5 Changes between 1.0.1o and 1.0.1p [xx XXX xxxx]
9 Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
11 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
12 incompatibility in the handling of HMAC. The previous ABI has now been
15 Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
17 *) Malformed ECParameters causes infinite loop
19 When processing an ECParameters structure OpenSSL enters an infinite loop
20 if the curve specified is over a specially malformed binary polynomial
23 This can be used to perform denial of service against any
24 system which processes public keys, certificate requests or
25 certificates. This includes TLS clients and TLS servers with
26 client authentication enabled.
28 This issue was reported to OpenSSL by Joseph Barr-Pixton.
32 *) Exploitable out-of-bounds read in X509_cmp_time
34 X509_cmp_time does not properly check the length of the ASN1_TIME
35 string and can read a few bytes out of bounds. In addition,
36 X509_cmp_time accepts an arbitrary number of fractional seconds in the
39 An attacker can use this to craft malformed certificates and CRLs of
40 various sizes and potentially cause a segmentation fault, resulting in
41 a DoS on applications that verify certificates or CRLs. TLS clients
42 that verify CRLs are affected. TLS clients and servers with client
43 authentication enabled may be affected if they use custom verification
46 This issue was reported to OpenSSL by Robert Swiecki (Google), and
47 independently by Hanno Böck.
51 *) PKCS7 crash with missing EnvelopedContent
53 The PKCS#7 parsing code does not handle missing inner EncryptedContent
54 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
55 with missing content and trigger a NULL pointer dereference on parsing.
57 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
58 structures from untrusted sources are affected. OpenSSL clients and
59 servers are not affected.
61 This issue was reported to OpenSSL by Michal Zalewski (Google).
65 *) CMS verify infinite loop with unknown hash function
67 When verifying a signedData message the CMS code can enter an infinite loop
68 if presented with an unknown hash function OID. This can be used to perform
69 denial of service against any system which verifies signedData messages using
71 This issue was reported to OpenSSL by Johannes Bauer.
75 *) Race condition handling NewSessionTicket
77 If a NewSessionTicket is received by a multi-threaded client when attempting to
78 reuse a previous ticket then a race condition can occur potentially leading to
79 a double free of the ticket data.
83 *) Reject DH handshakes with parameters shorter than 768 bits.
84 [Kurt Roeckx and Emilia Kasper]
86 Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
88 *) Segmentation fault in ASN1_TYPE_cmp fix
90 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
91 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
92 certificate signature algorithm consistency this can be used to crash any
93 certificate verification operation and exploited in a DoS attack. Any
94 application which performs certificate verification is vulnerable including
95 OpenSSL clients and servers which enable client authentication.
99 *) ASN.1 structure reuse memory corruption fix
101 Reusing a structure in ASN.1 parsing may allow an attacker to cause
102 memory corruption via an invalid write. Such reuse is and has been
103 strongly discouraged and is believed to be rare.
105 Applications that parse structures containing CHOICE or ANY DEFINED BY
106 components may be affected. Certificate parsing (d2i_X509 and related
107 functions) are however not affected. OpenSSL clients and servers are
112 *) PKCS7 NULL pointer dereferences fix
114 The PKCS#7 parsing code does not handle missing outer ContentInfo
115 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
116 missing content and trigger a NULL pointer dereference on parsing.
118 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
119 otherwise parse PKCS#7 structures from untrusted sources are
120 affected. OpenSSL clients and servers are not affected.
122 This issue was reported to OpenSSL by Michal Zalewski (Google).
126 *) DoS via reachable assert in SSLv2 servers fix
128 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
129 servers that both support SSLv2 and enable export cipher suites by sending
130 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
132 This issue was discovered by Sean Burford (Google) and Emilia Käsper
133 (OpenSSL development team).
137 *) Use After Free following d2i_ECPrivatekey error fix
139 A malformed EC private key file consumed via the d2i_ECPrivateKey function
140 could cause a use after free condition. This, in turn, could cause a double
141 free in several private key parsing functions (such as d2i_PrivateKey
142 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
143 for applications that receive EC private keys from untrusted
144 sources. This scenario is considered rare.
146 This issue was discovered by the BoringSSL project and fixed in their
151 *) X509_to_X509_REQ NULL pointer deref fix
153 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
154 the certificate key is invalid. This function is rarely used in practice.
156 This issue was discovered by Brian Carpenter.
160 *) Removed the export ciphers from the DEFAULT ciphers
163 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
165 *) Build fixes for the Windows and OpenVMS platforms
166 [Matt Caswell and Richard Levitte]
168 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
170 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
171 message can cause a segmentation fault in OpenSSL due to a NULL pointer
172 dereference. This could lead to a Denial Of Service attack. Thanks to
173 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
177 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
178 dtls1_buffer_record function under certain conditions. In particular this
179 could occur if an attacker sent repeated DTLS records with the same
180 sequence number but for the next epoch. The memory leak could be exploited
181 by an attacker in a Denial of Service attack through memory exhaustion.
182 Thanks to Chris Mueller for reporting this issue.
186 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
187 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
188 method would be set to NULL which could later result in a NULL pointer
189 dereference. Thanks to Frank Schmirler for reporting this issue.
193 *) Abort handshake if server key exchange message is omitted for ephemeral
196 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
197 reporting this issue.
201 *) Remove non-export ephemeral RSA code on client and server. This code
202 violated the TLS standard by allowing the use of temporary RSA keys in
203 non-export ciphersuites and could be used by a server to effectively
204 downgrade the RSA key length used to a value smaller than the server
205 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
206 INRIA or reporting this issue.
210 *) Fixed issue where DH client certificates are accepted without verification.
211 An OpenSSL server will accept a DH certificate for client authentication
212 without the certificate verify message. This effectively allows a client to
213 authenticate without the use of a private key. This only affects servers
214 which trust a client certificate authority which issues certificates
215 containing DH keys: these are extremely rare and hardly ever encountered.
216 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
221 *) Ensure that the session ID context of an SSL is updated when its
222 SSL_CTX is updated via SSL_set_SSL_CTX.
224 The session ID context is typically set from the parent SSL_CTX,
225 and can vary with the CTX.
228 *) Fix various certificate fingerprint issues.
230 By using non-DER or invalid encodings outside the signed portion of a
231 certificate the fingerprint can be changed without breaking the signature.
232 Although no details of the signed portion of the certificate can be changed
233 this can cause problems with some applications: e.g. those using the
234 certificate fingerprint for blacklists.
236 1. Reject signatures with non zero unused bits.
238 If the BIT STRING containing the signature has non zero unused bits reject
239 the signature. All current signature algorithms require zero unused bits.
241 2. Check certificate algorithm consistency.
243 Check the AlgorithmIdentifier inside TBS matches the one in the
244 certificate signature. NB: this will result in signature failure
245 errors for some broken certificates.
247 Thanks to Konrad Kraszewski from Google for reporting this issue.
249 3. Check DSA/ECDSA signatures use DER.
251 Reencode DSA/ECDSA signatures and compare with the original received
252 signature. Return an error if there is a mismatch.
254 This will reject various cases including garbage after signature
255 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
256 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
257 (negative or with leading zeroes).
259 Further analysis was conducted and fixes were developed by Stephen Henson
260 of the OpenSSL core team.
265 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
266 results on some platforms, including x86_64. This bug occurs at random
267 with a very low probability, and is not known to be exploitable in any
268 way, though its exact impact is difficult to determine. Thanks to Pieter
269 Wuille (Blockstream) who reported this issue and also suggested an initial
270 fix. Further analysis was conducted by the OpenSSL development team and
271 Adam Langley of Google. The final fix was developed by Andy Polyakov of
272 the OpenSSL core team.
276 *) Do not resume sessions on the server if the negotiated protocol
277 version does not match the session's version. Resuming with a different
278 version, while not strictly forbidden by the RFC, is of questionable
279 sanity and breaks all known clients.
280 [David Benjamin, Emilia Käsper]
282 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
283 early CCS messages during renegotiation. (Note that because
284 renegotiation is encrypted, this early CCS was not exploitable.)
287 *) Tighten client-side session ticket handling during renegotiation:
288 ensure that the client only accepts a session ticket if the server sends
289 the extension anew in the ServerHello. Previously, a TLS client would
290 reuse the old extension state and thus accept a session ticket if one was
291 announced in the initial ServerHello.
293 Similarly, ensure that the client requires a session ticket if one
294 was advertised in the ServerHello. Previously, a TLS client would
295 ignore a missing NewSessionTicket message.
298 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
302 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
303 sends a carefully crafted handshake message, to cause OpenSSL to fail
304 to free up to 64k of memory causing a memory leak. This could be
305 exploited in a Denial Of Service attack. This issue affects OpenSSL
306 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
307 whether SRTP is used or configured. Implementations of OpenSSL that
308 have been compiled with OPENSSL_NO_SRTP defined are not affected.
310 The fix was developed by the OpenSSL team.
314 *) Session Ticket Memory Leak.
316 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
317 integrity of that ticket is first verified. In the event of a session
318 ticket integrity check failing, OpenSSL will fail to free memory
319 causing a memory leak. By sending a large number of invalid session
320 tickets an attacker could exploit this issue in a Denial Of Service
325 *) Build option no-ssl3 is incomplete.
327 When OpenSSL is configured with "no-ssl3" as a build option, servers
328 could accept and complete a SSL 3.0 handshake, and clients could be
329 configured to send them.
331 [Akamai and the OpenSSL team]
333 *) Add support for TLS_FALLBACK_SCSV.
334 Client applications doing fallback retries should call
335 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
337 [Adam Langley, Bodo Moeller]
339 *) Add additional DigestInfo checks.
341 Reencode DigestInto in DER and check against the original when
342 verifying RSA signature: this will reject any improperly encoded
343 DigestInfo structures.
345 Note: this is a precautionary measure and no attacks are currently known.
349 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
351 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
352 SRP code can be overrun an internal buffer. Add sanity check that
353 g, A, B < N to SRP code.
355 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
356 Group for discovering this issue.
360 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
361 TLS 1.0 instead of higher protocol versions when the ClientHello message
362 is badly fragmented. This allows a man-in-the-middle attacker to force a
363 downgrade to TLS 1.0 even if both the server and the client support a
364 higher protocol version, by modifying the client's TLS records.
366 Thanks to David Benjamin and Adam Langley (Google) for discovering and
367 researching this issue.
371 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
372 to a denial of service attack. A malicious server can crash the client
373 with a null pointer dereference (read) by specifying an anonymous (EC)DH
374 ciphersuite and sending carefully crafted handshake messages.
376 Thanks to Felix Gröbert (Google) for discovering and researching this
381 *) By sending carefully crafted DTLS packets an attacker could cause openssl
382 to leak memory. This can be exploited through a Denial of Service attack.
383 Thanks to Adam Langley for discovering and researching this issue.
387 *) An attacker can force openssl to consume large amounts of memory whilst
388 processing DTLS handshake messages. This can be exploited through a
389 Denial of Service attack.
390 Thanks to Adam Langley for discovering and researching this issue.
394 *) An attacker can force an error condition which causes openssl to crash
395 whilst processing DTLS packets due to memory being freed twice. This
396 can be exploited through a Denial of Service attack.
397 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
402 *) If a multithreaded client connects to a malicious server using a resumed
403 session and the server sends an ec point format extension it could write
404 up to 255 bytes to freed memory.
406 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
411 *) A malicious server can crash an OpenSSL client with a null pointer
412 dereference (read) by specifying an SRP ciphersuite even though it was not
413 properly negotiated with the client. This can be exploited through a
414 Denial of Service attack.
416 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
417 discovering and researching this issue.
421 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
422 X509_name_oneline, X509_name_print_ex et al. to leak some information
423 from the stack. Applications may be affected if they echo pretty printing
424 output to the attacker.
426 Thanks to Ivan Fratric (Google) for discovering this issue.
428 [Emilia Käsper, and Steve Henson]
430 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
431 for corner cases. (Certain input points at infinity could lead to
432 bogus results, with non-infinity inputs mapped to infinity too.)
435 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
437 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
438 handshake can force the use of weak keying material in OpenSSL
439 SSL/TLS clients and servers.
441 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
442 researching this issue. (CVE-2014-0224)
443 [KIKUCHI Masashi, Steve Henson]
445 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
446 OpenSSL DTLS client the code can be made to recurse eventually crashing
449 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
451 [Imre Rad, Steve Henson]
453 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
454 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
455 client or server. This is potentially exploitable to run arbitrary
456 code on a vulnerable client or server.
458 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
459 [Jüri Aedla, Steve Henson]
461 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
462 are subject to a denial of service attack.
464 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
465 this issue. (CVE-2014-3470)
466 [Felix Gröbert, Ivan Fratric, Steve Henson]
468 *) Harmonize version and its documentation. -f flag is used to display
470 [mancha <mancha1@zoho.com>]
472 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
474 [mancha <mancha1@zoho.com>]
476 *) Fix some double frees. These are not thought to be exploitable.
477 [mancha <mancha1@zoho.com>]
479 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
481 *) A missing bounds check in the handling of the TLS heartbeat extension
482 can be used to reveal up to 64k of memory to a connected client or
485 Thanks for Neel Mehta of Google Security for discovering this bug and to
486 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
487 preparing the fix (CVE-2014-0160)
488 [Adam Langley, Bodo Moeller]
490 *) Fix for the attack described in the paper "Recovering OpenSSL
491 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
492 by Yuval Yarom and Naomi Benger. Details can be obtained from:
493 http://eprint.iacr.org/2014/140
495 Thanks to Yuval Yarom and Naomi Benger for discovering this
496 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
497 [Yuval Yarom and Naomi Benger]
499 *) TLS pad extension: draft-agl-tls-padding-03
501 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
502 TLS client Hello record length value would otherwise be > 255 and
503 less that 512 pad with a dummy extension containing zeroes so it
504 is at least 512 bytes long.
506 [Adam Langley, Steve Henson]
508 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
510 *) Fix for TLS record tampering bug. A carefully crafted invalid
511 handshake could crash OpenSSL with a NULL pointer exception.
512 Thanks to Anton Johansson for reporting this issues.
515 *) Keep original DTLS digest and encryption contexts in retransmission
516 structures so we can use the previous session parameters if they need
517 to be resent. (CVE-2013-6450)
520 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
521 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
522 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
523 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
524 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
525 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
526 [Rob Stradling, Adam Langley]
528 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
530 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
531 supporting platforms or when small records were transferred.
532 [Andy Polyakov, Steve Henson]
534 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
536 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
538 This addresses the flaw in CBC record processing discovered by
539 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
540 at: http://www.isg.rhul.ac.uk/tls/
542 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
543 Security Group at Royal Holloway, University of London
544 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
545 Emilia Käsper for the initial patch.
547 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
549 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
550 ciphersuites which can be exploited in a denial of service attack.
551 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
552 and detecting this bug and to Wolfgang Ettlinger
553 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
557 *) Return an error when checking OCSP signatures when key is NULL.
558 This fixes a DoS attack. (CVE-2013-0166)
561 *) Make openssl verify return errors.
562 [Chris Palmer <palmer@google.com> and Ben Laurie]
564 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
565 the right response is stapled. Also change SSL_get_certificate()
566 so it returns the certificate actually sent.
567 See http://rt.openssl.org/Ticket/Display.html?id=2836.
568 [Rob Stradling <rob.stradling@comodo.com>]
570 *) Fix possible deadlock when decoding public keys.
573 *) Don't use TLS 1.0 record version number in initial client hello
577 Changes between 1.0.1b and 1.0.1c [10 May 2012]
579 *) Sanity check record length before skipping explicit IV in TLS
580 1.2, 1.1 and DTLS to fix DoS attack.
582 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
583 fuzzing as a service testing platform.
587 *) Initialise tkeylen properly when encrypting CMS messages.
588 Thanks to Solar Designer of Openwall for reporting this issue.
591 *) In FIPS mode don't try to use composite ciphers as they are not
595 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
597 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
598 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
599 mean any application compiled against OpenSSL 1.0.0 headers setting
600 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
601 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
602 0x10000000L Any application which was previously compiled against
603 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
604 will need to be recompiled as a result. Letting be results in
605 inability to disable specifically TLS 1.1 and in client context,
606 in unlike event, limit maximum offered version to TLS 1.0 [see below].
609 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
610 disable just protocol X, but all protocols above X *if* there are
611 protocols *below* X still enabled. In more practical terms it means
612 that if application wants to disable TLS1.0 in favor of TLS1.1 and
613 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
614 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
618 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
620 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
621 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
622 in CRYPTO_realloc_clean.
624 Thanks to Tavis Ormandy, Google Security Team, for discovering this
625 issue and to Adam Langley <agl@chromium.org> for fixing it.
627 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
629 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
632 *) Workarounds for some broken servers that "hang" if a client hello
633 record length exceeds 255 bytes.
635 1. Do not use record version number > TLS 1.0 in initial client
636 hello: some (but not all) hanging servers will now work.
637 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
638 the number of ciphers sent in the client hello. This should be
639 set to an even number, such as 50, for example by passing:
640 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
641 Most broken servers should now work.
642 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
643 TLS 1.2 client support entirely.
646 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
649 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
651 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
652 STRING form instead of a DigestInfo.
655 *) The format used for MDC2 RSA signatures is inconsistent between EVP
656 and the RSA_sign/RSA_verify functions. This was made more apparent when
657 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
658 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
659 the correct format in RSA_verify so both forms transparently work.
662 *) Some servers which support TLS 1.0 can choke if we initially indicate
663 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
664 encrypted premaster secret. As a workaround use the maximum pemitted
665 client version in client hello, this should keep such servers happy
666 and still work with previous versions of OpenSSL.
669 *) Add support for TLS/DTLS heartbeats.
670 [Robin Seggelmann <seggelmann@fh-muenster.de>]
672 *) Add support for SCTP.
673 [Robin Seggelmann <seggelmann@fh-muenster.de>]
675 *) Improved PRNG seeding for VOS.
676 [Paul Green <Paul.Green@stratus.com>]
678 *) Extensive assembler packs updates, most notably:
680 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
681 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
682 - x86_64: bit-sliced AES implementation;
683 - ARM: NEON support, contemporary platforms optimizations;
684 - s390x: z196 support;
685 - *: GHASH and GF(2^m) multiplication implementations;
689 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
690 (removal of unnecessary code)
691 [Peter Sylvester <peter.sylvester@edelweb.fr>]
693 *) Add TLS key material exporter from RFC 5705.
696 *) Add DTLS-SRTP negotiation from RFC 5764.
699 *) Add Next Protocol Negotiation,
700 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
701 disabled with a no-npn flag to config or Configure. Code donated
703 [Adam Langley <agl@google.com> and Ben Laurie]
705 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
706 NIST-P256, NIST-P521, with constant-time single point multiplication on
707 typical inputs. Compiler support for the nonstandard type __uint128_t is
708 required to use this (present in gcc 4.4 and later, for 64-bit builds).
709 Code made available under Apache License version 2.0.
711 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
712 line to include this in your build of OpenSSL, and run "make depend" (or
713 "make update"). This enables the following EC_METHODs:
715 EC_GFp_nistp224_method()
716 EC_GFp_nistp256_method()
717 EC_GFp_nistp521_method()
719 EC_GROUP_new_by_curve_name() will automatically use these (while
720 EC_GROUP_new_curve_GFp() currently prefers the more flexible
722 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
724 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
725 all platforms. Move ssize_t definition from e_os.h to the public
726 header file e_os2.h as it now appears in public header file cms.h
729 *) New -sigopt option to the ca, req and x509 utilities. Additional
730 signature parameters can be passed using this option and in
734 *) Add RSA PSS signing function. This will generate and set the
735 appropriate AlgorithmIdentifiers for PSS based on those in the
736 corresponding EVP_MD_CTX structure. No application support yet.
739 *) Support for companion algorithm specific ASN1 signing routines.
740 New function ASN1_item_sign_ctx() signs a pre-initialised
741 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
742 the appropriate parameters.
745 *) Add new algorithm specific ASN1 verification initialisation function
746 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
747 handling will be the same no matter what EVP_PKEY_METHOD is used.
748 Add a PSS handler to support verification of PSS signatures: checked
749 against a number of sample certificates.
752 *) Add signature printing for PSS. Add PSS OIDs.
753 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
755 *) Add algorithm specific signature printing. An individual ASN1 method
756 can now print out signatures instead of the standard hex dump.
758 More complex signatures (e.g. PSS) can print out more meaningful
759 information. Include DSA version that prints out the signature
763 *) Password based recipient info support for CMS library: implementing
767 *) Split password based encryption into PBES2 and PBKDF2 functions. This
768 neatly separates the code into cipher and PBE sections and is required
769 for some algorithms that split PBES2 into separate pieces (such as
773 *) Session-handling fixes:
774 - Fix handling of connections that are resuming with a session ID,
775 but also support Session Tickets.
776 - Fix a bug that suppressed issuing of a new ticket if the client
777 presented a ticket with an expired session.
778 - Try to set the ticket lifetime hint to something reasonable.
779 - Make tickets shorter by excluding irrelevant information.
780 - On the client side, don't ignore renewed tickets.
781 [Adam Langley, Bodo Moeller (Google)]
783 *) Fix PSK session representation.
786 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
788 This work was sponsored by Intel.
791 *) Add GCM support to TLS library. Some custom code is needed to split
792 the IV between the fixed (from PRF) and explicit (from TLS record)
793 portions. This adds all GCM ciphersuites supported by RFC5288 and
794 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
795 add a special AESGCM string for GCM only.
798 *) Expand range of ctrls for AES GCM. Permit setting invocation
799 field on decrypt and retrieval of invocation field only on encrypt.
802 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
803 As required by RFC5289 these ciphersuites cannot be used if for
804 versions of TLS earlier than 1.2.
807 *) For FIPS capable OpenSSL interpret a NULL default public key method
808 as unset and return the appopriate default but do *not* set the default.
809 This means we can return the appopriate method in applications that
810 swicth between FIPS and non-FIPS modes.
813 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
814 ENGINE is used then we cannot handle that in the FIPS module so we
815 keep original code iff non-FIPS operations are allowed.
818 *) Add -attime option to openssl utilities.
819 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
821 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
824 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
825 FIPS EC methods unconditionally for now.
828 *) New build option no-ec2m to disable characteristic 2 code.
831 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
832 all cases can be covered as some introduce binary incompatibilities.
835 *) Redirect RSA operations to FIPS module including keygen,
836 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
839 *) Add similar low level API blocking to ciphers.
842 *) Low level digest APIs are not approved in FIPS mode: any attempt
843 to use these will cause a fatal error. Applications that *really* want
844 to use them can use the private_* version instead.
847 *) Redirect cipher operations to FIPS module for FIPS builds.
850 *) Redirect digest operations to FIPS module for FIPS builds.
853 *) Update build system to add "fips" flag which will link in fipscanister.o
854 for static and shared library builds embedding a signature if needed.
857 *) Output TLS supported curves in preference order instead of numerical
858 order. This is currently hardcoded for the highest order curves first.
859 This should be configurable so applications can judge speed vs strength.
862 *) Add TLS v1.2 server support for client authentication.
865 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
869 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
870 FIPS modules versions.
873 *) Add TLS v1.2 client side support for client authentication. Keep cache
874 of handshake records longer as we don't know the hash algorithm to use
875 until after the certificate request message is received.
878 *) Initial TLS v1.2 client support. Add a default signature algorithms
879 extension including all the algorithms we support. Parse new signature
880 format in client key exchange. Relax some ECC signing restrictions for
881 TLS v1.2 as indicated in RFC5246.
884 *) Add server support for TLS v1.2 signature algorithms extension. Switch
885 to new signature format when needed using client digest preference.
886 All server ciphersuites should now work correctly in TLS v1.2. No client
887 support yet and no support for client certificates.
890 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
891 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
892 ciphersuites. At present only RSA key exchange ciphersuites work with
893 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
894 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
895 and version checking.
898 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
899 with this defined it will not be affected by any changes to ssl internal
900 structures. Add several utility functions to allow openssl application
901 to work with OPENSSL_NO_SSL_INTERN defined.
905 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
907 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
910 *) Permit abbreviated handshakes when renegotiating using the function
911 SSL_renegotiate_abbreviated().
912 [Robin Seggelmann <seggelmann@fh-muenster.de>]
914 *) Add call to ENGINE_register_all_complete() to
915 ENGINE_load_builtin_engines(), so some implementations get used
916 automatically instead of needing explicit application support.
919 *) Add support for TLS key exporter as described in RFC5705.
920 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
922 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
923 a few changes are required:
925 Add SSL_OP_NO_TLSv1_1 flag.
927 Update version checking logic to handle version 1.1.
928 Add explicit IV handling (ported from DTLS code).
929 Add command line options to s_client/s_server.
932 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
934 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
935 in CMS and PKCS7 code. When RSA decryption fails use a random key for
936 content decryption and always return the same error. Note: this attack
937 needs on average 2^20 messages so it only affects automated senders. The
938 old behaviour can be reenabled in the CMS code by setting the
939 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
940 an MMA defence is not necessary.
941 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
942 this issue. (CVE-2012-0884)
945 *) Fix CVE-2011-4619: make sure we really are receiving a
946 client hello before rejecting multiple SGC restarts. Thanks to
947 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
950 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
952 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
953 Thanks to Antonio Martin, Enterprise Secure Access Research and
954 Development, Cisco Systems, Inc. for discovering this bug and
955 preparing a fix. (CVE-2012-0050)
958 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
960 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
961 of the Vaudenay padding oracle attack on CBC mode encryption
962 which enables an efficient plaintext recovery attack against
963 the OpenSSL implementation of DTLS. Their attack exploits timing
964 differences arising during decryption processing. A research
965 paper describing this attack can be found at:
966 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
967 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
968 Security Group at Royal Holloway, University of London
969 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
970 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
971 for preparing the fix. (CVE-2011-4108)
972 [Robin Seggelmann, Michael Tuexen]
974 *) Clear bytes used for block padding of SSL 3.0 records.
976 [Adam Langley (Google)]
978 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
979 Kadianakis <desnacked@gmail.com> for discovering this issue and
980 Adam Langley for preparing the fix. (CVE-2011-4619)
981 [Adam Langley (Google)]
983 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
984 [Andrey Kulikov <amdeich@gmail.com>]
986 *) Prevent malformed RFC3779 data triggering an assertion failure.
987 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
988 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
989 [Rob Austein <sra@hactrn.net>]
991 *) Improved PRNG seeding for VOS.
992 [Paul Green <Paul.Green@stratus.com>]
994 *) Fix ssl_ciph.c set-up race.
995 [Adam Langley (Google)]
997 *) Fix spurious failures in ecdsatest.c.
998 [Emilia Käsper (Google)]
1000 *) Fix the BIO_f_buffer() implementation (which was mixing different
1001 interpretations of the '..._len' fields).
1002 [Adam Langley (Google)]
1004 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1005 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1006 threads won't reuse the same blinding coefficients.
1008 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1009 lock to call BN_BLINDING_invert_ex, and avoids one use of
1010 BN_BLINDING_update for each BN_BLINDING structure (previously,
1011 the last update always remained unused).
1012 [Emilia Käsper (Google)]
1014 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
1015 [Bob Buckholz (Google)]
1017 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
1019 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
1020 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
1021 [Kaspar Brand <ossl@velox.ch>]
1023 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1024 for multi-threaded use of ECDH. (CVE-2011-3210)
1025 [Adam Langley (Google)]
1027 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1030 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
1031 signature public key algorithm by using OID xref utilities instead.
1032 Before this you could only use some ECC ciphersuites with SHA1 only.
1035 *) Add protection against ECDSA timing attacks as mentioned in the paper
1036 by Billy Bob Brumley and Nicola Tuveri, see:
1038 http://eprint.iacr.org/2011/232.pdf
1040 [Billy Bob Brumley and Nicola Tuveri]
1042 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
1044 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1045 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1047 *) Fix bug in string printing code: if *any* escaping is enabled we must
1048 escape the escape character (backslash) or the resulting string is
1052 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
1054 *) Disable code workaround for ancient and obsolete Netscape browsers
1055 and servers: an attacker can use it in a ciphersuite downgrade attack.
1056 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1059 *) Fixed J-PAKE implementation error, originally discovered by
1060 Sebastien Martini, further info and confirmation from Stefan
1061 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1064 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
1066 *) Fix extension code to avoid race conditions which can result in a buffer
1067 overrun vulnerability: resumed sessions must not be modified as they can
1068 be shared by multiple threads. CVE-2010-3864
1071 *) Fix WIN32 build system to correctly link an ENGINE directory into
1075 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
1077 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
1079 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
1081 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
1083 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
1084 context. The operation can be customised via the ctrl mechanism in
1085 case ENGINEs want to include additional functionality.
1088 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
1091 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
1092 output hashes compatible with older versions of OpenSSL.
1093 [Willy Weisz <weisz@vcpc.univie.ac.at>]
1095 *) Fix compression algorithm handling: if resuming a session use the
1096 compression algorithm of the resumed session instead of determining
1097 it from client hello again. Don't allow server to change algorithm.
1100 *) Add load_crls() function to apps tidying load_certs() too. Add option
1101 to verify utility to allow additional CRLs to be included.
1104 *) Update OCSP request code to permit adding custom headers to the request:
1105 some responders need this.
1108 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
1110 [Julia Lawall <julia@diku.dk>]
1112 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
1113 needlessly dereferenced structures, used obsolete functions and
1114 didn't handle all updated verify codes correctly.
1117 *) Disable MD2 in the default configuration.
1120 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
1121 indicate the initial BIO being pushed or popped. This makes it possible
1122 to determine whether the BIO is the one explicitly called or as a result
1123 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
1124 it handles reference counts correctly and doesn't zero out the I/O bio
1125 when it is not being explicitly popped. WARNING: applications which
1126 included workarounds for the old buggy behaviour will need to be modified
1127 or they could free up already freed BIOs.
1130 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
1131 renaming to all platforms (within the 0.9.8 branch, this was
1132 done conditionally on Netware platforms to avoid a name clash).
1133 [Guenter <lists@gknw.net>]
1135 *) Add ECDHE and PSK support to DTLS.
1136 [Michael Tuexen <tuexen@fh-muenster.de>]
1138 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
1142 *) Add "missing" function EVP_MD_flags() (without this the only way to
1143 retrieve a digest flags is by accessing the structure directly. Update
1144 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
1145 or cipher is registered as in the "from" argument. Print out all
1146 registered digests in the dgst usage message instead of manually
1147 attempting to work them out.
1150 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
1151 this allows the use of compression and extensions. Change default cipher
1152 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
1153 by default unless an application cipher string requests it.
1156 *) Alter match criteria in PKCS12_parse(). It used to try to use local
1157 key ids to find matching certificates and keys but some PKCS#12 files
1158 don't follow the (somewhat unwritten) rules and this strategy fails.
1159 Now just gather all certificates together and the first private key
1160 then look for the first certificate that matches the key.
1163 *) Support use of registered digest and cipher names for dgst and cipher
1164 commands instead of having to add each one as a special case. So now
1171 openssl dgst -sha256 foo
1173 and this works for ENGINE based algorithms too.
1177 *) Update Gost ENGINE to support parameter files.
1178 [Victor B. Wagner <vitus@cryptocom.ru>]
1180 *) Support GeneralizedTime in ca utility.
1181 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
1183 *) Enhance the hash format used for certificate directory links. The new
1184 form uses the canonical encoding (meaning equivalent names will work
1185 even if they aren't identical) and uses SHA1 instead of MD5. This form
1186 is incompatible with the older format and as a result c_rehash should
1187 be used to rebuild symbolic links.
1190 *) Make PKCS#8 the default write format for private keys, replacing the
1191 traditional format. This form is standardised, more secure and doesn't
1192 include an implicit MD5 dependency.
1195 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
1196 committed to OpenSSL should pass this lot as a minimum.
1199 *) Add session ticket override functionality for use by EAP-FAST.
1200 [Jouni Malinen <j@w1.fi>]
1202 *) Modify HMAC functions to return a value. Since these can be implemented
1203 in an ENGINE errors can occur.
1206 *) Type-checked OBJ_bsearch_ex.
1209 *) Type-checked OBJ_bsearch. Also some constification necessitated
1210 by type-checking. Still to come: TXT_DB, bsearch(?),
1211 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
1215 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
1216 seconds to a tm structure directly, instead of going through OS
1217 specific date routines. This avoids any issues with OS routines such
1218 as the year 2038 bug. New *_adj() functions for ASN1 time structures
1219 and X509_time_adj_ex() to cover the extended range. The existing
1220 X509_time_adj() is still usable and will no longer have any date issues.
1223 *) Delta CRL support. New use deltas option which will attempt to locate
1224 and search any appropriate delta CRLs available.
1226 This work was sponsored by Google.
1229 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
1230 code and add additional score elements. Validate alternate CRL paths
1231 as part of the CRL checking and indicate a new error "CRL path validation
1232 error" in this case. Applications wanting additional details can use
1233 the verify callback and check the new "parent" field. If this is not
1234 NULL CRL path validation is taking place. Existing applications wont
1235 see this because it requires extended CRL support which is off by
1238 This work was sponsored by Google.
1241 *) Support for freshest CRL extension.
1243 This work was sponsored by Google.
1246 *) Initial indirect CRL support. Currently only supported in the CRLs
1247 passed directly and not via lookup. Process certificate issuer
1248 CRL entry extension and lookup CRL entries by bother issuer name
1249 and serial number. Check and process CRL issuer entry in IDP extension.
1251 This work was sponsored by Google.
1254 *) Add support for distinct certificate and CRL paths. The CRL issuer
1255 certificate is validated separately in this case. Only enabled if
1256 an extended CRL support flag is set: this flag will enable additional
1257 CRL functionality in future.
1259 This work was sponsored by Google.
1262 *) Add support for policy mappings extension.
1264 This work was sponsored by Google.
1267 *) Fixes to pathlength constraint, self issued certificate handling,
1268 policy processing to align with RFC3280 and PKITS tests.
1270 This work was sponsored by Google.
1273 *) Support for name constraints certificate extension. DN, email, DNS
1274 and URI types are currently supported.
1276 This work was sponsored by Google.
1279 *) To cater for systems that provide a pointer-based thread ID rather
1280 than numeric, deprecate the current numeric thread ID mechanism and
1281 replace it with a structure and associated callback type. This
1282 mechanism allows a numeric "hash" to be extracted from a thread ID in
1283 either case, and on platforms where pointers are larger than 'long',
1284 mixing is done to help ensure the numeric 'hash' is usable even if it
1285 can't be guaranteed unique. The default mechanism is to use "&errno"
1286 as a pointer-based thread ID to distinguish between threads.
1288 Applications that want to provide their own thread IDs should now use
1289 CRYPTO_THREADID_set_callback() to register a callback that will call
1290 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
1292 Note that ERR_remove_state() is now deprecated, because it is tied
1293 to the assumption that thread IDs are numeric. ERR_remove_state(0)
1294 to free the current thread's error state should be replaced by
1295 ERR_remove_thread_state(NULL).
1297 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
1298 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
1299 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
1300 application was previously providing a numeric thread callback that
1301 was inappropriate for distinguishing threads, then uniqueness might
1302 have been obtained with &errno that happened immediately in the
1303 intermediate development versions of OpenSSL; this is no longer the
1304 case, the numeric thread callback will now override the automatic use
1306 [Geoff Thorpe, with help from Bodo Moeller]
1308 *) Initial support for different CRL issuing certificates. This covers a
1309 simple case where the self issued certificates in the chain exist and
1310 the real CRL issuer is higher in the existing chain.
1312 This work was sponsored by Google.
1315 *) Removed effectively defunct crypto/store from the build.
1318 *) Revamp of STACK to provide stronger type-checking. Still to come:
1319 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
1320 ASN1_STRING, CONF_VALUE.
1323 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
1324 RAM on SSL connections. This option can save about 34k per idle SSL.
1327 *) Revamp of LHASH to provide stronger type-checking. Still to come:
1328 STACK, TXT_DB, bsearch, qsort.
1331 *) Initial support for Cryptographic Message Syntax (aka CMS) based
1332 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
1333 support for data, signedData, compressedData, digestedData and
1334 encryptedData, envelopedData types included. Scripts to check against
1335 RFC4134 examples draft and interop and consistency checks of many
1336 content types and variants.
1339 *) Add options to enc utility to support use of zlib compression BIO.
1342 *) Extend mk1mf to support importing of options and assembly language
1343 files from Configure script, currently only included in VC-WIN32.
1344 The assembly language rules can now optionally generate the source
1345 files from the associated perl scripts.
1348 *) Implement remaining functionality needed to support GOST ciphersuites.
1349 Interop testing has been performed using CryptoPro implementations.
1350 [Victor B. Wagner <vitus@cryptocom.ru>]
1352 *) s390x assembler pack.
1355 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
1359 *) Implement Opaque PRF Input TLS extension as specified in
1360 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
1361 official specification yet and no extension type assignment by
1362 IANA exists, this extension (for now) will have to be explicitly
1363 enabled when building OpenSSL by providing the extension number
1364 to use. For example, specify an option
1366 -DTLSEXT_TYPE_opaque_prf_input=0x9527
1368 to the "config" or "Configure" script to enable the extension,
1369 assuming extension number 0x9527 (which is a completely arbitrary
1370 and unofficial assignment based on the MD5 hash of the Internet
1371 Draft). Note that by doing so, you potentially lose
1372 interoperability with other TLS implementations since these might
1373 be using the same extension number for other purposes.
1375 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
1376 opaque PRF input value to use in the handshake. This will create
1377 an interal copy of the length-'len' string at 'src', and will
1378 return non-zero for success.
1380 To get more control and flexibility, provide a callback function
1383 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
1384 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
1388 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
1391 Callback function 'cb' will be called in handshakes, and is
1392 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
1393 Argument 'arg' is for application purposes (the value as given to
1394 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
1395 be provided to the callback function). The callback function
1396 has to return non-zero to report success: usually 1 to use opaque
1397 PRF input just if possible, or 2 to enforce use of the opaque PRF
1398 input. In the latter case, the library will abort the handshake
1399 if opaque PRF input is not successfully negotiated.
1401 Arguments 'peerinput' and 'len' given to the callback function
1402 will always be NULL and 0 in the case of a client. A server will
1403 see the client's opaque PRF input through these variables if
1404 available (NULL and 0 otherwise). Note that if the server
1405 provides an opaque PRF input, the length must be the same as the
1406 length of the client's opaque PRF input.
1408 Note that the callback function will only be called when creating
1409 a new session (session resumption can resume whatever was
1410 previously negotiated), and will not be called in SSL 2.0
1411 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
1412 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
1413 for applications that need to enforce opaque PRF input.
1417 *) Update ssl code to support digests other than SHA1+MD5 for handshake
1420 [Victor B. Wagner <vitus@cryptocom.ru>]
1422 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1423 RFC4507bis. The encrypted ticket format is an encrypted encoded
1424 SSL_SESSION structure, that way new session features are automatically
1427 If a client application caches session in an SSL_SESSION structure
1428 support is transparent because tickets are now stored in the encoded
1431 The SSL_CTX structure automatically generates keys for ticket
1432 protection in servers so again support should be possible
1433 with no application modification.
1435 If a client or server wishes to disable RFC4507 support then the option
1436 SSL_OP_NO_TICKET can be set.
1438 Add a TLS extension debugging callback to allow the contents of any client
1439 or server extensions to be examined.
1441 This work was sponsored by Google.
1444 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
1445 OpenSSL should now compile cleanly on gcc 4.2
1446 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
1448 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
1449 support including streaming MAC support: this is required for GOST
1450 ciphersuite support.
1451 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
1453 *) Add option -stream to use PKCS#7 streaming in smime utility. New
1454 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
1455 to output in BER and PEM format.
1458 *) Experimental support for use of HMAC via EVP_PKEY interface. This
1459 allows HMAC to be handled via the EVP_DigestSign*() interface. The
1460 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
1461 ENGINE support for HMAC keys which are unextractable. New -mac and
1462 -macopt options to dgst utility.
1465 *) New option -sigopt to dgst utility. Update dgst to use
1466 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
1467 alternative signing paramaters such as X9.31 or PSS in the dgst
1471 *) Change ssl_cipher_apply_rule(), the internal function that does
1472 the work each time a ciphersuite string requests enabling
1473 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
1474 removing ("!foo+bar") a class of ciphersuites: Now it maintains
1475 the order of disabled ciphersuites such that those ciphersuites
1476 that most recently went from enabled to disabled not only stay
1477 in order with respect to each other, but also have higher priority
1478 than other disabled ciphersuites the next time ciphersuites are
1481 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
1482 the same ciphersuites as with "HIGH" alone, but in a specific
1483 order where the PSK ciphersuites come first (since they are the
1484 most recently disabled ciphersuites when "HIGH" is parsed).
1486 Also, change ssl_create_cipher_list() (using this new
1487 funcionality) such that between otherwise identical
1488 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
1492 *) Change ssl_create_cipher_list() so that it automatically
1493 arranges the ciphersuites in reasonable order before starting
1494 to process the rule string. Thus, the definition for "DEFAULT"
1495 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
1496 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
1497 This makes it much easier to arrive at a reasonable default order
1498 in applications for which anonymous ciphers are OK (meaning
1499 that you can't actually use DEFAULT).
1500 [Bodo Moeller; suggested by Victor Duchovni]
1502 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
1503 processing) into multiple integers instead of setting
1504 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
1505 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
1506 (These masks as well as the individual bit definitions are hidden
1507 away into the non-exported interface ssl/ssl_locl.h, so this
1508 change to the definition of the SSL_CIPHER structure shouldn't
1509 affect applications.) This give us more bits for each of these
1510 categories, so there is no longer a need to coagulate AES128 and
1511 AES256 into a single algorithm bit, and to coagulate Camellia128
1512 and Camellia256 into a single algorithm bit, which has led to all
1515 Thus, among other things, the kludge introduced in 0.9.7m and
1516 0.9.8e for masking out AES256 independently of AES128 or masking
1517 out Camellia256 independently of AES256 is not needed here in 0.9.9.
1519 With the change, we also introduce new ciphersuite aliases that
1520 so far were missing: "AES128", "AES256", "CAMELLIA128", and
1524 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
1525 Use the leftmost N bytes of the signature input if the input is
1526 larger than the prime q (with N being the size in bytes of q).
1529 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
1530 it yet and it is largely untested.
1533 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
1536 *) Initial incomplete changes to avoid need for function casts in OpenSSL
1537 some compilers (gcc 4.2 and later) reject their use. Safestack is
1538 reimplemented. Update ASN1 to avoid use of legacy functions.
1541 *) Win32/64 targets are linked with Winsock2.
1544 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
1545 to external functions. This can be used to increase CRL handling
1546 efficiency especially when CRLs are very large by (for example) storing
1547 the CRL revoked certificates in a database.
1550 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
1551 new CRLs added to a directory can be used. New command line option
1552 -verify_return_error to s_client and s_server. This causes real errors
1553 to be returned by the verify callback instead of carrying on no matter
1554 what. This reflects the way a "real world" verify callback would behave.
1557 *) GOST engine, supporting several GOST algorithms and public key formats.
1558 Kindly donated by Cryptocom.
1561 *) Partial support for Issuing Distribution Point CRL extension. CRLs
1562 partitioned by DP are handled but no indirect CRL or reason partitioning
1563 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
1564 selected via a scoring technique which handles IDP and AKID in CRLs.
1567 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
1568 will ultimately be used for all verify operations: this will remove the
1569 X509_STORE dependency on certificate verification and allow alternative
1570 lookup methods. X509_STORE based implementations of these two callbacks.
1573 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
1574 Modify get_crl() to find a valid (unexpired) CRL if possible.
1577 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
1578 this would be called X509_CRL_cmp() but that name is already used by
1579 a function that just compares CRL issuer names. Cache several CRL
1580 extensions in X509_CRL structure and cache CRLDP in X509.
1583 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
1584 this maps equivalent X509_NAME structures into a consistent structure.
1585 Name comparison can then be performed rapidly using memcmp().
1588 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
1592 *) Allow digests to supply their own micalg string for S/MIME type using
1593 the ctrl EVP_MD_CTRL_MICALG.
1596 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
1597 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
1598 ctrl. It can then customise the structure before and/or after signing
1602 *) New function OBJ_add_sigid() to allow application defined signature OIDs
1603 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
1604 to free up any added signature OIDs.
1607 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
1608 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
1609 digest and cipher tables. New options added to openssl utility:
1610 list-message-digest-algorithms and list-cipher-algorithms.
1613 *) Change the array representation of binary polynomials: the list
1614 of degrees of non-zero coefficients is now terminated with -1.
1615 Previously it was terminated with 0, which was also part of the
1616 value; thus, the array representation was not applicable to
1617 polynomials where t^0 has coefficient zero. This change makes
1618 the array representation useful in a more general context.
1621 *) Various modifications and fixes to SSL/TLS cipher string
1622 handling. For ECC, the code now distinguishes between fixed ECDH
1623 with RSA certificates on the one hand and with ECDSA certificates
1624 on the other hand, since these are separate ciphersuites. The
1625 unused code for Fortezza ciphersuites has been removed.
1627 For consistency with EDH, ephemeral ECDH is now called "EECDH"
1628 (not "ECDHE"). For consistency with the code for DH
1629 certificates, use of ECDH certificates is now considered ECDH
1630 authentication, not RSA or ECDSA authentication (the latter is
1631 merely the CA's signing algorithm and not actively used in the
1634 The temporary ciphersuite alias "ECCdraft" is no longer
1635 available, and ECC ciphersuites are no longer excluded from "ALL"
1636 and "DEFAULT". The following aliases now exist for RFC 4492
1637 ciphersuites, most of these by analogy with the DH case:
1639 kECDHr - ECDH cert, signed with RSA
1640 kECDHe - ECDH cert, signed with ECDSA
1641 kECDH - ECDH cert (signed with either RSA or ECDSA)
1642 kEECDH - ephemeral ECDH
1643 ECDH - ECDH cert or ephemeral ECDH
1649 AECDH - anonymous ECDH
1650 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
1654 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
1655 Use correct micalg parameters depending on digest(s) in signed message.
1658 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
1659 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
1662 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
1663 an engine to register a method. Add ENGINE lookups for methods and
1664 functional reference processing.
1667 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
1668 EVP_{Sign,Verify}* which allow an application to customise the signature
1672 *) New -resign option to smime utility. This adds one or more signers
1673 to an existing PKCS#7 signedData structure. Also -md option to use an
1674 alternative message digest algorithm for signing.
1677 *) Tidy up PKCS#7 routines and add new functions to make it easier to
1678 create PKCS7 structures containing multiple signers. Update smime
1679 application to support multiple signers.
1682 *) New -macalg option to pkcs12 utility to allow setting of an alternative
1686 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
1687 Reorganize PBE internals to lookup from a static table using NIDs,
1688 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
1689 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
1690 PRF which will be automatically used with PBES2.
1693 *) Replace the algorithm specific calls to generate keys in "req" with the
1697 *) Update PKCS#7 enveloped data routines to use new API. This is now
1698 supported by any public key method supporting the encrypt operation. A
1699 ctrl is added to allow the public key algorithm to examine or modify
1700 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
1704 *) Add a ctrl to asn1 method to allow a public key algorithm to express
1705 a default digest type to use. In most cases this will be SHA1 but some
1706 algorithms (such as GOST) need to specify an alternative digest. The
1707 return value indicates how strong the prefernce is 1 means optional and
1708 2 is mandatory (that is it is the only supported type). Modify
1709 ASN1_item_sign() to accept a NULL digest argument to indicate it should
1710 use the default md. Update openssl utilities to use the default digest
1711 type for signing if it is not explicitly indicated.
1714 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
1715 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
1716 signing method from the key type. This effectively removes the link
1717 between digests and public key types.
1720 *) Add an OID cross reference table and utility functions. Its purpose is to
1721 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
1722 rsaEncryption. This will allow some of the algorithm specific hackery
1723 needed to use the correct OID to be removed.
1726 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
1727 structures for PKCS7_sign(). They are now set up by the relevant public
1731 *) Add provisional EC pkey method with support for ECDSA and ECDH.
1734 *) Add support for key derivation (agreement) in the API, DH method and
1738 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
1739 public and private key formats. As a side effect these add additional
1740 command line functionality not previously available: DSA signatures can be
1741 generated and verified using pkeyutl and DH key support and generation in
1746 [Oliver Tappe <zooey@hirschkaefer.de>]
1748 *) New make target "install_html_docs" installs HTML renditions of the
1750 [Oliver Tappe <zooey@hirschkaefer.de>]
1752 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
1753 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
1754 support key and parameter generation and add initial key generation
1755 functionality for RSA.
1758 *) Add functions for main EVP_PKEY_method operations. The undocumented
1759 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
1760 EVP_PKEY_{encrypt,decrypt}_old.
1763 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
1764 key API, doesn't do much yet.
1767 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
1768 public key algorithms. New option to openssl utility:
1769 "list-public-key-algorithms" to print out info.
1772 *) Implement the Supported Elliptic Curves Extension for
1773 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1776 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
1777 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
1780 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
1781 utilities such as rsa, dsa, dsaparam etc except they process any key
1785 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1786 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1787 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
1791 *) Initial support for pluggable public key ASN1.
1792 De-spaghettify the public key ASN1 handling. Move public and private
1793 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
1794 algorithm specific handling to a single module within the relevant
1795 algorithm directory. Add functions to allow (near) opaque processing
1796 of public and private key structures.
1799 *) Implement the Supported Point Formats Extension for
1800 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1803 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
1804 for the psk identity [hint] and the psk callback functions to the
1805 SSL_SESSION, SSL and SSL_CTX structure.
1808 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
1812 SSL_CTX_use_psk_identity_hint
1813 SSL_get_psk_identity_hint
1814 SSL_get_psk_identity
1815 SSL_use_psk_identity_hint
1817 [Mika Kousa and Pasi Eronen of Nokia Corporation]
1819 *) Add RFC 3161 compliant time stamp request creation, response generation
1820 and response verification functionality.
1821 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
1823 *) Add initial support for TLS extensions, specifically for the server_name
1824 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1825 have new members for a host name. The SSL data structure has an
1826 additional member SSL_CTX *initial_ctx so that new sessions can be
1827 stored in that context to allow for session resumption, even after the
1828 SSL has been switched to a new SSL_CTX in reaction to a client's
1829 server_name extension.
1831 New functions (subject to change):
1833 SSL_get_servername()
1834 SSL_get_servername_type()
1837 New CTRL codes and macros (subject to change):
1839 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1840 - SSL_CTX_set_tlsext_servername_callback()
1841 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1842 - SSL_CTX_set_tlsext_servername_arg()
1843 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1845 openssl s_client has a new '-servername ...' option.
1847 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1848 '-key2 ...', '-servername_fatal' (subject to change). This allows
1849 testing the HostName extension for a specific single host name ('-cert'
1850 and '-key' remain fallbacks for handshakes without HostName
1851 negotiation). If the unrecogninzed_name alert has to be sent, this by
1852 default is a warning; it becomes fatal with the '-servername_fatal'
1855 [Peter Sylvester, Remy Allais, Christophe Renou]
1857 *) Whirlpool hash implementation is added.
1860 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1861 bn(64,32). Because of instruction set limitations it doesn't have
1862 any negative impact on performance. This was done mostly in order
1863 to make it possible to share assembler modules, such as bn_mul_mont
1864 implementations, between 32- and 64-bit builds without hassle.
1867 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1868 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1872 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1873 dedicated Montgomery multiplication procedure, is introduced.
1874 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1875 "64-bit" performance on certain 32-bit targets.
1878 *) New option SSL_OP_NO_COMP to disable use of compression selectively
1879 in SSL structures. New SSL ctrl to set maximum send fragment size.
1880 Save memory by seeting the I/O buffer sizes dynamically instead of
1881 using the maximum available value.
1884 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
1885 in addition to the text details.
1888 *) Very, very preliminary EXPERIMENTAL support for printing of general
1889 ASN1 structures. This currently produces rather ugly output and doesn't
1890 handle several customised structures at all.
1893 *) Integrated support for PVK file format and some related formats such
1894 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
1895 these in the 'rsa' and 'dsa' utilities.
1898 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
1901 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
1902 place for the (very old) "NETSCAPE" format certificates which are now
1903 handled using new ASN1 code equivalents.
1906 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
1907 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
1908 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
1911 *) Modify CRL distribution points extension code to print out previously
1912 unsupported fields. Enhance extension setting code to allow setting of
1916 *) Add print and set support for Issuing Distribution Point CRL extension.
1919 *) Change 'Configure' script to enable Camellia by default.
1922 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
1924 *) When rejecting SSL/TLS records due to an incorrect version number, never
1925 update s->server with a new major version number. As of
1926 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
1927 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
1928 the previous behavior could result in a read attempt at NULL when
1929 receiving specific incorrect SSL/TLS records once record payload
1930 protection is active. (CVE-2010-0740)
1931 [Bodo Moeller, Adam Langley <agl@chromium.org>]
1933 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
1934 could be crashed if the relevant tables were not present (e.g. chrooted).
1935 [Tomas Hoger <thoger@redhat.com>]
1937 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
1939 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
1940 [Martin Olsson, Neel Mehta]
1942 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
1943 accommodate for stack sorting, always a write lock!).
1946 *) On some versions of WIN32 Heap32Next is very slow. This can cause
1947 excessive delays in the RAND_poll(): over a minute. As a workaround
1948 include a time check in the inner Heap32Next loop too.
1951 *) The code that handled flushing of data in SSL/TLS originally used the
1952 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
1953 the problem outlined in PR#1949. The fix suggested there however can
1954 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
1955 of Apache). So instead simplify the code to flush unconditionally.
1956 This should be fine since flushing with no data to flush is a no op.
1959 *) Handle TLS versions 2.0 and later properly and correctly use the
1960 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
1961 off ancient servers have a habit of sticking around for a while...
1964 *) Modify compression code so it frees up structures without using the
1965 ex_data callbacks. This works around a problem where some applications
1966 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
1967 restarting) then use compression (e.g. SSL with compression) later.
1968 This results in significant per-connection memory leaks and
1969 has caused some security issues including CVE-2008-1678 and
1973 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
1974 change when encrypting or decrypting.
1977 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
1978 connect and renegotiate with servers which do not support RI.
1979 Until RI is more widely deployed this option is enabled by default.
1982 *) Add "missing" ssl ctrls to clear options and mode.
1985 *) If client attempts to renegotiate and doesn't support RI respond with
1986 a no_renegotiation alert as required by RFC5746. Some renegotiating
1987 TLS clients will continue a connection gracefully when they receive
1988 the alert. Unfortunately OpenSSL mishandled this alert and would hang
1989 waiting for a server hello which it will never receive. Now we treat a
1990 received no_renegotiation alert as a fatal error. This is because
1991 applications requesting a renegotiation might well expect it to succeed
1992 and would have no code in place to handle the server denying it so the
1993 only safe thing to do is to terminate the connection.
1996 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
1997 peer supports secure renegotiation and 0 otherwise. Print out peer
1998 renegotiation support in s_client/s_server.
2001 *) Replace the highly broken and deprecated SPKAC certification method with
2002 the updated NID creation version. This should correctly handle UTF8.
2005 *) Implement RFC5746. Re-enable renegotiation but require the extension
2006 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
2007 turns out to be a bad idea. It has been replaced by
2008 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
2009 SSL_CTX_set_options(). This is really not recommended unless you
2010 know what you are doing.
2011 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
2013 *) Fixes to stateless session resumption handling. Use initial_ctx when
2014 issuing and attempting to decrypt tickets in case it has changed during
2015 servername handling. Use a non-zero length session ID when attempting
2016 stateless session resumption: this makes it possible to determine if
2017 a resumption has occurred immediately after receiving server hello
2018 (several places in OpenSSL subtly assume this) instead of later in
2022 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
2023 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
2024 fixes for a few places where the return code is not checked
2026 [Julia Lawall <julia@diku.dk>]
2028 *) Add --strict-warnings option to Configure script to include devteam
2029 warnings in other configurations.
2032 *) Add support for --libdir option and LIBDIR variable in makefiles. This
2033 makes it possible to install openssl libraries in locations which
2034 have names other than "lib", for example "/usr/lib64" which some
2036 [Steve Henson, based on patch from Jeremy Utley]
2038 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
2039 X690 8.9.12 and can produce some misleading textual output of OIDs.
2040 [Steve Henson, reported by Dan Kaminsky]
2042 *) Delete MD2 from algorithm tables. This follows the recommendation in
2043 several standards that it is not used in new applications due to
2044 several cryptographic weaknesses. For binary compatibility reasons
2045 the MD2 API is still compiled in by default.
2048 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
2052 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
2053 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
2055 [Guenter <lists@gknw.net>]
2057 *) Fix the server certificate chain building code to use X509_verify_cert(),
2058 it used to have an ad-hoc builder which was unable to cope with anything
2059 other than a simple chain.
2060 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
2062 *) Don't check self signed certificate signatures in X509_verify_cert()
2063 by default (a flag can override this): it just wastes time without
2064 adding any security. As a useful side effect self signed root CAs
2065 with non-FIPS digests are now usable in FIPS mode.
2068 *) In dtls1_process_out_of_seq_message() the check if the current message
2069 is already buffered was missing. For every new message was memory
2070 allocated, allowing an attacker to perform an denial of service attack
2071 with sending out of seq handshake messages until there is no memory
2072 left. Additionally every future messege was buffered, even if the
2073 sequence number made no sense and would be part of another handshake.
2074 So only messages with sequence numbers less than 10 in advance will be
2075 buffered. (CVE-2009-1378)
2076 [Robin Seggelmann, discovered by Daniel Mentz]
2078 *) Records are buffered if they arrive with a future epoch to be
2079 processed after finishing the corresponding handshake. There is
2080 currently no limitation to this buffer allowing an attacker to perform
2081 a DOS attack with sending records with future epochs until there is no
2082 memory left. This patch adds the pqueue_size() function to detemine
2083 the size of a buffer and limits the record buffer to 100 entries.
2085 [Robin Seggelmann, discovered by Daniel Mentz]
2087 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
2088 parent structure is freed. (CVE-2009-1379)
2091 *) Handle non-blocking I/O properly in SSL_shutdown() call.
2092 [Darryl Miles <darryl-mailinglists@netbauds.net>]
2095 [Ilya O. <vrghost@gmail.com>]
2097 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
2099 *) Disable renegotiation completely - this fixes a severe security
2100 problem (CVE-2009-3555) at the cost of breaking all
2101 renegotiation. Renegotiation can be re-enabled by setting
2102 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
2103 run-time. This is really not recommended unless you know what
2107 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
2109 *) Don't set val to NULL when freeing up structures, it is freed up by
2110 underlying code. If sizeof(void *) > sizeof(long) this can result in
2111 zeroing past the valid field. (CVE-2009-0789)
2112 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
2114 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
2115 checked correctly. This would allow some invalid signed attributes to
2116 appear to verify correctly. (CVE-2009-0591)
2117 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2119 *) Reject UniversalString and BMPString types with invalid lengths. This
2120 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
2121 a legal length. (CVE-2009-0590)
2124 *) Set S/MIME signing as the default purpose rather than setting it
2125 unconditionally. This allows applications to override it at the store
2129 *) Permit restricted recursion of ASN1 strings. This is needed in practice
2130 to handle some structures.
2133 *) Improve efficiency of mem_gets: don't search whole buffer each time
2135 [Jeremy Shapiro <jnshapir@us.ibm.com>]
2137 *) New -hex option for openssl rand.
2140 *) Print out UTF8String and NumericString when parsing ASN1.
2143 *) Support NumericString type for name components.
2146 *) Allow CC in the environment to override the automatically chosen
2147 compiler. Note that nothing is done to ensure flags work with the
2151 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
2153 *) Properly check EVP_VerifyFinal() and similar return values
2155 [Ben Laurie, Bodo Moeller, Google Security Team]
2157 *) Enable TLS extensions by default.
2160 *) Allow the CHIL engine to be loaded, whether the application is
2161 multithreaded or not. (This does not release the developer from the
2162 obligation to set up the dynamic locking callbacks.)
2163 [Sander Temme <sander@temme.net>]
2165 *) Use correct exit code if there is an error in dgst command.
2166 [Steve Henson; problem pointed out by Roland Dirlewanger]
2168 *) Tweak Configure so that you need to say "experimental-jpake" to enable
2169 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
2172 *) Add experimental JPAKE support, including demo authentication in
2173 s_client and s_server.
2176 *) Set the comparison function in v3_addr_canonize().
2177 [Rob Austein <sra@hactrn.net>]
2179 *) Add support for XMPP STARTTLS in s_client.
2180 [Philip Paeps <philip@freebsd.org>]
2182 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
2183 to ensure that even with this option, only ciphersuites in the
2184 server's preference list will be accepted. (Note that the option
2185 applies only when resuming a session, so the earlier behavior was
2186 just about the algorithm choice for symmetric cryptography.)
2189 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
2191 *) Fix NULL pointer dereference if a DTLS server received
2192 ChangeCipherSpec as first record (CVE-2009-1386).
2195 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
2196 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
2199 *) The fix in 0.9.8c that supposedly got rid of unsafe
2200 double-checked locking was incomplete for RSA blinding,
2201 addressing just one layer of what turns out to have been
2202 doubly unsafe triple-checked locking.
2204 So now fix this for real by retiring the MONT_HELPER macro
2205 in crypto/rsa/rsa_eay.c.
2207 [Bodo Moeller; problem pointed out by Marius Schilder]
2209 *) Various precautionary measures:
2211 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
2213 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
2214 (NB: This would require knowledge of the secret session ticket key
2215 to exploit, in which case you'd be SOL either way.)
2217 - Change bn_nist.c so that it will properly handle input BIGNUMs
2218 outside the expected range.
2220 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
2223 [Neel Mehta, Bodo Moeller]
2225 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
2226 the load fails. Useful for distros.
2227 [Ben Laurie and the FreeBSD team]
2229 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
2232 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
2235 *) Expand ENGINE to support engine supplied SSL client certificate functions.
2237 This work was sponsored by Logica.
2240 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
2241 keystores. Support for SSL/TLS client authentication too.
2242 Not compiled unless enable-capieng specified to Configure.
2244 This work was sponsored by Logica.
2247 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
2248 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
2249 attribute creation routines such as certifcate requests and PKCS#12
2253 Changes between 0.9.8g and 0.9.8h [28 May 2008]
2255 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
2256 handshake which could lead to a cilent crash as found using the
2257 Codenomicon TLS test suite (CVE-2008-1672)
2258 [Steve Henson, Mark Cox]
2260 *) Fix double free in TLS server name extensions which could lead to
2261 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
2264 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
2266 Clear the error queue to ensure that error entries left from
2267 older function calls do not interfere with the correct operation.
2268 [Lutz Jaenicke, Erik de Castro Lopo]
2270 *) Remove root CA certificates of commercial CAs:
2272 The OpenSSL project does not recommend any specific CA and does not
2273 have any policy with respect to including or excluding any CA.
2274 Therefore it does not make any sense to ship an arbitrary selection
2275 of root CA certificates with the OpenSSL software.
2278 *) RSA OAEP patches to fix two separate invalid memory reads.
2279 The first one involves inputs when 'lzero' is greater than
2280 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
2281 before the beginning of from). The second one involves inputs where
2282 the 'db' section contains nothing but zeroes (there is a one-byte
2283 invalid read after the end of 'db').
2284 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2286 *) Partial backport from 0.9.9-dev:
2288 Introduce bn_mul_mont (dedicated Montgomery multiplication
2289 procedure) as a candidate for BIGNUM assembler implementation.
2290 While 0.9.9-dev uses assembler for various architectures, only
2291 x86_64 is available by default here in the 0.9.8 branch, and
2292 32-bit x86 is available through a compile-time setting.
2294 To try the 32-bit x86 assembler implementation, use Configure
2295 option "enable-montasm" (which exists only for this backport).
2297 As "enable-montasm" for 32-bit x86 disclaims code stability
2298 anyway, in this constellation we activate additional code
2299 backported from 0.9.9-dev for further performance improvements,
2300 namely BN_from_montgomery_word. (To enable this otherwise,
2301 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
2303 [Andy Polyakov (backport partially by Bodo Moeller)]
2305 *) Add TLS session ticket callback. This allows an application to set
2306 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
2307 values. This is useful for key rollover for example where several key
2308 sets may exist with different names.
2311 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
2312 This was broken until now in 0.9.8 releases, such that the only way
2313 a registered ENGINE could be used (assuming it initialises
2314 successfully on the host) was to explicitly set it as the default
2315 for the relevant algorithms. This is in contradiction with 0.9.7
2316 behaviour and the documentation. With this fix, when an ENGINE is
2317 registered into a given algorithm's table of implementations, the
2318 'uptodate' flag is reset so that auto-discovery will be used next
2319 time a new context for that algorithm attempts to select an
2321 [Ian Lister (tweaked by Geoff Thorpe)]
2323 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
2324 implemention in the following ways:
2326 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
2329 Lack of BER streaming support means one pass streaming processing is
2330 only supported if data is detached: setting the streaming flag is
2331 ignored for embedded content.
2333 CMS support is disabled by default and must be explicitly enabled
2334 with the enable-cms configuration option.
2337 *) Update the GMP engine glue to do direct copies between BIGNUM and
2338 mpz_t when openssl and GMP use the same limb size. Otherwise the
2339 existing "conversion via a text string export" trick is still used.
2340 [Paul Sheer <paulsheer@gmail.com>]
2342 *) Zlib compression BIO. This is a filter BIO which compressed and
2343 uncompresses any data passed through it.
2346 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
2347 RFC3394 compatible AES key wrapping.
2350 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
2351 sets string data without copying. X509_ALGOR_set0() and
2352 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
2353 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
2354 from an X509_ATTRIBUTE structure optionally checking it occurs only
2355 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
2359 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
2360 to get the expected BN_FLG_CONSTTIME behavior.
2361 [Bodo Moeller (Google)]
2365 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
2366 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
2367 - added some more tests to do_tests.pl
2368 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
2369 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
2370 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
2371 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
2372 - various changes to netware.pl to enable gcc-cross builds on Win32
2374 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
2375 - various changes to fix missing prototype warnings
2376 - fixed x86nasm.pl to create correct asm files for NASM COFF output
2377 - added AES, WHIRLPOOL and CPUID assembler code to build files
2378 - added missing AES assembler make rules to mk1mf.pl
2379 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
2380 [Guenter Knauf <eflash@gmx.net>]
2382 *) Implement certificate status request TLS extension defined in RFC3546.
2383 A client can set the appropriate parameters and receive the encoded
2384 OCSP response via a callback. A server can query the supplied parameters
2385 and set the encoded OCSP response in the callback. Add simplified examples
2386 to s_client and s_server.
2389 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
2391 *) Fix various bugs:
2392 + Binary incompatibility of ssl_ctx_st structure
2393 + DTLS interoperation with non-compliant servers
2394 + Don't call get_session_cb() without proposed session
2395 + Fix ia64 assembler code
2396 [Andy Polyakov, Steve Henson]
2398 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
2400 *) DTLS Handshake overhaul. There were longstanding issues with
2401 OpenSSL DTLS implementation, which were making it impossible for
2402 RFC 4347 compliant client to communicate with OpenSSL server.
2403 Unfortunately just fixing these incompatibilities would "cut off"
2404 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
2405 server keeps tolerating non RFC compliant syntax. The opposite is
2406 not true, 0.9.8f client can not communicate with earlier server.
2407 This update even addresses CVE-2007-4995.
2410 *) Changes to avoid need for function casts in OpenSSL: some compilers
2411 (gcc 4.2 and later) reject their use.
2412 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
2415 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2416 RFC4507bis. The encrypted ticket format is an encrypted encoded
2417 SSL_SESSION structure, that way new session features are automatically
2420 If a client application caches session in an SSL_SESSION structure
2421 support is transparent because tickets are now stored in the encoded
2424 The SSL_CTX structure automatically generates keys for ticket
2425 protection in servers so again support should be possible
2426 with no application modification.
2428 If a client or server wishes to disable RFC4507 support then the option
2429 SSL_OP_NO_TICKET can be set.
2431 Add a TLS extension debugging callback to allow the contents of any client
2432 or server extensions to be examined.
2434 This work was sponsored by Google.
2437 *) Add initial support for TLS extensions, specifically for the server_name
2438 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2439 have new members for a host name. The SSL data structure has an
2440 additional member SSL_CTX *initial_ctx so that new sessions can be
2441 stored in that context to allow for session resumption, even after the
2442 SSL has been switched to a new SSL_CTX in reaction to a client's
2443 server_name extension.
2445 New functions (subject to change):
2447 SSL_get_servername()
2448 SSL_get_servername_type()
2451 New CTRL codes and macros (subject to change):
2453 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2454 - SSL_CTX_set_tlsext_servername_callback()
2455 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2456 - SSL_CTX_set_tlsext_servername_arg()
2457 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2459 openssl s_client has a new '-servername ...' option.
2461 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2462 '-key2 ...', '-servername_fatal' (subject to change). This allows
2463 testing the HostName extension for a specific single host name ('-cert'
2464 and '-key' remain fallbacks for handshakes without HostName
2465 negotiation). If the unrecogninzed_name alert has to be sent, this by
2466 default is a warning; it becomes fatal with the '-servername_fatal'
2469 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
2471 *) Add AES and SSE2 assembly language support to VC++ build.
2474 *) Mitigate attack on final subtraction in Montgomery reduction.
2477 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
2478 (which previously caused an internal error).
2481 *) Squeeze another 10% out of IGE mode when in != out.
2484 *) AES IGE mode speedup.
2485 [Dean Gaudet (Google)]
2487 *) Add the Korean symmetric 128-bit cipher SEED (see
2488 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
2489 add SEED ciphersuites from RFC 4162:
2491 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
2492 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
2493 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
2494 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
2496 To minimize changes between patchlevels in the OpenSSL 0.9.8
2497 series, SEED remains excluded from compilation unless OpenSSL
2498 is configured with 'enable-seed'.
2499 [KISA, Bodo Moeller]
2501 *) Mitigate branch prediction attacks, which can be practical if a
2502 single processor is shared, allowing a spy process to extract
2503 information. For detailed background information, see
2504 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
2505 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
2506 and Necessary Software Countermeasures"). The core of the change
2507 are new versions BN_div_no_branch() and
2508 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
2509 respectively, which are slower, but avoid the security-relevant
2510 conditional branches. These are automatically called by BN_div()
2511 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
2512 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
2513 remove a conditional branch.
2515 BN_FLG_CONSTTIME is the new name for the previous
2516 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
2517 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
2518 in the exponent causes BN_mod_exp_mont() to use the alternative
2519 implementation in BN_mod_exp_mont_consttime().) The old name
2520 remains as a deprecated alias.
2522 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
2523 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
2524 constant-time implementations for more than just exponentiation.
2525 Here too the old name is kept as a deprecated alias.
2527 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2528 the BN_BLINDING structure gets an independent copy of the
2529 modulus. This means that the previous "BIGNUM *m" argument to
2530 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2531 essentially becomes "const BIGNUM *m", although we can't actually
2532 change this in the header file before 0.9.9. It allows
2533 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2534 enable BN_FLG_CONSTTIME.
2536 [Matthew D Wood (Intel Corp)]
2538 *) In the SSL/TLS server implementation, be strict about session ID
2539 context matching (which matters if an application uses a single
2540 external cache for different purposes). Previously,
2541 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2542 set. This did ensure strict client verification, but meant that,
2543 with applications using a single external cache for quite
2544 different requirements, clients could circumvent ciphersuite
2545 restrictions for a given session ID context by starting a session
2546 in a different context.
2549 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2550 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2551 authentication-only ciphersuites.
2554 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
2555 not complete and could lead to a possible single byte overflow
2556 (CVE-2007-5135) [Ben Laurie]
2558 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
2560 *) Since AES128 and AES256 (and similarly Camellia128 and
2561 Camellia256) share a single mask bit in the logic of
2562 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2563 kludge to work properly if AES128 is available and AES256 isn't
2564 (or if Camellia128 is available and Camellia256 isn't).
2567 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
2568 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
2569 When a point or a seed is encoded in a BIT STRING, we need to
2570 prevent the removal of trailing zero bits to get the proper DER
2571 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
2572 of a NamedBitList, for which trailing 0 bits need to be removed.)
2575 *) Have SSL/TLS server implementation tolerate "mismatched" record
2576 protocol version while receiving ClientHello even if the
2577 ClientHello is fragmented. (The server can't insist on the
2578 particular protocol version it has chosen before the ServerHello
2579 message has informed the client about his choice.)
2582 *) Add RFC 3779 support.
2583 [Rob Austein for ARIN, Ben Laurie]
2585 *) Load error codes if they are not already present instead of using a
2586 static variable. This allows them to be cleanly unloaded and reloaded.
2587 Improve header file function name parsing.
2590 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
2591 or CAPABILITY handshake as required by RFCs.
2594 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
2596 *) Introduce limits to prevent malicious keys being able to
2597 cause a denial of service. (CVE-2006-2940)
2598 [Steve Henson, Bodo Moeller]
2600 *) Fix ASN.1 parsing of certain invalid structures that can result
2601 in a denial of service. (CVE-2006-2937) [Steve Henson]
2603 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2604 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2606 *) Fix SSL client code which could crash if connecting to a
2607 malicious SSLv2 server. (CVE-2006-4343)
2608 [Tavis Ormandy and Will Drewry, Google Security Team]
2610 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
2611 match only those. Before that, "AES256-SHA" would be interpreted
2612 as a pattern and match "AES128-SHA" too (since AES128-SHA got
2613 the same strength classification in 0.9.7h) as we currently only
2614 have a single AES bit in the ciphersuite description bitmap.
2615 That change, however, also applied to ciphersuite strings such as
2616 "RC4-MD5" that intentionally matched multiple ciphersuites --
2617 namely, SSL 2.0 ciphersuites in addition to the more common ones
2618 from SSL 3.0/TLS 1.0.
2620 So we change the selection algorithm again: Naming an explicit
2621 ciphersuite selects this one ciphersuite, and any other similar
2622 ciphersuite (same bitmap) from *other* protocol versions.
2623 Thus, "RC4-MD5" again will properly select both the SSL 2.0
2624 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
2626 Since SSL 2.0 does not have any ciphersuites for which the
2627 128/256 bit distinction would be relevant, this works for now.
2628 The proper fix will be to use different bits for AES128 and
2629 AES256, which would have avoided the problems from the beginning;
2630 however, bits are scarce, so we can only do this in a new release
2631 (not just a patchlevel) when we can change the SSL_CIPHER
2632 definition to split the single 'unsigned long mask' bitmap into
2633 multiple values to extend the available space.
2637 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2639 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2640 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2642 *) Add AES IGE and biIGE modes.
2645 *) Change the Unix randomness entropy gathering to use poll() when
2646 possible instead of select(), since the latter has some
2647 undesirable limitations.
2648 [Darryl Miles via Richard Levitte and Bodo Moeller]
2650 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2651 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2652 cannot be implicitly activated as part of, e.g., the "AES" alias.
2653 However, please upgrade to OpenSSL 0.9.9[-dev] for
2654 non-experimental use of the ECC ciphersuites to get TLS extension
2655 support, which is required for curve and point format negotiation
2656 to avoid potential handshake problems.
2659 *) Disable rogue ciphersuites:
2661 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2662 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2663 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2665 The latter two were purportedly from
2666 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2669 Also deactivate the remaining ciphersuites from
2670 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2671 unofficial, and the ID has long expired.
2674 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2675 dual-core machines) and other potential thread-safety issues.
2678 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
2679 versions), which is now available for royalty-free use
2680 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
2681 Also, add Camellia TLS ciphersuites from RFC 4132.
2683 To minimize changes between patchlevels in the OpenSSL 0.9.8
2684 series, Camellia remains excluded from compilation unless OpenSSL
2685 is configured with 'enable-camellia'.
2688 *) Disable the padding bug check when compression is in use. The padding
2689 bug check assumes the first packet is of even length, this is not
2690 necessarily true if compresssion is enabled and can result in false
2691 positives causing handshake failure. The actual bug test is ancient
2692 code so it is hoped that implementations will either have fixed it by
2693 now or any which still have the bug do not support compression.
2696 Changes between 0.9.8a and 0.9.8b [04 May 2006]
2698 *) When applying a cipher rule check to see if string match is an explicit
2699 cipher suite and only match that one cipher suite if it is.
2702 *) Link in manifests for VC++ if needed.
2703 [Austin Ziegler <halostatue@gmail.com>]
2705 *) Update support for ECC-based TLS ciphersuites according to
2706 draft-ietf-tls-ecc-12.txt with proposed changes (but without
2707 TLS extensions, which are supported starting with the 0.9.9
2708 branch, not in the OpenSSL 0.9.8 branch).
2711 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
2712 opaque EVP_CIPHER_CTX handling.
2715 *) Fixes and enhancements to zlib compression code. We now only use
2716 "zlib1.dll" and use the default __cdecl calling convention on Win32
2717 to conform with the standards mentioned here:
2718 http://www.zlib.net/DLL_FAQ.txt
2719 Static zlib linking now works on Windows and the new --with-zlib-include
2720 --with-zlib-lib options to Configure can be used to supply the location
2721 of the headers and library. Gracefully handle case where zlib library
2725 *) Several fixes and enhancements to the OID generation code. The old code
2726 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
2727 handle numbers larger than ULONG_MAX, truncated printing and had a
2728 non standard OBJ_obj2txt() behaviour.
2731 *) Add support for building of engines under engine/ as shared libraries
2732 under VC++ build system.
2735 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
2736 Hopefully, we will not see any false combination of paths any more.
2739 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
2741 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2742 (part of SSL_OP_ALL). This option used to disable the
2743 countermeasure against man-in-the-middle protocol-version
2744 rollback in the SSL 2.0 server implementation, which is a bad
2745 idea. (CVE-2005-2969)
2747 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2748 for Information Security, National Institute of Advanced Industrial
2749 Science and Technology [AIST], Japan)]
2751 *) Add two function to clear and return the verify parameter flags.
2754 *) Keep cipherlists sorted in the source instead of sorting them at
2755 runtime, thus removing the need for a lock.
2758 *) Avoid some small subgroup attacks in Diffie-Hellman.
2759 [Nick Mathewson and Ben Laurie]
2761 *) Add functions for well-known primes.
2764 *) Extended Windows CE support.
2765 [Satoshi Nakamura and Andy Polyakov]
2767 *) Initialize SSL_METHOD structures at compile time instead of during
2768 runtime, thus removing the need for a lock.
2771 *) Make PKCS7_decrypt() work even if no certificate is supplied by
2772 attempting to decrypt each encrypted key in turn. Add support to
2776 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
2778 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
2781 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
2784 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
2785 key into the same file any more.
2788 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
2791 *) Add -utf8 command line and config file option to 'ca'.
2792 [Stefan <stf@udoma.org]
2794 *) Removed the macro des_crypt(), as it seems to conflict with some
2795 libraries. Use DES_crypt().
2798 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
2799 involves renaming the source and generated shared-libs for
2800 both. The engines will accept the corrected or legacy ids
2801 ('ncipher' and '4758_cca' respectively) when binding. NB,
2802 this only applies when building 'shared'.
2803 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
2805 *) Add attribute functions to EVP_PKEY structure. Modify
2806 PKCS12_create() to recognize a CSP name attribute and
2807 use it. Make -CSP option work again in pkcs12 utility.
2810 *) Add new functionality to the bn blinding code:
2811 - automatic re-creation of the BN_BLINDING parameters after
2812 a fixed number of uses (currently 32)
2813 - add new function for parameter creation
2814 - introduce flags to control the update behaviour of the
2815 BN_BLINDING parameters
2816 - hide BN_BLINDING structure
2817 Add a second BN_BLINDING slot to the RSA structure to improve
2818 performance when a single RSA object is shared among several
2822 *) Add support for DTLS.
2823 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
2825 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
2826 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
2829 *) Remove buggy and incompletet DH cert support from
2830 ssl/ssl_rsa.c and ssl/s3_both.c
2833 *) Use SHA-1 instead of MD5 as the default digest algorithm for
2834 the apps/openssl applications.
2837 *) Compile clean with "-Wall -Wmissing-prototypes
2838 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
2839 DEBUG_SAFESTACK must also be set.
2842 *) Change ./Configure so that certain algorithms can be disabled by default.
2843 The new counterpiece to "no-xxx" is "enable-xxx".
2845 The patented RC5 and MDC2 algorithms will now be disabled unless
2846 "enable-rc5" and "enable-mdc2", respectively, are specified.
2848 (IDEA remains enabled despite being patented. This is because IDEA
2849 is frequently required for interoperability, and there is no license
2850 fee for non-commercial use. As before, "no-idea" can be used to
2851 avoid this algorithm.)
2855 *) Add processing of proxy certificates (see RFC 3820). This work was
2856 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
2857 EGEE (Enabling Grids for E-science in Europe).
2860 *) RC4 performance overhaul on modern architectures/implementations, such
2861 as Intel P4, IA-64 and AMD64.
2864 *) New utility extract-section.pl. This can be used specify an alternative
2865 section number in a pod file instead of having to treat each file as
2866 a separate case in Makefile. This can be done by adding two lines to the
2869 =for comment openssl_section:XXX
2871 The blank line is mandatory.
2875 *) New arguments -certform, -keyform and -pass for s_client and s_server
2876 to allow alternative format key and certificate files and passphrase
2880 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
2881 update associated structures and add various utility functions.
2883 Add new policy related verify parameters, include policy checking in
2884 standard verify code. Enhance 'smime' application with extra parameters
2885 to support policy checking and print out.
2888 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
2889 Nehemiah processors. These extensions support AES encryption in hardware
2890 as well as RNG (though RNG support is currently disabled).
2891 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
2893 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
2896 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
2897 [Andy Polyakov and a number of other people]
2899 *) Improved PowerPC platform support. Most notably BIGNUM assembler
2900 implementation contributed by IBM.
2901 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
2903 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
2904 exponent rather than 'unsigned long'. There is a corresponding change to
2905 the new 'rsa_keygen' element of the RSA_METHOD structure.
2906 [Jelte Jansen, Geoff Thorpe]
2908 *) Functionality for creating the initial serial number file is now
2909 moved from CA.pl to the 'ca' utility with a new option -create_serial.
2911 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
2912 number file to 1, which is bound to cause problems. To avoid
2913 the problems while respecting compatibility between different 0.9.7
2914 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
2915 CA.pl for serial number initialization. With the new release 0.9.8,
2916 we can fix the problem directly in the 'ca' utility.)
2919 *) Reduced header interdepencies by declaring more opaque objects in
2920 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
2921 give fewer recursive includes, which could break lazy source code - so
2922 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
2923 developers should define this symbol when building and using openssl to
2924 ensure they track the recommended behaviour, interfaces, [etc], but
2925 backwards-compatible behaviour prevails when this isn't defined.
2928 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
2931 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
2932 This will generate a random key of the appropriate length based on the
2933 cipher context. The EVP_CIPHER can provide its own random key generation
2934 routine to support keys of a specific form. This is used in the des and
2935 3des routines to generate a key of the correct parity. Update S/MIME
2936 code to use new functions and hence generate correct parity DES keys.
2937 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
2938 valid (weak or incorrect parity).
2941 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
2942 as looking them up. This is useful when the verified structure may contain
2943 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
2944 present unless the new PKCS7_NO_CRL flag is asserted.
2947 *) Extend ASN1 oid configuration module. It now additionally accepts the
2950 shortName = some long name, 1.2.3.4
2953 *) Reimplemented the BN_CTX implementation. There is now no more static
2954 limitation on the number of variables it can handle nor the depth of the
2955 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
2956 information can now expand as required, and rather than having a single
2957 static array of bignums, BN_CTX now uses a linked-list of such arrays
2958 allowing it to expand on demand whilst maintaining the usefulness of
2959 BN_CTX's "bundling".
2962 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
2963 to allow all RSA operations to function using a single BN_CTX.
2966 *) Preliminary support for certificate policy evaluation and checking. This
2967 is initially intended to pass the tests outlined in "Conformance Testing
2968 of Relying Party Client Certificate Path Processing Logic" v1.07.
2971 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
2972 remained unused and not that useful. A variety of other little bignum
2973 tweaks and fixes have also been made continuing on from the audit (see
2977 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
2978 associated ASN1, EVP and SSL functions and old ASN1 macros.
2981 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
2982 and this should never fail. So the return value from the use of
2983 BN_set_word() (which can fail due to needless expansion) is now deprecated;
2984 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
2987 *) BN_CTX_get() should return zero-valued bignums, providing the same
2988 initialised value as BN_new().
2989 [Geoff Thorpe, suggested by Ulf Möller]
2991 *) Support for inhibitAnyPolicy certificate extension.
2994 *) An audit of the BIGNUM code is underway, for which debugging code is
2995 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
2996 is considered valid when processing BIGNUMs, and causes execution to
2997 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
2998 further steps are taken to deliberately pollute unused data in BIGNUM
2999 structures to try and expose faulty code further on. For now, openssl will
3000 (in its default mode of operation) continue to tolerate the inconsistent
3001 forms that it has tolerated in the past, but authors and packagers should
3002 consider trying openssl and their own applications when compiled with
3003 these debugging symbols defined. It will help highlight potential bugs in
3004 their own code, and will improve the test coverage for OpenSSL itself. At
3005 some point, these tighter rules will become openssl's default to improve
3006 maintainability, though the assert()s and other overheads will remain only
3007 in debugging configurations. See bn.h for more details.
3008 [Geoff Thorpe, Nils Larsch, Ulf Möller]
3010 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
3011 that can only be obtained through BN_CTX_new() (which implicitly
3012 initialises it). The presence of this function only made it possible
3013 to overwrite an existing structure (and cause memory leaks).
3016 *) Because of the callback-based approach for implementing LHASH as a
3017 template type, lh_insert() adds opaque objects to hash-tables and
3018 lh_doall() or lh_doall_arg() are typically used with a destructor callback
3019 to clean up those corresponding objects before destroying the hash table
3020 (and losing the object pointers). So some over-zealous constifications in
3021 LHASH have been relaxed so that lh_insert() does not take (nor store) the
3022 objects as "const" and the lh_doall[_arg] callback wrappers are not
3023 prototyped to have "const" restrictions on the object pointers they are
3024 given (and so aren't required to cast them away any more).
3027 *) The tmdiff.h API was so ugly and minimal that our own timing utility
3028 (speed) prefers to use its own implementation. The two implementations
3029 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
3030 its object type properly exposed (MS_TM) instead of casting to/from "char
3031 *". This may still change yet if someone realises MS_TM and "ms_time_***"
3032 aren't necessarily the greatest nomenclatures - but this is what was used
3033 internally to the implementation so I've used that for now.
3036 *) Ensure that deprecated functions do not get compiled when
3037 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
3038 the self-tests were still using deprecated key-generation functions so
3039 these have been updated also.
3042 *) Reorganise PKCS#7 code to separate the digest location functionality
3043 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
3044 New function PKCS7_set_digest() to set the digest type for PKCS#7
3045 digestedData type. Add additional code to correctly generate the
3046 digestedData type and add support for this type in PKCS7 initialization
3050 *) New function PKCS7_set0_type_other() this initializes a PKCS7
3051 structure of type "other".
3054 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
3055 sure the loop does correctly stop and breaking ("division by zero")
3056 modulus operations are not performed. The (pre-generated) prime
3057 table crypto/bn/bn_prime.h was already correct, but it could not be
3058 re-generated on some platforms because of the "division by zero"
3059 situation in the script.
3060 [Ralf S. Engelschall]
3062 *) Update support for ECC-based TLS ciphersuites according to
3063 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
3064 SHA-1 now is only used for "small" curves (where the
3065 representation of a field element takes up to 24 bytes); for
3066 larger curves, the field element resulting from ECDH is directly
3067 used as premaster secret.
3068 [Douglas Stebila (Sun Microsystems Laboratories)]
3070 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
3071 curve secp160r1 to the tests.
3072 [Douglas Stebila (Sun Microsystems Laboratories)]
3074 *) Add the possibility to load symbols globally with DSO.
3075 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
3077 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
3078 control of the error stack.
3081 *) Add support for STORE in ENGINE.
3084 *) Add the STORE type. The intention is to provide a common interface
3085 to certificate and key stores, be they simple file-based stores, or
3086 HSM-type store, or LDAP stores, or...
3087 NOTE: The code is currently UNTESTED and isn't really used anywhere.
3090 *) Add a generic structure called OPENSSL_ITEM. This can be used to
3091 pass a list of arguments to any function as well as provide a way
3092 for a function to pass data back to the caller.
3095 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
3096 works like BUF_strdup() but can be used to duplicate a portion of
3097 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
3101 *) Add the function sk_find_ex() which works like sk_find(), but will
3102 return an index to an element even if an exact match couldn't be
3103 found. The index is guaranteed to point at the element where the
3104 searched-for key would be inserted to preserve sorting order.
3107 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
3108 takes an extra flags argument for optional functionality. Currently,
3109 the following flags are defined:
3111 OBJ_BSEARCH_VALUE_ON_NOMATCH
3112 This one gets OBJ_bsearch_ex() to return a pointer to the first
3113 element where the comparing function returns a negative or zero
3116 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
3117 This one gets OBJ_bsearch_ex() to return a pointer to the first
3118 element where the comparing function returns zero. This is useful
3119 if there are more than one element where the comparing function
3123 *) Make it possible to create self-signed certificates with 'openssl ca'
3124 in such a way that the self-signed certificate becomes part of the
3125 CA database and uses the same mechanisms for serial number generation
3126 as all other certificate signing. The new flag '-selfsign' enables
3127 this functionality. Adapt CA.sh and CA.pl.in.
3130 *) Add functionality to check the public key of a certificate request
3131 against a given private. This is useful to check that a certificate
3132 request can be signed by that key (self-signing).
3135 *) Make it possible to have multiple active certificates with the same
3136 subject in the CA index file. This is done only if the keyword
3137 'unique_subject' is set to 'no' in the main CA section (default
3138 if 'CA_default') of the configuration file. The value is saved
3139 with the database itself in a separate index attribute file,
3140 named like the index file with '.attr' appended to the name.
3143 *) Generate muti valued AVAs using '+' notation in config files for
3147 *) Support for nameConstraints certificate extension.
3150 *) Support for policyConstraints certificate extension.
3153 *) Support for policyMappings certificate extension.
3156 *) Make sure the default DSA_METHOD implementation only uses its
3157 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
3158 and change its own handlers to be NULL so as to remove unnecessary
3159 indirection. This lets alternative implementations fallback to the
3160 default implementation more easily.
3163 *) Support for directoryName in GeneralName related extensions
3167 *) Make it possible to link applications using Makefile.shared.
3168 Make that possible even when linking against static libraries!
3171 *) Support for single pass processing for S/MIME signing. This now
3172 means that S/MIME signing can be done from a pipe, in addition
3173 cleartext signing (multipart/signed type) is effectively streaming
3174 and the signed data does not need to be all held in memory.
3176 This is done with a new flag PKCS7_STREAM. When this flag is set
3177 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
3178 is done after the data is output (and digests calculated) in
3179 SMIME_write_PKCS7().
3182 *) Add full support for -rpath/-R, both in shared libraries and
3183 applications, at least on the platforms where it's known how
3187 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
3188 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
3189 will now compute a table of multiples of the generator that
3190 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
3191 faster (notably in the case of a single point multiplication,
3192 scalar * generator).
3193 [Nils Larsch, Bodo Moeller]
3195 *) IPv6 support for certificate extensions. The various extensions
3196 which use the IP:a.b.c.d can now take IPv6 addresses using the
3197 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
3201 *) Added an ENGINE that implements RSA by performing private key
3202 exponentiations with the GMP library. The conversions to and from
3203 GMP's mpz_t format aren't optimised nor are any montgomery forms
3204 cached, and on x86 it appears OpenSSL's own performance has caught up.
3205 However there are likely to be other architectures where GMP could
3206 provide a boost. This ENGINE is not built in by default, but it can be
3207 specified at Configure time and should be accompanied by the necessary
3208 linker additions, eg;
3209 ./config -DOPENSSL_USE_GMP -lgmp
3212 *) "openssl engine" will not display ENGINE/DSO load failure errors when
3213 testing availability of engines with "-t" - the old behaviour is
3214 produced by increasing the feature's verbosity with "-tt".
3217 *) ECDSA routines: under certain error conditions uninitialized BN objects
3218 could be freed. Solution: make sure initialization is performed early
3219 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
3223 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
3224 and DH_METHOD (eg. by ENGINE implementations) to override the normal
3225 software implementations. For DSA and DH, parameter generation can
3226 also be overriden by providing the appropriate method callbacks.
3229 *) Change the "progress" mechanism used in key-generation and
3230 primality testing to functions that take a new BN_GENCB pointer in
3231 place of callback/argument pairs. The new API functions have "_ex"
3232 postfixes and the older functions are reimplemented as wrappers for
3233 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
3234 declarations of the old functions to help (graceful) attempts to
3235 migrate to the new functions. Also, the new key-generation API
3236 functions operate on a caller-supplied key-structure and return
3237 success/failure rather than returning a key or NULL - this is to
3238 help make "keygen" another member function of RSA_METHOD etc.
3240 Example for using the new callback interface:
3242 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
3246 BN_GENCB_set(&my_cb, my_callback, my_arg);
3248 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
3249 /* For the meaning of a, b in calls to my_callback(), see the
3250 * documentation of the function that calls the callback.
3251 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
3252 * my_callback should return 1 if it wants BN_is_prime_ex()
3253 * to continue, or 0 to stop.
3258 *) Change the ZLIB compression method to be stateful, and make it
3259 available to TLS with the number defined in
3260 draft-ietf-tls-compression-04.txt.
3263 *) Add the ASN.1 structures and functions for CertificatePair, which
3264 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
3266 CertificatePair ::= SEQUENCE {
3267 forward [0] Certificate OPTIONAL,
3268 reverse [1] Certificate OPTIONAL,
3269 -- at least one of the pair shall be present -- }
3271 Also implement the PEM functions to read and write certificate
3272 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
3274 This needed to be defined, mostly for the sake of the LDAP
3275 attribute crossCertificatePair, but may prove useful elsewhere as
3279 *) Make it possible to inhibit symlinking of shared libraries in
3280 Makefile.shared, for Cygwin's sake.
3283 *) Extend the BIGNUM API by creating a function
3284 void BN_set_negative(BIGNUM *a, int neg);
3285 and a macro that behave like
3286 int BN_is_negative(const BIGNUM *a);
3288 to avoid the need to access 'a->neg' directly in applications.
3291 *) Implement fast modular reduction for pseudo-Mersenne primes
3292 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
3293 EC_GROUP_new_curve_GFp() will now automatically use this
3295 [Nils Larsch <nla@trustcenter.de>]
3297 *) Add new lock type (CRYPTO_LOCK_BN).
3300 *) Change the ENGINE framework to automatically load engines
3301 dynamically from specific directories unless they could be
3302 found to already be built in or loaded. Move all the
3303 current engines except for the cryptodev one to a new
3305 The engines in engines/ are built as shared libraries if
3306 the "shared" options was given to ./Configure or ./config.
3307 Otherwise, they are inserted in libcrypto.a.
3308 /usr/local/ssl/engines is the default directory for dynamic
3309 engines, but that can be overriden at configure time through
3310 the usual use of --prefix and/or --openssldir, and at run
3311 time with the environment variable OPENSSL_ENGINES.
3312 [Geoff Thorpe and Richard Levitte]
3314 *) Add Makefile.shared, a helper makefile to build shared
3315 libraries. Addapt Makefile.org.
3318 *) Add version info to Win32 DLLs.
3319 [Peter 'Luna' Runestig" <peter@runestig.com>]
3321 *) Add new 'medium level' PKCS#12 API. Certificates and keys
3322 can be added using this API to created arbitrary PKCS#12
3323 files while avoiding the low level API.
3325 New options to PKCS12_create(), key or cert can be NULL and
3326 will then be omitted from the output file. The encryption
3327 algorithm NIDs can be set to -1 for no encryption, the mac
3328 iteration count can be set to 0 to omit the mac.
3330 Enhance pkcs12 utility by making the -nokeys and -nocerts
3331 options work when creating a PKCS#12 file. New option -nomac
3332 to omit the mac, NONE can be set for an encryption algorithm.
3333 New code is modified to use the enhanced PKCS12_create()
3334 instead of the low level API.
3337 *) Extend ASN1 encoder to support indefinite length constructed
3338 encoding. This can output sequences tags and octet strings in
3339 this form. Modify pk7_asn1.c to support indefinite length
3340 encoding. This is experimental and needs additional code to
3341 be useful, such as an ASN1 bio and some enhanced streaming
3344 Extend template encode functionality so that tagging is passed
3345 down to the template encoder.
3348 *) Let 'openssl req' fail if an argument to '-newkey' is not
3349 recognized instead of using RSA as a default.
3352 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
3353 As these are not official, they are not included in "ALL";
3354 the "ECCdraft" ciphersuite group alias can be used to select them.
3355 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
3357 *) Add ECDH engine support.
3358 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
3360 *) Add ECDH in new directory crypto/ecdh/.
3361 [Douglas Stebila (Sun Microsystems Laboratories)]
3363 *) Let BN_rand_range() abort with an error after 100 iterations
3364 without success (which indicates a broken PRNG).
3367 *) Change BN_mod_sqrt() so that it verifies that the input value
3368 is really the square of the return value. (Previously,
3369 BN_mod_sqrt would show GIGO behaviour.)
3372 *) Add named elliptic curves over binary fields from X9.62, SECG,
3373 and WAP/WTLS; add OIDs that were still missing.
3375 [Sheueling Chang Shantz and Douglas Stebila
3376 (Sun Microsystems Laboratories)]
3378 *) Extend the EC library for elliptic curves over binary fields
3379 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
3382 EC_GF2m_simple_method
3386 EC_GROUP_new_curve_GF2m
3387 EC_GROUP_set_curve_GF2m
3388 EC_GROUP_get_curve_GF2m
3389 EC_POINT_set_affine_coordinates_GF2m
3390 EC_POINT_get_affine_coordinates_GF2m
3391 EC_POINT_set_compressed_coordinates_GF2m
3393 Point compression for binary fields is disabled by default for
3394 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
3397 As binary polynomials are represented as BIGNUMs, various members
3398 of the EC_GROUP and EC_POINT data structures can be shared
3399 between the implementations for prime fields and binary fields;
3400 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
3401 are essentially identical to their ..._GFp counterparts.
3402 (For simplicity, the '..._GFp' prefix has been dropped from
3403 various internal method names.)
3405 An internal 'field_div' method (similar to 'field_mul' and
3406 'field_sqr') has been added; this is used only for binary fields.
3408 [Sheueling Chang Shantz and Douglas Stebila
3409 (Sun Microsystems Laboratories)]
3411 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
3412 through methods ('mul', 'precompute_mult').
3414 The generic implementations (now internally called 'ec_wNAF_mul'
3415 and 'ec_wNAF_precomputed_mult') remain the default if these
3416 methods are undefined.
3418 [Sheueling Chang Shantz and Douglas Stebila
3419 (Sun Microsystems Laboratories)]
3421 *) New function EC_GROUP_get_degree, which is defined through
3422 EC_METHOD. For curves over prime fields, this returns the bit
3423 length of the modulus.
3425 [Sheueling Chang Shantz and Douglas Stebila
3426 (Sun Microsystems Laboratories)]
3428 *) New functions EC_GROUP_dup, EC_POINT_dup.
3429 (These simply call ..._new and ..._copy).
3431 [Sheueling Chang Shantz and Douglas Stebila
3432 (Sun Microsystems Laboratories)]
3434 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
3435 Polynomials are represented as BIGNUMs (where the sign bit is not
3436 used) in the following functions [macros]:
3439 BN_GF2m_sub [= BN_GF2m_add]
3440 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
3441 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
3442 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
3444 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
3445 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
3446 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
3447 BN_GF2m_cmp [= BN_ucmp]
3449 (Note that only the 'mod' functions are actually for fields GF(2^m).
3450 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
3452 For some functions, an the irreducible polynomial defining a
3453 field can be given as an 'unsigned int[]' with strictly
3454 decreasing elements giving the indices of those bits that are set;
3455 i.e., p[] represents the polynomial
3456 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
3458 p[0] > p[1] > ... > p[k] = 0.
3459 This applies to the following functions:
3464 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
3465 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
3467 BN_GF2m_mod_sqrt_arr
3468 BN_GF2m_mod_solve_quad_arr
3472 Conversion can be performed by the following functions:
3477 bntest.c has additional tests for binary polynomial arithmetic.
3479 Two implementations for BN_GF2m_mod_div() are available.
3480 The default algorithm simply uses BN_GF2m_mod_inv() and
3481 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
3482 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
3483 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
3485 [Sheueling Chang Shantz and Douglas Stebila
3486 (Sun Microsystems Laboratories)]
3488 *) Add new error code 'ERR_R_DISABLED' that can be used when some
3489 functionality is disabled at compile-time.
3490 [Douglas Stebila <douglas.stebila@sun.com>]
3492 *) Change default behaviour of 'openssl asn1parse' so that more
3493 information is visible when viewing, e.g., a certificate:
3495 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
3496 mode the content of non-printable OCTET STRINGs is output in a
3497 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
3498 avoid the appearance of a printable string.
3499 [Nils Larsch <nla@trustcenter.de>]
3501 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
3503 EC_GROUP_set_asn1_flag()
3504 EC_GROUP_get_asn1_flag()
3505 EC_GROUP_set_point_conversion_form()
3506 EC_GROUP_get_point_conversion_form()
3507 These control ASN1 encoding details:
3508 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
3509 has been set to OPENSSL_EC_NAMED_CURVE.
3510 - Points are encoded in uncompressed form by default; options for
3511 asn1_for are as for point2oct, namely
3512 POINT_CONVERSION_COMPRESSED
3513 POINT_CONVERSION_UNCOMPRESSED
3514 POINT_CONVERSION_HYBRID
3516 Also add 'seed' and 'seed_len' members to EC_GROUP with access
3519 EC_GROUP_get0_seed()
3520 EC_GROUP_get_seed_len()
3521 This is used only for ASN1 purposes (so far).
3522 [Nils Larsch <nla@trustcenter.de>]
3524 *) Add 'field_type' member to EC_METHOD, which holds the NID
3525 of the appropriate field type OID. The new function
3526 EC_METHOD_get_field_type() returns this value.
3527 [Nils Larsch <nla@trustcenter.de>]
3532 EC_POINT_point2hex()
3533 EC_POINT_hex2point()
3534 providing useful interfaces to EC_POINT_point2oct() and
3535 EC_POINT_oct2point().
3536 [Nils Larsch <nla@trustcenter.de>]
3538 *) Change internals of the EC library so that the functions
3539 EC_GROUP_set_generator()
3540 EC_GROUP_get_generator()
3541 EC_GROUP_get_order()
3542 EC_GROUP_get_cofactor()
3543 are implemented directly in crypto/ec/ec_lib.c and not dispatched
3544 to methods, which would lead to unnecessary code duplication when
3545 adding different types of curves.
3546 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
3548 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
3549 arithmetic, and such that modified wNAFs are generated
3550 (which avoid length expansion in many cases).
3553 *) Add a function EC_GROUP_check_discriminant() (defined via
3554 EC_METHOD) that verifies that the curve discriminant is non-zero.
3556 Add a function EC_GROUP_check() that makes some sanity tests
3557 on a EC_GROUP, its generator and order. This includes
3558 EC_GROUP_check_discriminant().
3559 [Nils Larsch <nla@trustcenter.de>]
3561 *) Add ECDSA in new directory crypto/ecdsa/.
3563 Add applications 'openssl ecparam' and 'openssl ecdsa'
3564 (these are based on 'openssl dsaparam' and 'openssl dsa').
3566 ECDSA support is also included in various other files across the
3567 library. Most notably,
3568 - 'openssl req' now has a '-newkey ecdsa:file' option;
3569 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
3570 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
3571 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
3572 them suitable for ECDSA where domain parameters must be
3573 extracted before the specific public key;
3574 - ECDSA engine support has been added.
3575 [Nils Larsch <nla@trustcenter.de>]
3577 *) Include some named elliptic curves, and add OIDs from X9.62,
3578 SECG, and WAP/WTLS. Each curve can be obtained from the new
3580 EC_GROUP_new_by_curve_name(),
3581 and the list of available named curves can be obtained with
3582 EC_get_builtin_curves().
3583 Also add a 'curve_name' member to EC_GROUP objects, which can be
3585 EC_GROUP_set_curve_name()
3586 EC_GROUP_get_curve_name()
3587 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
3589 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3590 was actually never needed) and in BN_mul(). The removal in BN_mul()
3591 required a small change in bn_mul_part_recursive() and the addition
3592 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3593 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3594 bn_sub_words() and bn_add_words() except they take arrays with
3598 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
3600 *) Cleanse PEM buffers before freeing them since they may contain
3602 [Benjamin Bennett <ben@psc.edu>]
3604 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3605 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3606 authentication-only ciphersuites.
3609 *) Since AES128 and AES256 share a single mask bit in the logic of
3610 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3611 kludge to work properly if AES128 is available and AES256 isn't.
3614 *) Expand security boundary to match 1.1.1 module.
3617 *) Remove redundant features: hash file source, editing of test vectors
3618 modify fipsld to use external fips_premain.c signature.
3621 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
3622 run algorithm test programs.
3625 *) Make algorithm test programs more tolerant of whitespace.
3628 *) Have SSL/TLS server implementation tolerate "mismatched" record
3629 protocol version while receiving ClientHello even if the
3630 ClientHello is fragmented. (The server can't insist on the
3631 particular protocol version it has chosen before the ServerHello
3632 message has informed the client about his choice.)
3635 *) Load error codes if they are not already present instead of using a
3636 static variable. This allows them to be cleanly unloaded and reloaded.
3639 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3641 *) Introduce limits to prevent malicious keys being able to
3642 cause a denial of service. (CVE-2006-2940)
3643 [Steve Henson, Bodo Moeller]
3645 *) Fix ASN.1 parsing of certain invalid structures that can result
3646 in a denial of service. (CVE-2006-2937) [Steve Henson]
3648 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3649 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3651 *) Fix SSL client code which could crash if connecting to a
3652 malicious SSLv2 server. (CVE-2006-4343)
3653 [Tavis Ormandy and Will Drewry, Google Security Team]
3655 *) Change ciphersuite string processing so that an explicit
3656 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3657 will no longer include "AES128-SHA"), and any other similar
3658 ciphersuite (same bitmap) from *other* protocol versions (so that
3659 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3660 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3661 changes from 0.9.8b and 0.9.8d.
3664 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
3666 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3667 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3669 *) Change the Unix randomness entropy gathering to use poll() when
3670 possible instead of select(), since the latter has some
3671 undesirable limitations.
3672 [Darryl Miles via Richard Levitte and Bodo Moeller]
3674 *) Disable rogue ciphersuites:
3676 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3677 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3678 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3680 The latter two were purportedly from
3681 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3684 Also deactive the remaining ciphersuites from
3685 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3686 unofficial, and the ID has long expired.
3689 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3690 dual-core machines) and other potential thread-safety issues.
3693 Changes between 0.9.7i and 0.9.7j [04 May 2006]
3695 *) Adapt fipsld and the build system to link against the validated FIPS
3696 module in FIPS mode.
3699 *) Fixes for VC++ 2005 build under Windows.
3702 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
3703 from a Windows bash shell such as MSYS. It is autodetected from the
3704 "config" script when run from a VC++ environment. Modify standard VC++
3705 build to use fipscanister.o from the GNU make build.
3708 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
3710 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
3711 The value now differs depending on if you build for FIPS or not.
3712 BEWARE! A program linked with a shared FIPSed libcrypto can't be
3713 safely run with a non-FIPSed libcrypto, as it may crash because of
3714 the difference induced by this change.
3717 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
3719 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3720 (part of SSL_OP_ALL). This option used to disable the
3721 countermeasure against man-in-the-middle protocol-version
3722 rollback in the SSL 2.0 server implementation, which is a bad
3723 idea. (CVE-2005-2969)
3725 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3726 for Information Security, National Institute of Advanced Industrial
3727 Science and Technology [AIST], Japan)]
3729 *) Minimal support for X9.31 signatures and PSS padding modes. This is
3730 mainly for FIPS compliance and not fully integrated at this stage.
3733 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
3734 the exponentiation using a fixed-length exponent. (Otherwise,
3735 the information leaked through timing could expose the secret key
3736 after many signatures; cf. Bleichenbacher's attack on DSA with
3740 *) Make a new fixed-window mod_exp implementation the default for
3741 RSA, DSA, and DH private-key operations so that the sequence of
3742 squares and multiplies and the memory access pattern are
3743 independent of the particular secret key. This will mitigate
3744 cache-timing and potential related attacks.
3746 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
3747 and this is automatically used by BN_mod_exp_mont() if the new flag
3748 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
3749 will use this BN flag for private exponents unless the flag
3750 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
3751 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
3753 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
3755 *) Change the client implementation for SSLv23_method() and
3756 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
3757 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
3758 (Previously, the SSL 2.0 backwards compatible Client Hello
3759 message format would be used even with SSL_OP_NO_SSLv2.)
3762 *) Add support for smime-type MIME parameter in S/MIME messages which some
3766 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
3767 a threadsafe manner. Modify rsa code to use new function and add calls
3768 to dsa and dh code (which had race conditions before).
3771 *) Include the fixed error library code in the C error file definitions
3772 instead of fixing them up at runtime. This keeps the error code
3773 structures constant.
3776 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
3778 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
3781 *) Fixes for newer kerberos headers. NB: the casts are needed because
3782 the 'length' field is signed on one version and unsigned on another
3783 with no (?) obvious way to tell the difference, without these VC++
3784 complains. Also the "definition" of FAR (blank) is no longer included
3785 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
3786 some needed definitions.
3789 *) Undo Cygwin change.
3792 *) Added support for proxy certificates according to RFC 3820.
3793 Because they may be a security thread to unaware applications,
3794 they must be explicitely allowed in run-time. See
3795 docs/HOWTO/proxy_certificates.txt for further information.
3798 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
3800 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
3801 server and client random values. Previously
3802 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
3803 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
3805 This change has negligible security impact because:
3807 1. Server and client random values still have 24 bytes of pseudo random
3810 2. Server and client random values are sent in the clear in the initial
3813 3. The master secret is derived using the premaster secret (48 bytes in
3814 size for static RSA ciphersuites) as well as client server and random
3817 The OpenSSL team would like to thank the UK NISCC for bringing this issue
3820 [Stephen Henson, reported by UK NISCC]
3822 *) Use Windows randomness collection on Cygwin.
3825 *) Fix hang in EGD/PRNGD query when communication socket is closed
3826 prematurely by EGD/PRNGD.
3827 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
3829 *) Prompt for pass phrases when appropriate for PKCS12 input format.
3832 *) Back-port of selected performance improvements from development
3833 branch, as well as improved support for PowerPC platforms.
3836 *) Add lots of checks for memory allocation failure, error codes to indicate
3837 failure and freeing up memory if a failure occurs.
3838 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
3840 *) Add new -passin argument to dgst.
3843 *) Perform some character comparisons of different types in X509_NAME_cmp:
3844 this is needed for some certificates that reencode DNs into UTF8Strings
3845 (in violation of RFC3280) and can't or wont issue name rollover
3849 *) Make an explicit check during certificate validation to see that
3850 the CA setting in each certificate on the chain is correct. As a
3851 side effect always do the following basic checks on extensions,
3852 not just when there's an associated purpose to the check:
3854 - if there is an unhandled critical extension (unless the user
3855 has chosen to ignore this fault)
3856 - if the path length has been exceeded (if one is set at all)
3857 - that certain extensions fit the associated purpose (if one has
3861 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
3863 *) Avoid a race condition when CRLs are checked in a multi threaded
3864 environment. This would happen due to the reordering of the revoked
3865 entries during signature checking and serial number lookup. Now the
3866 encoding is cached and the serial number sort performed under a lock.
3867 Add new STACK function sk_is_sorted().
3870 *) Add Delta CRL to the extension code.
3873 *) Various fixes to s3_pkt.c so alerts are sent properly.
3874 [David Holmes <d.holmes@f5.com>]
3876 *) Reduce the chances of duplicate issuer name and serial numbers (in
3877 violation of RFC3280) using the OpenSSL certificate creation utilities.
3878 This is done by creating a random 64 bit value for the initial serial
3879 number when a serial number file is created or when a self signed
3880 certificate is created using 'openssl req -x509'. The initial serial
3881 number file is created using 'openssl x509 -next_serial' in CA.pl
3882 rather than being initialized to 1.
3885 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
3887 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3888 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3889 [Joe Orton, Steve Henson]
3891 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
3893 [Joe Orton, Steve Henson]
3895 *) Make it possible to have multiple active certificates with the same
3896 subject in the CA index file. This is done only if the keyword
3897 'unique_subject' is set to 'no' in the main CA section (default
3898 if 'CA_default') of the configuration file. The value is saved
3899 with the database itself in a separate index attribute file,
3900 named like the index file with '.attr' appended to the name.
3903 *) X509 verify fixes. Disable broken certificate workarounds when
3904 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
3905 keyUsage extension present. Don't accept CRLs with unhandled critical
3906 extensions: since verify currently doesn't process CRL extensions this
3907 rejects a CRL with *any* critical extensions. Add new verify error codes
3911 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
3912 A clarification of RFC2560 will require the use of OCTET STRINGs and
3913 some implementations cannot handle the current raw format. Since OpenSSL
3914 copies and compares OCSP nonces as opaque blobs without any attempt at
3915 parsing them this should not create any compatibility issues.
3918 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
3919 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
3920 this HMAC (and other) operations are several times slower than OpenSSL
3924 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
3925 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3927 *) Use the correct content when signing type "other".
3930 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
3932 *) Fix various bugs revealed by running the NISCC test suite:
3934 Stop out of bounds reads in the ASN1 code when presented with
3935 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3937 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
3939 If verify callback ignores invalid public key errors don't try to check
3940 certificate signature with the NULL public key.
3944 *) New -ignore_err option in ocsp application to stop the server
3945 exiting on the first error in a request.
3948 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3949 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3953 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3954 extra data after the compression methods not only for TLS 1.0
3955 but also for SSL 3.0 (as required by the specification).
3956 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3958 *) Change X509_certificate_type() to mark the key as exported/exportable
3959 when it's 512 *bits* long, not 512 bytes.
3962 *) Change AES_cbc_encrypt() so it outputs exact multiple of
3963 blocks during encryption.
3966 *) Various fixes to base64 BIO and non blocking I/O. On write
3967 flushes were not handled properly if the BIO retried. On read
3968 data was not being buffered properly and had various logic bugs.
3969 This also affects blocking I/O when the data being decoded is a
3973 *) Various S/MIME bugfixes and compatibility changes:
3974 output correct application/pkcs7 MIME type if
3975 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
3976 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
3977 of files as .eml work). Correctly handle very long lines in MIME
3981 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
3983 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3984 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3985 a protocol version number mismatch like a decryption error
3986 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3989 *) Turn on RSA blinding by default in the default implementation
3990 to avoid a timing attack. Applications that don't want it can call
3991 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3992 They would be ill-advised to do so in most cases.
3993 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3995 *) Change RSA blinding code so that it works when the PRNG is not
3996 seeded (in this case, the secret RSA exponent is abused as
3997 an unpredictable seed -- if it is not unpredictable, there
3998 is no point in blinding anyway). Make RSA blinding thread-safe
3999 by remembering the creator's thread ID in rsa->blinding and
4000 having all other threads use local one-time blinding factors
4001 (this requires more computation than sharing rsa->blinding, but
4002 avoids excessive locking; and if an RSA object is not shared
4003 between threads, blinding will still be very fast).
4006 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
4007 ENGINE as defaults for all supported algorithms irrespective of
4008 the 'flags' parameter. 'flags' is now honoured, so applications
4009 should make sure they are passing it correctly.
4012 *) Target "mingw" now allows native Windows code to be generated in
4013 the Cygwin environment as well as with the MinGW compiler.
4016 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
4018 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4019 via timing by performing a MAC computation even if incorrrect
4020 block cipher padding has been found. This is a countermeasure
4021 against active attacks where the attacker has to distinguish
4022 between bad padding and a MAC verification error. (CVE-2003-0078)
4024 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4025 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4026 Martin Vuagnoux (EPFL, Ilion)]
4028 *) Make the no-err option work as intended. The intention with no-err
4029 is not to have the whole error stack handling routines removed from
4030 libcrypto, it's only intended to remove all the function name and
4031 reason texts, thereby removing some of the footprint that may not
4032 be interesting if those errors aren't displayed anyway.
4034 NOTE: it's still possible for any application or module to have it's
4035 own set of error texts inserted. The routines are there, just not
4036 used by default when no-err is given.
4039 *) Add support for FreeBSD on IA64.
4040 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
4042 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
4043 Kerberos function mit_des_cbc_cksum(). Before this change,
4044 the value returned by DES_cbc_cksum() was like the one from
4045 mit_des_cbc_cksum(), except the bytes were swapped.
4046 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
4048 *) Allow an application to disable the automatic SSL chain building.
4049 Before this a rather primitive chain build was always performed in
4050 ssl3_output_cert_chain(): an application had no way to send the
4051 correct chain if the automatic operation produced an incorrect result.
4053 Now the chain builder is disabled if either:
4055 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
4057 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
4059 The reasoning behind this is that an application would not want the
4060 auto chain building to take place if extra chain certificates are
4061 present and it might also want a means of sending no additional
4062 certificates (for example the chain has two certificates and the
4066 *) Add the possibility to build without the ENGINE framework.
4067 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4069 *) Under Win32 gmtime() can return NULL: check return value in
4070 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
4073 *) DSA routines: under certain error conditions uninitialized BN objects
4074 could be freed. Solution: make sure initialization is performed early
4075 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
4076 Nils Larsch <nla@trustcenter.de> via PR#459)
4079 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
4080 checked on reconnect on the client side, therefore session resumption
4081 could still fail with a "ssl session id is different" error. This
4082 behaviour is masked when SSL_OP_ALL is used due to
4083 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
4084 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4085 followup to PR #377.
4088 *) IA-32 assembler support enhancements: unified ELF targets, support
4089 for SCO/Caldera platforms, fix for Cygwin shared build.
4092 *) Add support for FreeBSD on sparc64. As a consequence, support for
4093 FreeBSD on non-x86 processors is separate from x86 processors on
4094 the config script, much like the NetBSD support.
4095 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
4097 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
4099 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
4102 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
4103 code (06) was taken as the first octet of the session ID and the last
4104 octet was ignored consequently. As a result SSLv2 client side session
4105 caching could not have worked due to the session ID mismatch between
4107 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4111 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
4112 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
4116 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
4117 seems that in spite of existing for more than a year, many application
4118 author have done nothing to provide the necessary callbacks, which
4119 means that this particular engine will not work properly anywhere.
4120 This is a very unfortunate situation which forces us, in the name
4121 of usability, to give the hw_ncipher.c a static lock, which is part
4123 NOTE: This is for the 0.9.7 series ONLY. This hack will never
4124 appear in 0.9.8 or later. We EXPECT application authors to have
4125 dealt properly with this when 0.9.8 is released (unless we actually
4126 make such changes in the libcrypto locking code that changes will
4127 have to be made anyway).
4130 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
4131 octets have been read, EOF or an error occurs. Without this change
4132 some truncated ASN1 structures will not produce an error.
4135 *) Disable Heimdal support, since it hasn't been fully implemented.
4136 Still give the possibility to force the use of Heimdal, but with
4137 warnings and a request that patches get sent to openssl-dev.
4140 *) Add the VC-CE target, introduce the WINCE sysname, and add
4141 INSTALL.WCE and appropriate conditionals to make it build.
4142 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4144 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
4145 cygssl-x.y.z.dll, where x, y and z are the major, minor and
4146 edit numbers of the version.
4147 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4149 *) Introduce safe string copy and catenation functions
4150 (BUF_strlcpy() and BUF_strlcat()).
4151 [Ben Laurie (CHATS) and Richard Levitte]
4153 *) Avoid using fixed-size buffers for one-line DNs.
4154 [Ben Laurie (CHATS)]
4156 *) Add BUF_MEM_grow_clean() to avoid information leakage when
4157 resizing buffers containing secrets, and use where appropriate.
4158 [Ben Laurie (CHATS)]
4160 *) Avoid using fixed size buffers for configuration file location.
4161 [Ben Laurie (CHATS)]
4163 *) Avoid filename truncation for various CA files.
4164 [Ben Laurie (CHATS)]
4166 *) Use sizeof in preference to magic numbers.
4167 [Ben Laurie (CHATS)]
4169 *) Avoid filename truncation in cert requests.
4170 [Ben Laurie (CHATS)]
4172 *) Add assertions to check for (supposedly impossible) buffer
4174 [Ben Laurie (CHATS)]
4176 *) Don't cache truncated DNS entries in the local cache (this could
4177 potentially lead to a spoofing attack).
4178 [Ben Laurie (CHATS)]
4180 *) Fix various buffers to be large enough for hex/decimal
4181 representations in a platform independent manner.
4182 [Ben Laurie (CHATS)]
4184 *) Add CRYPTO_realloc_clean() to avoid information leakage when
4185 resizing buffers containing secrets, and use where appropriate.
4186 [Ben Laurie (CHATS)]
4188 *) Add BIO_indent() to avoid much slightly worrying code to do
4190 [Ben Laurie (CHATS)]
4192 *) Convert sprintf()/BIO_puts() to BIO_printf().
4193 [Ben Laurie (CHATS)]
4195 *) buffer_gets() could terminate with the buffer only half
4197 [Ben Laurie (CHATS)]
4199 *) Add assertions to prevent user-supplied crypto functions from
4200 overflowing internal buffers by having large block sizes, etc.
4201 [Ben Laurie (CHATS)]
4203 *) New OPENSSL_assert() macro (similar to assert(), but enabled
4205 [Ben Laurie (CHATS)]
4207 *) Eliminate unused copy of key in RC4.
4208 [Ben Laurie (CHATS)]
4210 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
4211 [Ben Laurie (CHATS)]
4213 *) Fix off-by-one error in EGD path.
4214 [Ben Laurie (CHATS)]
4216 *) If RANDFILE path is too long, ignore instead of truncating.
4217 [Ben Laurie (CHATS)]
4219 *) Eliminate unused and incorrectly sized X.509 structure
4221 [Ben Laurie (CHATS)]
4223 *) Eliminate unused and dangerous function knumber().
4224 [Ben Laurie (CHATS)]
4226 *) Eliminate unused and dangerous structure, KSSL_ERR.
4227 [Ben Laurie (CHATS)]
4229 *) Protect against overlong session ID context length in an encoded
4230 session object. Since these are local, this does not appear to be
4232 [Ben Laurie (CHATS)]
4234 *) Change from security patch (see 0.9.6e below) that did not affect
4235 the 0.9.6 release series:
4237 Remote buffer overflow in SSL3 protocol - an attacker could
4238 supply an oversized master key in Kerberos-enabled versions.
4240 [Ben Laurie (CHATS)]
4242 *) Change the SSL kerb5 codes to match RFC 2712.
4245 *) Make -nameopt work fully for req and add -reqopt switch.
4246 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
4248 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
4249 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
4251 *) Make sure tests can be performed even if the corresponding algorithms
4252 have been removed entirely. This was also the last step to make
4253 OpenSSL compilable with DJGPP under all reasonable conditions.
4254 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
4256 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
4257 to allow version independent disabling of normally unselected ciphers,
4258 which may be activated as a side-effect of selecting a single cipher.
4260 (E.g., cipher list string "RSA" enables ciphersuites that are left
4261 out of "ALL" because they do not provide symmetric encryption.
4262 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
4263 [Lutz Jaenicke, Bodo Moeller]
4265 *) Add appropriate support for separate platform-dependent build
4266 directories. The recommended way to make a platform-dependent
4267 build directory is the following (tested on Linux), maybe with
4270 # Place yourself outside of the OpenSSL source tree. In
4271 # this example, the environment variable OPENSSL_SOURCE
4272 # is assumed to contain the absolute OpenSSL source directory.
4273 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
4274 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
4275 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
4276 mkdir -p `dirname $F`
4277 ln -s $OPENSSL_SOURCE/$F $F
4280 To be absolutely sure not to disturb the source tree, a "make clean"
4281 is a good thing. If it isn't successfull, don't worry about it,
4282 it probably means the source directory is very clean.
4285 *) Make sure any ENGINE control commands make local copies of string
4286 pointers passed to them whenever necessary. Otherwise it is possible
4287 the caller may have overwritten (or deallocated) the original string
4288 data when a later ENGINE operation tries to use the stored values.
4289 [Götz Babin-Ebell <babinebell@trustcenter.de>]
4291 *) Improve diagnostics in file reading and command-line digests.
4292 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
4294 *) Add AES modes CFB and OFB to the object database. Correct an
4295 error in AES-CFB decryption.
4298 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
4299 allows existing EVP_CIPHER_CTX structures to be reused after
4300 calling EVP_*Final(). This behaviour is used by encryption
4301 BIOs and some applications. This has the side effect that
4302 applications must explicitly clean up cipher contexts with
4303 EVP_CIPHER_CTX_cleanup() or they will leak memory.
4306 *) Check the values of dna and dnb in bn_mul_recursive before calling
4307 bn_mul_comba (a non zero value means the a or b arrays do not contain
4308 n2 elements) and fallback to bn_mul_normal if either is not zero.
4311 *) Fix escaping of non-ASCII characters when using the -subj option
4312 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
4315 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
4316 form for "surname", serialNumber has no short form.
4317 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
4318 therefore remove "mail" short name for "internet 7".
4319 The OID for unique identifiers in X509 certificates is
4320 x500UniqueIdentifier, not uniqueIdentifier.
4321 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
4324 *) Add an "init" command to the ENGINE config module and auto initialize
4325 ENGINEs. Without any "init" command the ENGINE will be initialized
4326 after all ctrl commands have been executed on it. If init=1 the
4327 ENGINE is initailized at that point (ctrls before that point are run
4328 on the uninitialized ENGINE and after on the initialized one). If
4329 init=0 then the ENGINE will not be iniatialized at all.
4332 *) Fix the 'app_verify_callback' interface so that the user-defined
4333 argument is actually passed to the callback: In the
4334 SSL_CTX_set_cert_verify_callback() prototype, the callback
4335 declaration has been changed from
4338 int (*cb)(X509_STORE_CTX *,void *);
4339 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
4340 i=s->ctx->app_verify_callback(&ctx)
4341 has been changed into
4342 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
4344 To update applications using SSL_CTX_set_cert_verify_callback(),
4345 a dummy argument can be added to their callback functions.
4346 [D. K. Smetters <smetters@parc.xerox.com>]
4348 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
4349 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
4351 *) Add and OPENSSL_LOAD_CONF define which will cause
4352 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
4353 This allows older applications to transparently support certain
4354 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
4355 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
4356 load the config file and OPENSSL_add_all_algorithms_conf() which will
4357 always load it have also been added.
4360 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
4361 Adjust NIDs and EVP layer.
4362 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4364 *) Config modules support in openssl utility.
4366 Most commands now load modules from the config file,
4367 though in a few (such as version) this isn't done
4368 because it couldn't be used for anything.
4370 In the case of ca and req the config file used is
4371 the same as the utility itself: that is the -config
4372 command line option can be used to specify an
4376 *) Move default behaviour from OPENSSL_config(). If appname is NULL
4377 use "openssl_conf" if filename is NULL use default openssl config file.
4380 *) Add an argument to OPENSSL_config() to allow the use of an alternative
4381 config section name. Add a new flag to tolerate a missing config file
4382 and move code to CONF_modules_load_file().
4385 *) Support for crypto accelerator cards from Accelerated Encryption
4386 Processing, www.aep.ie. (Use engine 'aep')
4387 The support was copied from 0.9.6c [engine] and adapted/corrected
4388 to work with the new engine framework.
4389 [AEP Inc. and Richard Levitte]
4391 *) Support for SureWare crypto accelerator cards from Baltimore
4392 Technologies. (Use engine 'sureware')
4393 The support was copied from 0.9.6c [engine] and adapted
4394 to work with the new engine framework.
4397 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
4398 make the newer ENGINE framework commands for the CHIL engine work.
4399 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
4401 *) Make it possible to produce shared libraries on ReliantUNIX.
4402 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
4404 *) Add the configuration target debug-linux-ppro.
4405 Make 'openssl rsa' use the general key loading routines
4406 implemented in apps.c, and make those routines able to
4407 handle the key format FORMAT_NETSCAPE and the variant
4409 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4411 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4412 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4414 *) Add -keyform to rsautl, and document -engine.
4415 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
4417 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
4418 BIO_R_NO_SUCH_FILE error code rather than the generic
4419 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
4422 *) Add new functions
4424 ERR_peek_last_error_line
4425 ERR_peek_last_error_line_data.
4426 These are similar to
4429 ERR_peek_error_line_data,
4430 but report on the latest error recorded rather than the first one
4431 still in the error queue.
4432 [Ben Laurie, Bodo Moeller]
4434 *) default_algorithms option in ENGINE config module. This allows things
4436 default_algorithms = ALL
4437 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
4440 *) Prelminary ENGINE config module.
4443 *) New experimental application configuration code.
4446 *) Change the AES code to follow the same name structure as all other
4447 symmetric ciphers, and behave the same way. Move everything to
4448 the directory crypto/aes, thereby obsoleting crypto/rijndael.
4449 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4451 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
4452 [Ben Laurie and Theo de Raadt]
4454 *) Add option to output public keys in req command.
4455 [Massimiliano Pala madwolf@openca.org]
4457 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
4458 (up to about 10% better than before for P-192 and P-224).
4461 *) New functions/macros
4463 SSL_CTX_set_msg_callback(ctx, cb)
4464 SSL_CTX_set_msg_callback_arg(ctx, arg)
4465 SSL_set_msg_callback(ssl, cb)
4466 SSL_set_msg_callback_arg(ssl, arg)
4468 to request calling a callback function
4470 void cb(int write_p, int version, int content_type,
4471 const void *buf, size_t len, SSL *ssl, void *arg)
4473 whenever a protocol message has been completely received
4474 (write_p == 0) or sent (write_p == 1). Here 'version' is the
4475 protocol version according to which the SSL library interprets
4476 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
4477 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
4478 the content type as defined in the SSL 3.0/TLS 1.0 protocol
4479 specification (change_cipher_spec(20), alert(21), handshake(22)).
4480 'buf' and 'len' point to the actual message, 'ssl' to the
4481 SSL object, and 'arg' is the application-defined value set by
4482 SSL[_CTX]_set_msg_callback_arg().
4484 'openssl s_client' and 'openssl s_server' have new '-msg' options
4485 to enable a callback that displays all protocol messages.
4488 *) Change the shared library support so shared libraries are built as
4489 soon as the corresponding static library is finished, and thereby get
4490 openssl and the test programs linked against the shared library.
4491 This still only happens when the keyword "shard" has been given to
4492 the configuration scripts.
4494 NOTE: shared library support is still an experimental thing, and
4495 backward binary compatibility is still not guaranteed.
4496 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
4498 *) Add support for Subject Information Access extension.
4499 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4501 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
4502 additional bytes when new memory had to be allocated, not just
4503 when reusing an existing buffer.
4506 *) New command line and configuration option 'utf8' for the req command.
4507 This allows field values to be specified as UTF8 strings.
4510 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
4511 runs for the former and machine-readable output for the latter.
4514 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
4515 of the e-mail address in the DN (i.e., it will go into a certificate
4516 extension only). The new configuration file option 'email_in_dn = no'
4517 has the same effect.
4518 [Massimiliano Pala madwolf@openca.org]
4520 *) Change all functions with names starting with des_ to be starting
4521 with DES_ instead. Add wrappers that are compatible with libdes,
4522 but are named _ossl_old_des_*. Finally, add macros that map the
4523 des_* symbols to the corresponding _ossl_old_des_* if libdes
4524 compatibility is desired. If OpenSSL 0.9.6c compatibility is
4525 desired, the des_* symbols will be mapped to DES_*, with one
4528 Since we provide two compatibility mappings, the user needs to
4529 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
4530 compatibility is desired. The default (i.e., when that macro
4531 isn't defined) is OpenSSL 0.9.6c compatibility.
4533 There are also macros that enable and disable the support of old
4534 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
4535 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
4536 are defined, the default will apply: to support the old des routines.
4538 In either case, one must include openssl/des.h to get the correct
4539 definitions. Do not try to just include openssl/des_old.h, that
4542 NOTE: This is a major break of an old API into a new one. Software
4543 authors are encouraged to switch to the DES_ style functions. Some
4544 time in the future, des_old.h and the libdes compatibility functions
4545 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
4546 default), and then completely removed.
4549 *) Test for certificates which contain unsupported critical extensions.
4550 If such a certificate is found during a verify operation it is
4551 rejected by default: this behaviour can be overridden by either
4552 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
4553 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
4554 X509_supported_extension() has also been added which returns 1 if a
4555 particular extension is supported.
4558 *) Modify the behaviour of EVP cipher functions in similar way to digests
4559 to retain compatibility with existing code.
4562 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
4563 compatibility with existing code. In particular the 'ctx' parameter does
4564 not have to be to be initialized before the call to EVP_DigestInit() and
4565 it is tidied up after a call to EVP_DigestFinal(). New function
4566 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
4567 EVP_MD_CTX_copy() changed to not require the destination to be
4568 initialized valid and new function EVP_MD_CTX_copy_ex() added which
4569 requires the destination to be valid.
4571 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
4572 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
4575 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
4576 so that complete 'Handshake' protocol structures are kept in memory
4577 instead of overwriting 'msg_type' and 'length' with 'body' data.
4580 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
4581 [Massimo Santin via Richard Levitte]
4583 *) Major restructuring to the underlying ENGINE code. This includes
4584 reduction of linker bloat, separation of pure "ENGINE" manipulation
4585 (initialisation, etc) from functionality dealing with implementations
4586 of specific crypto iterfaces. This change also introduces integrated
4587 support for symmetric ciphers and digest implementations - so ENGINEs
4588 can now accelerate these by providing EVP_CIPHER and EVP_MD
4589 implementations of their own. This is detailed in crypto/engine/README
4590 as it couldn't be adequately described here. However, there are a few
4591 API changes worth noting - some RSA, DSA, DH, and RAND functions that
4592 were changed in the original introduction of ENGINE code have now
4593 reverted back - the hooking from this code to ENGINE is now a good
4594 deal more passive and at run-time, operations deal directly with
4595 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
4596 dereferencing through an ENGINE pointer any more. Also, the ENGINE
4597 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
4598 they were not being used by the framework as there is no concept of a
4599 BIGNUM_METHOD and they could not be generalised to the new
4600 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
4601 ENGINE_cpy() has been removed as it cannot be consistently defined in
4605 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
4608 *) Change mkdef.pl to sort symbols that get the same entry number,
4609 and make sure the automatically generated functions ERR_load_*
4610 become part of libeay.num as well.
4613 *) New function SSL_renegotiate_pending(). This returns true once
4614 renegotiation has been requested (either SSL_renegotiate() call
4615 or HelloRequest/ClientHello receveived from the peer) and becomes
4616 false once a handshake has been completed.
4617 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
4618 sends a HelloRequest, but does not ensure that a handshake takes
4619 place. SSL_renegotiate_pending() is useful for checking if the
4620 client has followed the request.)
4623 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
4624 By default, clients may request session resumption even during
4625 renegotiation (if session ID contexts permit); with this option,
4626 session resumption is possible only in the first handshake.
4628 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4629 more bits available for options that should not be part of
4630 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4633 *) Add some demos for certificate and certificate request creation.
4636 *) Make maximum certificate chain size accepted from the peer application
4637 settable (SSL*_get/set_max_cert_list()), as proposed by
4638 "Douglas E. Engert" <deengert@anl.gov>.
4641 *) Add support for shared libraries for Unixware-7
4642 (Boyd Lynn Gerber <gerberb@zenez.com>).
4645 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4646 be done prior to destruction. Use this to unload error strings from
4647 ENGINEs that load their own error strings. NB: This adds two new API
4648 functions to "get" and "set" this destroy handler in an ENGINE.
4651 *) Alter all existing ENGINE implementations (except "openssl" and
4652 "openbsd") to dynamically instantiate their own error strings. This
4653 makes them more flexible to be built both as statically-linked ENGINEs
4654 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4655 Also, add stub code to each that makes building them as self-contained
4656 shared-libraries easier (see README.ENGINE).
4659 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4660 implementations into applications that are completely implemented in
4661 self-contained shared-libraries. The "dynamic" ENGINE exposes control
4662 commands that can be used to configure what shared-library to load and
4663 to control aspects of the way it is handled. Also, made an update to
4664 the README.ENGINE file that brings its information up-to-date and
4665 provides some information and instructions on the "dynamic" ENGINE
4666 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
4669 *) Make it possible to unload ranges of ERR strings with a new
4670 "ERR_unload_strings" function.
4673 *) Add a copy() function to EVP_MD.
4676 *) Make EVP_MD routines take a context pointer instead of just the
4677 md_data void pointer.
4680 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
4681 that the digest can only process a single chunk of data
4682 (typically because it is provided by a piece of
4683 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
4684 is only going to provide a single chunk of data, and hence the
4685 framework needn't accumulate the data for oneshot drivers.
4688 *) As with "ERR", make it possible to replace the underlying "ex_data"
4689 functions. This change also alters the storage and management of global
4690 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
4691 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
4692 index counters. The API functions that use this state have been changed
4693 to take a "class_index" rather than pointers to the class's local STACK
4694 and counter, and there is now an API function to dynamically create new
4695 classes. This centralisation allows us to (a) plug a lot of the
4696 thread-safety problems that existed, and (b) makes it possible to clean
4697 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
4698 such data would previously have always leaked in application code and
4699 workarounds were in place to make the memory debugging turn a blind eye
4700 to it. Application code that doesn't use this new function will still
4701 leak as before, but their memory debugging output will announce it now
4702 rather than letting it slide.
4704 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
4705 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
4706 has a return value to indicate success or failure.
4709 *) Make it possible to replace the underlying "ERR" functions such that the
4710 global state (2 LHASH tables and 2 locks) is only used by the "default"
4711 implementation. This change also adds two functions to "get" and "set"
4712 the implementation prior to it being automatically set the first time
4713 any other ERR function takes place. Ie. an application can call "get",
4714 pass the return value to a module it has just loaded, and that module
4715 can call its own "set" function using that value. This means the
4716 module's "ERR" operations will use (and modify) the error state in the
4717 application and not in its own statically linked copy of OpenSSL code.
4720 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
4721 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
4722 the operation, and provides a more encapsulated way for external code
4723 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
4724 to use these functions rather than manually incrementing the counts.
4726 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
4729 *) Add EVP test program.
4732 *) Add symmetric cipher support to ENGINE. Expect the API to change!
4735 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
4736 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
4737 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
4738 These allow a CRL to be built without having to access X509_CRL fields
4739 directly. Modify 'ca' application to use new functions.
4742 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
4743 bug workarounds. Rollback attack detection is a security feature.
4744 The problem will only arise on OpenSSL servers when TLSv1 is not
4745 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
4746 Software authors not wanting to support TLSv1 will have special reasons
4747 for their choice and can explicitly enable this option.
4748 [Bodo Moeller, Lutz Jaenicke]
4750 *) Rationalise EVP so it can be extended: don't include a union of
4751 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
4752 (similar to those existing for EVP_CIPHER_CTX).
4757 EVP_MD_CTX_init(&md); /* new function call */
4758 EVP_DigestInit(&md, EVP_sha1());
4759 EVP_DigestUpdate(&md, in, len);
4760 EVP_DigestFinal(&md, out, NULL);
4761 EVP_MD_CTX_cleanup(&md); /* new function call */
4765 *) Make DES key schedule conform to the usual scheme, as well as
4766 correcting its structure. This means that calls to DES functions
4767 now have to pass a pointer to a des_key_schedule instead of a
4768 plain des_key_schedule (which was actually always a pointer
4771 des_key_schedule ks;
4773 des_set_key_checked(..., &ks);
4774 des_ncbc_encrypt(..., &ks, ...);
4776 (Note that a later change renames 'des_...' into 'DES_...'.)
4779 *) Initial reduction of linker bloat: the use of some functions, such as
4780 PEM causes large amounts of unused functions to be linked in due to
4781 poor organisation. For example pem_all.c contains every PEM function
4782 which has a knock on effect of linking in large amounts of (unused)
4783 ASN1 code. Grouping together similar functions and splitting unrelated
4784 functions prevents this.
4787 *) Cleanup of EVP macros.
4790 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
4791 correct _ecb suffix.
4794 *) Add initial OCSP responder support to ocsp application. The
4795 revocation information is handled using the text based index
4796 use by the ca application. The responder can either handle
4797 requests generated internally, supplied in files (for example
4798 via a CGI script) or using an internal minimal server.
4801 *) Add configuration choices to get zlib compression for TLS.
4804 *) Changes to Kerberos SSL for RFC 2712 compliance:
4805 1. Implemented real KerberosWrapper, instead of just using
4806 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
4807 2. Implemented optional authenticator field of KerberosWrapper.
4809 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
4810 and authenticator structs; see crypto/krb5/.
4812 Generalized Kerberos calls to support multiple Kerberos libraries.
4813 [Vern Staats <staatsvr@asc.hpc.mil>,
4814 Jeffrey Altman <jaltman@columbia.edu>
4815 via Richard Levitte]
4817 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
4818 already does with RSA. testdsa.h now has 'priv_key/pub_key'
4819 values for each of the key sizes rather than having just
4820 parameters (and 'speed' generating keys each time).
4823 *) Speed up EVP routines.
4826 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
4827 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
4828 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
4829 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
4831 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
4832 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
4833 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
4836 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
4838 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
4841 *) Added the OS2-EMX target.
4842 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
4844 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
4845 to support NCONF routines in extension code. New function CONF_set_nconf()
4846 to allow functions which take an NCONF to also handle the old LHASH
4847 structure: this means that the old CONF compatible routines can be
4848 retained (in particular wrt extensions) without having to duplicate the
4849 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
4852 *) Enhance the general user interface with mechanisms for inner control
4853 and with possibilities to have yes/no kind of prompts.
4856 *) Change all calls to low level digest routines in the library and
4857 applications to use EVP. Add missing calls to HMAC_cleanup() and
4858 don't assume HMAC_CTX can be copied using memcpy().
4859 [Verdon Walker <VWalker@novell.com>, Steve Henson]
4861 *) Add the possibility to control engines through control names but with
4862 arbitrary arguments instead of just a string.
4863 Change the key loaders to take a UI_METHOD instead of a callback
4864 function pointer. NOTE: this breaks binary compatibility with earlier
4865 versions of OpenSSL [engine].
4866 Adapt the nCipher code for these new conditions and add a card insertion
4870 *) Enhance the general user interface with mechanisms to better support
4871 dialog box interfaces, application-defined prompts, the possibility
4872 to use defaults (for example default passwords from somewhere else)
4873 and interrupts/cancellations.
4876 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
4877 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
4880 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
4881 tidy up some unnecessarily weird code in 'sk_new()').
4882 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
4884 *) Change the key loading routines for ENGINEs to use the same kind
4885 callback (pem_password_cb) as all other routines that need this
4889 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
4890 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
4891 than this minimum value is recommended.
4894 *) New random seeder for OpenVMS, using the system process statistics
4895 that are easily reachable.
4898 *) Windows apparently can't transparently handle global
4899 variables defined in DLLs. Initialisations such as:
4901 const ASN1_ITEM *it = &ASN1_INTEGER_it;
4903 wont compile. This is used by the any applications that need to
4904 declare their own ASN1 modules. This was fixed by adding the option
4905 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
4906 needed for static libraries under Win32.
4909 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
4910 setting of purpose and trust fields. New X509_STORE trust and
4911 purpose functions and tidy up setting in other SSL functions.
4914 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
4915 structure. These are inherited by X509_STORE_CTX when it is
4916 initialised. This allows various defaults to be set in the
4917 X509_STORE structure (such as flags for CRL checking and custom
4918 purpose or trust settings) for functions which only use X509_STORE_CTX
4919 internally such as S/MIME.
4921 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
4922 trust settings if they are not set in X509_STORE. This allows X509_STORE
4923 purposes and trust (in S/MIME for example) to override any set by default.
4925 Add command line options for CRL checking to smime, s_client and s_server
4929 *) Initial CRL based revocation checking. If the CRL checking flag(s)
4930 are set then the CRL is looked up in the X509_STORE structure and
4931 its validity and signature checked, then if the certificate is found
4932 in the CRL the verify fails with a revoked error.
4934 Various new CRL related callbacks added to X509_STORE_CTX structure.
4936 Command line options added to 'verify' application to support this.
4938 This needs some additional work, such as being able to handle multiple
4939 CRLs with different times, extension based lookup (rather than just
4940 by subject name) and ultimately more complete V2 CRL extension
4944 *) Add a general user interface API (crypto/ui/). This is designed
4945 to replace things like des_read_password and friends (backward
4946 compatibility functions using this new API are provided).
4947 The purpose is to remove prompting functions from the DES code
4948 section as well as provide for prompting through dialog boxes in
4949 a window system and the like.
4952 *) Add "ex_data" support to ENGINE so implementations can add state at a
4953 per-structure level rather than having to store it globally.
4956 *) Make it possible for ENGINE structures to be copied when retrieved by
4957 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
4958 This causes the "original" ENGINE structure to act like a template,
4959 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
4960 operational state can be localised to each ENGINE structure, despite the
4961 fact they all share the same "methods". New ENGINE structures returned in
4962 this case have no functional references and the return value is the single
4963 structural reference. This matches the single structural reference returned
4964 by ENGINE_by_id() normally, when it is incremented on the pre-existing
4968 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
4969 needs to match any other type at all we need to manually clear the
4973 *) Changes to the "openssl engine" utility to include;
4974 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
4975 about an ENGINE's available control commands.
4976 - executing control commands from command line arguments using the
4977 '-pre' and '-post' switches. '-post' is only used if '-t' is
4978 specified and the ENGINE is successfully initialised. The syntax for
4979 the individual commands are colon-separated, for example;
4980 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
4983 *) New dynamic control command support for ENGINEs. ENGINEs can now
4984 declare their own commands (numbers), names (strings), descriptions,
4985 and input types for run-time discovery by calling applications. A
4986 subset of these commands are implicitly classed as "executable"
4987 depending on their input type, and only these can be invoked through
4988 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
4989 can be based on user input, config files, etc). The distinction is
4990 that "executable" commands cannot return anything other than a boolean
4991 result and can only support numeric or string input, whereas some
4992 discoverable commands may only be for direct use through
4993 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
4994 pointers, or other custom uses. The "executable" commands are to
4995 support parameterisations of ENGINE behaviour that can be
4996 unambiguously defined by ENGINEs and used consistently across any
4997 OpenSSL-based application. Commands have been added to all the
4998 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
4999 control over shared-library paths without source code alterations.
5002 *) Changed all ENGINE implementations to dynamically allocate their
5003 ENGINEs rather than declaring them statically. Apart from this being
5004 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
5005 this also allows the implementations to compile without using the
5006 internal engine_int.h header.
5009 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
5010 'const' value. Any code that should be able to modify a RAND_METHOD
5011 should already have non-const pointers to it (ie. they should only
5012 modify their own ones).
5015 *) Made a variety of little tweaks to the ENGINE code.
5016 - "atalla" and "ubsec" string definitions were moved from header files
5017 to C code. "nuron" string definitions were placed in variables
5018 rather than hard-coded - allowing parameterisation of these values
5019 later on via ctrl() commands.
5020 - Removed unused "#if 0"'d code.
5021 - Fixed engine list iteration code so it uses ENGINE_free() to release
5022 structural references.
5023 - Constified the RAND_METHOD element of ENGINE structures.
5024 - Constified various get/set functions as appropriate and added
5025 missing functions (including a catch-all ENGINE_cpy that duplicates
5026 all ENGINE values onto a new ENGINE except reference counts/state).
5027 - Removed NULL parameter checks in get/set functions. Setting a method
5028 or function to NULL is a way of cancelling out a previously set
5029 value. Passing a NULL ENGINE parameter is just plain stupid anyway
5030 and doesn't justify the extra error symbols and code.
5031 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
5032 flags from engine_int.h to engine.h.
5033 - Changed prototypes for ENGINE handler functions (init(), finish(),
5034 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
5037 *) Implement binary inversion algorithm for BN_mod_inverse in addition
5038 to the algorithm using long division. The binary algorithm can be
5039 used only if the modulus is odd. On 32-bit systems, it is faster
5040 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
5041 roughly 5-15% for 256-bit moduli), so we use it only for moduli
5042 up to 450 bits. In 64-bit environments, the binary algorithm
5043 appears to be advantageous for much longer moduli; here we use it
5044 for moduli up to 2048 bits.
5047 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
5048 could not support the combine flag in choice fields.
5051 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
5052 extensions from a certificate request to the certificate.
5055 *) Allow multiple 'certopt' and 'nameopt' options to be separated
5056 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
5057 file: this allows the display of the certificate about to be
5058 signed to be customised, to allow certain fields to be included
5059 or excluded and extension details. The old system didn't display
5060 multicharacter strings properly, omitted fields not in the policy
5061 and couldn't display additional details such as extensions.
5064 *) Function EC_POINTs_mul for multiple scalar multiplication
5065 of an arbitrary number of elliptic curve points
5066 \sum scalars[i]*points[i],
5067 optionally including the generator defined for the EC_GROUP:
5068 scalar*generator + \sum scalars[i]*points[i].
5070 EC_POINT_mul is a simple wrapper function for the typical case
5071 that the point list has just one item (besides the optional
5075 *) First EC_METHODs for curves over GF(p):
5077 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
5078 operations and provides various method functions that can also
5079 operate with faster implementations of modular arithmetic.
5081 EC_GFp_mont_method() reuses most functions that are part of
5082 EC_GFp_simple_method, but uses Montgomery arithmetic.
5084 [Bodo Moeller; point addition and point doubling
5085 implementation directly derived from source code provided by
5086 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
5088 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
5089 crypto/ec/ec_lib.c):
5091 Curves are EC_GROUP objects (with an optional group generator)
5092 based on EC_METHODs that are built into the library.
5094 Points are EC_POINT objects based on EC_GROUP objects.
5096 Most of the framework would be able to handle curves over arbitrary
5097 finite fields, but as there are no obvious types for fields other
5098 than GF(p), some functions are limited to that for now.
5101 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
5102 that the file contains a complete HTTP response.
5105 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
5106 change the def and num file printf format specifier from "%-40sXXX"
5107 to "%-39s XXX". The latter will always guarantee a space after the
5108 field while the former will cause them to run together if the field
5109 is 40 of more characters long.
5112 *) Constify the cipher and digest 'method' functions and structures
5113 and modify related functions to take constant EVP_MD and EVP_CIPHER
5117 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
5118 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
5121 *) Modify EVP_Digest*() routines so they now return values. Although the
5122 internal software routines can never fail additional hardware versions
5126 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
5128 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
5129 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
5132 ERR_R_NESTED_ASN1_ERROR
5134 ERR_R_MISSING_ASN1_EOS
5135 were 4 .. 9, conflicting with
5136 ERR_LIB_RSA (= ERR_R_RSA_LIB)
5138 ERR_LIB_PEM (= ERR_R_PEM_LIB).
5139 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
5141 Add new error code 'ERR_R_INTERNAL_ERROR'.
5144 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
5148 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
5149 sets the subject name for a new request or supersedes the
5150 subject name in a given request. Formats that can be parsed are
5151 'CN=Some Name, OU=myOU, C=IT'
5153 'CN=Some Name/OU=myOU/C=IT'.
5155 Add options '-batch' and '-verbose' to 'openssl req'.
5156 [Massimiliano Pala <madwolf@hackmasters.net>]
5158 *) Introduce the possibility to access global variables through
5159 functions on platform were that's the best way to handle exporting
5160 global variables in shared libraries. To enable this functionality,
5161 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
5162 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
5163 is normally done by Configure or something similar).
5165 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
5166 in the source file (foo.c) like this:
5168 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
5169 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
5171 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
5172 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
5174 OPENSSL_DECLARE_GLOBAL(int,foo);
5175 #define foo OPENSSL_GLOBAL_REF(foo)
5176 OPENSSL_DECLARE_GLOBAL(double,bar);
5177 #define bar OPENSSL_GLOBAL_REF(bar)
5179 The #defines are very important, and therefore so is including the
5180 header file everywhere where the defined globals are used.
5182 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
5183 of ASN.1 items, but that structure is a bit different.
5185 The largest change is in util/mkdef.pl which has been enhanced with
5186 better and easier to understand logic to choose which symbols should
5187 go into the Windows .def files as well as a number of fixes and code
5188 cleanup (among others, algorithm keywords are now sorted
5189 lexicographically to avoid constant rewrites).
5192 *) In BN_div() keep a copy of the sign of 'num' before writing the
5193 result to 'rm' because if rm==num the value will be overwritten
5194 and produce the wrong result if 'num' is negative: this caused
5195 problems with BN_mod() and BN_nnmod().
5198 *) Function OCSP_request_verify(). This checks the signature on an
5199 OCSP request and verifies the signer certificate. The signer
5200 certificate is just checked for a generic purpose and OCSP request
5204 *) Add OCSP_check_validity() function to check the validity of OCSP
5205 responses. OCSP responses are prepared in real time and may only
5206 be a few seconds old. Simply checking that the current time lies
5207 between thisUpdate and nextUpdate max reject otherwise valid responses
5208 caused by either OCSP responder or client clock inaccuracy. Instead
5209 we allow thisUpdate and nextUpdate to fall within a certain period of
5210 the current time. The age of the response can also optionally be
5211 checked. Two new options -validity_period and -status_age added to
5215 *) If signature or public key algorithm is unrecognized print out its
5216 OID rather that just UNKNOWN.
5219 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
5220 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
5221 ID to be generated from the issuer certificate alone which can then be
5222 passed to OCSP_id_issuer_cmp().
5225 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
5226 ASN1 modules to export functions returning ASN1_ITEM pointers
5227 instead of the ASN1_ITEM structures themselves. This adds several
5228 new macros which allow the underlying ASN1 function/structure to
5229 be accessed transparently. As a result code should not use ASN1_ITEM
5230 references directly (such as &X509_it) but instead use the relevant
5231 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
5232 use of the new ASN1 code on platforms where exporting structures
5233 is problematical (for example in shared libraries) but exporting
5234 functions returning pointers to structures is not.
5237 *) Add support for overriding the generation of SSL/TLS session IDs.
5238 These callbacks can be registered either in an SSL_CTX or per SSL.
5239 The purpose of this is to allow applications to control, if they wish,
5240 the arbitrary values chosen for use as session IDs, particularly as it
5241 can be useful for session caching in multiple-server environments. A
5242 command-line switch for testing this (and any client code that wishes
5243 to use such a feature) has been added to "s_server".
5244 [Geoff Thorpe, Lutz Jaenicke]
5246 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
5247 of the form '#if defined(...) || defined(...) || ...' and
5248 '#if !defined(...) && !defined(...) && ...'. This also avoids
5249 the growing number of special cases it was previously handling.
5252 *) Make all configuration macros available for application by making
5253 sure they are available in opensslconf.h, by giving them names starting
5254 with "OPENSSL_" to avoid conflicts with other packages and by making
5255 sure e_os2.h will cover all platform-specific cases together with
5257 Additionally, it is now possible to define configuration/platform-
5258 specific names (called "system identities"). In the C code, these
5259 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
5260 macro with the name beginning with "OPENSSL_SYS_", which is determined
5261 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
5265 *) New option -set_serial to 'req' and 'x509' this allows the serial
5266 number to use to be specified on the command line. Previously self
5267 signed certificates were hard coded with serial number 0 and the
5268 CA options of 'x509' had to use a serial number in a file which was
5272 *) New options to 'ca' utility to support V2 CRL entry extensions.
5273 Currently CRL reason, invalidity date and hold instruction are
5274 supported. Add new CRL extensions to V3 code and some new objects.
5277 *) New function EVP_CIPHER_CTX_set_padding() this is used to
5278 disable standard block padding (aka PKCS#5 padding) in the EVP
5279 API, which was previously mandatory. This means that the data is
5280 not padded in any way and so the total length much be a multiple
5281 of the block size, otherwise an error occurs.
5284 *) Initial (incomplete) OCSP SSL support.
5287 *) New function OCSP_parse_url(). This splits up a URL into its host,
5288 port and path components: primarily to parse OCSP URLs. New -url
5289 option to ocsp utility.
5292 *) New nonce behavior. The return value of OCSP_check_nonce() now
5293 reflects the various checks performed. Applications can decide
5294 whether to tolerate certain situations such as an absent nonce
5295 in a response when one was present in a request: the ocsp application
5296 just prints out a warning. New function OCSP_add1_basic_nonce()
5297 this is to allow responders to include a nonce in a response even if
5298 the request is nonce-less.
5301 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
5302 skipped when using openssl x509 multiple times on a single input file,
5303 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
5306 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
5307 set string type: to handle setting ASN1_TIME structures. Fix ca
5308 utility to correctly initialize revocation date of CRLs.
5311 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
5312 the clients preferred ciphersuites and rather use its own preferences.
5313 Should help to work around M$ SGC (Server Gated Cryptography) bug in
5314 Internet Explorer by ensuring unchanged hash method during stepup.
5315 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
5318 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
5319 to aes and add a new 'exist' option to print out symbols that don't
5323 *) Additional options to ocsp utility to allow flags to be set and
5324 additional certificates supplied.
5327 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
5328 OCSP client a number of certificate to only verify the response
5332 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
5333 handle the new API. Currently only ECB, CBC modes supported. Add new
5336 Add TLS AES ciphersuites as described in RFC3268, "Advanced
5337 Encryption Standard (AES) Ciphersuites for Transport Layer
5338 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
5339 not enabled by default and were not part of the "ALL" ciphersuite
5340 alias because they were not yet official; they could be
5341 explicitly requested by specifying the "AESdraft" ciphersuite
5342 group alias. In the final release of OpenSSL 0.9.7, the group
5343 alias is called "AES" and is part of "ALL".)
5344 [Ben Laurie, Steve Henson, Bodo Moeller]
5346 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
5347 request to response.
5350 *) Functions for OCSP responders. OCSP_request_onereq_count(),
5351 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
5352 extract information from a certificate request. OCSP_response_create()
5353 creates a response and optionally adds a basic response structure.
5354 OCSP_basic_add1_status() adds a complete single response to a basic
5355 response and returns the OCSP_SINGLERESP structure just added (to allow
5356 extensions to be included for example). OCSP_basic_add1_cert() adds a
5357 certificate to a basic response and OCSP_basic_sign() signs a basic
5358 response with various flags. New helper functions ASN1_TIME_check()
5359 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
5360 (converts ASN1_TIME to GeneralizedTime).
5363 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
5364 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
5365 structure from a certificate. X509_pubkey_digest() digests the public_key
5366 contents: this is used in various key identifiers.
5369 *) Make sk_sort() tolerate a NULL argument.
5370 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
5372 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
5373 passed by the function are trusted implicitly. If any of them signed the
5374 response then it is assumed to be valid and is not verified.
5377 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
5378 to data. This was previously part of the PKCS7 ASN1 code. This
5379 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
5380 [Steve Henson, reported by Kenneth R. Robinette
5381 <support@securenetterm.com>]
5383 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
5384 routines: without these tracing memory leaks is very painful.
5385 Fix leaks in PKCS12 and PKCS7 routines.
5388 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
5389 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
5390 effectively meant GeneralizedTime would never be used. Now it
5391 is initialised to -1 but X509_time_adj() now has to check the value
5392 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
5393 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
5394 [Steve Henson, reported by Kenneth R. Robinette
5395 <support@securenetterm.com>]
5397 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
5398 result in a zero length in the ASN1_INTEGER structure which was
5399 not consistent with the structure when d2i_ASN1_INTEGER() was used
5400 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
5401 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
5402 where it did not print out a minus for negative ASN1_INTEGER.
5405 *) Add summary printout to ocsp utility. The various functions which
5406 convert status values to strings have been renamed to:
5407 OCSP_response_status_str(), OCSP_cert_status_str() and
5408 OCSP_crl_reason_str() and are no longer static. New options
5409 to verify nonce values and to disable verification. OCSP response
5410 printout format cleaned up.
5413 *) Add additional OCSP certificate checks. These are those specified
5414 in RFC2560. This consists of two separate checks: the CA of the
5415 certificate being checked must either be the OCSP signer certificate
5416 or the issuer of the OCSP signer certificate. In the latter case the
5417 OCSP signer certificate must contain the OCSP signing extended key
5418 usage. This check is performed by attempting to match the OCSP
5419 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
5420 in the OCSP_CERTID structures of the response.
5423 *) Initial OCSP certificate verification added to OCSP_basic_verify()
5424 and related routines. This uses the standard OpenSSL certificate
5425 verify routines to perform initial checks (just CA validity) and
5426 to obtain the certificate chain. Then additional checks will be
5427 performed on the chain. Currently the root CA is checked to see
5428 if it is explicitly trusted for OCSP signing. This is used to set
5429 a root CA as a global signing root: that is any certificate that
5430 chains to that CA is an acceptable OCSP signing certificate.
5433 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
5434 extensions from a separate configuration file.
5435 As when reading extensions from the main configuration file,
5436 the '-extensions ...' option may be used for specifying the
5438 [Massimiliano Pala <madwolf@comune.modena.it>]
5440 *) New OCSP utility. Allows OCSP requests to be generated or
5441 read. The request can be sent to a responder and the output
5442 parsed, outputed or printed in text form. Not complete yet:
5443 still needs to check the OCSP response validity.
5446 *) New subcommands for 'openssl ca':
5447 'openssl ca -status <serial>' prints the status of the cert with
5448 the given serial number (according to the index file).
5449 'openssl ca -updatedb' updates the expiry status of certificates
5451 [Massimiliano Pala <madwolf@comune.modena.it>]
5453 *) New '-newreq-nodes' command option to CA.pl. This is like
5454 '-newreq', but calls 'openssl req' with the '-nodes' option
5455 so that the resulting key is not encrypted.
5456 [Damien Miller <djm@mindrot.org>]
5458 *) New configuration for the GNU Hurd.
5459 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
5461 *) Initial code to implement OCSP basic response verify. This
5462 is currently incomplete. Currently just finds the signer's
5463 certificate and verifies the signature on the response.
5466 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
5467 value of OPENSSLDIR. This is available via the new '-d' option
5468 to 'openssl version', and is also included in 'openssl version -a'.
5471 *) Allowing defining memory allocation callbacks that will be given
5472 file name and line number information in additional arguments
5473 (a const char* and an int). The basic functionality remains, as
5474 well as the original possibility to just replace malloc(),
5475 realloc() and free() by functions that do not know about these
5476 additional arguments. To register and find out the current
5477 settings for extended allocation functions, the following
5478 functions are provided:
5480 CRYPTO_set_mem_ex_functions
5481 CRYPTO_set_locked_mem_ex_functions
5482 CRYPTO_get_mem_ex_functions
5483 CRYPTO_get_locked_mem_ex_functions
5485 These work the same way as CRYPTO_set_mem_functions and friends.
5486 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
5487 extended allocation function is enabled.
5488 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
5489 a conventional allocation function is enabled.
5490 [Richard Levitte, Bodo Moeller]
5492 *) Finish off removing the remaining LHASH function pointer casts.
5493 There should no longer be any prototype-casting required when using
5494 the LHASH abstraction, and any casts that remain are "bugs". See
5495 the callback types and macros at the head of lhash.h for details
5496 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
5499 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
5500 If /dev/[u]random devices are not available or do not return enough
5501 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
5503 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
5504 /etc/entropy will be queried once each in this sequence, quering stops
5505 when enough entropy was collected without querying more sockets.
5508 *) Change the Unix RAND_poll() variant to be able to poll several
5509 random devices, as specified by DEVRANDOM, until a sufficient amount
5510 of data has been collected. We spend at most 10 ms on each file
5511 (select timeout) and read in non-blocking mode. DEVRANDOM now
5512 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
5513 (previously it was just the string "/dev/urandom"), so on typical
5514 platforms the 10 ms delay will never occur.
5515 Also separate out the Unix variant to its own file, rand_unix.c.
5516 For VMS, there's a currently-empty rand_vms.c.
5519 *) Move OCSP client related routines to ocsp_cl.c. These
5520 provide utility functions which an application needing
5521 to issue a request to an OCSP responder and analyse the
5522 response will typically need: as opposed to those which an
5523 OCSP responder itself would need which will be added later.
5525 OCSP_request_sign() signs an OCSP request with an API similar
5526 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
5527 response. OCSP_response_get1_basic() extracts basic response
5528 from response. OCSP_resp_find_status(): finds and extracts status
5529 information from an OCSP_CERTID structure (which will be created
5530 when the request structure is built). These are built from lower
5531 level functions which work on OCSP_SINGLERESP structures but
5532 wont normally be used unless the application wishes to examine
5533 extensions in the OCSP response for example.
5535 Replace nonce routines with a pair of functions.
5536 OCSP_request_add1_nonce() adds a nonce value and optionally
5537 generates a random value. OCSP_check_nonce() checks the
5538 validity of the nonce in an OCSP response.
5541 *) Change function OCSP_request_add() to OCSP_request_add0_id().
5542 This doesn't copy the supplied OCSP_CERTID and avoids the
5543 need to free up the newly created id. Change return type
5544 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
5545 This can then be used to add extensions to the request.
5546 Deleted OCSP_request_new(), since most of its functionality
5547 is now in OCSP_REQUEST_new() (and the case insensitive name
5548 clash) apart from the ability to set the request name which
5549 will be added elsewhere.
5552 *) Update OCSP API. Remove obsolete extensions argument from
5553 various functions. Extensions are now handled using the new
5554 OCSP extension code. New simple OCSP HTTP function which
5555 can be used to send requests and parse the response.
5558 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
5559 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
5560 uses the special reorder version of SET OF to sort the attributes
5561 and reorder them to match the encoded order. This resolves a long
5562 standing problem: a verify on a PKCS7 structure just after signing
5563 it used to fail because the attribute order did not match the
5564 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
5565 it uses the received order. This is necessary to tolerate some broken
5566 software that does not order SET OF. This is handled by encoding
5567 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
5568 to produce the required SET OF.
5571 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
5572 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
5573 files to get correct declarations of the ASN.1 item variables.
5576 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
5577 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
5578 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
5579 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
5580 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
5581 ASN1_ITEM and no wrapper functions.
5584 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
5585 replace the old function pointer based I/O routines. Change most of
5586 the *_d2i_bio() and *_d2i_fp() functions to use these.
5589 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
5590 lines, recognice more "algorithms" that can be deselected, and make
5591 it complain about algorithm deselection that isn't recognised.
5594 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
5595 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
5596 to use new functions. Add NO_ASN1_OLD which can be set to remove
5597 some old style ASN1 functions: this can be used to determine if old
5598 code will still work when these eventually go away.
5601 *) New extension functions for OCSP structures, these follow the
5602 same conventions as certificates and CRLs.
5605 *) New function X509V3_add1_i2d(). This automatically encodes and
5606 adds an extension. Its behaviour can be customised with various
5607 flags to append, replace or delete. Various wrappers added for
5608 certifcates and CRLs.
5611 *) Fix to avoid calling the underlying ASN1 print routine when
5612 an extension cannot be parsed. Correct a typo in the
5613 OCSP_SERVICELOC extension. Tidy up print OCSP format.
5616 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
5617 entries for variables.
5620 *) Add functionality to apps/openssl.c for detecting locking
5621 problems: As the program is single-threaded, all we have
5622 to do is register a locking callback using an array for
5623 storing which locks are currently held by the program.
5626 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
5627 SSL_get_ex_data_X509_STORE_idx(), which is used in
5628 ssl_verify_cert_chain() and thus can be called at any time
5629 during TLS/SSL handshakes so that thread-safety is essential.
5630 Unfortunately, the ex_data design is not at all suited
5631 for multi-threaded use, so it probably should be abolished.
5634 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5635 [Broadcom, tweaked and integrated by Geoff Thorpe]
5637 *) Move common extension printing code to new function
5638 X509V3_print_extensions(). Reorganise OCSP print routines and
5639 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5642 *) New function X509_signature_print() to remove duplication in some
5646 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5647 set (this was treated exactly the same as SET OF previously). This
5648 is used to reorder the STACK representing the structure to match the
5649 encoding. This will be used to get round a problem where a PKCS7
5650 structure which was signed could not be verified because the STACK
5651 order did not reflect the encoded order.
5654 *) Reimplement the OCSP ASN1 module using the new code.
5657 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5658 for its ASN1 operations. The old style function pointers still exist
5659 for now but they will eventually go away.
5662 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
5663 completely replaces the old ASN1 functionality with a table driven
5664 encoder and decoder which interprets an ASN1_ITEM structure describing
5665 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
5666 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
5667 has also been converted to the new form.
5670 *) Change BN_mod_exp_recp so that negative moduli are tolerated
5671 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
5672 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
5673 for negative moduli.
5676 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
5677 of not touching the result's sign bit.
5680 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
5684 *) Changed the LHASH code to use prototypes for callbacks, and created
5685 macros to declare and implement thin (optionally static) functions
5686 that provide type-safety and avoid function pointer casting for the
5687 type-specific callbacks.
5690 *) Added Kerberos Cipher Suites to be used with TLS, as written in
5692 [Veers Staats <staatsvr@asc.hpc.mil>,
5693 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
5695 *) Reformat the FAQ so the different questions and answers can be divided
5696 in sections depending on the subject.
5699 *) Have the zlib compression code load ZLIB.DLL dynamically under
5703 *) New function BN_mod_sqrt for computing square roots modulo a prime
5704 (using the probabilistic Tonelli-Shanks algorithm unless
5705 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
5706 be handled deterministically).
5707 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5709 *) Make BN_mod_inverse faster by explicitly handling small quotients
5710 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
5711 512 bits], about 30% for larger ones [1024 or 2048 bits].)
5714 *) New function BN_kronecker.
5717 *) Fix BN_gcd so that it works on negative inputs; the result is
5718 positive unless both parameters are zero.
5719 Previously something reasonably close to an infinite loop was
5720 possible because numbers could be growing instead of shrinking
5721 in the implementation of Euclid's algorithm.
5724 *) Fix BN_is_word() and BN_is_one() macros to take into account the
5725 sign of the number in question.
5727 Fix BN_is_word(a,w) to work correctly for w == 0.
5729 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
5730 because its test if the absolute value of 'a' equals 'w'.
5731 Note that BN_abs_is_word does *not* handle w == 0 reliably;
5732 it exists mostly for use in the implementations of BN_is_zero(),
5733 BN_is_one(), and BN_is_word().
5736 *) New function BN_swap.
5739 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
5740 the exponentiation functions are more likely to produce reasonable
5741 results on negative inputs.
5744 *) Change BN_mod_mul so that the result is always non-negative.
5745 Previously, it could be negative if one of the factors was negative;
5746 I don't think anyone really wanted that behaviour.
5749 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
5750 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
5751 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
5752 and add new functions:
5761 BN_mod_lshift1_quick
5765 These functions always generate non-negative results.
5767 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
5768 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
5770 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
5771 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
5772 be reduced modulo m.
5773 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5776 The following entry accidentily appeared in the CHANGES file
5777 distributed with OpenSSL 0.9.7. The modifications described in
5778 it do *not* apply to OpenSSL 0.9.7.
5780 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5781 was actually never needed) and in BN_mul(). The removal in BN_mul()
5782 required a small change in bn_mul_part_recursive() and the addition
5783 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5784 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5785 bn_sub_words() and bn_add_words() except they take arrays with
5790 *) In 'openssl passwd', verify passwords read from the terminal
5791 unless the '-salt' option is used (which usually means that
5792 verification would just waste user's time since the resulting
5793 hash is going to be compared with some given password hash)
5794 or the new '-noverify' option is used.
5796 This is an incompatible change, but it does not affect
5797 non-interactive use of 'openssl passwd' (passwords on the command
5798 line, '-stdin' option, '-in ...' option) and thus should not
5802 *) Remove all references to RSAref, since there's no more need for it.
5805 *) Make DSO load along a path given through an environment variable
5806 (SHLIB_PATH) with shl_load().
5809 *) Constify the ENGINE code as a result of BIGNUM constification.
5810 Also constify the RSA code and most things related to it. In a
5811 few places, most notable in the depth of the ASN.1 code, ugly
5812 casts back to non-const were required (to be solved at a later
5816 *) Make it so the openssl application has all engines loaded by default.
5819 *) Constify the BIGNUM routines a little more.
5822 *) Add the following functions:
5824 ENGINE_load_cswift()
5826 ENGINE_load_atalla()
5828 ENGINE_load_builtin_engines()
5830 That way, an application can itself choose if external engines that
5831 are built-in in OpenSSL shall ever be used or not. The benefit is
5832 that applications won't have to be linked with libdl or other dso
5833 libraries unless it's really needed.
5835 Changed 'openssl engine' to load all engines on demand.
5836 Changed the engine header files to avoid the duplication of some
5837 declarations (they differed!).
5840 *) 'openssl engine' can now list capabilities.
5843 *) Better error reporting in 'openssl engine'.
5846 *) Never call load_dh_param(NULL) in s_server.
5849 *) Add engine application. It can currently list engines by name and
5850 identity, and test if they are actually available.
5853 *) Improve RPM specification file by forcing symbolic linking and making
5854 sure the installed documentation is also owned by root.root.
5855 [Damien Miller <djm@mindrot.org>]
5857 *) Give the OpenSSL applications more possibilities to make use of
5858 keys (public as well as private) handled by engines.
5861 *) Add OCSP code that comes from CertCo.
5864 *) Add VMS support for the Rijndael code.
5867 *) Added untested support for Nuron crypto accelerator.
5870 *) Add support for external cryptographic devices. This code was
5871 previously distributed separately as the "engine" branch.
5872 [Geoff Thorpe, Richard Levitte]
5874 *) Rework the filename-translation in the DSO code. It is now possible to
5875 have far greater control over how a "name" is turned into a filename
5876 depending on the operating environment and any oddities about the
5877 different shared library filenames on each system.
5880 *) Support threads on FreeBSD-elf in Configure.
5883 *) Fix for SHA1 assembly problem with MASM: it produces
5884 warnings about corrupt line number information when assembling
5885 with debugging information. This is caused by the overlapping
5887 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
5890 NCONF_get_number() has no error checking at all. As a replacement,
5891 NCONF_get_number_e() is defined (_e for "error checking") and is
5892 promoted strongly. The old NCONF_get_number is kept around for
5893 binary backward compatibility.
5894 Make it possible for methods to load from something other than a BIO,
5895 by providing a function pointer that is given a name instead of a BIO.
5896 For example, this could be used to load configuration data from an
5900 *) Fix for non blocking accept BIOs. Added new I/O special reason
5901 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
5902 with non blocking I/O was not possible because no retry code was
5903 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
5907 *) Added the beginnings of Rijndael support.
5910 *) Fix for bug in DirectoryString mask setting. Add support for
5911 X509_NAME_print_ex() in 'req' and X509_print_ex() function
5912 to allow certificate printing to more controllable, additional
5913 'certopt' option to 'x509' to allow new printing options to be
5917 *) Clean old EAY MD5 hack from e_os.h.
5920 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
5922 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5923 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5924 [Joe Orton, Steve Henson]
5926 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
5928 *) Fix additional bug revealed by the NISCC test suite:
5930 Stop bug triggering large recursion when presented with
5931 certain ASN.1 tags (CVE-2003-0851)
5934 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
5936 *) Fix various bugs revealed by running the NISCC test suite:
5938 Stop out of bounds reads in the ASN1 code when presented with
5939 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5941 If verify callback ignores invalid public key errors don't try to check
5942 certificate signature with the NULL public key.
5946 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5947 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5951 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5952 extra data after the compression methods not only for TLS 1.0
5953 but also for SSL 3.0 (as required by the specification).
5954 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5956 *) Change X509_certificate_type() to mark the key as exported/exportable
5957 when it's 512 *bits* long, not 512 bytes.
5960 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
5962 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5963 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5964 a protocol version number mismatch like a decryption error
5965 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5968 *) Turn on RSA blinding by default in the default implementation
5969 to avoid a timing attack. Applications that don't want it can call
5970 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5971 They would be ill-advised to do so in most cases.
5972 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5974 *) Change RSA blinding code so that it works when the PRNG is not
5975 seeded (in this case, the secret RSA exponent is abused as
5976 an unpredictable seed -- if it is not unpredictable, there
5977 is no point in blinding anyway). Make RSA blinding thread-safe
5978 by remembering the creator's thread ID in rsa->blinding and
5979 having all other threads use local one-time blinding factors
5980 (this requires more computation than sharing rsa->blinding, but
5981 avoids excessive locking; and if an RSA object is not shared
5982 between threads, blinding will still be very fast).
5985 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
5987 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5988 via timing by performing a MAC computation even if incorrrect
5989 block cipher padding has been found. This is a countermeasure
5990 against active attacks where the attacker has to distinguish
5991 between bad padding and a MAC verification error. (CVE-2003-0078)
5993 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5994 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5995 Martin Vuagnoux (EPFL, Ilion)]
5997 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
5999 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
6000 memory from it's contents. This is done with a counter that will
6001 place alternating values in each byte. This can be used to solve
6002 two issues: 1) the removal of calls to memset() by highly optimizing
6003 compilers, and 2) cleansing with other values than 0, since those can
6004 be read through on certain media, for example a swap space on disk.
6007 *) Bugfix: client side session caching did not work with external caching,
6008 because the session->cipher setting was not restored when reloading
6009 from the external cache. This problem was masked, when
6010 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
6011 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
6014 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
6015 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
6016 [Zeev Lieber <zeev-l@yahoo.com>]
6018 *) Undo an undocumented change introduced in 0.9.6e which caused
6019 repeated calls to OpenSSL_add_all_ciphers() and
6020 OpenSSL_add_all_digests() to be ignored, even after calling
6024 *) Change the default configuration reader to deal with last line not
6025 being properly terminated.
6028 *) Change X509_NAME_cmp() so it applies the special rules on handling
6029 DN values that are of type PrintableString, as well as RDNs of type
6030 emailAddress where the value has the type ia5String.
6031 [stefank@valicert.com via Richard Levitte]
6033 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
6034 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
6035 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
6036 the bitwise-OR of the two for use by the majority of applications
6037 wanting this behaviour, and update the docs. The documented
6038 behaviour and actual behaviour were inconsistent and had been
6039 changing anyway, so this is more a bug-fix than a behavioural
6041 [Geoff Thorpe, diagnosed by Nadav Har'El]
6043 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
6044 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
6047 *) Fix initialization code race conditions in
6048 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
6049 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
6050 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
6051 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
6052 ssl2_get_cipher_by_char(),
6053 ssl3_get_cipher_by_char().
6054 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
6056 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
6057 the cached sessions are flushed, as the remove_cb() might use ex_data
6058 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
6059 (see [openssl.org #212]).
6060 [Geoff Thorpe, Lutz Jaenicke]
6062 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
6063 length, instead of the encoding length to d2i_ASN1_OBJECT.
6066 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
6068 *) [In 0.9.6g-engine release:]
6069 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
6070 [Lynn Gazis <lgazis@rainbow.com>]
6072 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
6074 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
6075 and get fix the header length calculation.
6076 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
6077 Alon Kantor <alonk@checkpoint.com> (and others),
6080 *) Use proper error handling instead of 'assertions' in buffer
6081 overflow checks added in 0.9.6e. This prevents DoS (the
6082 assertions could call abort()).
6083 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
6085 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
6087 *) Add various sanity checks to asn1_get_length() to reject
6088 the ASN1 length bytes if they exceed sizeof(long), will appear
6089 negative or the content length exceeds the length of the
6091 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6093 *) Fix cipher selection routines: ciphers without encryption had no flags
6094 for the cipher strength set and where therefore not handled correctly
6095 by the selection routines (PR #130).
6098 *) Fix EVP_dsa_sha macro.
6102 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
6103 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
6104 that was added in OpenSSL 0.9.6d.
6106 As the countermeasure turned out to be incompatible with some
6107 broken SSL implementations, the new option is part of SSL_OP_ALL.
6108 SSL_OP_ALL is usually employed when compatibility with weird SSL
6109 implementations is desired (e.g. '-bugs' option to 's_client' and
6110 's_server'), so the new option is automatically set in many
6114 *) Changes in security patch:
6116 Changes marked "(CHATS)" were sponsored by the Defense Advanced
6117 Research Projects Agency (DARPA) and Air Force Research Laboratory,
6118 Air Force Materiel Command, USAF, under agreement number
6121 *) Add various sanity checks to asn1_get_length() to reject
6122 the ASN1 length bytes if they exceed sizeof(long), will appear
6123 negative or the content length exceeds the length of the
6124 supplied buffer. (CVE-2002-0659)
6125 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6127 *) Assertions for various potential buffer overflows, not known to
6129 [Ben Laurie (CHATS)]
6131 *) Various temporary buffers to hold ASCII versions of integers were
6132 too small for 64 bit platforms. (CVE-2002-0655)
6133 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
6135 *) Remote buffer overflow in SSL3 protocol - an attacker could
6136 supply an oversized session ID to a client. (CVE-2002-0656)
6137 [Ben Laurie (CHATS)]
6139 *) Remote buffer overflow in SSL2 protocol - an attacker could
6140 supply an oversized client master key. (CVE-2002-0656)
6141 [Ben Laurie (CHATS)]
6143 Changes between 0.9.6c and 0.9.6d [9 May 2002]
6145 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
6146 encoded as NULL) with id-dsa-with-sha1.
6147 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
6149 *) Check various X509_...() return values in apps/req.c.
6150 [Nils Larsch <nla@trustcenter.de>]
6152 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
6153 an end-of-file condition would erronously be flagged, when the CRLF
6154 was just at the end of a processed block. The bug was discovered when
6155 processing data through a buffering memory BIO handing the data to a
6156 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
6157 <ptsekov@syntrex.com> and Nedelcho Stanev.
6160 *) Implement a countermeasure against a vulnerability recently found
6161 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
6162 before application data chunks to avoid the use of known IVs
6163 with data potentially chosen by the attacker.
6166 *) Fix length checks in ssl3_get_client_hello().
6169 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
6170 to prevent ssl3_read_internal() from incorrectly assuming that
6171 ssl3_read_bytes() found application data while handshake
6172 processing was enabled when in fact s->s3->in_read_app_data was
6173 merely automatically cleared during the initial handshake.
6174 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
6176 *) Fix object definitions for Private and Enterprise: they were not
6177 recognized in their shortname (=lowercase) representation. Extend
6178 obj_dat.pl to issue an error when using undefined keywords instead
6179 of silently ignoring the problem (Svenning Sorensen
6180 <sss@sss.dnsalias.net>).
6183 *) Fix DH_generate_parameters() so that it works for 'non-standard'
6184 generators, i.e. generators other than 2 and 5. (Previously, the
6185 code did not properly initialise the 'add' and 'rem' values to
6186 BN_generate_prime().)
6188 In the new general case, we do not insist that 'generator' is
6189 actually a primitive root: This requirement is rather pointless;
6190 a generator of the order-q subgroup is just as good, if not
6194 *) Map new X509 verification errors to alerts. Discovered and submitted by
6195 Tom Wu <tom@arcot.com>.
6198 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
6199 returning non-zero before the data has been completely received
6200 when using non-blocking I/O.
6201 [Bodo Moeller; problem pointed out by John Hughes]
6203 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
6204 [Ben Laurie, Lutz Jaenicke]
6206 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
6207 Yoram Zahavi <YoramZ@gilian.com>).
6210 *) Add information about CygWin 1.3 and on, and preserve proper
6211 configuration for the versions before that.
6212 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6214 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
6215 check whether we deal with a copy of a session and do not delete from
6216 the cache in this case. Problem reported by "Izhar Shoshani Levi"
6217 <izhar@checkpoint.com>.
6220 *) Do not store session data into the internal session cache, if it
6221 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
6222 flag is set). Proposed by Aslam <aslam@funk.com>.
6225 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
6229 *) [In 0.9.6d-engine release:]
6230 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6231 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6233 *) Add the configuration target linux-s390x.
6234 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
6236 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
6237 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
6238 variable as an indication that a ClientHello message has been
6239 received. As the flag value will be lost between multiple
6240 invocations of ssl3_accept when using non-blocking I/O, the
6241 function may not be aware that a handshake has actually taken
6242 place, thus preventing a new session from being added to the
6245 To avoid this problem, we now set s->new_session to 2 instead of
6246 using a local variable.
6247 [Lutz Jaenicke, Bodo Moeller]
6249 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
6250 if the SSL_R_LENGTH_MISMATCH error is detected.
6251 [Geoff Thorpe, Bodo Moeller]
6253 *) New 'shared_ldflag' column in Configure platform table.
6256 *) Fix EVP_CIPHER_mode macro.
6257 ["Dan S. Camper" <dan@bti.net>]
6259 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
6260 type, we must throw them away by setting rr->length to 0.
6261 [D P Chang <dpc@qualys.com>]
6263 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
6265 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
6266 <Dominikus.Scherkl@biodata.com>. (The previous implementation
6267 worked incorrectly for those cases where range = 10..._2 and
6268 3*range is two bits longer than range.)
6271 *) Only add signing time to PKCS7 structures if it is not already
6275 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
6276 OBJ_ld_ce should be OBJ_id_ce.
6277 Also some ip-pda OIDs in crypto/objects/objects.txt were
6278 incorrect (cf. RFC 3039).
6279 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
6281 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
6282 returns early because it has nothing to do.
6283 [Andy Schneider <andy.schneider@bjss.co.uk>]
6285 *) [In 0.9.6c-engine release:]
6286 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
6287 [Andy Schneider <andy.schneider@bjss.co.uk>]
6289 *) [In 0.9.6c-engine release:]
6290 Add support for Cryptographic Appliance's keyserver technology.
6291 (Use engine 'keyclient')
6292 [Cryptographic Appliances and Geoff Thorpe]
6294 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
6295 is called via tools/c89.sh because arguments have to be
6296 rearranged (all '-L' options must appear before the first object
6298 [Richard Shapiro <rshapiro@abinitio.com>]
6300 *) [In 0.9.6c-engine release:]
6301 Add support for Broadcom crypto accelerator cards, backported
6303 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
6305 *) [In 0.9.6c-engine release:]
6306 Add support for SureWare crypto accelerator cards from
6307 Baltimore Technologies. (Use engine 'sureware')
6308 [Baltimore Technologies and Mark Cox]
6310 *) [In 0.9.6c-engine release:]
6311 Add support for crypto accelerator cards from Accelerated
6312 Encryption Processing, www.aep.ie. (Use engine 'aep')
6313 [AEP Inc. and Mark Cox]
6315 *) Add a configuration entry for gcc on UnixWare.
6316 [Gary Benson <gbenson@redhat.com>]
6318 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
6319 messages are stored in a single piece (fixed-length part and
6320 variable-length part combined) and fix various bugs found on the way.
6323 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
6324 instead. BIO_gethostbyname() does not know what timeouts are
6325 appropriate, so entries would stay in cache even when they have
6327 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
6329 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
6330 faced with a pathologically small ClientHello fragment that does
6331 not contain client_version: Instead of aborting with an error,
6332 simply choose the highest available protocol version (i.e.,
6333 TLS 1.0 unless it is disabled). In practice, ClientHello
6334 messages are never sent like this, but this change gives us
6335 strictly correct behaviour at least for TLS.
6338 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
6339 never resets s->method to s->ctx->method when called from within
6340 one of the SSL handshake functions.
6341 [Bodo Moeller; problem pointed out by Niko Baric]
6343 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
6344 (sent using the client's version number) if client_version is
6345 smaller than the protocol version in use. Also change
6346 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
6347 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
6348 the client will at least see that alert.
6351 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
6355 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
6356 client receives HelloRequest while in a handshake.
6357 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
6359 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
6360 should end in 'break', not 'goto end' which circuments various
6361 cleanups done in state SSL_ST_OK. But session related stuff
6362 must be disabled for SSL_ST_OK in the case that we just sent a
6365 Also avoid some overhead by not calling ssl_init_wbio_buffer()
6366 before just sending a HelloRequest.
6367 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
6369 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
6370 reveal whether illegal block cipher padding was found or a MAC
6371 verification error occured. (Neither SSLerr() codes nor alerts
6372 are directly visible to potential attackers, but the information
6373 may leak via logfiles.)
6375 Similar changes are not required for the SSL 2.0 implementation
6376 because the number of padding bytes is sent in clear for SSL 2.0,
6377 and the extra bytes are just ignored. However ssl/s2_pkt.c
6378 failed to verify that the purported number of padding bytes is in
6382 *) Add OpenUNIX-8 support including shared libraries
6383 (Boyd Lynn Gerber <gerberb@zenez.com>).
6386 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
6387 'wristwatch attack' using huge encoding parameters (cf.
6388 James H. Manger's CRYPTO 2001 paper). Note that the
6389 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
6390 encoding parameters and hence was not vulnerable.
6393 *) BN_sqr() bug fix.
6394 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
6396 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
6397 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
6398 followed by modular reduction.
6399 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
6401 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
6402 equivalent based on BN_pseudo_rand() instead of BN_rand().
6405 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
6406 This function was broken, as the check for a new client hello message
6407 to handle SGC did not allow these large messages.
6408 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
6411 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
6414 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
6415 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
6418 *) Rework the configuration and shared library support for Tru64 Unix.
6419 The configuration part makes use of modern compiler features and
6420 still retains old compiler behavior for those that run older versions
6421 of the OS. The shared library support part includes a variant that
6422 uses the RPATH feature, and is available through the special
6423 configuration target "alpha-cc-rpath", which will never be selected
6425 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
6427 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
6428 with the same message size as in ssl3_get_certificate_request().
6429 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
6430 messages might inadvertently be reject as too long.
6431 [Petr Lampa <lampa@fee.vutbr.cz>]
6433 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
6436 *) Modified SSL library such that the verify_callback that has been set
6437 specificly for an SSL object with SSL_set_verify() is actually being
6438 used. Before the change, a verify_callback set with this function was
6439 ignored and the verify_callback() set in the SSL_CTX at the time of
6440 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
6441 to allow the necessary settings.
6444 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
6445 explicitly to NULL, as at least on Solaris 8 this seems not always to be
6446 done automatically (in contradiction to the requirements of the C
6447 standard). This made problems when used from OpenSSH.
6450 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
6451 dh->length and always used
6453 BN_rand_range(priv_key, dh->p).
6455 BN_rand_range() is not necessary for Diffie-Hellman, and this
6456 specific range makes Diffie-Hellman unnecessarily inefficient if
6457 dh->length (recommended exponent length) is much smaller than the
6458 length of dh->p. We could use BN_rand_range() if the order of
6459 the subgroup was stored in the DH structure, but we only have
6464 BN_rand(priv_key, l, ...)
6466 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
6472 RSA_eay_public_encrypt
6473 RSA_eay_private_decrypt
6474 RSA_eay_private_encrypt (signing)
6475 RSA_eay_public_decrypt (signature verification)
6477 (default implementations for RSA_public_encrypt,
6478 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
6479 always reject numbers >= n.
6482 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
6483 to synchronize access to 'locking_thread'. This is necessary on
6484 systems where access to 'locking_thread' (an 'unsigned long'
6485 variable) is not atomic.
6488 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
6489 *before* setting the 'crypto_lock_rand' flag. The previous code had
6490 a race condition if 0 is a valid thread ID.
6491 [Travis Vitek <vitek@roguewave.com>]
6493 *) Add support for shared libraries under Irix.
6494 [Albert Chin-A-Young <china@thewrittenword.com>]
6496 *) Add configuration option to build on Linux on both big-endian and
6498 [Ralf Baechle <ralf@uni-koblenz.de>]
6500 *) Add the possibility to create shared libraries on HP-UX.
6503 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
6505 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
6506 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
6507 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
6508 PRNG state recovery was possible based on the output of
6509 one PRNG request appropriately sized to gain knowledge on
6510 'md' followed by enough consecutive 1-byte PRNG requests
6511 to traverse all of 'state'.
6513 1. When updating 'md_local' (the current thread's copy of 'md')
6514 during PRNG output generation, hash all of the previous
6515 'md_local' value, not just the half used for PRNG output.
6517 2. Make the number of bytes from 'state' included into the hash
6518 independent from the number of PRNG bytes requested.
6520 The first measure alone would be sufficient to avoid
6521 Markku-Juhani's attack. (Actually it had never occurred
6522 to me that the half of 'md_local' used for chaining was the
6523 half from which PRNG output bytes were taken -- I had always
6524 assumed that the secret half would be used.) The second
6525 measure makes sure that additional data from 'state' is never
6526 mixed into 'md_local' in small portions; this heuristically
6527 further strengthens the PRNG.
6530 *) Fix crypto/bn/asm/mips3.s.
6533 *) When only the key is given to "enc", the IV is undefined. Print out
6534 an error message in this case.
6537 *) Handle special case when X509_NAME is empty in X509 printing routines.
6540 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
6541 positive and less than q.
6544 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
6545 used: it isn't thread safe and the add_lock_callback should handle
6547 [Paul Rose <Paul.Rose@bridge.com>]
6549 *) Verify that incoming data obeys the block size in
6550 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
6554 [Ulf Möller, Bodo Möller]
6556 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
6557 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
6558 when fixing the server behaviour for backwards-compatible 'client
6559 hello' messages. (Note that the attack is impractical against
6560 SSL 3.0 and TLS 1.0 anyway because length and version checking
6561 means that the probability of guessing a valid ciphertext is
6562 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
6565 Before 0.9.5, the countermeasure (hide the error by generating a
6566 random 'decryption result') did not work properly because
6567 ERR_clear_error() was missing, meaning that SSL_get_error() would
6568 detect the supposedly ignored error.
6570 Both problems are now fixed.
6573 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
6574 (previously it was 1024).
6577 *) Fix for compatibility mode trust settings: ignore trust settings
6578 unless some valid trust or reject settings are present.
6581 *) Fix for blowfish EVP: its a variable length cipher.
6584 *) Fix various bugs related to DSA S/MIME verification. Handle missing
6585 parameters in DSA public key structures and return an error in the
6586 DSA routines if parameters are absent.
6589 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
6590 in the current directory if neither $RANDFILE nor $HOME was set.
6591 RAND_file_name() in 0.9.6a returned NULL in this case. This has
6592 caused some confusion to Windows users who haven't defined $HOME.
6593 Thus RAND_file_name() is changed again: e_os.h can define a
6594 DEFAULT_HOME, which will be used if $HOME is not set.
6595 For Windows, we use "C:"; on other platforms, we still require
6596 environment variables.
6598 *) Move 'if (!initialized) RAND_poll()' into regions protected by
6599 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
6600 having multiple threads call RAND_poll() concurrently.
6603 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
6604 combination of a flag and a thread ID variable.
6605 Otherwise while one thread is in ssleay_rand_bytes (which sets the
6606 flag), *other* threads can enter ssleay_add_bytes without obeying
6607 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
6608 that they do not hold after the first thread unsets add_do_not_lock).
6611 *) Change bctest again: '-x' expressions are not available in all
6615 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
6617 *) Fix a couple of memory leaks in PKCS7_dataDecode()
6618 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
6620 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
6621 the default extension for executables, if any. Also, make the perl
6622 scripts that use symlink() to test if it really exists and use "cp"
6623 if it doesn't. All this made OpenSSL compilable and installable in
6627 *) Fix for asn1_GetSequence() for indefinite length constructed data.
6628 If SEQUENCE is length is indefinite just set c->slen to the total
6629 amount of data available.
6630 [Steve Henson, reported by shige@FreeBSD.org]
6631 [This change does not apply to 0.9.7.]
6633 *) Change bctest to avoid here-documents inside command substitution
6634 (workaround for FreeBSD /bin/sh bug).
6635 For compatibility with Ultrix, avoid shell functions (introduced
6636 in the bctest version that searches along $PATH).
6639 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6640 with des_encrypt() defined on some operating systems, like Solaris
6644 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6645 On the Importance of Eliminating Errors in Cryptographic
6646 Computations, J. Cryptology 14 (2001) 2, 101-119,
6647 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6650 *) MIPS assembler BIGNUM division bug fix.
6653 *) Disabled incorrect Alpha assembler code.
6656 *) Fix PKCS#7 decode routines so they correctly update the length
6657 after reading an EOC for the EXPLICIT tag.
6659 [This change does not apply to 0.9.7.]
6661 *) Fix bug in PKCS#12 key generation routines. This was triggered
6662 if a 3DES key was generated with a 0 initial byte. Include
6663 PKCS12_BROKEN_KEYGEN compilation option to retain the old
6664 (but broken) behaviour.
6667 *) Enhance bctest to search for a working bc along $PATH and print
6669 [Tim Rice <tim@multitalents.net> via Richard Levitte]
6671 *) Fix memory leaks in err.c: free err_data string if necessary;
6672 don't write to the wrong index in ERR_set_error_data.
6675 *) Implement ssl23_peek (analogous to ssl23_read), which previously
6679 *) Replace rdtsc with _emit statements for VC++ version 5.
6680 [Jeremy Cooper <jeremy@baymoo.org>]
6682 *) Make it possible to reuse SSLv2 sessions.
6685 *) In copy_email() check for >= 0 as a return value for
6686 X509_NAME_get_index_by_NID() since 0 is a valid index.
6687 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
6689 *) Avoid coredump with unsupported or invalid public keys by checking if
6690 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
6691 PKCS7_verify() fails with non detached data.
6694 *) Don't use getenv in library functions when run as setuid/setgid.
6695 New function OPENSSL_issetugid().
6698 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
6699 due to incorrect handling of multi-threading:
6701 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
6703 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
6705 3. Count how many times MemCheck_off() has been called so that
6706 nested use can be treated correctly. This also avoids
6707 inband-signalling in the previous code (which relied on the
6708 assumption that thread ID 0 is impossible).
6711 *) Add "-rand" option also to s_client and s_server.
6714 *) Fix CPU detection on Irix 6.x.
6715 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
6716 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6718 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
6721 [This change does not apply to 0.9.7.]
6723 *) Use the cached encoding of an X509_NAME structure rather than
6724 copying it. This is apparently the reason for the libsafe "errors"
6725 but the code is actually correct.
6728 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
6729 Bleichenbacher's DSA attack.
6730 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
6731 to be set and top=0 forces the highest bit to be set; top=-1 is new
6732 and leaves the highest bit random.
6733 [Ulf Moeller, Bodo Moeller]
6735 *) In the NCONF_...-based implementations for CONF_... queries
6736 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
6737 a temporary CONF structure with the data component set to NULL
6738 (which gives segmentation faults in lh_retrieve).
6739 Instead, use NULL for the CONF pointer in CONF_get_string and
6740 CONF_get_number (which may use environment variables) and directly
6741 return NULL from CONF_get_section.
6744 *) Fix potential buffer overrun for EBCDIC.
6747 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
6748 keyUsage if basicConstraints absent for a CA.
6751 *) Make SMIME_write_PKCS7() write mail header values with a format that
6752 is more generally accepted (no spaces before the semicolon), since
6753 some programs can't parse those values properly otherwise. Also make
6754 sure BIO's that break lines after each write do not create invalid
6758 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
6759 macros previously used would not encode an empty SEQUENCE OF
6760 and break the signature.
6762 [This change does not apply to 0.9.7.]
6764 *) Zero the premaster secret after deriving the master secret in
6768 *) Add some EVP_add_digest_alias registrations (as found in
6769 OpenSSL_add_all_digests()) to SSL_library_init()
6770 aka OpenSSL_add_ssl_algorithms(). This provides improved
6771 compatibility with peers using X.509 certificates
6772 with unconventional AlgorithmIdentifier OIDs.
6775 *) Fix for Irix with NO_ASM.
6776 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6778 *) ./config script fixes.
6779 [Ulf Moeller, Richard Levitte]
6781 *) Fix 'openssl passwd -1'.
6784 *) Change PKCS12_key_gen_asc() so it can cope with non null
6785 terminated strings whose length is passed in the passlen
6786 parameter, for example from PEM callbacks. This was done
6787 by adding an extra length parameter to asc2uni().
6788 [Steve Henson, reported by <oddissey@samsung.co.kr>]
6790 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
6791 call failed, free the DSA structure.
6794 *) Fix to uni2asc() to cope with zero length Unicode strings.
6795 These are present in some PKCS#12 files.
6798 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
6799 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
6800 when writing a 32767 byte record.
6801 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
6803 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
6804 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
6806 (RSA objects have a reference count access to which is protected
6807 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
6808 so they are meant to be shared between threads.)
6809 [Bodo Moeller, Geoff Thorpe; original patch submitted by
6810 "Reddie, Steven" <Steven.Reddie@ca.com>]
6812 *) Fix a deadlock in CRYPTO_mem_leaks().
6815 *) Use better test patterns in bntest.
6818 *) rand_win.c fix for Borland C.
6821 *) BN_rshift bugfix for n == 0.
6824 *) Add a 'bctest' script that checks for some known 'bc' bugs
6825 so that 'make test' does not abort just because 'bc' is broken.
6828 *) Store verify_result within SSL_SESSION also for client side to
6829 avoid potential security hole. (Re-used sessions on the client side
6830 always resulted in verify_result==X509_V_OK, not using the original
6831 result of the server certificate verification.)
6834 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
6835 SSL3_RT_APPLICATION_DATA, return 0.
6836 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
6840 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
6841 releases, have been re-implemented by renaming the previous
6842 implementations of ssl2_read and ssl3_read to ssl2_read_internal
6843 and ssl3_read_internal, respectively, and adding 'peek' parameters
6844 to them. The new ssl[23]_{read,peek} functions are calls to
6845 ssl[23]_read_internal with the 'peek' flag set appropriately.
6846 A 'peek' parameter has also been added to ssl3_read_bytes, which
6847 does the actual work for ssl3_read_internal.
6850 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
6851 the method-specific "init()" handler. Also clean up ex_data after
6852 calling the method-specific "finish()" handler. Previously, this was
6853 happening the other way round.
6856 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
6857 The previous value, 12, was not always sufficient for BN_mod_exp().
6860 *) Make sure that shared libraries get the internal name engine with
6861 the full version number and not just 0. This should mark the
6862 shared libraries as not backward compatible. Of course, this should
6863 be changed again when we can guarantee backward binary compatibility.
6866 *) Fix typo in get_cert_by_subject() in by_dir.c
6867 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
6869 *) Rework the system to generate shared libraries:
6871 - Make note of the expected extension for the shared libraries and
6872 if there is a need for symbolic links from for example libcrypto.so.0
6873 to libcrypto.so.0.9.7. There is extended info in Configure for
6876 - Make as few rebuilds of the shared libraries as possible.
6878 - Still avoid linking the OpenSSL programs with the shared libraries.
6880 - When installing, install the shared libraries separately from the
6884 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
6886 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
6887 and not in SSL_clear because the latter is also used by the
6888 accept/connect functions; previously, the settings made by
6889 SSL_set_read_ahead would be lost during the handshake.
6890 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
6892 *) Correct util/mkdef.pl to be selective about disabled algorithms.
6893 Previously, it would create entries for disableed algorithms no
6897 *) Added several new manual pages for SSL_* function.
6900 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
6902 *) In ssl23_get_client_hello, generate an error message when faced
6903 with an initial SSL 3.0/TLS record that is too small to contain the
6904 first two bytes of the ClientHello message, i.e. client_version.
6905 (Note that this is a pathologic case that probably has never happened
6906 in real life.) The previous approach was to use the version number
6907 from the record header as a substitute; but our protocol choice
6908 should not depend on that one because it is not authenticated
6909 by the Finished messages.
6912 *) More robust randomness gathering functions for Windows.
6913 [Jeffrey Altman <jaltman@columbia.edu>]
6915 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
6916 not set then we don't setup the error code for issuer check errors
6917 to avoid possibly overwriting other errors which the callback does
6918 handle. If an application does set the flag then we assume it knows
6919 what it is doing and can handle the new informational codes
6923 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
6924 a general "ANY" type, as such it should be able to decode anything
6925 including tagged types. However it didn't check the class so it would
6926 wrongly interpret tagged types in the same way as their universal
6927 counterpart and unknown types were just rejected. Changed so that the
6928 tagged and unknown types are handled in the same way as a SEQUENCE:
6929 that is the encoding is stored intact. There is also a new type
6930 "V_ASN1_OTHER" which is used when the class is not universal, in this
6931 case we have no idea what the actual type is so we just lump them all
6935 *) On VMS, stdout may very well lead to a file that is written to
6936 in a record-oriented fashion. That means that every write() will
6937 write a separate record, which will be read separately by the
6938 programs trying to read from it. This can be very confusing.
6940 The solution is to put a BIO filter in the way that will buffer
6941 text until a linefeed is reached, and then write everything a
6942 line at a time, so every record written will be an actual line,
6943 not chunks of lines and not (usually doesn't happen, but I've
6944 seen it once) several lines in one record. BIO_f_linebuffer() is
6947 Currently, it's a VMS-only method, because that's where it has
6948 been tested well enough.
6951 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
6952 it can return incorrect results.
6953 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
6954 but it was in 0.9.6-beta[12].)
6957 *) Disable the check for content being present when verifying detached
6958 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
6959 include zero length content when signing messages.
6962 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
6963 BIO_ctrl (for BIO pairs).
6966 *) Add DSO method for VMS.
6969 *) Bug fix: Montgomery multiplication could produce results with the
6973 *) Add RPM specification openssl.spec and modify it to build three
6974 packages. The default package contains applications, application
6975 documentation and run-time libraries. The devel package contains
6976 include files, static libraries and function documentation. The
6977 doc package contains the contents of the doc directory. The original
6978 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
6981 *) Add a large number of documentation files for many SSL routines.
6982 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6984 *) Add a configuration entry for Sony News 4.
6985 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
6987 *) Don't set the two most significant bits to one when generating a
6988 random number < q in the DSA library.
6991 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
6992 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
6993 the underlying transport is blocking) if a handshake took place.
6994 (The default behaviour is needed by applications such as s_client
6995 and s_server that use select() to determine when to use SSL_read;
6996 but for applications that know in advance when to expect data, it
6997 just makes things more complicated.)
7000 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
7004 *) Add a few more EBCDIC conditionals that make `req' and `x509'
7005 work better on such systems.
7006 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7008 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
7009 Update PKCS12_parse() so it copies the friendlyName and the
7010 keyid to the certificates aux info.
7013 *) Fix bug in PKCS7_verify() which caused an infinite loop
7014 if there was more than one signature.
7015 [Sven Uszpelkat <su@celocom.de>]
7017 *) Major change in util/mkdef.pl to include extra information
7018 about each symbol, as well as presentig variables as well
7019 as functions. This change means that there's n more need
7020 to rebuild the .num files when some algorithms are excluded.
7023 *) Allow the verify time to be set by an application,
7024 rather than always using the current time.
7027 *) Phase 2 verify code reorganisation. The certificate
7028 verify code now looks up an issuer certificate by a
7029 number of criteria: subject name, authority key id
7030 and key usage. It also verifies self signed certificates
7031 by the same criteria. The main comparison function is
7032 X509_check_issued() which performs these checks.
7034 Lot of changes were necessary in order to support this
7035 without completely rewriting the lookup code.
7037 Authority and subject key identifier are now cached.
7039 The LHASH 'certs' is X509_STORE has now been replaced
7040 by a STACK_OF(X509_OBJECT). This is mainly because an
7041 LHASH can't store or retrieve multiple objects with
7042 the same hash value.
7044 As a result various functions (which were all internal
7045 use only) have changed to handle the new X509_STORE
7046 structure. This will break anything that messed round
7047 with X509_STORE internally.
7049 The functions X509_STORE_add_cert() now checks for an
7050 exact match, rather than just subject name.
7052 The X509_STORE API doesn't directly support the retrieval
7053 of multiple certificates matching a given criteria, however
7054 this can be worked round by performing a lookup first
7055 (which will fill the cache with candidate certificates)
7056 and then examining the cache for matches. This is probably
7057 the best we can do without throwing out X509_LOOKUP
7058 entirely (maybe later...).
7060 The X509_VERIFY_CTX structure has been enhanced considerably.
7062 All certificate lookup operations now go via a get_issuer()
7063 callback. Although this currently uses an X509_STORE it
7064 can be replaced by custom lookups. This is a simple way
7065 to bypass the X509_STORE hackery necessary to make this
7066 work and makes it possible to use more efficient techniques
7067 in future. A very simple version which uses a simple
7068 STACK for its trusted certificate store is also provided
7069 using X509_STORE_CTX_trusted_stack().
7071 The verify_cb() and verify() callbacks now have equivalents
7072 in the X509_STORE_CTX structure.
7074 X509_STORE_CTX also has a 'flags' field which can be used
7075 to customise the verify behaviour.
7078 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
7079 excludes S/MIME capabilities.
7082 *) When a certificate request is read in keep a copy of the
7083 original encoding of the signed data and use it when outputing
7084 again. Signatures then use the original encoding rather than
7085 a decoded, encoded version which may cause problems if the
7086 request is improperly encoded.
7089 *) For consistency with other BIO_puts implementations, call
7090 buffer_write(b, ...) directly in buffer_puts instead of calling
7093 In BIO_puts, increment b->num_write as in BIO_write.
7094 [Peter.Sylvester@EdelWeb.fr]
7096 *) Fix BN_mul_word for the case where the word is 0. (We have to use
7097 BN_zero, we may not return a BIGNUM with an array consisting of
7101 *) Avoid calling abort() from within the library when problems are
7102 detected, except if preprocessor symbols have been defined
7103 (such as REF_CHECK, BN_DEBUG etc.).
7106 *) New openssl application 'rsautl'. This utility can be
7107 used for low level RSA operations. DER public key
7108 BIO/fp routines also added.
7111 *) New Configure entry and patches for compiling on QNX 4.
7112 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
7114 *) A demo state-machine implementation was sponsored by
7115 Nuron (http://www.nuron.com/) and is now available in
7116 demos/state_machine.
7119 *) New options added to the 'dgst' utility for signature
7120 generation and verification.
7123 *) Unrecognized PKCS#7 content types are now handled via a
7124 catch all ASN1_TYPE structure. This allows unsupported
7125 types to be stored as a "blob" and an application can
7126 encode and decode it manually.
7129 *) Fix various signed/unsigned issues to make a_strex.c
7131 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
7133 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
7134 length if passed a buffer. ASN1_INTEGER_to_BN failed
7135 if passed a NULL BN and its argument was negative.
7136 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
7138 *) Modification to PKCS#7 encoding routines to output definite
7139 length encoding. Since currently the whole structures are in
7140 memory there's not real point in using indefinite length
7141 constructed encoding. However if OpenSSL is compiled with
7142 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
7145 *) Added BIO_vprintf() and BIO_vsnprintf().
7148 *) Added more prefixes to parse for in the the strings written
7149 through a logging bio, to cover all the levels that are available
7150 through syslog. The prefixes are now:
7152 PANIC, EMERG, EMR => LOG_EMERG
7153 ALERT, ALR => LOG_ALERT
7154 CRIT, CRI => LOG_CRIT
7155 ERROR, ERR => LOG_ERR
7156 WARNING, WARN, WAR => LOG_WARNING
7157 NOTICE, NOTE, NOT => LOG_NOTICE
7158 INFO, INF => LOG_INFO
7159 DEBUG, DBG => LOG_DEBUG
7161 and as before, if none of those prefixes are present at the
7162 beginning of the string, LOG_ERR is chosen.
7164 On Win32, the LOG_* levels are mapped according to this:
7166 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
7167 LOG_WARNING => EVENTLOG_WARNING_TYPE
7168 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
7172 *) Made it possible to reconfigure with just the configuration
7173 argument "reconf" or "reconfigure". The command line arguments
7174 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
7175 and are retrieved from there when reconfiguring.
7179 [Assar Westerlund <assar@sics.se>, Richard Levitte]
7181 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
7184 *) The obj_dat.pl script was messing up the sorting of object
7185 names. The reason was that it compared the quoted version
7186 of strings as a result "OCSP" > "OCSP Signing" because
7187 " > SPACE. Changed script to store unquoted versions of
7188 names and add quotes on output. It was also omitting some
7189 names from the lookup table if they were given a default
7190 value (that is if SN is missing it is given the same
7191 value as LN and vice versa), these are now added on the
7192 grounds that if an object has a name we should be able to
7193 look it up. Finally added warning output when duplicate
7194 short or long names are found.
7197 *) Changes needed for Tandem NSK.
7198 [Scott Uroff <scott@xypro.com>]
7200 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
7201 RSA_padding_check_SSLv23(), special padding was never detected
7202 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
7203 version rollback attacks was not effective.
7205 In s23_clnt.c, don't use special rollback-attack detection padding
7206 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
7207 client; similarly, in s23_srvr.c, don't do the rollback check if
7208 SSL 2.0 is the only protocol enabled in the server.
7211 *) Make it possible to get hexdumps of unprintable data with 'openssl
7212 asn1parse'. By implication, the functions ASN1_parse_dump() and
7213 BIO_dump_indent() are added.
7216 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
7217 these print out strings and name structures based on various
7218 flags including RFC2253 support and proper handling of
7219 multibyte characters. Added options to the 'x509' utility
7220 to allow the various flags to be set.
7223 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
7224 Also change the functions X509_cmp_current_time() and
7225 X509_gmtime_adj() work with an ASN1_TIME structure,
7226 this will enable certificates using GeneralizedTime in validity
7227 dates to be checked.
7230 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
7231 negative public key encodings) on by default,
7232 NO_NEG_PUBKEY_BUG can be set to disable it.
7235 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
7236 content octets. An i2c_ASN1_OBJECT is unnecessary because
7237 the encoding can be trivially obtained from the structure.
7240 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
7241 not read locks (CRYPTO_r_[un]lock).
7244 *) A first attempt at creating official support for shared
7245 libraries through configuration. I've kept it so the
7246 default is static libraries only, and the OpenSSL programs
7247 are always statically linked for now, but there are
7248 preparations for dynamic linking in place.
7249 This has been tested on Linux and Tru64.
7252 *) Randomness polling function for Win9x, as described in:
7253 Peter Gutmann, Software Generation of Practically Strong
7257 *) Fix so PRNG is seeded in req if using an already existing
7261 *) New options to smime application. -inform and -outform
7262 allow alternative formats for the S/MIME message including
7263 PEM and DER. The -content option allows the content to be
7264 specified separately. This should allow things like Netscape
7265 form signing output easier to verify.
7268 *) Fix the ASN1 encoding of tags using the 'long form'.
7271 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
7272 STRING types. These convert content octets to and from the
7273 underlying type. The actual tag and length octets are
7274 already assumed to have been read in and checked. These
7275 are needed because all other string types have virtually
7276 identical handling apart from the tag. By having versions
7277 of the ASN1 functions that just operate on content octets
7278 IMPLICIT tagging can be handled properly. It also allows
7279 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
7280 and ASN1_INTEGER are identical apart from the tag.
7283 *) Change the handling of OID objects as follows:
7285 - New object identifiers are inserted in objects.txt, following
7286 the syntax given in objects.README.
7287 - objects.pl is used to process obj_mac.num and create a new
7289 - obj_dat.pl is used to create a new obj_dat.h, using the data in
7292 This is currently kind of a hack, and the perl code in objects.pl
7293 isn't very elegant, but it works as I intended. The simplest way
7294 to check that it worked correctly is to look in obj_dat.h and
7295 check the array nid_objs and make sure the objects haven't moved
7296 around (this is important!). Additions are OK, as well as
7297 consistent name changes.
7300 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
7303 *) Addition of the command line parameter '-rand file' to 'openssl req'.
7304 The given file adds to whatever has already been seeded into the
7305 random pool through the RANDFILE configuration file option or
7306 environment variable, or the default random state file.
7309 *) mkstack.pl now sorts each macro group into lexical order.
7310 Previously the output order depended on the order the files
7311 appeared in the directory, resulting in needless rewriting
7315 *) Patches to make OpenSSL compile under Win32 again. Mostly
7316 work arounds for the VC++ problem that it treats func() as
7317 func(void). Also stripped out the parts of mkdef.pl that
7318 added extra typesafe functions: these no longer exist.
7321 *) Reorganisation of the stack code. The macros are now all
7322 collected in safestack.h . Each macro is defined in terms of
7323 a "stack macro" of the form SKM_<name>(type, a, b). The
7324 DEBUG_SAFESTACK is now handled in terms of function casts,
7325 this has the advantage of retaining type safety without the
7326 use of additional functions. If DEBUG_SAFESTACK is not defined
7327 then the non typesafe macros are used instead. Also modified the
7328 mkstack.pl script to handle the new form. Needs testing to see
7329 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
7330 the default if no major problems. Similar behaviour for ASN1_SET_OF
7331 and PKCS12_STACK_OF.
7334 *) When some versions of IIS use the 'NET' form of private key the
7335 key derivation algorithm is different. Normally MD5(password) is
7336 used as a 128 bit RC4 key. In the modified case
7337 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
7338 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
7339 as the old Netscape_RSA functions except they have an additional
7340 'sgckey' parameter which uses the modified algorithm. Also added
7341 an -sgckey command line option to the rsa utility. Thanks to
7342 Adrian Peck <bertie@ncipher.com> for posting details of the modified
7343 algorithm to openssl-dev.
7346 *) The evp_local.h macros were using 'c.##kname' which resulted in
7347 invalid expansion on some systems (SCO 5.0.5 for example).
7348 Corrected to 'c.kname'.
7349 [Phillip Porch <root@theporch.com>]
7351 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
7352 a STACK of email addresses from a certificate or request, these look
7353 in the subject name and the subject alternative name extensions and
7354 omit any duplicate addresses.
7357 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
7358 This makes DSA verification about 2 % faster.
7361 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
7362 (meaning that now 2^5 values will be precomputed, which is only 4 KB
7363 plus overhead for 1024 bit moduli).
7364 This makes exponentiations about 0.5 % faster for 1024 bit
7365 exponents (as measured by "openssl speed rsa2048").
7368 *) Rename memory handling macros to avoid conflicts with other
7370 Malloc => OPENSSL_malloc
7371 Malloc_locked => OPENSSL_malloc_locked
7372 Realloc => OPENSSL_realloc
7373 Free => OPENSSL_free
7376 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
7377 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
7380 *) CygWin32 support.
7381 [John Jarvie <jjarvie@newsguy.com>]
7383 *) The type-safe stack code has been rejigged. It is now only compiled
7384 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
7385 by default all type-specific stack functions are "#define"d back to
7386 standard stack functions. This results in more streamlined output
7387 but retains the type-safety checking possibilities of the original
7391 *) The STACK code has been cleaned up, and certain type declarations
7392 that didn't make a lot of sense have been brought in line. This has
7393 also involved a cleanup of sorts in safestack.h to more correctly
7394 map type-safe stack functions onto their plain stack counterparts.
7395 This work has also resulted in a variety of "const"ifications of
7396 lots of the code, especially "_cmp" operations which should normally
7397 be prototyped with "const" parameters anyway.
7400 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
7401 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
7402 (The PRNG state consists of two parts, the large pool 'state' and 'md',
7403 where all of 'md' is used each time the PRNG is used, but 'state'
7404 is used only indexed by a cyclic counter. As entropy may not be
7405 well distributed from the beginning, 'md' is important as a
7406 chaining variable. However, the output function chains only half
7407 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
7408 all of 'md', and seeding with STATE_SIZE dummy bytes will result
7409 in all of 'state' being rewritten, with the new values depending
7410 on virtually all of 'md'. This overcomes the 80 bit limitation.)
7413 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
7414 the handshake is continued after ssl_verify_cert_chain();
7415 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
7416 can lead to 'unexplainable' connection aborts later.
7417 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
7419 *) Major EVP API cipher revision.
7420 Add hooks for extra EVP features. This allows various cipher
7421 parameters to be set in the EVP interface. Support added for variable
7422 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
7423 setting of RC2 and RC5 parameters.
7425 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
7428 Remove lots of duplicated code from the EVP library. For example *every*
7429 cipher init() function handles the 'iv' in the same way according to the
7430 cipher mode. They also all do nothing if the 'key' parameter is NULL and
7431 for CFB and OFB modes they zero ctx->num.
7433 New functionality allows removal of S/MIME code RC2 hack.
7435 Most of the routines have the same form and so can be declared in terms
7438 By shifting this to the top level EVP_CipherInit() it can be removed from
7439 all individual ciphers. If the cipher wants to handle IVs or keys
7440 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
7443 Change lots of functions like EVP_EncryptUpdate() to now return a
7444 value: although software versions of the algorithms cannot fail
7445 any installed hardware versions can.
7448 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
7449 this option is set, tolerate broken clients that send the negotiated
7450 protocol version number instead of the requested protocol version
7454 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
7455 i.e. non-zero for export ciphersuites, zero otherwise.
7456 Previous versions had this flag inverted, inconsistent with
7457 rsa_tmp_cb (..._TMP_RSA_CB).
7458 [Bodo Moeller; problem reported by Amit Chopra]
7460 *) Add missing DSA library text string. Work around for some IIS
7461 key files with invalid SEQUENCE encoding.
7464 *) Add a document (doc/standards.txt) that list all kinds of standards
7465 and so on that are implemented in OpenSSL.
7468 *) Enhance c_rehash script. Old version would mishandle certificates
7469 with the same subject name hash and wouldn't handle CRLs at all.
7470 Added -fingerprint option to crl utility, to support new c_rehash
7474 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
7477 *) Fix for SSL server purpose checking. Server checking was
7478 rejecting certificates which had extended key usage present
7479 but no ssl client purpose.
7480 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
7482 *) Make PKCS#12 code work with no password. The PKCS#12 spec
7483 is a little unclear about how a blank password is handled.
7484 Since the password in encoded as a BMPString with terminating
7485 double NULL a zero length password would end up as just the
7486 double NULL. However no password at all is different and is
7487 handled differently in the PKCS#12 key generation code. NS
7488 treats a blank password as zero length. MSIE treats it as no
7489 password on export: but it will try both on import. We now do
7490 the same: PKCS12_parse() tries zero length and no password if
7491 the password is set to "" or NULL (NULL is now a valid password:
7492 it wasn't before) as does the pkcs12 application.
7495 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
7496 perror when PEM_read_bio_X509_REQ fails, the error message must
7497 be obtained from the error queue.
7500 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
7501 it in ERR_remove_state if appropriate, and change ERR_get_state
7502 accordingly to avoid race conditions (this is necessary because
7503 thread_hash is no longer constant once set).
7506 *) Bugfix for linux-elf makefile.one.
7509 *) RSA_get_default_method() will now cause a default
7510 RSA_METHOD to be chosen if one doesn't exist already.
7511 Previously this was only set during a call to RSA_new()
7512 or RSA_new_method(NULL) meaning it was possible for
7513 RSA_get_default_method() to return NULL.
7516 *) Added native name translation to the existing DSO code
7517 that will convert (if the flag to do so is set) filenames
7518 that are sufficiently small and have no path information
7519 into a canonical native form. Eg. "blah" converted to
7520 "libblah.so" or "blah.dll" etc.
7523 *) New function ERR_error_string_n(e, buf, len) which is like
7524 ERR_error_string(e, buf), but writes at most 'len' bytes
7525 including the 0 terminator. For ERR_error_string_n, 'buf'
7527 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
7529 *) CONF library reworked to become more general. A new CONF
7530 configuration file reader "class" is implemented as well as a
7531 new functions (NCONF_*, for "New CONF") to handle it. The now
7532 old CONF_* functions are still there, but are reimplemented to
7533 work in terms of the new functions. Also, a set of functions
7534 to handle the internal storage of the configuration data is
7535 provided to make it easier to write new configuration file
7536 reader "classes" (I can definitely see something reading a
7537 configuration file in XML format, for example), called _CONF_*,
7538 or "the configuration storage API"...
7540 The new configuration file reading functions are:
7542 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
7543 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
7545 NCONF_default, NCONF_WIN32
7547 NCONF_dump_fp, NCONF_dump_bio
7549 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
7550 NCONF_new creates a new CONF object. This works in the same way
7551 as other interfaces in OpenSSL, like the BIO interface.
7552 NCONF_dump_* dump the internal storage of the configuration file,
7553 which is useful for debugging. All other functions take the same
7554 arguments as the old CONF_* functions wth the exception of the
7555 first that must be a `CONF *' instead of a `LHASH *'.
7557 To make it easer to use the new classes with the old CONF_* functions,
7558 the function CONF_set_default_method is provided.
7561 *) Add '-tls1' option to 'openssl ciphers', which was already
7562 mentioned in the documentation but had not been implemented.
7563 (This option is not yet really useful because even the additional
7564 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
7567 *) Initial DSO code added into libcrypto for letting OpenSSL (and
7568 OpenSSL-based applications) load shared libraries and bind to
7569 them in a portable way.
7570 [Geoff Thorpe, with contributions from Richard Levitte]
7572 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
7574 *) Make sure _lrotl and _lrotr are only used with MSVC.
7576 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
7577 (the default implementation of RAND_status).
7579 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
7580 to '-clrext' (= clear extensions), as intended and documented.
7581 [Bodo Moeller; inconsistency pointed out by Michael Attili
7584 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
7585 was larger than the MD block size.
7586 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
7588 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
7589 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
7590 using the passed key: if the passed key was a private key the result
7591 of X509_print(), for example, would be to print out all the private key
7595 *) des_quad_cksum() byte order bug fix.
7596 [Ulf Möller, using the problem description in krb4-0.9.7, where
7597 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
7599 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
7601 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
7603 *) For easily testing in shell scripts whether some command
7604 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
7605 returns with exit code 0 iff no command of the given name is available.
7606 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
7607 the output goes to stdout and nothing is printed to stderr.
7608 Additional arguments are always ignored.
7610 Since for each cipher there is a command of the same name,
7611 the 'no-cipher' compilation switches can be tested this way.
7613 ('openssl no-XXX' is not able to detect pseudo-commands such
7614 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
7617 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
7620 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
7621 is set; it will be thrown away anyway because each handshake creates
7623 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
7624 to parameters -- in previous versions (since OpenSSL 0.9.3) the
7625 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
7626 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7629 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7630 'Q' and 'R' lose their special meanings (quit/renegotiate).
7631 This is part of what -quiet does; unlike -quiet, -ign_eof
7632 does not suppress any output.
7635 *) Add compatibility options to the purpose and trust code. The
7636 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7637 accepts a certificate or CA, this was the previous behaviour,
7638 with all the associated security issues.
7640 X509_TRUST_COMPAT is the old trust behaviour: only and
7641 automatically trust self signed roots in certificate store. A
7642 new trust setting X509_TRUST_DEFAULT is used to specify that
7643 a purpose has no associated trust setting and it should instead
7644 use the value in the default purpose.
7647 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7648 and fix a memory leak.
7651 *) In util/mkerr.pl (which implements 'make errors'), preserve
7652 reason strings from the previous version of the .c file, as
7653 the default to have only downcase letters (and digits) in
7654 automatically generated reasons codes is not always appropriate.
7657 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7658 using strerror. Previously, ERR_reason_error_string() returned
7659 library names as reason strings for SYSerr; but SYSerr is a special
7660 case where small numbers are errno values, not library numbers.
7663 *) Add '-dsaparam' option to 'openssl dhparam' application. This
7664 converts DSA parameters into DH parameters. (When creating parameters,
7665 DSA_generate_parameters is used.)
7668 *) Include 'length' (recommended exponent length) in C code generated
7669 by 'openssl dhparam -C'.
7672 *) The second argument to set_label in perlasm was already being used
7673 so couldn't be used as a "file scope" flag. Moved to third argument
7677 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
7678 instead of RAND_bytes for encryption IVs and salts.
7681 *) Include RAND_status() into RAND_METHOD instead of implementing
7682 it only for md_rand.c Otherwise replacing the PRNG by calling
7683 RAND_set_rand_method would be impossible.
7686 *) Don't let DSA_generate_key() enter an infinite loop if the random
7687 number generation fails.
7690 *) New 'rand' application for creating pseudo-random output.
7693 *) Added configuration support for Linux/IA64
7694 [Rolf Haberrecker <rolf@suse.de>]
7696 *) Assembler module support for Mingw32.
7699 *) Shared library support for HPUX (in shlib/).
7700 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
7702 *) Shared library support for Solaris gcc.
7703 [Lutz Behnke <behnke@trustcenter.de>]
7705 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
7707 *) PKCS7_encrypt() was adding text MIME headers twice because they
7708 were added manually and by SMIME_crlf_copy().
7711 *) In bntest.c don't call BN_rand with zero bits argument.
7712 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
7714 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
7715 case was implemented. This caused BN_div_recp() to fail occasionally.
7718 *) Add an optional second argument to the set_label() in the perl
7719 assembly language builder. If this argument exists and is set
7720 to 1 it signals that the assembler should use a symbol whose
7721 scope is the entire file, not just the current function. This
7722 is needed with MASM which uses the format label:: for this scope.
7723 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
7725 *) Change the ASN1 types so they are typedefs by default. Before
7726 almost all types were #define'd to ASN1_STRING which was causing
7727 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
7731 *) Change names of new functions to the new get1/get0 naming
7732 convention: After 'get1', the caller owns a reference count
7733 and has to call ..._free; 'get0' returns a pointer to some
7734 data structure without incrementing reference counters.
7735 (Some of the existing 'get' functions increment a reference
7736 counter, some don't.)
7737 Similarly, 'set1' and 'add1' functions increase reference
7738 counters or duplicate objects.
7741 *) Allow for the possibility of temp RSA key generation failure:
7742 the code used to assume it always worked and crashed on failure.
7745 *) Fix potential buffer overrun problem in BIO_printf().
7746 [Ulf Möller, using public domain code by Patrick Powell; problem
7747 pointed out by David Sacerdote <das33@cornell.edu>]
7749 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
7750 RAND_egd() and RAND_status(). In the command line application,
7751 the EGD socket can be specified like a seed file using RANDFILE
7755 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
7756 Some CAs (e.g. Verisign) distribute certificates in this form.
7759 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
7760 list to exclude them. This means that no special compilation option
7761 is needed to use anonymous DH: it just needs to be included in the
7765 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
7766 EVP_MD_type. The old functionality is available in a new macro called
7767 EVP_MD_md(). Change code that uses it and update docs.
7770 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
7771 where the 'void *' argument is replaced by a function pointer argument.
7772 Previously 'void *' was abused to point to functions, which works on
7773 many platforms, but is not correct. As these functions are usually
7774 called by macros defined in OpenSSL header files, most source code
7775 should work without changes.
7778 *) <openssl/opensslconf.h> (which is created by Configure) now contains
7779 sections with information on -D... compiler switches used for
7780 compiling the library so that applications can see them. To enable
7781 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
7782 must be defined. E.g.,
7783 #define OPENSSL_ALGORITHM_DEFINES
7784 #include <openssl/opensslconf.h>
7785 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
7786 [Richard Levitte, Ulf and Bodo Möller]
7788 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
7792 *) Change the 'other' type in certificate aux info to a STACK_OF
7793 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
7794 the required ASN1 format: arbitrary types determined by an OID.
7797 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
7798 argument to 'req'. This is not because the function is newer or
7799 better than others it just uses the work 'NEW' in the certificate
7800 request header lines. Some software needs this.
7803 *) Reorganise password command line arguments: now passwords can be
7804 obtained from various sources. Delete the PEM_cb function and make
7805 it the default behaviour: i.e. if the callback is NULL and the
7806 usrdata argument is not NULL interpret it as a null terminated pass
7807 phrase. If usrdata and the callback are NULL then the pass phrase
7808 is prompted for as usual.
7811 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
7812 the support is automatically enabled. The resulting binaries will
7813 autodetect the card and use it if present.
7814 [Ben Laurie and Compaq Inc.]
7816 *) Work around for Netscape hang bug. This sends certificate request
7817 and server done in one record. Since this is perfectly legal in the
7818 SSL/TLS protocol it isn't a "bug" option and is on by default. See
7819 the bugs/SSLv3 entry for more info.
7822 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
7825 *) Add -rand argument to smime and pkcs12 applications and read/write
7829 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
7832 *) Add command line password options to the remaining applications.
7835 *) Bug fix for BN_div_recp() for numerators with an even number of
7839 *) More tests in bntest.c, and changed test_bn output.
7842 *) ./config recognizes MacOS X now.
7845 *) Bug fix for BN_div() when the first words of num and divsor are
7846 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
7849 *) Add support for various broken PKCS#8 formats, and command line
7850 options to produce them.
7853 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
7854 get temporary BIGNUMs from a BN_CTX.
7857 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
7861 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
7862 include a #define from the old name to the new. The original intent
7863 was that statically linked binaries could for example just call
7864 SSLeay_add_all_ciphers() to just add ciphers to the table and not
7865 link with digests. This never worked becayse SSLeay_add_all_digests()
7866 and SSLeay_add_all_ciphers() were in the same source file so calling
7867 one would link with the other. They are now in separate source files.
7870 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
7873 *) Use a less unusual form of the Miller-Rabin primality test (it used
7874 a binary algorithm for exponentiation integrated into the Miller-Rabin
7875 loop, our standard modexp algorithms are faster).
7878 *) Support for the EBCDIC character set completed.
7879 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
7881 *) Source code cleanups: use const where appropriate, eliminate casts,
7882 use void * instead of char * in lhash.
7885 *) Bugfix: ssl3_send_server_key_exchange was not restartable
7886 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
7887 this the server could overwrite ephemeral keys that the client
7891 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
7892 using 50 iterations of the Rabin-Miller test.
7894 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
7895 iterations of the Rabin-Miller test as required by the appendix
7896 to FIPS PUB 186[-1]) instead of DSA_is_prime.
7897 As BN_is_prime_fasttest includes trial division, DSA parameter
7898 generation becomes much faster.
7900 This implies a change for the callback functions in DSA_is_prime
7901 and DSA_generate_parameters: The callback function is called once
7902 for each positive witness in the Rabin-Miller test, not just
7903 occasionally in the inner loop; and the parameters to the
7904 callback function now provide an iteration count for the outer
7905 loop rather than for the current invocation of the inner loop.
7906 DSA_generate_parameters additionally can call the callback
7907 function with an 'iteration count' of -1, meaning that a
7908 candidate has passed the trial division test (when q is generated
7909 from an application-provided seed, trial division is skipped).
7912 *) New function BN_is_prime_fasttest that optionally does trial
7913 division before starting the Rabin-Miller test and has
7914 an additional BN_CTX * argument (whereas BN_is_prime always
7915 has to allocate at least one BN_CTX).
7916 'callback(1, -1, cb_arg)' is called when a number has passed the
7917 trial division stage.
7920 *) Fix for bug in CRL encoding. The validity dates weren't being handled
7924 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
7927 *) New function BN_pseudo_rand().
7930 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
7931 bignum version of BN_from_montgomery() with the working code from
7932 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
7936 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
7937 made it impossible to use the same SSL_SESSION data structure in
7938 SSL2 clients in multiple threads.
7941 *) The return value of RAND_load_file() no longer counts bytes obtained
7942 by stat(). RAND_load_file(..., -1) is new and uses the complete file
7943 to seed the PRNG (previously an explicit byte count was required).
7944 [Ulf Möller, Bodo Möller]
7946 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
7947 used (char *) instead of (void *) and had casts all over the place.
7950 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
7953 *) Retain source code compatibility for BN_prime_checks macro:
7954 BN_is_prime(..., BN_prime_checks, ...) now uses
7955 BN_prime_checks_for_size to determine the appropriate number of
7956 Rabin-Miller iterations.
7959 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
7960 DH_CHECK_P_NOT_SAFE_PRIME.
7961 (Check if this is true? OpenPGP calls them "strong".)
7964 *) Merge the functionality of "dh" and "gendh" programs into a new program
7965 "dhparam". The old programs are retained for now but will handle DH keys
7966 (instead of parameters) in future.
7969 *) Make the ciphers, s_server and s_client programs check the return values
7970 when a new cipher list is set.
7973 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
7974 ciphers. Before when the 56bit ciphers were enabled the sorting was
7977 The syntax for the cipher sorting has been extended to support sorting by
7978 cipher-strength (using the strength_bits hard coded in the tables).
7979 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
7981 Fix a bug in the cipher-command parser: when supplying a cipher command
7982 string with an "undefined" symbol (neither command nor alphanumeric
7983 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
7984 an error is flagged.
7986 Due to the strength-sorting extension, the code of the
7987 ssl_create_cipher_list() function was completely rearranged. I hope that
7988 the readability was also increased :-)
7989 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7991 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
7992 for the first serial number and places 2 in the serial number file. This
7993 avoids problems when the root CA is created with serial number zero and
7994 the first user certificate has the same issuer name and serial number
7998 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
7999 the new code. Add documentation for this stuff.
8002 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
8003 X509_*() to X509at_*() on the grounds that they don't handle X509
8004 structures and behave in an analagous way to the X509v3 functions:
8005 they shouldn't be called directly but wrapper functions should be used
8008 So we also now have some wrapper functions that call the X509at functions
8009 when passed certificate requests. (TO DO: similar things can be done with
8010 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
8011 things. Some of these need some d2i or i2d and print functionality
8012 because they handle more complex structures.)
8015 *) Add missing #ifndefs that caused missing symbols when building libssl
8016 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
8017 NO_RSA in ssl/s2*.c.
8018 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
8020 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
8021 has a return value which indicates the quality of the random data
8022 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
8023 error queue. New function RAND_pseudo_bytes() generates output that is
8024 guaranteed to be unique but not unpredictable. RAND_add is like
8025 RAND_seed, but takes an extra argument for an entropy estimate
8026 (RAND_seed always assumes full entropy).
8029 *) Do more iterations of Rabin-Miller probable prime test (specifically,
8030 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
8031 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
8032 in crypto/bn/bn_prime.c for the complete table). This guarantees a
8033 false-positive rate of at most 2^-80 for random input.
8036 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
8039 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
8040 in the 0.9.5 release), this returns the chain
8041 from an X509_CTX structure with a dup of the stack and all
8042 the X509 reference counts upped: so the stack will exist
8043 after X509_CTX_cleanup() has been called. Modify pkcs12.c
8046 Also make SSL_SESSION_print() print out the verify return
8050 *) Add manpage for the pkcs12 command. Also change the default
8051 behaviour so MAC iteration counts are used unless the new
8052 -nomaciter option is used. This improves file security and
8053 only older versions of MSIE (4.0 for example) need it.
8056 *) Honor the no-xxx Configure options when creating .DEF files.
8059 *) Add PKCS#10 attributes to field table: challengePassword,
8060 unstructuredName and unstructuredAddress. These are taken from
8061 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
8062 international characters are used.
8064 More changes to X509_ATTRIBUTE code: allow the setting of types
8065 based on strings. Remove the 'loc' parameter when adding
8066 attributes because these will be a SET OF encoding which is sorted
8070 *) Initial changes to the 'req' utility to allow request generation
8071 automation. This will allow an application to just generate a template
8072 file containing all the field values and have req construct the
8075 Initial support for X509_ATTRIBUTE handling. Stacks of these are
8076 used all over the place including certificate requests and PKCS#7
8077 structures. They are currently handled manually where necessary with
8078 some primitive wrappers for PKCS#7. The new functions behave in a
8079 manner analogous to the X509 extension functions: they allow
8080 attributes to be looked up by NID and added.
8082 Later something similar to the X509V3 code would be desirable to
8083 automatically handle the encoding, decoding and printing of the
8084 more complex types. The string types like challengePassword can
8085 be handled by the string table functions.
8087 Also modified the multi byte string table handling. Now there is
8088 a 'global mask' which masks out certain types. The table itself
8089 can use the flag STABLE_NO_MASK to ignore the mask setting: this
8090 is useful when for example there is only one permissible type
8091 (as in countryName) and using the mask might result in no valid
8095 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
8096 SSL_get_peer_finished to allow applications to obtain the latest
8097 Finished messages sent to the peer or expected from the peer,
8098 respectively. (SSL_get_peer_finished is usually the Finished message
8099 actually received from the peer, otherwise the protocol will be aborted.)
8101 As the Finished message are message digests of the complete handshake
8102 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
8103 be used for external authentication procedures when the authentication
8104 provided by SSL/TLS is not desired or is not enough.
8107 *) Enhanced support for Alpha Linux is added. Now ./config checks if
8108 the host supports BWX extension and if Compaq C is present on the
8109 $PATH. Just exploiting of the BWX extension results in 20-30%
8110 performance kick for some algorithms, e.g. DES and RC4 to mention
8111 a couple. Compaq C in turn generates ~20% faster code for MD5 and
8115 *) Add support for MS "fast SGC". This is arguably a violation of the
8116 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
8117 weak crypto and after checking the certificate is SGC a second one
8118 with strong crypto. MS SGC stops the first handshake after receiving
8119 the server certificate message and sends a second client hello. Since
8120 a server will typically do all the time consuming operations before
8121 expecting any further messages from the client (server key exchange
8122 is the most expensive) there is little difference between the two.
8124 To get OpenSSL to support MS SGC we have to permit a second client
8125 hello message after we have sent server done. In addition we have to
8126 reset the MAC if we do get this second client hello.
8129 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
8130 if a DER encoded private key is RSA or DSA traditional format. Changed
8131 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
8132 format DER encoded private key. Newer code should use PKCS#8 format which
8133 has the key type encoded in the ASN1 structure. Added DER private key
8134 support to pkcs8 application.
8137 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
8138 ciphersuites has been selected (as required by the SSL 3/TLS 1
8139 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
8140 is set, we interpret this as a request to violate the specification
8141 (the worst that can happen is a handshake failure, and 'correct'
8142 behaviour would result in a handshake failure anyway).
8145 *) In SSL_CTX_add_session, take into account that there might be multiple
8146 SSL_SESSION structures with the same session ID (e.g. when two threads
8147 concurrently obtain them from an external cache).
8148 The internal cache can handle only one SSL_SESSION with a given ID,
8149 so if there's a conflict, we now throw out the old one to achieve
8153 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
8154 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
8155 some routines that use cipher OIDs: some ciphers do not have OIDs
8156 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
8160 *) Simplify the trust setting structure and code. Now we just have
8161 two sequences of OIDs for trusted and rejected settings. These will
8162 typically have values the same as the extended key usage extension
8163 and any application specific purposes.
8165 The trust checking code now has a default behaviour: it will just
8166 check for an object with the same NID as the passed id. Functions can
8167 be provided to override either the default behaviour or the behaviour
8168 for a given id. SSL client, server and email already have functions
8169 in place for compatibility: they check the NID and also return "trusted"
8170 if the certificate is self signed.
8173 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
8174 traditional format into an EVP_PKEY structure.
8177 *) Add a password callback function PEM_cb() which either prompts for
8178 a password if usr_data is NULL or otherwise assumes it is a null
8179 terminated password. Allow passwords to be passed on command line
8180 environment or config files in a few more utilities.
8183 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
8184 keys. Add some short names for PKCS#8 PBE algorithms and allow them
8185 to be specified on the command line for the pkcs8 and pkcs12 utilities.
8186 Update documentation.
8189 *) Support for ASN1 "NULL" type. This could be handled before by using
8190 ASN1_TYPE but there wasn't any function that would try to read a NULL
8191 and produce an error if it couldn't. For compatibility we also have
8192 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
8193 don't allocate anything because they don't need to.
8196 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
8198 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
8200 *) Rebuild of the memory allocation routines used by OpenSSL code and
8201 possibly others as well. The purpose is to make an interface that
8202 provide hooks so anyone can build a separate set of allocation and
8203 deallocation routines to be used by OpenSSL, for example memory
8204 pool implementations, or something else, which was previously hard
8205 since Malloc(), Realloc() and Free() were defined as macros having
8206 the values malloc, realloc and free, respectively (except for Win32
8207 compilations). The same is provided for memory debugging code.
8208 OpenSSL already comes with functionality to find memory leaks, but
8209 this gives people a chance to debug other memory problems.
8211 With these changes, a new set of functions and macros have appeared:
8213 CRYPTO_set_mem_debug_functions() [F]
8214 CRYPTO_get_mem_debug_functions() [F]
8215 CRYPTO_dbg_set_options() [F]
8216 CRYPTO_dbg_get_options() [F]
8217 CRYPTO_malloc_debug_init() [M]
8219 The memory debug functions are NULL by default, unless the library
8220 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
8221 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
8222 gives the standard debugging functions that come with OpenSSL) or
8223 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
8224 provided by the library user) must be used. When the standard
8225 debugging functions are used, CRYPTO_dbg_set_options can be used to
8226 request additional information:
8227 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
8228 the CRYPTO_MDEBUG_xxx macro when compiling the library.
8230 Also, things like CRYPTO_set_mem_functions will always give the
8231 expected result (the new set of functions is used for allocation
8232 and deallocation) at all times, regardless of platform and compiler
8235 To finish it up, some functions that were never use in any other
8236 way than through macros have a new API and new semantic:
8239 CRYPTO_dbg_realloc()
8242 All macros of value have retained their old syntax.
8243 [Richard Levitte and Bodo Moeller]
8245 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
8246 ordering of SMIMECapabilities wasn't in "strength order" and there
8247 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
8251 *) Some ASN1 types with illegal zero length encoding (INTEGER,
8252 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
8253 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
8255 *) Merge in my S/MIME library for OpenSSL. This provides a simple
8256 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
8257 functionality to handle multipart/signed properly) and a utility
8258 called 'smime' to call all this stuff. This is based on code I
8259 originally wrote for Celo who have kindly allowed it to be
8260 included in OpenSSL.
8263 *) Add variants des_set_key_checked and des_set_key_unchecked of
8264 des_set_key (aka des_key_sched). Global variable des_check_key
8265 decides which of these is called by des_set_key; this way
8266 des_check_key behaves as it always did, but applications and
8267 the library itself, which was buggy for des_check_key == 1,
8268 have a cleaner way to pick the version they need.
8271 *) New function PKCS12_newpass() which changes the password of a
8275 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
8276 dynamic mix. In both cases the ids can be used as an index into the
8277 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
8278 functions so they accept a list of the field values and the
8279 application doesn't need to directly manipulate the X509_TRUST
8283 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
8287 *) Modify the way the V3 extension code looks up extensions. This now
8288 works in a similar way to the object code: we have some "standard"
8289 extensions in a static table which is searched with OBJ_bsearch()
8290 and the application can add dynamic ones if needed. The file
8291 crypto/x509v3/ext_dat.h now has the info: this file needs to be
8292 updated whenever a new extension is added to the core code and kept
8293 in ext_nid order. There is a simple program 'tabtest.c' which checks
8294 this. New extensions are not added too often so this file can readily
8295 be maintained manually.
8297 There are two big advantages in doing things this way. The extensions
8298 can be looked up immediately and no longer need to be "added" using
8299 X509V3_add_standard_extensions(): this function now does nothing.
8300 [Side note: I get *lots* of email saying the extension code doesn't
8301 work because people forget to call this function]
8302 Also no dynamic allocation is done unless new extensions are added:
8303 so if we don't add custom extensions there is no need to call
8304 X509V3_EXT_cleanup().
8307 *) Modify enc utility's salting as follows: make salting the default. Add a
8308 magic header, so unsalted files fail gracefully instead of just decrypting
8309 to garbage. This is because not salting is a big security hole, so people
8310 should be discouraged from doing it.
8313 *) Fixes and enhancements to the 'x509' utility. It allowed a message
8314 digest to be passed on the command line but it only used this
8315 parameter when signing a certificate. Modified so all relevant
8316 operations are affected by the digest parameter including the
8317 -fingerprint and -x509toreq options. Also -x509toreq choked if a
8318 DSA key was used because it didn't fix the digest.
8321 *) Initial certificate chain verify code. Currently tests the untrusted
8322 certificates for consistency with the verify purpose (which is set
8323 when the X509_STORE_CTX structure is set up) and checks the pathlength.
8325 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
8326 this is because it will reject chains with invalid extensions whereas
8327 every previous version of OpenSSL and SSLeay made no checks at all.
8329 Trust code: checks the root CA for the relevant trust settings. Trust
8330 settings have an initial value consistent with the verify purpose: e.g.
8331 if the verify purpose is for SSL client use it expects the CA to be
8332 trusted for SSL client use. However the default value can be changed to
8333 permit custom trust settings: one example of this would be to only trust
8334 certificates from a specific "secure" set of CAs.
8336 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
8337 which should be used for version portability: especially since the
8338 verify structure is likely to change more often now.
8340 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
8341 to set them. If not set then assume SSL clients will verify SSL servers
8344 Two new options to the verify program: -untrusted allows a set of
8345 untrusted certificates to be passed in and -purpose which sets the
8346 intended purpose of the certificate. If a purpose is set then the
8347 new chain verify code is used to check extension consistency.
8350 *) Support for the authority information access extension.
8353 *) Modify RSA and DSA PEM read routines to transparently handle
8354 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
8355 public keys in a format compatible with certificate
8356 SubjectPublicKeyInfo structures. Unfortunately there were already
8357 functions called *_PublicKey_* which used various odd formats so
8358 these are retained for compatibility: however the DSA variants were
8359 never in a public release so they have been deleted. Changed dsa/rsa
8360 utilities to handle the new format: note no releases ever handled public
8361 keys so we should be OK.
8363 The primary motivation for this change is to avoid the same fiasco
8364 that dogs private keys: there are several incompatible private key
8365 formats some of which are standard and some OpenSSL specific and
8366 require various evil hacks to allow partial transparent handling and
8367 even then it doesn't work with DER formats. Given the option anything
8368 other than PKCS#8 should be dumped: but the other formats have to
8369 stay in the name of compatibility.
8371 With public keys and the benefit of hindsight one standard format
8372 is used which works with EVP_PKEY, RSA or DSA structures: though
8373 it clearly returns an error if you try to read the wrong kind of key.
8375 Added a -pubkey option to the 'x509' utility to output the public key.
8376 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
8377 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
8378 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
8379 that do the same as the EVP_PKEY_assign_*() except they up the
8380 reference count of the added key (they don't "swallow" the
8384 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
8385 CRLs would fail if the file contained no certificates or no CRLs:
8386 added a new function to read in both types and return the number
8387 read: this means that if none are read it will be an error. The
8388 DER versions of the certificate and CRL reader would always fail
8389 because it isn't possible to mix certificates and CRLs in DER format
8390 without choking one or the other routine. Changed this to just read
8391 a certificate: this is the best we can do. Also modified the code
8392 in apps/verify.c to take notice of return codes: it was previously
8393 attempting to read in certificates from NULL pointers and ignoring
8394 any errors: this is one reason why the cert and CRL reader seemed
8395 to work. It doesn't check return codes from the default certificate
8396 routines: these may well fail if the certificates aren't installed.
8399 *) Code to support otherName option in GeneralName.
8402 *) First update to verify code. Change the verify utility
8403 so it warns if it is passed a self signed certificate:
8404 for consistency with the normal behaviour. X509_verify
8405 has been modified to it will now verify a self signed
8406 certificate if *exactly* the same certificate appears
8407 in the store: it was previously impossible to trust a
8408 single self signed certificate. This means that:
8409 openssl verify ss.pem
8410 now gives a warning about a self signed certificate but
8411 openssl verify -CAfile ss.pem ss.pem
8415 *) For servers, store verify_result in SSL_SESSION data structure
8416 (and add it to external session representation).
8417 This is needed when client certificate verifications fails,
8418 but an application-provided verification callback (set by
8419 SSL_CTX_set_cert_verify_callback) allows accepting the session
8420 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
8421 but returns 1): When the session is reused, we have to set
8422 ssl->verify_result to the appropriate error code to avoid
8424 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
8426 *) Fix a bug in the new PKCS#7 code: it didn't consider the
8427 case in PKCS7_dataInit() where the signed PKCS7 structure
8428 didn't contain any existing data because it was being created.
8429 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
8431 *) Add a salt to the key derivation routines in enc.c. This
8432 forms the first 8 bytes of the encrypted file. Also add a
8433 -S option to allow a salt to be input on the command line.
8436 *) New function X509_cmp(). Oddly enough there wasn't a function
8437 to compare two certificates. We do this by working out the SHA1
8438 hash and comparing that. X509_cmp() will be needed by the trust
8442 *) SSL_get1_session() is like SSL_get_session(), but increments
8443 the reference count in the SSL_SESSION returned.
8444 [Geoff Thorpe <geoff@eu.c2.net>]
8446 *) Fix for 'req': it was adding a null to request attributes.
8447 Also change the X509_LOOKUP and X509_INFO code to handle
8448 certificate auxiliary information.
8451 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
8455 *) Add the possibility to add extra information to the memory leak
8456 detecting output, to form tracebacks, showing from where each
8457 allocation was originated: CRYPTO_push_info("constant string") adds
8458 the string plus current file name and line number to a per-thread
8459 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
8460 is like calling CYRPTO_pop_info() until the stack is empty.
8461 Also updated memory leak detection code to be multi-thread-safe.
8464 *) Add options -text and -noout to pkcs7 utility and delete the
8465 encryption options which never did anything. Update docs.
8468 *) Add options to some of the utilities to allow the pass phrase
8469 to be included on either the command line (not recommended on
8470 OSes like Unix) or read from the environment. Update the
8471 manpages and fix a few bugs.
8474 *) Add a few manpages for some of the openssl commands.
8477 *) Fix the -revoke option in ca. It was freeing up memory twice,
8478 leaking and not finding already revoked certificates.
8481 *) Extensive changes to support certificate auxiliary information.
8482 This involves the use of X509_CERT_AUX structure and X509_AUX
8483 functions. An X509_AUX function such as PEM_read_X509_AUX()
8484 can still read in a certificate file in the usual way but it
8485 will also read in any additional "auxiliary information". By
8486 doing things this way a fair degree of compatibility can be
8487 retained: existing certificates can have this information added
8488 using the new 'x509' options.
8490 Current auxiliary information includes an "alias" and some trust
8491 settings. The trust settings will ultimately be used in enhanced
8492 certificate chain verification routines: currently a certificate
8493 can only be trusted if it is self signed and then it is trusted
8497 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
8498 The problem was that one of the replacement routines had not been working
8499 since SSLeay releases. For now the offending routine has been replaced
8500 with non-optimised assembler. Even so, this now gives around 95%
8501 performance improvement for 1024 bit RSA signs.
8504 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
8505 handling. Most clients have the effective key size in bits equal to
8506 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
8507 A few however don't do this and instead use the size of the decrypted key
8508 to determine the RC2 key length and the AlgorithmIdentifier to determine
8509 the effective key length. In this case the effective key length can still
8510 be 40 bits but the key length can be 168 bits for example. This is fixed
8511 by manually forcing an RC2 key into the EVP_PKEY structure because the
8512 EVP code can't currently handle unusual RC2 key sizes: it always assumes
8513 the key length and effective key length are equal.
8516 *) Add a bunch of functions that should simplify the creation of
8517 X509_NAME structures. Now you should be able to do:
8518 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
8519 and have it automatically work out the correct field type and fill in
8520 the structures. The more adventurous can try:
8521 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
8522 and it will (hopefully) work out the correct multibyte encoding.
8525 *) Change the 'req' utility to use the new field handling and multibyte
8526 copy routines. Before the DN field creation was handled in an ad hoc
8527 way in req, ca, and x509 which was rather broken and didn't support
8528 BMPStrings or UTF8Strings. Since some software doesn't implement
8529 BMPStrings or UTF8Strings yet, they can be enabled using the config file
8530 using the dirstring_type option. See the new comment in the default
8531 openssl.cnf for more info.
8534 *) Make crypto/rand/md_rand.c more robust:
8535 - Assure unique random numbers after fork().
8536 - Make sure that concurrent threads access the global counter and
8537 md serializably so that we never lose entropy in them
8538 or use exactly the same state in multiple threads.
8539 Access to the large state is not always serializable because
8540 the additional locking could be a performance killer, and
8541 md should be large enough anyway.
8544 *) New file apps/app_rand.c with commonly needed functionality
8545 for handling the random seed file.
8547 Use the random seed file in some applications that previously did not:
8549 dsaparam -genkey (which also ignored its '-rand' option),
8552 x509 (when signing).
8553 Except on systems with /dev/urandom, it is crucial to have a random
8554 seed file at least for key creation, DSA signing, and for DH exchanges;
8555 for RSA signatures we could do without one.
8557 gendh and gendsa (unlike genrsa) used to read only the first byte
8558 of each file listed in the '-rand' option. The function as previously
8559 found in genrsa is now in app_rand.c and is used by all programs
8560 that support '-rand'.
8563 *) In RAND_write_file, use mode 0600 for creating files;
8564 don't just chmod when it may be too late.
8567 *) Report an error from X509_STORE_load_locations
8568 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
8571 *) New function ASN1_mbstring_copy() this copies a string in either
8572 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
8573 into an ASN1_STRING type. A mask of permissible types is passed
8574 and it chooses the "minimal" type to use or an error if not type
8578 *) Add function equivalents to the various macros in asn1.h. The old
8579 macros are retained with an M_ prefix. Code inside the library can
8580 use the M_ macros. External code (including the openssl utility)
8581 should *NOT* in order to be "shared library friendly".
8584 *) Add various functions that can check a certificate's extensions
8585 to see if it usable for various purposes such as SSL client,
8586 server or S/MIME and CAs of these types. This is currently
8587 VERY EXPERIMENTAL but will ultimately be used for certificate chain
8588 verification. Also added a -purpose flag to x509 utility to
8589 print out all the purposes.
8592 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
8596 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
8597 for, obtain and decode and extension and obtain its critical flag.
8598 This allows all the necessary extension code to be handled in a
8599 single function call.
8602 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
8603 platforms. See crypto/rc4/rc4_enc.c for further details.
8606 *) New -noout option to asn1parse. This causes no output to be produced
8607 its main use is when combined with -strparse and -out to extract data
8608 from a file (which may not be in ASN.1 format).
8611 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
8612 when producing the local key id.
8613 [Richard Levitte <levitte@stacken.kth.se>]
8615 *) New option -dhparam in s_server. This allows a DH parameter file to be
8616 stated explicitly. If it is not stated then it tries the first server
8617 certificate file. The previous behaviour hard coded the filename
8621 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
8622 a public key to be input or output. For example:
8623 openssl rsa -in key.pem -pubout -out pubkey.pem
8624 Also added necessary DSA public key functions to handle this.
8627 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
8628 in the message. This was handled by allowing
8629 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8630 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8632 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8633 to the end of the strings whereas this didn't. This would cause problems
8634 if strings read with d2i_ASN1_bytes() were later modified.
8635 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8637 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8638 data and it contains EOF it will end up returning an error. This is
8639 caused by input 46 bytes long. The cause is due to the way base64
8640 BIOs find the start of base64 encoded data. They do this by trying a
8641 trial decode on each line until they find one that works. When they
8642 do a flag is set and it starts again knowing it can pass all the
8643 data directly through the decoder. Unfortunately it doesn't reset
8644 the context it uses. This means that if EOF is reached an attempt
8645 is made to pass two EOFs through the context and this causes the
8646 resulting error. This can also cause other problems as well. As is
8647 usual with these problems it takes *ages* to find and the fix is
8648 trivial: move one line.
8649 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8651 *) Ugly workaround to get s_client and s_server working under Windows. The
8652 old code wouldn't work because it needed to select() on sockets and the
8653 tty (for keypresses and to see if data could be written). Win32 only
8654 supports select() on sockets so we select() with a 1s timeout on the
8655 sockets and then see if any characters are waiting to be read, if none
8656 are present then we retry, we also assume we can always write data to
8657 the tty. This isn't nice because the code then blocks until we've
8658 received a complete line of data and it is effectively polling the
8659 keyboard at 1s intervals: however it's quite a bit better than not
8660 working at all :-) A dedicated Windows application might handle this
8661 with an event loop for example.
8664 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
8665 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
8666 will be called when RSA_sign() and RSA_verify() are used. This is useful
8667 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
8668 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
8669 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
8670 This necessitated the support of an extra signature type NID_md5_sha1
8671 for SSL signatures and modifications to the SSL library to use it instead
8672 of calling RSA_public_decrypt() and RSA_private_encrypt().
8675 *) Add new -verify -CAfile and -CApath options to the crl program, these
8676 will lookup a CRL issuers certificate and verify the signature in a
8677 similar way to the verify program. Tidy up the crl program so it
8678 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
8679 less strict. It will now permit CRL extensions even if it is not
8680 a V2 CRL: this will allow it to tolerate some broken CRLs.
8683 *) Initialize all non-automatic variables each time one of the openssl
8684 sub-programs is started (this is necessary as they may be started
8685 multiple times from the "OpenSSL>" prompt).
8686 [Lennart Bang, Bodo Moeller]
8688 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
8689 removing all other RSA functionality (this is what NO_RSA does). This
8690 is so (for example) those in the US can disable those operations covered
8691 by the RSA patent while allowing storage and parsing of RSA keys and RSA
8695 *) Non-copying interface to BIO pairs.
8696 (still largely untested)
8699 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
8700 ASCII string. This was handled independently in various places before.
8703 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
8704 UTF8 strings a character at a time.
8707 *) Use client_version from client hello to select the protocol
8708 (s23_srvr.c) and for RSA client key exchange verification
8709 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
8712 *) Add various utility functions to handle SPKACs, these were previously
8713 handled by poking round in the structure internals. Added new function
8714 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
8715 print, verify and generate SPKACs. Based on an original idea from
8716 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
8719 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
8722 *) Allow the config file extension section to be overwritten on the
8723 command line. Based on an original idea from Massimiliano Pala
8724 <madwolf@comune.modena.it>. The new option is called -extensions
8725 and can be applied to ca, req and x509. Also -reqexts to override
8726 the request extensions in req and -crlexts to override the crl extensions
8730 *) Add new feature to the SPKAC handling in ca. Now you can include
8731 the same field multiple times by preceding it by "XXXX." for example:
8734 this is the same syntax as used in the req config file.
8737 *) Allow certificate extensions to be added to certificate requests. These
8738 are specified in a 'req_extensions' option of the req section of the
8739 config file. They can be printed out with the -text option to req but
8740 are otherwise ignored at present.
8743 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
8744 data read consists of only the final block it would not decrypted because
8745 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
8746 A misplaced 'break' also meant the decrypted final block might not be
8747 copied until the next read.
8750 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
8751 a few extra parameters to the DH structure: these will be useful if
8752 for example we want the value of 'q' or implement X9.42 DH.
8755 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
8756 provides hooks that allow the default DSA functions or functions on a
8757 "per key" basis to be replaced. This allows hardware acceleration and
8758 hardware key storage to be handled without major modification to the
8759 library. Also added low level modexp hooks and CRYPTO_EX structure and
8760 associated functions.
8763 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
8764 as "read only": it can't be written to and the buffer it points to will
8765 not be freed. Reading from a read only BIO is much more efficient than
8766 a normal memory BIO. This was added because there are several times when
8767 an area of memory needs to be read from a BIO. The previous method was
8768 to create a memory BIO and write the data to it, this results in two
8769 copies of the data and an O(n^2) reading algorithm. There is a new
8770 function BIO_new_mem_buf() which creates a read only memory BIO from
8771 an area of memory. Also modified the PKCS#7 routines to use read only
8775 *) Bugfix: ssl23_get_client_hello did not work properly when called in
8776 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
8777 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
8778 but a retry condition occured while trying to read the rest.
8781 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
8782 NID_pkcs7_encrypted by default: this was wrong since this should almost
8783 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
8784 the encrypted data type: this is a more sensible place to put it and it
8785 allows the PKCS#12 code to be tidied up that duplicated this
8789 *) Changed obj_dat.pl script so it takes its input and output files on
8790 the command line. This should avoid shell escape redirection problems
8794 *) Initial support for certificate extension requests, these are included
8795 in things like Xenroll certificate requests. Included functions to allow
8796 extensions to be obtained and added.
8799 *) -crlf option to s_client and s_server for sending newlines as
8800 CRLF (as required by many protocols).
8803 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
8805 *) Install libRSAglue.a when OpenSSL is built with RSAref.
8806 [Ralf S. Engelschall]
8808 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
8809 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
8811 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
8815 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
8816 DH parameters/keys (q is lost during that conversion, but the resulting
8817 DH parameters contain its length).
8819 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
8820 much faster than DH_generate_parameters (which creates parameters
8821 where p = 2*q + 1), and also the smaller q makes DH computations
8822 much more efficient (160-bit exponentiation instead of 1024-bit
8823 exponentiation); so this provides a convenient way to support DHE
8824 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
8825 utter importance to use
8826 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8828 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8829 when such DH parameters are used, because otherwise small subgroup
8830 attacks may become possible!
8833 *) Avoid memory leak in i2d_DHparams.
8836 *) Allow the -k option to be used more than once in the enc program:
8837 this allows the same encrypted message to be read by multiple recipients.
8840 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
8841 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
8842 it will always use the numerical form of the OID, even if it has a short
8846 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
8847 method only got called if p,q,dmp1,dmq1,iqmp components were present,
8848 otherwise bn_mod_exp was called. In the case of hardware keys for example
8849 no private key components need be present and it might store extra data
8850 in the RSA structure, which cannot be accessed from bn_mod_exp.
8851 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
8852 private key operations.
8855 *) Added support for SPARC Linux.
8858 *) pem_password_cb function type incompatibly changed from
8859 typedef int pem_password_cb(char *buf, int size, int rwflag);
8861 ....(char *buf, int size, int rwflag, void *userdata);
8862 so that applications can pass data to their callbacks:
8863 The PEM[_ASN1]_{read,write}... functions and macros now take an
8864 additional void * argument, which is just handed through whenever
8865 the password callback is called.
8866 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
8868 New function SSL_CTX_set_default_passwd_cb_userdata.
8870 Compatibility note: As many C implementations push function arguments
8871 onto the stack in reverse order, the new library version is likely to
8872 interoperate with programs that have been compiled with the old
8873 pem_password_cb definition (PEM_whatever takes some data that
8874 happens to be on the stack as its last argument, and the callback
8875 just ignores this garbage); but there is no guarantee whatsoever that
8878 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
8879 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
8880 problems not only on Windows, but also on some Unix platforms.
8881 To avoid problematic command lines, these definitions are now in an
8882 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
8883 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
8886 *) MIPS III/IV assembler module is reimplemented.
8889 *) More DES library cleanups: remove references to srand/rand and
8890 delete an unused file.
8893 *) Add support for the the free Netwide assembler (NASM) under Win32,
8894 since not many people have MASM (ml) and it can be hard to obtain.
8895 This is currently experimental but it seems to work OK and pass all
8896 the tests. Check out INSTALL.W32 for info.
8899 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
8900 without temporary keys kept an extra copy of the server key,
8901 and connections with temporary keys did not free everything in case
8905 *) New function RSA_check_key and new openssl rsa option -check
8906 for verifying the consistency of RSA keys.
8907 [Ulf Moeller, Bodo Moeller]
8909 *) Various changes to make Win32 compile work:
8910 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
8911 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
8912 comparison" warnings.
8913 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
8916 *) Add a debugging option to PKCS#5 v2 key generation function: when
8917 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
8918 derived keys are printed to stderr.
8921 *) Copy the flags in ASN1_STRING_dup().
8922 [Roman E. Pavlov <pre@mo.msk.ru>]
8924 *) The x509 application mishandled signing requests containing DSA
8925 keys when the signing key was also DSA and the parameters didn't match.
8927 It was supposed to omit the parameters when they matched the signing key:
8928 the verifying software was then supposed to automatically use the CA's
8929 parameters if they were absent from the end user certificate.
8931 Omitting parameters is no longer recommended. The test was also
8932 the wrong way round! This was probably due to unusual behaviour in
8933 EVP_cmp_parameters() which returns 1 if the parameters match.
8934 This meant that parameters were omitted when they *didn't* match and
8935 the certificate was useless. Certificates signed with 'ca' didn't have
8937 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
8939 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
8940 The interface is as follows:
8941 Applications can use
8942 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
8943 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
8944 "off" is now the default.
8945 The library internally uses
8946 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
8947 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
8948 to disable memory-checking temporarily.
8950 Some inconsistent states that previously were possible (and were
8951 even the default) are now avoided.
8953 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
8954 with each memory chunk allocated; this is occasionally more helpful
8955 than just having a counter.
8957 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
8959 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
8963 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
8964 which largely parallels "options", but is for changing API behaviour,
8965 whereas "options" are about protocol behaviour.
8966 Initial "mode" flags are:
8968 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
8969 a single record has been written.
8970 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
8971 retries use the same buffer location.
8972 (But all of the contents must be
8976 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
8979 *) Fix problems with no-hmac etc.
8980 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
8982 *) New functions RSA_get_default_method(), RSA_set_method() and
8983 RSA_get_method(). These allows replacement of RSA_METHODs without having
8984 to mess around with the internals of an RSA structure.
8987 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
8988 Also really enable memory leak checks in openssl.c and in some
8990 [Chad C. Mulligan, Bodo Moeller]
8992 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
8993 up the length of negative integers. This has now been simplified to just
8994 store the length when it is first determined and use it later, rather
8995 than trying to keep track of where data is copied and updating it to
8997 [Steve Henson, reported by Brien Wheeler
8998 <bwheeler@authentica-security.com>]
9000 *) Add a new function PKCS7_signatureVerify. This allows the verification
9001 of a PKCS#7 signature but with the signing certificate passed to the
9002 function itself. This contrasts with PKCS7_dataVerify which assumes the
9003 certificate is present in the PKCS#7 structure. This isn't always the
9004 case: certificates can be omitted from a PKCS#7 structure and be
9005 distributed by "out of band" means (such as a certificate database).
9008 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
9009 function prototypes in pem.h, also change util/mkdef.pl to add the
9010 necessary function names.
9013 *) mk1mf.pl (used by Windows builds) did not properly read the
9014 options set by Configure in the top level Makefile, and Configure
9015 was not even able to write more than one option correctly.
9016 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
9019 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
9020 file to be loaded from a BIO or FILE pointer. The BIO version will
9021 for example allow memory BIOs to contain config info.
9024 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
9025 Whoever hopes to achieve shared-library compatibility across versions
9026 must use this, not the compile-time macro.
9027 (Exercise 0.9.4: Which is the minimum library version required by
9029 Note: All this applies only to multi-threaded programs, others don't
9033 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
9034 through a BIO pair triggered the default case, i.e.
9035 SSLerr(...,SSL_R_UNKNOWN_STATE).
9038 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
9039 can use the SSL library even if none of the specific BIOs is
9043 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
9044 for the encoded length.
9045 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
9047 *) Add initial documentation of the X509V3 functions.
9050 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
9051 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
9052 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
9053 secure PKCS#8 private key format with a high iteration count.
9056 *) Fix determination of Perl interpreter: A perl or perl5
9057 _directory_ in $PATH was also accepted as the interpreter.
9058 [Ralf S. Engelschall]
9060 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
9061 wrong with it but it was very old and did things like calling
9062 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
9066 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
9067 to use the new extension code.
9070 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
9071 with macros. This should make it easier to change their form, add extra
9072 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
9076 *) Add to configuration table a new entry that can specify an alternative
9077 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
9078 according to Mark Crispin <MRC@Panda.COM>.
9082 *) DES CBC did not update the IV. Weird.
9085 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
9086 Changing the behaviour of the former might break existing programs --
9087 where IV updating is needed, des_ncbc_encrypt can be used.
9090 *) When bntest is run from "make test" it drives bc to check its
9091 calculations, as well as internally checking them. If an internal check
9092 fails, it needs to cause bc to give a non-zero result or make test carries
9093 on without noticing the failure. Fixed.
9096 *) DES library cleanups.
9099 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
9100 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
9101 ciphers. NOTE: although the key derivation function has been verified
9102 against some published test vectors it has not been extensively tested
9103 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
9107 *) Instead of "mkdir -p", which is not fully portable, use new
9108 Perl script "util/mkdir-p.pl".
9111 *) Rewrite the way password based encryption (PBE) is handled. It used to
9112 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
9113 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
9114 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
9115 the 'parameter' field of the AlgorithmIdentifier is passed to the
9116 underlying key generation function so it must do its own ASN1 parsing.
9117 This has also changed the EVP_PBE_CipherInit() function which now has a
9118 'parameter' argument instead of literal salt and iteration count values
9119 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
9122 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
9123 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
9124 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
9125 KEY" because this clashed with PKCS#8 unencrypted string. Since this
9126 value was just used as a "magic string" and not used directly its
9127 value doesn't matter.
9130 *) Introduce some semblance of const correctness to BN. Shame C doesn't
9134 *) "linux-sparc64" configuration (ultrapenguin).
9135 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
9136 "linux-sparc" configuration.
9137 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
9139 *) config now generates no-xxx options for missing ciphers.
9142 *) Support the EBCDIC character set (work in progress).
9143 File ebcdic.c not yet included because it has a different license.
9144 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9146 *) Support BS2000/OSD-POSIX.
9147 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9149 *) Make callbacks for key generation use void * instead of char *.
9152 *) Make S/MIME samples compile (not yet tested).
9155 *) Additional typesafe stacks.
9158 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
9162 Changes between 0.9.3 and 0.9.3a [29 May 1999]
9164 *) New configuration variant "sco5-gcc".
9166 *) Updated some demos.
9167 [Sean O Riordain, Wade Scholine]
9169 *) Add missing BIO_free at exit of pkcs12 application.
9172 *) Fix memory leak in conf.c.
9175 *) Updates for Win32 to assembler version of MD5.
9178 *) Set #! path to perl in apps/der_chop to where we found it
9179 instead of using a fixed path.
9182 *) SHA library changes for irix64-mips4-cc.
9185 *) Improvements for VMS support.
9189 Changes between 0.9.2b and 0.9.3 [24 May 1999]
9191 *) Bignum library bug fix. IRIX 6 passes "make test" now!
9192 This also avoids the problems with SC4.2 and unpatched SC5.
9193 [Andy Polyakov <appro@fy.chalmers.se>]
9195 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
9196 These are required because of the typesafe stack would otherwise break
9197 existing code. If old code used a structure member which used to be STACK
9198 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
9199 sk_num or sk_value it would produce an error because the num, data members
9200 are not present in STACK_OF. Now it just produces a warning. sk_set
9201 replaces the old method of assigning a value to sk_value
9202 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
9203 that does this will no longer work (and should use sk_set instead) but
9204 this could be regarded as a "questionable" behaviour anyway.
9207 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
9208 correctly handle encrypted S/MIME data.
9211 *) Change type of various DES function arguments from des_cblock
9212 (which means, in function argument declarations, pointer to char)
9213 to des_cblock * (meaning pointer to array with 8 char elements),
9214 which allows the compiler to do more typechecking; it was like
9215 that back in SSLeay, but with lots of ugly casts.
9217 Introduce new type const_des_cblock.
9220 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
9221 problems: find RecipientInfo structure that matches recipient certificate
9222 and initialise the ASN1 structures properly based on passed cipher.
9225 *) Belatedly make the BN tests actually check the results.
9228 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
9229 to and from BNs: it was completely broken. New compilation option
9230 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
9231 key elements as negative integers.
9234 *) Reorganize and speed up MD5.
9235 [Andy Polyakov <appro@fy.chalmers.se>]
9238 [Richard Levitte <richard@levitte.org>]
9240 *) New option -out to asn1parse to allow the parsed structure to be
9241 output to a file. This is most useful when combined with the -strparse
9242 option to examine the output of things like OCTET STRINGS.
9245 *) Make SSL library a little more fool-proof by not requiring any longer
9246 that SSL_set_{accept,connect}_state be called before
9247 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
9248 in many applications because usually everything *appeared* to work as
9249 intended anyway -- now it really works as intended).
9252 *) Move openssl.cnf out of lib/.
9255 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
9256 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
9257 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
9258 [Ralf S. Engelschall]
9260 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
9261 handle PKCS#7 enveloped data properly.
9262 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
9264 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
9265 copying pointers. The cert_st handling is changed by this in
9266 various ways (and thus what used to be known as ctx->default_cert
9267 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
9268 any longer when s->cert does not give us what we need).
9269 ssl_cert_instantiate becomes obsolete by this change.
9270 As soon as we've got the new code right (possibly it already is?),
9271 we have solved a couple of bugs of the earlier code where s->cert
9272 was used as if it could not have been shared with other SSL structures.
9274 Note that using the SSL API in certain dirty ways now will result
9275 in different behaviour than observed with earlier library versions:
9276 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
9277 does not influence s as it used to.
9279 In order to clean up things more thoroughly, inside SSL_SESSION
9280 we don't use CERT any longer, but a new structure SESS_CERT
9281 that holds per-session data (if available); currently, this is
9282 the peer's certificate chain and, for clients, the server's certificate
9283 and temporary key. CERT holds only those values that can have
9284 meaningful defaults in an SSL_CTX.
9287 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
9288 from the internal representation. Various PKCS#7 fixes: remove some
9289 evil casts and set the enc_dig_alg field properly based on the signing
9293 *) Allow PKCS#12 password to be set from the command line or the
9294 environment. Let 'ca' get its config file name from the environment
9295 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
9299 *) Allow certificate policies extension to use an IA5STRING for the
9300 organization field. This is contrary to the PKIX definition but
9301 VeriSign uses it and IE5 only recognises this form. Document 'x509'
9305 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
9306 without disallowing inline assembler and the like for non-pedantic builds.
9309 *) Support Borland C++ builder.
9310 [Janez Jere <jj@void.si>, modified by Ulf Möller]
9315 *) SHA-1 cleanups and performance enhancements.
9316 [Andy Polyakov <appro@fy.chalmers.se>]
9318 *) Sparc v8plus assembler for the bignum library.
9319 [Andy Polyakov <appro@fy.chalmers.se>]
9321 *) Accept any -xxx and +xxx compiler options in Configure.
9324 *) Update HPUX configuration.
9327 *) Add missing sk_<type>_unshift() function to safestack.h
9328 [Ralf S. Engelschall]
9330 *) New function SSL_CTX_use_certificate_chain_file that sets the
9331 "extra_cert"s in addition to the certificate. (This makes sense
9332 only for "PEM" format files, as chains as a whole are not
9336 *) Support verify_depth from the SSL API.
9337 x509_vfy.c had what can be considered an off-by-one-error:
9338 Its depth (which was not part of the external interface)
9339 was actually counting the number of certificates in a chain;
9340 now it really counts the depth.
9343 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
9344 instead of X509err, which often resulted in confusing error
9345 messages since the error codes are not globally unique
9346 (e.g. an alleged error in ssl3_accept when a certificate
9347 didn't match the private key).
9349 *) New function SSL_CTX_set_session_id_context that allows to set a default
9350 value (so that you don't need SSL_set_session_id_context for each
9351 connection using the SSL_CTX).
9354 *) OAEP decoding bug fix.
9357 *) Support INSTALL_PREFIX for package builders, as proposed by
9361 *) New Configure options "threads" and "no-threads". For systems
9362 where the proper compiler options are known (currently Solaris
9363 and Linux), "threads" is the default.
9366 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
9369 *) Install various scripts to $(OPENSSLDIR)/misc, not to
9370 $(INSTALLTOP)/bin -- they shouldn't clutter directories
9371 such as /usr/local/bin.
9374 *) "make linux-shared" to build shared libraries.
9375 [Niels Poppe <niels@netbox.org>]
9377 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
9380 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
9381 extension adding in x509 utility.
9384 *) Remove NOPROTO sections and error code comments.
9387 *) Partial rewrite of the DEF file generator to now parse the ANSI
9391 *) New Configure options --prefix=DIR and --openssldir=DIR.
9394 *) Complete rewrite of the error code script(s). It is all now handled
9395 by one script at the top level which handles error code gathering,
9396 header rewriting and C source file generation. It should be much better
9397 than the old method: it now uses a modified version of Ulf's parser to
9398 read the ANSI prototypes in all header files (thus the old K&R definitions
9399 aren't needed for error creation any more) and do a better job of
9400 translating function codes into names. The old 'ASN1 error code imbedded
9401 in a comment' is no longer necessary and it doesn't use .err files which
9402 have now been deleted. Also the error code call doesn't have to appear all
9403 on one line (which resulted in some large lines...).
9406 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
9409 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
9410 0 (which usually indicates a closed connection), but continue reading.
9413 *) Fix some race conditions.
9416 *) Add support for CRL distribution points extension. Add Certificate
9417 Policies and CRL distribution points documentation.
9420 *) Move the autogenerated header file parts to crypto/opensslconf.h.
9423 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
9424 8 of keying material. Merlin has also confirmed interop with this fix
9425 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
9426 [Merlin Hughes <merlin@baltimore.ie>]
9428 *) Fix lots of warnings.
9429 [Richard Levitte <levitte@stacken.kth.se>]
9431 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
9432 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
9433 [Richard Levitte <levitte@stacken.kth.se>]
9435 *) Fix problems with sizeof(long) == 8.
9436 [Andy Polyakov <appro@fy.chalmers.se>]
9438 *) Change functions to ANSI C.
9441 *) Fix typos in error codes.
9442 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
9444 *) Remove defunct assembler files from Configure.
9447 *) SPARC v8 assembler BIGNUM implementation.
9448 [Andy Polyakov <appro@fy.chalmers.se>]
9450 *) Support for Certificate Policies extension: both print and set.
9451 Various additions to support the r2i method this uses.
9454 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
9455 return a const string when you are expecting an allocated buffer.
9458 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
9459 types DirectoryString and DisplayText.
9462 *) Add code to allow r2i extensions to access the configuration database,
9463 add an LHASH database driver and add several ctx helper functions.
9466 *) Fix an evil bug in bn_expand2() which caused various BN functions to
9467 fail when they extended the size of a BIGNUM.
9470 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
9471 support typesafe stack.
9474 *) Fix typo in SSL_[gs]et_options().
9475 [Nils Frostberg <nils@medcom.se>]
9477 *) Delete various functions and files that belonged to the (now obsolete)
9478 old X509V3 handling code.
9481 *) New Configure option "rsaref".
9484 *) Don't auto-generate pem.h.
9487 *) Introduce type-safe ASN.1 SETs.
9490 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
9491 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
9493 *) Introduce type-safe STACKs. This will almost certainly break lots of code
9494 that links with OpenSSL (well at least cause lots of warnings), but fear
9495 not: the conversion is trivial, and it eliminates loads of evil casts. A
9496 few STACKed things have been converted already. Feel free to convert more.
9497 In the fullness of time, I'll do away with the STACK type altogether.
9500 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
9501 specified in <certfile> by updating the entry in the index.txt file.
9502 This way one no longer has to edit the index.txt file manually for
9503 revoking a certificate. The -revoke option does the gory details now.
9504 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
9506 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
9507 `-text' option at all and this way the `-noout -text' combination was
9508 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
9509 [Ralf S. Engelschall]
9511 *) Make sure a corresponding plain text error message exists for the
9512 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
9513 verify callback function determined that a certificate was revoked.
9514 [Ralf S. Engelschall]
9516 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
9517 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
9518 all available cipers including rc5, which was forgotten until now.
9519 In order to let the testing shell script know which algorithms
9520 are available, a new (up to now undocumented) command
9521 "openssl list-cipher-commands" is used.
9524 *) Bugfix: s_client occasionally would sleep in select() when
9525 it should have checked SSL_pending() first.
9528 *) New functions DSA_do_sign and DSA_do_verify to provide access to
9529 the raw DSA values prior to ASN.1 encoding.
9532 *) Tweaks to Configure
9533 [Niels Poppe <niels@netbox.org>]
9535 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
9539 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
9542 *) New config option to avoid instructions that are illegal on the 80386.
9543 The default code is faster, but requires at least a 486.
9546 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
9547 SSL2_SERVER_VERSION (not used at all) macros, which are now the
9548 same as SSL2_VERSION anyway.
9551 *) New "-showcerts" option for s_client.
9554 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
9555 application. Various cleanups and fixes.
9558 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
9559 modify error routines to work internally. Add error codes and PBE init
9560 to library startup routines.
9563 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
9564 packing functions to asn1 and evp. Changed function names and error
9565 codes along the way.
9568 *) PKCS12 integration: and so it begins... First of several patches to
9569 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
9570 objects to objects.h
9573 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
9574 and display support for Thawte strong extranet extension.
9577 *) Add LinuxPPC support.
9578 [Jeff Dubrule <igor@pobox.org>]
9580 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
9581 bn_div_words in alpha.s.
9582 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
9584 *) Make sure the RSA OAEP test is skipped under -DRSAref because
9585 OAEP isn't supported when OpenSSL is built with RSAref.
9586 [Ulf Moeller <ulf@fitug.de>]
9588 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
9589 so they no longer are missing under -DNOPROTO.
9590 [Soren S. Jorvang <soren@t.dk>]
9593 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
9595 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
9596 doesn't work when the session is reused. Coming soon!
9599 *) Fix a security hole, that allows sessions to be reused in the wrong
9600 context thus bypassing client cert protection! All software that uses
9601 client certs and session caches in multiple contexts NEEDS PATCHING to
9602 allow session reuse! A fuller solution is in the works.
9603 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
9605 *) Some more source tree cleanups (removed obsolete files
9606 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
9607 permission on "config" script to be executable) and a fix for the INSTALL
9609 [Ulf Moeller <ulf@fitug.de>]
9611 *) Remove some legacy and erroneous uses of malloc, free instead of
9613 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
9615 *) Make rsa_oaep_test return non-zero on error.
9616 [Ulf Moeller <ulf@fitug.de>]
9618 *) Add support for native Solaris shared libraries. Configure
9619 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
9620 if someone would make that last step automatic.
9621 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
9623 *) ctx_size was not built with the right compiler during "make links". Fixed.
9626 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
9627 except NULL ciphers". This means the default cipher list will no longer
9628 enable NULL ciphers. They need to be specifically enabled e.g. with
9629 the string "DEFAULT:eNULL".
9632 *) Fix to RSA private encryption routines: if p < q then it would
9633 occasionally produce an invalid result. This will only happen with
9634 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9637 *) Be less restrictive and allow also `perl util/perlpath.pl
9638 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9639 because this way one can also use an interpreter named `perl5' (which is
9640 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9641 installed as `perl').
9642 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9644 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9645 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9647 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9648 advapi32.lib to Win32 build and change the pem test comparision
9649 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9650 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9651 and crypto/des/ede_cbcm_enc.c.
9654 *) DES quad checksum was broken on big-endian architectures. Fixed.
9657 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9658 Win32 test batch file so it (might) work again. The Win32 test batch file
9659 is horrible: I feel ill....
9662 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
9663 in e_os.h. Audit of header files to check ANSI and non ANSI
9664 sections: 10 functions were absent from non ANSI section and not exported
9665 from Windows DLLs. Fixed up libeay.num for new functions.
9668 *) Make `openssl version' output lines consistent.
9669 [Ralf S. Engelschall]
9671 *) Fix Win32 symbol export lists for BIO functions: Added
9672 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
9673 to ms/libeay{16,32}.def.
9674 [Ralf S. Engelschall]
9676 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
9677 fine under Unix and passes some trivial tests I've now added. But the
9678 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
9679 added to make sure no one expects that this stuff really works in the
9680 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
9681 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
9683 [Ralf S. Engelschall]
9685 *) Fix the generation of two part addresses in perl.
9686 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
9688 *) Add config entry for Linux on MIPS.
9689 [John Tobey <jtobey@channel1.com>]
9691 *) Make links whenever Configure is run, unless we are on Windoze.
9694 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
9695 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
9699 *) Add a useful kludge to allow package maintainers to specify compiler and
9700 other platforms details on the command line without having to patch the
9701 Configure script everytime: One now can use ``perl Configure
9702 <id>:<details>'', i.e. platform ids are allowed to have details appended
9703 to them (seperated by colons). This is treated as there would be a static
9704 pre-configured entry in Configure's %table under key <id> with value
9705 <details> and ``perl Configure <id>'' is called. So, when you want to
9706 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
9707 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
9708 now, which overrides the FreeBSD-elf entry on-the-fly.
9709 [Ralf S. Engelschall]
9711 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
9714 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
9715 on the `perl Configure ...' command line. This way one can compile
9716 OpenSSL libraries with Position Independent Code (PIC) which is needed
9717 for linking it into DSOs.
9718 [Ralf S. Engelschall]
9720 *) Remarkably, export ciphers were totally broken and no-one had noticed!
9724 *) Cleaned up the LICENSE document: The official contact for any license
9725 questions now is the OpenSSL core team under openssl-core@openssl.org.
9726 And add a paragraph about the dual-license situation to make sure people
9727 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
9728 to the OpenSSL toolkit.
9729 [Ralf S. Engelschall]
9731 *) General source tree makefile cleanups: Made `making xxx in yyy...'
9732 display consistent in the source tree and replaced `/bin/rm' by `rm'.
9733 Additonally cleaned up the `make links' target: Remove unnecessary
9734 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
9735 to speed processing and no longer clutter the display with confusing
9736 stuff. Instead only the actually done links are displayed.
9737 [Ralf S. Engelschall]
9739 *) Permit null encryption ciphersuites, used for authentication only. It used
9740 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
9741 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
9745 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
9746 signed attributes when verifying signatures (this would break them),
9747 the detached data encoding was wrong and public keys obtained using
9748 X509_get_pubkey() weren't freed.
9751 *) Add text documentation for the BUFFER functions. Also added a work around
9752 to a Win95 console bug. This was triggered by the password read stuff: the
9753 last character typed gets carried over to the next fread(). If you were
9754 generating a new cert request using 'req' for example then the last
9755 character of the passphrase would be CR which would then enter the first
9759 *) Added the new `Includes OpenSSL Cryptography Software' button as
9760 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
9761 button and can be used by applications based on OpenSSL to show the
9762 relationship to the OpenSSL project.
9763 [Ralf S. Engelschall]
9765 *) Remove confusing variables in function signatures in files
9766 ssl/ssl_lib.c and ssl/ssl.h.
9767 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9769 *) Don't install bss_file.c under PREFIX/include/
9770 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9772 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
9773 functions that return function pointers and has support for NT specific
9774 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
9775 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
9776 unsigned to signed types: this was killing the Win32 compile.
9779 *) Add new certificate file to stack functions,
9780 SSL_add_dir_cert_subjects_to_stack() and
9781 SSL_add_file_cert_subjects_to_stack(). These largely supplant
9782 SSL_load_client_CA_file(), and can be used to add multiple certs easily
9783 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
9784 This means that Apache-SSL and similar packages don't have to mess around
9785 to add as many CAs as they want to the preferred list.
9788 *) Experiment with doxygen documentation. Currently only partially applied to
9790 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
9791 openssl.doxy as the configuration file.
9794 *) Get rid of remaining C++-style comments which strict C compilers hate.
9795 [Ralf S. Engelschall, pointed out by Carlos Amengual]
9797 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
9798 compiled in by default: it has problems with large keys.
9801 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
9802 DH private keys and/or callback functions which directly correspond to
9803 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
9804 is needed for applications which have to configure certificates on a
9805 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
9807 For the RSA certificate situation is makes no difference, but
9808 for the DSA certificate situation this fixes the "no shared cipher"
9809 problem where the OpenSSL cipher selection procedure failed because the
9810 temporary keys were not overtaken from the context and the API provided
9811 no way to reconfigure them.
9812 The new functions now let applications reconfigure the stuff and they
9813 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
9814 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
9815 non-public-API function ssl_cert_instantiate() is used as a helper
9816 function and also to reduce code redundancy inside ssl_rsa.c.
9817 [Ralf S. Engelschall]
9819 *) Move s_server -dcert and -dkey options out of the undocumented feature
9820 area because they are useful for the DSA situation and should be
9821 recognized by the users.
9822 [Ralf S. Engelschall]
9824 *) Fix the cipher decision scheme for export ciphers: the export bits are
9825 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
9826 SSL_EXP_MASK. So, the original variable has to be used instead of the
9827 already masked variable.
9828 [Richard Levitte <levitte@stacken.kth.se>]
9830 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
9831 [Richard Levitte <levitte@stacken.kth.se>]
9833 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
9834 from `int' to `unsigned int' because it's a length and initialized by
9835 EVP_DigestFinal() which expects an `unsigned int *'.
9836 [Richard Levitte <levitte@stacken.kth.se>]
9838 *) Don't hard-code path to Perl interpreter on shebang line of Configure
9839 script. Instead use the usual Shell->Perl transition trick.
9840 [Ralf S. Engelschall]
9842 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
9843 (in addition to RSA certificates) to match the behaviour of `openssl dsa
9844 -noout -modulus' as it's already the case for `openssl rsa -noout
9845 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
9846 currently the public key is printed (a decision which was already done by
9847 `openssl dsa -modulus' in the past) which serves a similar purpose.
9848 Additionally the NO_RSA no longer completely removes the whole -modulus
9849 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
9851 [Ralf S. Engelschall]
9853 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
9854 BIO. See the source (crypto/evp/bio_ok.c) for more info.
9855 [Arne Ansper <arne@ats.cyber.ee>]
9857 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
9858 to be added. Now both 'req' and 'ca' can use new objects defined in the
9862 *) Add cool BIO that does syslog (or event log on NT).
9863 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
9865 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
9866 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
9867 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
9868 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
9871 *) Add preliminary config info for new extension code.
9874 *) Make RSA_NO_PADDING really use no padding.
9875 [Ulf Moeller <ulf@fitug.de>]
9877 *) Generate errors when private/public key check is done.
9880 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
9881 for some CRL extensions and new objects added.
9884 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
9885 key usage extension and fuller support for authority key id.
9888 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
9889 padding method for RSA, which is recommended for new applications in PKCS
9890 #1 v2.0 (RFC 2437, October 1998).
9891 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
9892 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
9893 against Bleichbacher's attack on RSA.
9894 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
9897 *) Updates to the new SSL compression code
9898 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9900 *) Fix so that the version number in the master secret, when passed
9901 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
9902 (because the server will not accept higher), that the version number
9903 is 0x03,0x01, not 0x03,0x00
9904 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9906 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
9907 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
9908 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
9911 *) Support for RAW extensions where an arbitrary extension can be
9912 created by including its DER encoding. See apps/openssl.cnf for
9916 *) Make sure latest Perl versions don't interpret some generated C array
9917 code as Perl array code in the crypto/err/err_genc.pl script.
9918 [Lars Weber <3weber@informatik.uni-hamburg.de>]
9920 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
9921 not many people have the assembler. Various Win32 compilation fixes and
9922 update to the INSTALL.W32 file with (hopefully) more accurate Win32
9926 *) Modify configure script 'Configure' to automatically create crypto/date.h
9927 file under Win32 and also build pem.h from pem.org. New script
9928 util/mkfiles.pl to create the MINFO file on environments that can't do a
9929 'make files': perl util/mkfiles.pl >MINFO should work.
9932 *) Major rework of DES function declarations, in the pursuit of correctness
9933 and purity. As a result, many evil casts evaporated, and some weirdness,
9934 too. You may find this causes warnings in your code. Zapping your evil
9935 casts will probably fix them. Mostly.
9938 *) Fix for a typo in asn1.h. Bug fix to object creation script
9939 obj_dat.pl. It considered a zero in an object definition to mean
9940 "end of object": none of the objects in objects.h have any zeros
9941 so it wasn't spotted.
9942 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
9944 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
9945 Masking (CBCM). In the absence of test vectors, the best I have been able
9946 to do is check that the decrypt undoes the encrypt, so far. Send me test
9947 vectors if you have them.
9950 *) Correct calculation of key length for export ciphers (too much space was
9951 allocated for null ciphers). This has not been tested!
9954 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
9955 message is now correct (it understands "crypto" and "ssl" on its
9956 command line). There is also now an "update" option. This will update
9957 the util/ssleay.num and util/libeay.num files with any new functions.
9959 perl util/mkdef.pl crypto ssl update
9960 it will update them.
9963 *) Overhauled the Perl interface (perl/*):
9964 - ported BN stuff to OpenSSL's different BN library
9965 - made the perl/ source tree CVS-aware
9966 - renamed the package from SSLeay to OpenSSL (the files still contain
9967 their history because I've copied them in the repository)
9968 - removed obsolete files (the test scripts will be replaced
9969 by better Test::Harness variants in the future)
9970 [Ralf S. Engelschall]
9972 *) First cut for a very conservative source tree cleanup:
9973 1. merge various obsolete readme texts into doc/ssleay.txt
9974 where we collect the old documents and readme texts.
9975 2. remove the first part of files where I'm already sure that we no
9976 longer need them because of three reasons: either they are just temporary
9977 files which were left by Eric or they are preserved original files where
9978 I've verified that the diff is also available in the CVS via "cvs diff
9979 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
9980 the crypto/md/ stuff).
9981 [Ralf S. Engelschall]
9983 *) More extension code. Incomplete support for subject and issuer alt
9984 name, issuer and authority key id. Change the i2v function parameters
9985 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
9986 what that's for :-) Fix to ASN1 macro which messed up
9987 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
9990 *) Preliminary support for ENUMERATED type. This is largely copied from the
9994 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
9995 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9997 *) Make sure `make rehash' target really finds the `openssl' program.
9998 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10000 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
10001 like to hear about it if this slows down other processors.
10004 *) Add CygWin32 platform information to Configure script.
10005 [Alan Batie <batie@aahz.jf.intel.com>]
10007 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
10008 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
10010 *) New program nseq to manipulate netscape certificate sequences
10013 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
10017 *) Fixes to BN code. Previously the default was to define BN_RECURSION
10018 but the BN code had some problems that would cause failures when
10019 doing certificate verification and some other functions.
10020 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10022 *) Add ASN1 and PEM code to support netscape certificate sequences.
10025 *) Add ASN1 and PEM code to support netscape certificate sequences.
10028 *) Add several PKIX and private extended key usage OIDs.
10031 *) Modify the 'ca' program to handle the new extension code. Modify
10032 openssl.cnf for new extension format, add comments.
10035 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
10036 and add a sample to openssl.cnf so req -x509 now adds appropriate
10040 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
10041 error code, add initial support to X509_print() and x509 application.
10044 *) Takes a deep breath and start addding X509 V3 extension support code. Add
10045 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
10046 stuff is currently isolated and isn't even compiled yet.
10049 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
10050 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
10051 Removed the versions check from X509 routines when loading extensions:
10052 this allows certain broken certificates that don't set the version
10053 properly to be processed.
10056 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
10057 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
10058 can still be regenerated with "make depend".
10061 *) Spelling mistake in C version of CAST-128.
10062 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
10064 *) Changes to the error generation code. The perl script err-code.pl
10065 now reads in the old error codes and retains the old numbers, only
10066 adding new ones if necessary. It also only changes the .err files if new
10067 codes are added. The makefiles have been modified to only insert errors
10068 when needed (to avoid needlessly modifying header files). This is done
10069 by only inserting errors if the .err file is newer than the auto generated
10070 C file. To rebuild all the error codes from scratch (the old behaviour)
10071 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
10072 or delete all the .err files.
10075 *) CAST-128 was incorrectly implemented for short keys. The C version has
10076 been fixed, but is untested. The assembler versions are also fixed, but
10077 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
10078 to regenerate it if needed.
10079 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
10080 Hagino <itojun@kame.net>]
10082 *) File was opened incorrectly in randfile.c.
10083 [Ulf Möller <ulf@fitug.de>]
10085 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
10086 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
10087 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
10088 al: it's just almost always a UTCTime. Note this patch adds new error
10089 codes so do a "make errors" if there are problems.
10092 *) Correct Linux 1 recognition in config.
10093 [Ulf Möller <ulf@fitug.de>]
10095 *) Remove pointless MD5 hash when using DSA keys in ca.
10096 [Anonymous <nobody@replay.com>]
10098 *) Generate an error if given an empty string as a cert directory. Also
10099 generate an error if handed NULL (previously returned 0 to indicate an
10100 error, but didn't set one).
10101 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
10103 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
10106 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
10107 parameters. This was causing a warning which killed off the Win32 compile.
10110 *) Remove C++ style comments from crypto/bn/bn_local.h.
10111 [Neil Costigan <neil.costigan@celocom.com>]
10113 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
10114 based on a text string, looking up short and long names and finally
10115 "dot" format. The "dot" format stuff didn't work. Added new function
10116 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
10117 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
10118 OID is not part of the table.
10121 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
10122 X509_LOOKUP_by_alias().
10125 *) Sort openssl functions by name.
10128 *) Get the gendsa program working (hopefully) and add it to app list. Remove
10129 encryption from sample DSA keys (in case anyone is interested the password
10133 *) Make _all_ *_free functions accept a NULL pointer.
10134 [Frans Heymans <fheymans@isaserver.be>]
10136 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
10138 [Anonymous <nobody@replay.com>]
10140 *) s_server should send the CAfile as acceptable CAs, not its own cert.
10141 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10143 *) Don't blow it for numeric -newkey arguments to apps/req.
10144 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10146 *) Temp key "for export" tests were wrong in s3_srvr.c.
10147 [Anonymous <nobody@replay.com>]
10149 *) Add prototype for temp key callback functions
10150 SSL_CTX_set_tmp_{rsa,dh}_callback().
10153 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
10154 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
10157 *) X509_name_add_entry() freed the wrong thing after an error.
10158 [Arne Ansper <arne@ats.cyber.ee>]
10160 *) rsa_eay.c would attempt to free a NULL context.
10161 [Arne Ansper <arne@ats.cyber.ee>]
10163 *) BIO_s_socket() had a broken should_retry() on Windoze.
10164 [Arne Ansper <arne@ats.cyber.ee>]
10166 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
10167 [Arne Ansper <arne@ats.cyber.ee>]
10169 *) Make sure the already existing X509_STORE->depth variable is initialized
10170 in X509_STORE_new(), but document the fact that this variable is still
10171 unused in the certificate verification process.
10172 [Ralf S. Engelschall]
10174 *) Fix the various library and apps files to free up pkeys obtained from
10175 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
10178 *) Fix reference counting in X509_PUBKEY_get(). This makes
10179 demos/maurice/example2.c work, amongst others, probably.
10180 [Steve Henson and Ben Laurie]
10182 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
10183 `openssl' and second, the shortcut symlinks for the `openssl <command>'
10184 are no longer created. This way we have a single and consistent command
10185 line interface `openssl <command>', similar to `cvs <command>'.
10186 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
10188 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
10189 BIT STRING wrapper always have zero unused bits.
10192 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
10195 *) Make the top-level INSTALL documentation easier to understand.
10198 *) Makefiles updated to exit if an error occurs in a sub-directory
10199 make (including if user presses ^C) [Paul Sutton]
10201 *) Make Montgomery context stuff explicit in RSA data structure.
10204 *) Fix build order of pem and err to allow for generated pem.h.
10207 *) Fix renumbering bug in X509_NAME_delete_entry().
10210 *) Enhanced the err-ins.pl script so it makes the error library number
10211 global and can add a library name. This is needed for external ASN1 and
10212 other error libraries.
10215 *) Fixed sk_insert which never worked properly.
10218 *) Fix ASN1 macros so they can handle indefinite length construted
10219 EXPLICIT tags. Some non standard certificates use these: they can now
10223 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
10224 into a single doc/ssleay.txt bundle. This way the information is still
10225 preserved but no longer messes up this directory. Now it's new room for
10226 the new set of documenation files.
10227 [Ralf S. Engelschall]
10229 *) SETs were incorrectly DER encoded. This was a major pain, because they
10230 shared code with SEQUENCEs, which aren't coded the same. This means that
10231 almost everything to do with SETs or SEQUENCEs has either changed name or
10232 number of arguments.
10233 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
10235 *) Fix test data to work with the above.
10238 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
10239 was already fixed by Eric for 0.9.1 it seems.
10240 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
10242 *) Autodetect FreeBSD3.
10245 *) Fix various bugs in Configure. This affects the following platforms:
10249 unixware-2.0-pentium
10253 *) Eliminate generated files from CVS. Reorder tests to regenerate files
10254 before they are needed.
10257 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
10261 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
10263 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
10264 changed SSLeay to OpenSSL in version strings.
10265 [Ralf S. Engelschall]
10267 *) Some fixups to the top-level documents.
10270 *) Fixed the nasty bug where rsaref.h was not found under compile-time
10271 because the symlink to include/ was missing.
10272 [Ralf S. Engelschall]
10274 *) Incorporated the popular no-RSA/DSA-only patches
10275 which allow to compile a RSA-free SSLeay.
10276 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
10278 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
10279 when "ssleay" is still not found.
10280 [Ralf S. Engelschall]
10282 *) Added more platforms to Configure: Cray T3E, HPUX 11,
10283 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
10285 *) Updated the README file.
10286 [Ralf S. Engelschall]
10288 *) Added various .cvsignore files in the CVS repository subdirs
10289 to make a "cvs update" really silent.
10290 [Ralf S. Engelschall]
10292 *) Recompiled the error-definition header files and added
10293 missing symbols to the Win32 linker tables.
10294 [Ralf S. Engelschall]
10296 *) Cleaned up the top-level documents;
10297 o new files: CHANGES and LICENSE
10298 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
10299 o merged COPYRIGHT into LICENSE
10300 o removed obsolete TODO file
10301 o renamed MICROSOFT to INSTALL.W32
10302 [Ralf S. Engelschall]
10304 *) Removed dummy files from the 0.9.1b source tree:
10305 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
10306 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
10307 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
10308 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
10309 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
10310 [Ralf S. Engelschall]
10312 *) Added various platform portability fixes.
10315 *) The Genesis of the OpenSSL rpject:
10316 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
10317 Young and Tim J. Hudson created while they were working for C2Net until
10319 [The OpenSSL Project]
10322 Changes between 0.9.0b and 0.9.1b [not released]
10324 *) Updated a few CA certificates under certs/
10327 *) Changed some BIGNUM api stuff.
10330 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
10331 DGUX x86, Linux Alpha, etc.
10334 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
10335 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
10339 *) Add -strparse option to asn1pars program which parses nested
10341 [Dr Stephen Henson <shenson@bigfoot.com>]
10343 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
10346 *) DSA fix for "ca" program.
10349 *) Added "-genkey" option to "dsaparam" program.
10352 *) Added RIPE MD160 (rmd160) message digest.
10355 *) Added -a (all) option to "ssleay version" command.
10358 *) Added PLATFORM define which is the id given to Configure.
10361 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
10364 *) Extended the ASN.1 parser routines.
10367 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
10370 *) Added a BN_CTX to the BN library.
10373 *) Fixed the weak key values in DES library
10376 *) Changed API in EVP library for cipher aliases.
10379 *) Added support for RC2/64bit cipher.
10382 *) Converted the lhash library to the crypto/mem.c functions.
10385 *) Added more recognized ASN.1 object ids.
10388 *) Added more RSA padding checks for SSL/TLS.
10391 *) Added BIO proxy/filter functionality.
10394 *) Added extra_certs to SSL_CTX which can be used
10395 send extra CA certificates to the client in the CA cert chain sending
10396 process. It can be configured with SSL_CTX_add_extra_chain_cert().
10399 *) Now Fortezza is denied in the authentication phase because
10400 this is key exchange mechanism is not supported by SSLeay at all.
10403 *) Additional PKCS1 checks.
10406 *) Support the string "TLSv1" for all TLS v1 ciphers.
10409 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
10410 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
10413 *) Fixed a few memory leaks.
10416 *) Fixed various code and comment typos.
10419 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
10420 bytes sent in the client random.
10421 [Edward Bishop <ebishop@spyglass.com>]