and
9846 update to the INSTALL.W32 file with (hopefully) more accurate Win32
9850 *) Modify configure script 'Configure' to automatically create crypto/date.h
9851 file under Win32 and also build pem.h from pem.org. New script
9852 util/mkfiles.pl to create the MINFO file on environments that can't do a
9853 'make files': perl util/mkfiles.pl >MINFO should work.
9856 *) Major rework of DES function declarations, in the pursuit of correctness
9857 and purity. As a result, many evil casts evaporated, and some weirdness,
9858 too. You may find this causes warnings in your code. Zapping your evil
9859 casts will probably fix them. Mostly.
9862 *) Fix for a typo in asn1.h. Bug fix to object creation script
9863 obj_dat.pl. It considered a zero in an object definition to mean
9864 "end of object": none of the objects in objects.h have any zeros
9865 so it wasn't spotted.
9866 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
9868 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
9869 Masking (CBCM). In the absence of test vectors, the best I have been able
9870 to do is check that the decrypt undoes the encrypt, so far. Send me test
9871 vectors if you have them.
9874 *) Correct calculation of key length for export ciphers (too much space was
9875 allocated for null ciphers). This has not been tested!
9878 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
9879 message is now correct (it understands "crypto" and "ssl" on its
9880 command line). There is also now an "update" option. This will update
9881 the util/ssleay.num and util/libeay.num files with any new functions.
9883 perl util/mkdef.pl crypto ssl update
9884 it will update them.
9887 *) Overhauled the Perl interface (perl/*):
9888 - ported BN stuff to OpenSSL's different BN library
9889 - made the perl/ source tree CVS-aware
9890 - renamed the package from SSLeay to OpenSSL (the files still contain
9891 their history because I've copied them in the repository)
9892 - removed obsolete files (the test scripts will be replaced
9893 by better Test::Harness variants in the future)
9894 [Ralf S. Engelschall]
9896 *) First cut for a very conservative source tree cleanup:
9897 1. merge various obsolete readme texts into doc/ssleay.txt
9898 where we collect the old documents and readme texts.
9899 2. remove the first part of files where I'm already sure that we no
9900 longer need them because of three reasons: either they are just temporary
9901 files which were left by Eric or they are preserved original files where
9902 I've verified that the diff is also available in the CVS via "cvs diff
9903 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
9904 the crypto/md/ stuff).
9905 [Ralf S. Engelschall]
9907 *) More extension code. Incomplete support for subject and issuer alt
9908 name, issuer and authority key id. Change the i2v function parameters
9909 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
9910 what that's for :-) Fix to ASN1 macro which messed up
9911 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
9914 *) Preliminary support for ENUMERATED type. This is largely copied from the
9918 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
9919 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9921 *) Make sure `make rehash' target really finds the `openssl' program.
9922 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9924 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
9925 like to hear about it if this slows down other processors.
9928 *) Add CygWin32 platform information to Configure script.
9929 [Alan Batie <batie@aahz.jf.intel.com>]
9931 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
9932 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
9934 *) New program nseq to manipulate netscape certificate sequences
9937 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
9941 *) Fixes to BN code. Previously the default was to define BN_RECURSION
9942 but the BN code had some problems that would cause failures when
9943 doing certificate verification and some other functions.
9944 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9946 *) Add ASN1 and PEM code to support netscape certificate sequences.
9949 *) Add ASN1 and PEM code to support netscape certificate sequences.
9952 *) Add several PKIX and private extended key usage OIDs.
9955 *) Modify the 'ca' program to handle the new extension code. Modify
9956 openssl.cnf for new extension format, add comments.
9959 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
9960 and add a sample to openssl.cnf so req -x509 now adds appropriate
9964 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
9965 error code, add initial support to X509_print() and x509 application.
9968 *) Takes a deep breath and start addding X509 V3 extension support code. Add
9969 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
9970 stuff is currently isolated and isn't even compiled yet.
9973 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
9974 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
9975 Removed the versions check from X509 routines when loading extensions:
9976 this allows certain broken certificates that don't set the version
9977 properly to be processed.
9980 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
9981 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
9982 can still be regenerated with "make depend".
9985 *) Spelling mistake in C version of CAST-128.
9986 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
9988 *) Changes to the error generation code. The perl script err-code.pl
9989 now reads in the old error codes and retains the old numbers, only
9990 adding new ones if necessary. It also only changes the .err files if new
9991 codes are added. The makefiles have been modified to only insert errors
9992 when needed (to avoid needlessly modifying header files). This is done
9993 by only inserting errors if the .err file is newer than the auto generated
9994 C file. To rebuild all the error codes from scratch (the old behaviour)
9995 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
9996 or delete all the .err files.
9999 *) CAST-128 was incorrectly implemented for short keys. The C version has
10000 been fixed, but is untested. The assembler versions are also fixed, but
10001 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
10002 to regenerate it if needed.
10003 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
10004 Hagino <itojun@kame.net>]
10006 *) File was opened incorrectly in randfile.c.
10007 [Ulf Möller <ulf@fitug.de>]
10009 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
10010 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
10011 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
10012 al: it's just almost always a UTCTime. Note this patch adds new error
10013 codes so do a "make errors" if there are problems.
10016 *) Correct Linux 1 recognition in config.
10017 [Ulf Möller <ulf@fitug.de>]
10019 *) Remove pointless MD5 hash when using DSA keys in ca.
10020 [Anonymous <nobody@replay.com>]
10022 *) Generate an error if given an empty string as a cert directory. Also
10023 generate an error if handed NULL (previously returned 0 to indicate an
10024 error, but didn't set one).
10025 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
10027 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
10030 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
10031 parameters. This was causing a warning which killed off the Win32 compile.
10034 *) Remove C++ style comments from crypto/bn/bn_local.h.
10035 [Neil Costigan <neil.costigan@celocom.com>]
10037 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
10038 based on a text string, looking up short and long names and finally
10039 "dot" format. The "dot" format stuff didn't work. Added new function
10040 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
10041 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
10042 OID is not part of the table.
10045 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
10046 X509_LOOKUP_by_alias().
10049 *) Sort openssl functions by name.
10052 *) Get the gendsa program working (hopefully) and add it to app list. Remove
10053 encryption from sample DSA keys (in case anyone is interested the password
10057 *) Make _all_ *_free functions accept a NULL pointer.
10058 [Frans Heymans <fheymans@isaserver.be>]
10060 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
10062 [Anonymous <nobody@replay.com>]
10064 *) s_server should send the CAfile as acceptable CAs, not its own cert.
10065 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10067 *) Don't blow it for numeric -newkey arguments to apps/req.
10068 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10070 *) Temp key "for export" tests were wrong in s3_srvr.c.
10071 [Anonymous <nobody@replay.com>]
10073 *) Add prototype for temp key callback functions
10074 SSL_CTX_set_tmp_{rsa,dh}_callback().
10077 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
10078 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
10081 *) X509_name_add_entry() freed the wrong thing after an error.
10082 [Arne Ansper <arne@ats.cyber.ee>]
10084 *) rsa_eay.c would attempt to free a NULL context.
10085 [Arne Ansper <arne@ats.cyber.ee>]
10087 *) BIO_s_socket() had a broken should_retry() on Windoze.
10088 [Arne Ansper <arne@ats.cyber.ee>]
10090 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
10091 [Arne Ansper <arne@ats.cyber.ee>]
10093 *) Make sure the already existing X509_STORE->depth variable is initialized
10094 in X509_STORE_new(), but document the fact that this variable is still
10095 unused in the certificate verification process.
10096 [Ralf S. Engelschall]
10098 *) Fix the various library and apps files to free up pkeys obtained from
10099 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
10102 *) Fix reference counting in X509_PUBKEY_get(). This makes
10103 demos/maurice/example2.c work, amongst others, probably.
10104 [Steve Henson and Ben Laurie]
10106 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
10107 `openssl' and second, the shortcut symlinks for the `openssl <command>'
10108 are no longer created. This way we have a single and consistent command
10109 line interface `openssl <command>', similar to `cvs <command>'.
10110 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
10112 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
10113 BIT STRING wrapper always have zero unused bits.
10116 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
10119 *) Make the top-level INSTALL documentation easier to understand.
10122 *) Makefiles updated to exit if an error occurs in a sub-directory
10123 make (including if user presses ^C) [Paul Sutton]
10125 *) Make Montgomery context stuff explicit in RSA data structure.
10128 *) Fix build order of pem and err to allow for generated pem.h.
10131 *) Fix renumbering bug in X509_NAME_delete_entry().
10134 *) Enhanced the err-ins.pl script so it makes the error library number
10135 global and can add a library name. This is needed for external ASN1 and
10136 other error libraries.
10139 *) Fixed sk_insert which never worked properly.
10142 *) Fix ASN1 macros so they can handle indefinite length construted
10143 EXPLICIT tags. Some non standard certificates use these: they can now
10147 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
10148 into a single doc/ssleay.txt bundle. This way the information is still
10149 preserved but no longer messes up this directory. Now it's new room for
10150 the new set of documenation files.
10151 [Ralf S. Engelschall]
10153 *) SETs were incorrectly DER encoded. This was a major pain, because they
10154 shared code with SEQUENCEs, which aren't coded the same. This means that
10155 almost everything to do with SETs or SEQUENCEs has either changed name or
10156 number of arguments.
10157 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
10159 *) Fix test data to work with the above.
10162 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
10163 was already fixed by Eric for 0.9.1 it seems.
10164 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
10166 *) Autodetect FreeBSD3.
10169 *) Fix various bugs in Configure. This affects the following platforms:
10173 unixware-2.0-pentium
10177 *) Eliminate generated files from CVS. Reorder tests to regenerate files
10178 before they are needed.
10181 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
10185 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
10187 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
10188 changed SSLeay to OpenSSL in version strings.
10189 [Ralf S. Engelschall]
10191 *) Some fixups to the top-level documents.
10194 *) Fixed the nasty bug where rsaref.h was not found under compile-time
10195 because the symlink to include/ was missing.
10196 [Ralf S. Engelschall]
10198 *) Incorporated the popular no-RSA/DSA-only patches
10199 which allow to compile a RSA-free SSLeay.
10200 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
10202 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
10203 when "ssleay" is still not found.
10204 [Ralf S. Engelschall]
10206 *) Added more platforms to Configure: Cray T3E, HPUX 11,
10207 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
10209 *) Updated the README file.
10210 [Ralf S. Engelschall]
10212 *) Added various .cvsignore files in the CVS repository subdirs
10213 to make a "cvs update" really silent.
10214 [Ralf S. Engelschall]
10216 *) Recompiled the error-definition header files and added
10217 missing symbols to the Win32 linker tables.
10218 [Ralf S. Engelschall]
10220 *) Cleaned up the top-level documents;
10221 o new files: CHANGES and LICENSE
10222 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
10223 o merged COPYRIGHT into LICENSE
10224 o removed obsolete TODO file
10225 o renamed MICROSOFT to INSTALL.W32
10226 [Ralf S. Engelschall]
10228 *) Removed dummy files from the 0.9.1b source tree:
10229 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
10230 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
10231 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
10232 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
10233 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
10234 [Ralf S. Engelschall]
10236 *) Added various platform portability fixes.
10239 *) The Genesis of the OpenSSL rpject:
10240 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
10241 Young and Tim J. Hudson created while they were working for C2Net until
10243 [The OpenSSL Project]
10246 Changes between 0.9.0b and 0.9.1b [not released]
10248 *) Updated a few CA certificates under certs/
10251 *) Changed some BIGNUM api stuff.
10254 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
10255 DGUX x86, Linux Alpha, etc.
10258 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
10259 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
10263 *) Add -strparse option to asn1pars program which parses nested
10265 [Dr Stephen Henson <shenson@bigfoot.com>]
10267 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
10270 *) DSA fix for "ca" program.
10273 *) Added "-genkey" option to "dsaparam" program.
10276 *) Added RIPE MD160 (rmd160) message digest.
10279 *) Added -a (all) option to "ssleay version" command.
10282 *) Added PLATFORM define which is the id given to Configure.
10285 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
10288 *) Extended the ASN.1 parser routines.
10291 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
10294 *) Added a BN_CTX to the BN library.
10297 *) Fixed the weak key values in DES library
10300 *) Changed API in EVP library for cipher aliases.
10303 *) Added support for RC2/64bit cipher.
10306 *) Converted the lhash library to the crypto/mem.c functions.
10309 *) Added more recognized ASN.1 object ids.
10312 *) Added more RSA padding checks for SSL/TLS.
10315 *) Added BIO proxy/filter functionality.
10318 *) Added extra_certs to SSL_CTX which can be used
10319 send extra CA certificates to the client in the CA cert chain sending
10320 process. It can be configured with SSL_CTX_add_extra_chain_cert().
10323 *) Now Fortezza is denied in the authentication phase because
10324 this is key exchange mechanism is not supported by SSLeay at all.
10327 *) Additional PKCS1 checks.
10330 *) Support the string "TLSv1" for all TLS v1 ciphers.
10333 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
10334 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
10337 *) Fixed a few memory leaks.
10340 *) Fixed various code and comment typos.
10343 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
10344 bytes sent in the client random.
10345 [Edward Bishop <ebishop@spyglass.com>]