5 Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
7 *) Alternate chains certificate forgery
9 During certificate verfification, OpenSSL will attempt to find an
10 alternative certificate chain if the first attempt to build such a chain
11 fails. An error in the implementation of this logic can mean that an
12 attacker could cause certain checks on untrusted certificates to be
13 bypassed, such as the CA flag, enabling them to use a valid leaf
14 certificate to act as a CA and "issue" an invalid certificate.
16 This issue was reported to OpenSSL by Adam Langley/David Benjamin
20 Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
22 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
23 incompatibility in the handling of HMAC. The previous ABI has now been
26 Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
28 *) Malformed ECParameters causes infinite loop
30 When processing an ECParameters structure OpenSSL enters an infinite loop
31 if the curve specified is over a specially malformed binary polynomial
34 This can be used to perform denial of service against any
35 system which processes public keys, certificate requests or
36 certificates. This includes TLS clients and TLS servers with
37 client authentication enabled.
39 This issue was reported to OpenSSL by Joseph Barr-Pixton.
43 *) Exploitable out-of-bounds read in X509_cmp_time
45 X509_cmp_time does not properly check the length of the ASN1_TIME
46 string and can read a few bytes out of bounds. In addition,
47 X509_cmp_time accepts an arbitrary number of fractional seconds in the
50 An attacker can use this to craft malformed certificates and CRLs of
51 various sizes and potentially cause a segmentation fault, resulting in
52 a DoS on applications that verify certificates or CRLs. TLS clients
53 that verify CRLs are affected. TLS clients and servers with client
54 authentication enabled may be affected if they use custom verification
57 This issue was reported to OpenSSL by Robert Swiecki (Google), and
58 independently by Hanno Böck.
62 *) PKCS7 crash with missing EnvelopedContent
64 The PKCS#7 parsing code does not handle missing inner EncryptedContent
65 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
66 with missing content and trigger a NULL pointer dereference on parsing.
68 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
69 structures from untrusted sources are affected. OpenSSL clients and
70 servers are not affected.
72 This issue was reported to OpenSSL by Michal Zalewski (Google).
76 *) CMS verify infinite loop with unknown hash function
78 When verifying a signedData message the CMS code can enter an infinite loop
79 if presented with an unknown hash function OID. This can be used to perform
80 denial of service against any system which verifies signedData messages using
82 This issue was reported to OpenSSL by Johannes Bauer.
86 *) Race condition handling NewSessionTicket
88 If a NewSessionTicket is received by a multi-threaded client when attempting to
89 reuse a previous ticket then a race condition can occur potentially leading to
90 a double free of the ticket data.
94 *) Reject DH handshakes with parameters shorter than 768 bits.
95 [Kurt Roeckx and Emilia Kasper]
97 Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
99 *) Segmentation fault in ASN1_TYPE_cmp fix
101 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
102 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
103 certificate signature algorithm consistency this can be used to crash any
104 certificate verification operation and exploited in a DoS attack. Any
105 application which performs certificate verification is vulnerable including
106 OpenSSL clients and servers which enable client authentication.
110 *) ASN.1 structure reuse memory corruption fix
112 Reusing a structure in ASN.1 parsing may allow an attacker to cause
113 memory corruption via an invalid write. Such reuse is and has been
114 strongly discouraged and is believed to be rare.
116 Applications that parse structures containing CHOICE or ANY DEFINED BY
117 components may be affected. Certificate parsing (d2i_X509 and related
118 functions) are however not affected. OpenSSL clients and servers are
123 *) PKCS7 NULL pointer dereferences fix
125 The PKCS#7 parsing code does not handle missing outer ContentInfo
126 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
127 missing content and trigger a NULL pointer dereference on parsing.
129 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
130 otherwise parse PKCS#7 structures from untrusted sources are
131 affected. OpenSSL clients and servers are not affected.
133 This issue was reported to OpenSSL by Michal Zalewski (Google).
137 *) DoS via reachable assert in SSLv2 servers fix
139 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
140 servers that both support SSLv2 and enable export cipher suites by sending
141 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
143 This issue was discovered by Sean Burford (Google) and Emilia Käsper
144 (OpenSSL development team).
148 *) Use After Free following d2i_ECPrivatekey error fix
150 A malformed EC private key file consumed via the d2i_ECPrivateKey function
151 could cause a use after free condition. This, in turn, could cause a double
152 free in several private key parsing functions (such as d2i_PrivateKey
153 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
154 for applications that receive EC private keys from untrusted
155 sources. This scenario is considered rare.
157 This issue was discovered by the BoringSSL project and fixed in their
162 *) X509_to_X509_REQ NULL pointer deref fix
164 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
165 the certificate key is invalid. This function is rarely used in practice.
167 This issue was discovered by Brian Carpenter.
171 *) Removed the export ciphers from the DEFAULT ciphers
174 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
176 *) Build fixes for the Windows and OpenVMS platforms
177 [Matt Caswell and Richard Levitte]
179 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
181 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
182 message can cause a segmentation fault in OpenSSL due to a NULL pointer
183 dereference. This could lead to a Denial Of Service attack. Thanks to
184 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
188 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
189 dtls1_buffer_record function under certain conditions. In particular this
190 could occur if an attacker sent repeated DTLS records with the same
191 sequence number but for the next epoch. The memory leak could be exploited
192 by an attacker in a Denial of Service attack through memory exhaustion.
193 Thanks to Chris Mueller for reporting this issue.
197 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
198 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
199 method would be set to NULL which could later result in a NULL pointer
200 dereference. Thanks to Frank Schmirler for reporting this issue.
204 *) Abort handshake if server key exchange message is omitted for ephemeral
207 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
208 reporting this issue.
212 *) Remove non-export ephemeral RSA code on client and server. This code
213 violated the TLS standard by allowing the use of temporary RSA keys in
214 non-export ciphersuites and could be used by a server to effectively
215 downgrade the RSA key length used to a value smaller than the server
216 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
217 INRIA or reporting this issue.
221 *) Fixed issue where DH client certificates are accepted without verification.
222 An OpenSSL server will accept a DH certificate for client authentication
223 without the certificate verify message. This effectively allows a client to
224 authenticate without the use of a private key. This only affects servers
225 which trust a client certificate authority which issues certificates
226 containing DH keys: these are extremely rare and hardly ever encountered.
227 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
232 *) Ensure that the session ID context of an SSL is updated when its
233 SSL_CTX is updated via SSL_set_SSL_CTX.
235 The session ID context is typically set from the parent SSL_CTX,
236 and can vary with the CTX.
239 *) Fix various certificate fingerprint issues.
241 By using non-DER or invalid encodings outside the signed portion of a
242 certificate the fingerprint can be changed without breaking the signature.
243 Although no details of the signed portion of the certificate can be changed
244 this can cause problems with some applications: e.g. those using the
245 certificate fingerprint for blacklists.
247 1. Reject signatures with non zero unused bits.
249 If the BIT STRING containing the signature has non zero unused bits reject
250 the signature. All current signature algorithms require zero unused bits.
252 2. Check certificate algorithm consistency.
254 Check the AlgorithmIdentifier inside TBS matches the one in the
255 certificate signature. NB: this will result in signature failure
256 errors for some broken certificates.
258 Thanks to Konrad Kraszewski from Google for reporting this issue.
260 3. Check DSA/ECDSA signatures use DER.
262 Reencode DSA/ECDSA signatures and compare with the original received
263 signature. Return an error if there is a mismatch.
265 This will reject various cases including garbage after signature
266 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
267 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
268 (negative or with leading zeroes).
270 Further analysis was conducted and fixes were developed by Stephen Henson
271 of the OpenSSL core team.
276 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
277 results on some platforms, including x86_64. This bug occurs at random
278 with a very low probability, and is not known to be exploitable in any
279 way, though its exact impact is difficult to determine. Thanks to Pieter
280 Wuille (Blockstream) who reported this issue and also suggested an initial
281 fix. Further analysis was conducted by the OpenSSL development team and
282 Adam Langley of Google. The final fix was developed by Andy Polyakov of
283 the OpenSSL core team.
287 *) Do not resume sessions on the server if the negotiated protocol
288 version does not match the session's version. Resuming with a different
289 version, while not strictly forbidden by the RFC, is of questionable
290 sanity and breaks all known clients.
291 [David Benjamin, Emilia Käsper]
293 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
294 early CCS messages during renegotiation. (Note that because
295 renegotiation is encrypted, this early CCS was not exploitable.)
298 *) Tighten client-side session ticket handling during renegotiation:
299 ensure that the client only accepts a session ticket if the server sends
300 the extension anew in the ServerHello. Previously, a TLS client would
301 reuse the old extension state and thus accept a session ticket if one was
302 announced in the initial ServerHello.
304 Similarly, ensure that the client requires a session ticket if one
305 was advertised in the ServerHello. Previously, a TLS client would
306 ignore a missing NewSessionTicket message.
309 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
313 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
314 sends a carefully crafted handshake message, to cause OpenSSL to fail
315 to free up to 64k of memory causing a memory leak. This could be
316 exploited in a Denial Of Service attack. This issue affects OpenSSL
317 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
318 whether SRTP is used or configured. Implementations of OpenSSL that
319 have been compiled with OPENSSL_NO_SRTP defined are not affected.
321 The fix was developed by the OpenSSL team.
325 *) Session Ticket Memory Leak.
327 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
328 integrity of that ticket is first verified. In the event of a session
329 ticket integrity check failing, OpenSSL will fail to free memory
330 causing a memory leak. By sending a large number of invalid session
331 tickets an attacker could exploit this issue in a Denial Of Service
336 *) Build option no-ssl3 is incomplete.
338 When OpenSSL is configured with "no-ssl3" as a build option, servers
339 could accept and complete a SSL 3.0 handshake, and clients could be
340 configured to send them.
342 [Akamai and the OpenSSL team]
344 *) Add support for TLS_FALLBACK_SCSV.
345 Client applications doing fallback retries should call
346 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
348 [Adam Langley, Bodo Moeller]
350 *) Add additional DigestInfo checks.
352 Reencode DigestInto in DER and check against the original when
353 verifying RSA signature: this will reject any improperly encoded
354 DigestInfo structures.
356 Note: this is a precautionary measure and no attacks are currently known.
360 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
362 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
363 SRP code can be overrun an internal buffer. Add sanity check that
364 g, A, B < N to SRP code.
366 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
367 Group for discovering this issue.
371 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
372 TLS 1.0 instead of higher protocol versions when the ClientHello message
373 is badly fragmented. This allows a man-in-the-middle attacker to force a
374 downgrade to TLS 1.0 even if both the server and the client support a
375 higher protocol version, by modifying the client's TLS records.
377 Thanks to David Benjamin and Adam Langley (Google) for discovering and
378 researching this issue.
382 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
383 to a denial of service attack. A malicious server can crash the client
384 with a null pointer dereference (read) by specifying an anonymous (EC)DH
385 ciphersuite and sending carefully crafted handshake messages.
387 Thanks to Felix Gröbert (Google) for discovering and researching this
392 *) By sending carefully crafted DTLS packets an attacker could cause openssl
393 to leak memory. This can be exploited through a Denial of Service attack.
394 Thanks to Adam Langley for discovering and researching this issue.
398 *) An attacker can force openssl to consume large amounts of memory whilst
399 processing DTLS handshake messages. This can be exploited through a
400 Denial of Service attack.
401 Thanks to Adam Langley for discovering and researching this issue.
405 *) An attacker can force an error condition which causes openssl to crash
406 whilst processing DTLS packets due to memory being freed twice. This
407 can be exploited through a Denial of Service attack.
408 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
413 *) If a multithreaded client connects to a malicious server using a resumed
414 session and the server sends an ec point format extension it could write
415 up to 255 bytes to freed memory.
417 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
422 *) A malicious server can crash an OpenSSL client with a null pointer
423 dereference (read) by specifying an SRP ciphersuite even though it was not
424 properly negotiated with the client. This can be exploited through a
425 Denial of Service attack.
427 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
428 discovering and researching this issue.
432 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
433 X509_name_oneline, X509_name_print_ex et al. to leak some information
434 from the stack. Applications may be affected if they echo pretty printing
435 output to the attacker.
437 Thanks to Ivan Fratric (Google) for discovering this issue.
439 [Emilia Käsper, and Steve Henson]
441 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
442 for corner cases. (Certain input points at infinity could lead to
443 bogus results, with non-infinity inputs mapped to infinity too.)
446 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
448 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
449 handshake can force the use of weak keying material in OpenSSL
450 SSL/TLS clients and servers.
452 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
453 researching this issue. (CVE-2014-0224)
454 [KIKUCHI Masashi, Steve Henson]
456 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
457 OpenSSL DTLS client the code can be made to recurse eventually crashing
460 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
462 [Imre Rad, Steve Henson]
464 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
465 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
466 client or server. This is potentially exploitable to run arbitrary
467 code on a vulnerable client or server.
469 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
470 [Jüri Aedla, Steve Henson]
472 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
473 are subject to a denial of service attack.
475 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
476 this issue. (CVE-2014-3470)
477 [Felix Gröbert, Ivan Fratric, Steve Henson]
479 *) Harmonize version and its documentation. -f flag is used to display
481 [mancha <mancha1@zoho.com>]
483 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
485 [mancha <mancha1@zoho.com>]
487 *) Fix some double frees. These are not thought to be exploitable.
488 [mancha <mancha1@zoho.com>]
490 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
492 *) A missing bounds check in the handling of the TLS heartbeat extension
493 can be used to reveal up to 64k of memory to a connected client or
496 Thanks for Neel Mehta of Google Security for discovering this bug and to
497 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
498 preparing the fix (CVE-2014-0160)
499 [Adam Langley, Bodo Moeller]
501 *) Fix for the attack described in the paper "Recovering OpenSSL
502 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
503 by Yuval Yarom and Naomi Benger. Details can be obtained from:
504 http://eprint.iacr.org/2014/140
506 Thanks to Yuval Yarom and Naomi Benger for discovering this
507 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
508 [Yuval Yarom and Naomi Benger]
510 *) TLS pad extension: draft-agl-tls-padding-03
512 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
513 TLS client Hello record length value would otherwise be > 255 and
514 less that 512 pad with a dummy extension containing zeroes so it
515 is at least 512 bytes long.
517 [Adam Langley, Steve Henson]
519 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
521 *) Fix for TLS record tampering bug. A carefully crafted invalid
522 handshake could crash OpenSSL with a NULL pointer exception.
523 Thanks to Anton Johansson for reporting this issues.
526 *) Keep original DTLS digest and encryption contexts in retransmission
527 structures so we can use the previous session parameters if they need
528 to be resent. (CVE-2013-6450)
531 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
532 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
533 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
534 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
535 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
536 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
537 [Rob Stradling, Adam Langley]
539 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
541 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
542 supporting platforms or when small records were transferred.
543 [Andy Polyakov, Steve Henson]
545 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
547 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
549 This addresses the flaw in CBC record processing discovered by
550 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
551 at: http://www.isg.rhul.ac.uk/tls/
553 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
554 Security Group at Royal Holloway, University of London
555 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
556 Emilia Käsper for the initial patch.
558 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
560 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
561 ciphersuites which can be exploited in a denial of service attack.
562 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
563 and detecting this bug and to Wolfgang Ettlinger
564 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
568 *) Return an error when checking OCSP signatures when key is NULL.
569 This fixes a DoS attack. (CVE-2013-0166)
572 *) Make openssl verify return errors.
573 [Chris Palmer <palmer@google.com> and Ben Laurie]
575 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
576 the right response is stapled. Also change SSL_get_certificate()
577 so it returns the certificate actually sent.
578 See http://rt.openssl.org/Ticket/Display.html?id=2836.
579 [Rob Stradling <rob.stradling@comodo.com>]
581 *) Fix possible deadlock when decoding public keys.
584 *) Don't use TLS 1.0 record version number in initial client hello
588 Changes between 1.0.1b and 1.0.1c [10 May 2012]
590 *) Sanity check record length before skipping explicit IV in TLS
591 1.2, 1.1 and DTLS to fix DoS attack.
593 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
594 fuzzing as a service testing platform.
598 *) Initialise tkeylen properly when encrypting CMS messages.
599 Thanks to Solar Designer of Openwall for reporting this issue.
602 *) In FIPS mode don't try to use composite ciphers as they are not
606 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
608 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
609 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
610 mean any application compiled against OpenSSL 1.0.0 headers setting
611 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
612 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
613 0x10000000L Any application which was previously compiled against
614 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
615 will need to be recompiled as a result. Letting be results in
616 inability to disable specifically TLS 1.1 and in client context,
617 in unlike event, limit maximum offered version to TLS 1.0 [see below].
620 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
621 disable just protocol X, but all protocols above X *if* there are
622 protocols *below* X still enabled. In more practical terms it means
623 that if application wants to disable TLS1.0 in favor of TLS1.1 and
624 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
625 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
629 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
631 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
632 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
633 in CRYPTO_realloc_clean.
635 Thanks to Tavis Ormandy, Google Security Team, for discovering this
636 issue and to Adam Langley <agl@chromium.org> for fixing it.
638 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
640 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
643 *) Workarounds for some broken servers that "hang" if a client hello
644 record length exceeds 255 bytes.
646 1. Do not use record version number > TLS 1.0 in initial client
647 hello: some (but not all) hanging servers will now work.
648 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
649 the number of ciphers sent in the client hello. This should be
650 set to an even number, such as 50, for example by passing:
651 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
652 Most broken servers should now work.
653 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
654 TLS 1.2 client support entirely.
657 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
660 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
662 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
663 STRING form instead of a DigestInfo.
666 *) The format used for MDC2 RSA signatures is inconsistent between EVP
667 and the RSA_sign/RSA_verify functions. This was made more apparent when
668 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
669 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
670 the correct format in RSA_verify so both forms transparently work.
673 *) Some servers which support TLS 1.0 can choke if we initially indicate
674 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
675 encrypted premaster secret. As a workaround use the maximum pemitted
676 client version in client hello, this should keep such servers happy
677 and still work with previous versions of OpenSSL.
680 *) Add support for TLS/DTLS heartbeats.
681 [Robin Seggelmann <seggelmann@fh-muenster.de>]
683 *) Add support for SCTP.
684 [Robin Seggelmann <seggelmann@fh-muenster.de>]
686 *) Improved PRNG seeding for VOS.
687 [Paul Green <Paul.Green@stratus.com>]
689 *) Extensive assembler packs updates, most notably:
691 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
692 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
693 - x86_64: bit-sliced AES implementation;
694 - ARM: NEON support, contemporary platforms optimizations;
695 - s390x: z196 support;
696 - *: GHASH and GF(2^m) multiplication implementations;
700 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
701 (removal of unnecessary code)
702 [Peter Sylvester <peter.sylvester@edelweb.fr>]
704 *) Add TLS key material exporter from RFC 5705.
707 *) Add DTLS-SRTP negotiation from RFC 5764.
710 *) Add Next Protocol Negotiation,
711 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
712 disabled with a no-npn flag to config or Configure. Code donated
714 [Adam Langley <agl@google.com> and Ben Laurie]
716 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
717 NIST-P256, NIST-P521, with constant-time single point multiplication on
718 typical inputs. Compiler support for the nonstandard type __uint128_t is
719 required to use this (present in gcc 4.4 and later, for 64-bit builds).
720 Code made available under Apache License version 2.0.
722 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
723 line to include this in your build of OpenSSL, and run "make depend" (or
724 "make update"). This enables the following EC_METHODs:
726 EC_GFp_nistp224_method()
727 EC_GFp_nistp256_method()
728 EC_GFp_nistp521_method()
730 EC_GROUP_new_by_curve_name() will automatically use these (while
731 EC_GROUP_new_curve_GFp() currently prefers the more flexible
733 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
735 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
736 all platforms. Move ssize_t definition from e_os.h to the public
737 header file e_os2.h as it now appears in public header file cms.h
740 *) New -sigopt option to the ca, req and x509 utilities. Additional
741 signature parameters can be passed using this option and in
745 *) Add RSA PSS signing function. This will generate and set the
746 appropriate AlgorithmIdentifiers for PSS based on those in the
747 corresponding EVP_MD_CTX structure. No application support yet.
750 *) Support for companion algorithm specific ASN1 signing routines.
751 New function ASN1_item_sign_ctx() signs a pre-initialised
752 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
753 the appropriate parameters.
756 *) Add new algorithm specific ASN1 verification initialisation function
757 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
758 handling will be the same no matter what EVP_PKEY_METHOD is used.
759 Add a PSS handler to support verification of PSS signatures: checked
760 against a number of sample certificates.
763 *) Add signature printing for PSS. Add PSS OIDs.
764 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
766 *) Add algorithm specific signature printing. An individual ASN1 method
767 can now print out signatures instead of the standard hex dump.
769 More complex signatures (e.g. PSS) can print out more meaningful
770 information. Include DSA version that prints out the signature
774 *) Password based recipient info support for CMS library: implementing
778 *) Split password based encryption into PBES2 and PBKDF2 functions. This
779 neatly separates the code into cipher and PBE sections and is required
780 for some algorithms that split PBES2 into separate pieces (such as
784 *) Session-handling fixes:
785 - Fix handling of connections that are resuming with a session ID,
786 but also support Session Tickets.
787 - Fix a bug that suppressed issuing of a new ticket if the client
788 presented a ticket with an expired session.
789 - Try to set the ticket lifetime hint to something reasonable.
790 - Make tickets shorter by excluding irrelevant information.
791 - On the client side, don't ignore renewed tickets.
792 [Adam Langley, Bodo Moeller (Google)]
794 *) Fix PSK session representation.
797 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
799 This work was sponsored by Intel.
802 *) Add GCM support to TLS library. Some custom code is needed to split
803 the IV between the fixed (from PRF) and explicit (from TLS record)
804 portions. This adds all GCM ciphersuites supported by RFC5288 and
805 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
806 add a special AESGCM string for GCM only.
809 *) Expand range of ctrls for AES GCM. Permit setting invocation
810 field on decrypt and retrieval of invocation field only on encrypt.
813 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
814 As required by RFC5289 these ciphersuites cannot be used if for
815 versions of TLS earlier than 1.2.
818 *) For FIPS capable OpenSSL interpret a NULL default public key method
819 as unset and return the appopriate default but do *not* set the default.
820 This means we can return the appopriate method in applications that
821 swicth between FIPS and non-FIPS modes.
824 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
825 ENGINE is used then we cannot handle that in the FIPS module so we
826 keep original code iff non-FIPS operations are allowed.
829 *) Add -attime option to openssl utilities.
830 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
832 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
835 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
836 FIPS EC methods unconditionally for now.
839 *) New build option no-ec2m to disable characteristic 2 code.
842 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
843 all cases can be covered as some introduce binary incompatibilities.
846 *) Redirect RSA operations to FIPS module including keygen,
847 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
850 *) Add similar low level API blocking to ciphers.
853 *) Low level digest APIs are not approved in FIPS mode: any attempt
854 to use these will cause a fatal error. Applications that *really* want
855 to use them can use the private_* version instead.
858 *) Redirect cipher operations to FIPS module for FIPS builds.
861 *) Redirect digest operations to FIPS module for FIPS builds.
864 *) Update build system to add "fips" flag which will link in fipscanister.o
865 for static and shared library builds embedding a signature if needed.
868 *) Output TLS supported curves in preference order instead of numerical
869 order. This is currently hardcoded for the highest order curves first.
870 This should be configurable so applications can judge speed vs strength.
873 *) Add TLS v1.2 server support for client authentication.
876 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
880 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
881 FIPS modules versions.
884 *) Add TLS v1.2 client side support for client authentication. Keep cache
885 of handshake records longer as we don't know the hash algorithm to use
886 until after the certificate request message is received.
889 *) Initial TLS v1.2 client support. Add a default signature algorithms
890 extension including all the algorithms we support. Parse new signature
891 format in client key exchange. Relax some ECC signing restrictions for
892 TLS v1.2 as indicated in RFC5246.
895 *) Add server support for TLS v1.2 signature algorithms extension. Switch
896 to new signature format when needed using client digest preference.
897 All server ciphersuites should now work correctly in TLS v1.2. No client
898 support yet and no support for client certificates.
901 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
902 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
903 ciphersuites. At present only RSA key exchange ciphersuites work with
904 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
905 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
906 and version checking.
909 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
910 with this defined it will not be affected by any changes to ssl internal
911 structures. Add several utility functions to allow openssl application
912 to work with OPENSSL_NO_SSL_INTERN defined.
916 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
918 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
921 *) Permit abbreviated handshakes when renegotiating using the function
922 SSL_renegotiate_abbreviated().
923 [Robin Seggelmann <seggelmann@fh-muenster.de>]
925 *) Add call to ENGINE_register_all_complete() to
926 ENGINE_load_builtin_engines(), so some implementations get used
927 automatically instead of needing explicit application support.
930 *) Add support for TLS key exporter as described in RFC5705.
931 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
933 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
934 a few changes are required:
936 Add SSL_OP_NO_TLSv1_1 flag.
938 Update version checking logic to handle version 1.1.
939 Add explicit IV handling (ported from DTLS code).
940 Add command line options to s_client/s_server.
943 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
945 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
946 in CMS and PKCS7 code. When RSA decryption fails use a random key for
947 content decryption and always return the same error. Note: this attack
948 needs on average 2^20 messages so it only affects automated senders. The
949 old behaviour can be reenabled in the CMS code by setting the
950 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
951 an MMA defence is not necessary.
952 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
953 this issue. (CVE-2012-0884)
956 *) Fix CVE-2011-4619: make sure we really are receiving a
957 client hello before rejecting multiple SGC restarts. Thanks to
958 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
961 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
963 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
964 Thanks to Antonio Martin, Enterprise Secure Access Research and
965 Development, Cisco Systems, Inc. for discovering this bug and
966 preparing a fix. (CVE-2012-0050)
969 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
971 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
972 of the Vaudenay padding oracle attack on CBC mode encryption
973 which enables an efficient plaintext recovery attack against
974 the OpenSSL implementation of DTLS. Their attack exploits timing
975 differences arising during decryption processing. A research
976 paper describing this attack can be found at:
977 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
978 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
979 Security Group at Royal Holloway, University of London
980 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
981 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
982 for preparing the fix. (CVE-2011-4108)
983 [Robin Seggelmann, Michael Tuexen]
985 *) Clear bytes used for block padding of SSL 3.0 records.
987 [Adam Langley (Google)]
989 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
990 Kadianakis <desnacked@gmail.com> for discovering this issue and
991 Adam Langley for preparing the fix. (CVE-2011-4619)
992 [Adam Langley (Google)]
994 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
995 [Andrey Kulikov <amdeich@gmail.com>]
997 *) Prevent malformed RFC3779 data triggering an assertion failure.
998 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
999 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1000 [Rob Austein <sra@hactrn.net>]
1002 *) Improved PRNG seeding for VOS.
1003 [Paul Green <Paul.Green@stratus.com>]
1005 *) Fix ssl_ciph.c set-up race.
1006 [Adam Langley (Google)]
1008 *) Fix spurious failures in ecdsatest.c.
1009 [Emilia Käsper (Google)]
1011 *) Fix the BIO_f_buffer() implementation (which was mixing different
1012 interpretations of the '..._len' fields).
1013 [Adam Langley (Google)]
1015 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1016 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1017 threads won't reuse the same blinding coefficients.
1019 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1020 lock to call BN_BLINDING_invert_ex, and avoids one use of
1021 BN_BLINDING_update for each BN_BLINDING structure (previously,
1022 the last update always remained unused).
1023 [Emilia Käsper (Google)]
1025 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
1026 [Bob Buckholz (Google)]
1028 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
1030 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
1031 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
1032 [Kaspar Brand <ossl@velox.ch>]
1034 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1035 for multi-threaded use of ECDH. (CVE-2011-3210)
1036 [Adam Langley (Google)]
1038 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1041 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
1042 signature public key algorithm by using OID xref utilities instead.
1043 Before this you could only use some ECC ciphersuites with SHA1 only.
1046 *) Add protection against ECDSA timing attacks as mentioned in the paper
1047 by Billy Bob Brumley and Nicola Tuveri, see:
1049 http://eprint.iacr.org/2011/232.pdf
1051 [Billy Bob Brumley and Nicola Tuveri]
1053 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
1055 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1056 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1058 *) Fix bug in string printing code: if *any* escaping is enabled we must
1059 escape the escape character (backslash) or the resulting string is
1063 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
1065 *) Disable code workaround for ancient and obsolete Netscape browsers
1066 and servers: an attacker can use it in a ciphersuite downgrade attack.
1067 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1070 *) Fixed J-PAKE implementation error, originally discovered by
1071 Sebastien Martini, further info and confirmation from Stefan
1072 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1075 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
1077 *) Fix extension code to avoid race conditions which can result in a buffer
1078 overrun vulnerability: resumed sessions must not be modified as they can
1079 be shared by multiple threads. CVE-2010-3864
1082 *) Fix WIN32 build system to correctly link an ENGINE directory into
1086 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
1088 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
1090 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
1092 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
1094 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
1095 context. The operation can be customised via the ctrl mechanism in
1096 case ENGINEs want to include additional functionality.
1099 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
1102 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
1103 output hashes compatible with older versions of OpenSSL.
1104 [Willy Weisz <weisz@vcpc.univie.ac.at>]
1106 *) Fix compression algorithm handling: if resuming a session use the
1107 compression algorithm of the resumed session instead of determining
1108 it from client hello again. Don't allow server to change algorithm.
1111 *) Add load_crls() function to apps tidying load_certs() too. Add option
1112 to verify utility to allow additional CRLs to be included.
1115 *) Update OCSP request code to permit adding custom headers to the request:
1116 some responders need this.
1119 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
1121 [Julia Lawall <julia@diku.dk>]
1123 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
1124 needlessly dereferenced structures, used obsolete functions and
1125 didn't handle all updated verify codes correctly.
1128 *) Disable MD2 in the default configuration.
1131 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
1132 indicate the initial BIO being pushed or popped. This makes it possible
1133 to determine whether the BIO is the one explicitly called or as a result
1134 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
1135 it handles reference counts correctly and doesn't zero out the I/O bio
1136 when it is not being explicitly popped. WARNING: applications which
1137 included workarounds for the old buggy behaviour will need to be modified
1138 or they could free up already freed BIOs.
1141 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
1142 renaming to all platforms (within the 0.9.8 branch, this was
1143 done conditionally on Netware platforms to avoid a name clash).
1144 [Guenter <lists@gknw.net>]
1146 *) Add ECDHE and PSK support to DTLS.
1147 [Michael Tuexen <tuexen@fh-muenster.de>]
1149 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
1153 *) Add "missing" function EVP_MD_flags() (without this the only way to
1154 retrieve a digest flags is by accessing the structure directly. Update
1155 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
1156 or cipher is registered as in the "from" argument. Print out all
1157 registered digests in the dgst usage message instead of manually
1158 attempting to work them out.
1161 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
1162 this allows the use of compression and extensions. Change default cipher
1163 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
1164 by default unless an application cipher string requests it.
1167 *) Alter match criteria in PKCS12_parse(). It used to try to use local
1168 key ids to find matching certificates and keys but some PKCS#12 files
1169 don't follow the (somewhat unwritten) rules and this strategy fails.
1170 Now just gather all certificates together and the first private key
1171 then look for the first certificate that matches the key.
1174 *) Support use of registered digest and cipher names for dgst and cipher
1175 commands instead of having to add each one as a special case. So now
1182 openssl dgst -sha256 foo
1184 and this works for ENGINE based algorithms too.
1188 *) Update Gost ENGINE to support parameter files.
1189 [Victor B. Wagner <vitus@cryptocom.ru>]
1191 *) Support GeneralizedTime in ca utility.
1192 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
1194 *) Enhance the hash format used for certificate directory links. The new
1195 form uses the canonical encoding (meaning equivalent names will work
1196 even if they aren't identical) and uses SHA1 instead of MD5. This form
1197 is incompatible with the older format and as a result c_rehash should
1198 be used to rebuild symbolic links.
1201 *) Make PKCS#8 the default write format for private keys, replacing the
1202 traditional format. This form is standardised, more secure and doesn't
1203 include an implicit MD5 dependency.
1206 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
1207 committed to OpenSSL should pass this lot as a minimum.
1210 *) Add session ticket override functionality for use by EAP-FAST.
1211 [Jouni Malinen <j@w1.fi>]
1213 *) Modify HMAC functions to return a value. Since these can be implemented
1214 in an ENGINE errors can occur.
1217 *) Type-checked OBJ_bsearch_ex.
1220 *) Type-checked OBJ_bsearch. Also some constification necessitated
1221 by type-checking. Still to come: TXT_DB, bsearch(?),
1222 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
1226 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
1227 seconds to a tm structure directly, instead of going through OS
1228 specific date routines. This avoids any issues with OS routines such
1229 as the year 2038 bug. New *_adj() functions for ASN1 time structures
1230 and X509_time_adj_ex() to cover the extended range. The existing
1231 X509_time_adj() is still usable and will no longer have any date issues.
1234 *) Delta CRL support. New use deltas option which will attempt to locate
1235 and search any appropriate delta CRLs available.
1237 This work was sponsored by Google.
1240 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
1241 code and add additional score elements. Validate alternate CRL paths
1242 as part of the CRL checking and indicate a new error "CRL path validation
1243 error" in this case. Applications wanting additional details can use
1244 the verify callback and check the new "parent" field. If this is not
1245 NULL CRL path validation is taking place. Existing applications wont
1246 see this because it requires extended CRL support which is off by
1249 This work was sponsored by Google.
1252 *) Support for freshest CRL extension.
1254 This work was sponsored by Google.
1257 *) Initial indirect CRL support. Currently only supported in the CRLs
1258 passed directly and not via lookup. Process certificate issuer
1259 CRL entry extension and lookup CRL entries by bother issuer name
1260 and serial number. Check and process CRL issuer entry in IDP extension.
1262 This work was sponsored by Google.
1265 *) Add support for distinct certificate and CRL paths. The CRL issuer
1266 certificate is validated separately in this case. Only enabled if
1267 an extended CRL support flag is set: this flag will enable additional
1268 CRL functionality in future.
1270 This work was sponsored by Google.
1273 *) Add support for policy mappings extension.
1275 This work was sponsored by Google.
1278 *) Fixes to pathlength constraint, self issued certificate handling,
1279 policy processing to align with RFC3280 and PKITS tests.
1281 This work was sponsored by Google.
1284 *) Support for name constraints certificate extension. DN, email, DNS
1285 and URI types are currently supported.
1287 This work was sponsored by Google.
1290 *) To cater for systems that provide a pointer-based thread ID rather
1291 than numeric, deprecate the current numeric thread ID mechanism and
1292 replace it with a structure and associated callback type. This
1293 mechanism allows a numeric "hash" to be extracted from a thread ID in
1294 either case, and on platforms where pointers are larger than 'long',
1295 mixing is done to help ensure the numeric 'hash' is usable even if it
1296 can't be guaranteed unique. The default mechanism is to use "&errno"
1297 as a pointer-based thread ID to distinguish between threads.
1299 Applications that want to provide their own thread IDs should now use
1300 CRYPTO_THREADID_set_callback() to register a callback that will call
1301 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
1303 Note that ERR_remove_state() is now deprecated, because it is tied
1304 to the assumption that thread IDs are numeric. ERR_remove_state(0)
1305 to free the current thread's error state should be replaced by
1306 ERR_remove_thread_state(NULL).
1308 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
1309 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
1310 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
1311 application was previously providing a numeric thread callback that
1312 was inappropriate for distinguishing threads, then uniqueness might
1313 have been obtained with &errno that happened immediately in the
1314 intermediate development versions of OpenSSL; this is no longer the
1315 case, the numeric thread callback will now override the automatic use
1317 [Geoff Thorpe, with help from Bodo Moeller]
1319 *) Initial support for different CRL issuing certificates. This covers a
1320 simple case where the self issued certificates in the chain exist and
1321 the real CRL issuer is higher in the existing chain.
1323 This work was sponsored by Google.
1326 *) Removed effectively defunct crypto/store from the build.
1329 *) Revamp of STACK to provide stronger type-checking. Still to come:
1330 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
1331 ASN1_STRING, CONF_VALUE.
1334 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
1335 RAM on SSL connections. This option can save about 34k per idle SSL.
1338 *) Revamp of LHASH to provide stronger type-checking. Still to come:
1339 STACK, TXT_DB, bsearch, qsort.
1342 *) Initial support for Cryptographic Message Syntax (aka CMS) based
1343 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
1344 support for data, signedData, compressedData, digestedData and
1345 encryptedData, envelopedData types included. Scripts to check against
1346 RFC4134 examples draft and interop and consistency checks of many
1347 content types and variants.
1350 *) Add options to enc utility to support use of zlib compression BIO.
1353 *) Extend mk1mf to support importing of options and assembly language
1354 files from Configure script, currently only included in VC-WIN32.
1355 The assembly language rules can now optionally generate the source
1356 files from the associated perl scripts.
1359 *) Implement remaining functionality needed to support GOST ciphersuites.
1360 Interop testing has been performed using CryptoPro implementations.
1361 [Victor B. Wagner <vitus@cryptocom.ru>]
1363 *) s390x assembler pack.
1366 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
1370 *) Implement Opaque PRF Input TLS extension as specified in
1371 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
1372 official specification yet and no extension type assignment by
1373 IANA exists, this extension (for now) will have to be explicitly
1374 enabled when building OpenSSL by providing the extension number
1375 to use. For example, specify an option
1377 -DTLSEXT_TYPE_opaque_prf_input=0x9527
1379 to the "config" or "Configure" script to enable the extension,
1380 assuming extension number 0x9527 (which is a completely arbitrary
1381 and unofficial assignment based on the MD5 hash of the Internet
1382 Draft). Note that by doing so, you potentially lose
1383 interoperability with other TLS implementations since these might
1384 be using the same extension number for other purposes.
1386 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
1387 opaque PRF input value to use in the handshake. This will create
1388 an interal copy of the length-'len' string at 'src', and will
1389 return non-zero for success.
1391 To get more control and flexibility, provide a callback function
1394 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
1395 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
1399 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
1402 Callback function 'cb' will be called in handshakes, and is
1403 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
1404 Argument 'arg' is for application purposes (the value as given to
1405 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
1406 be provided to the callback function). The callback function
1407 has to return non-zero to report success: usually 1 to use opaque
1408 PRF input just if possible, or 2 to enforce use of the opaque PRF
1409 input. In the latter case, the library will abort the handshake
1410 if opaque PRF input is not successfully negotiated.
1412 Arguments 'peerinput' and 'len' given to the callback function
1413 will always be NULL and 0 in the case of a client. A server will
1414 see the client's opaque PRF input through these variables if
1415 available (NULL and 0 otherwise). Note that if the server
1416 provides an opaque PRF input, the length must be the same as the
1417 length of the client's opaque PRF input.
1419 Note that the callback function will only be called when creating
1420 a new session (session resumption can resume whatever was
1421 previously negotiated), and will not be called in SSL 2.0
1422 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
1423 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
1424 for applications that need to enforce opaque PRF input.
1428 *) Update ssl code to support digests other than SHA1+MD5 for handshake
1431 [Victor B. Wagner <vitus@cryptocom.ru>]
1433 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1434 RFC4507bis. The encrypted ticket format is an encrypted encoded
1435 SSL_SESSION structure, that way new session features are automatically
1438 If a client application caches session in an SSL_SESSION structure
1439 support is transparent because tickets are now stored in the encoded
1442 The SSL_CTX structure automatically generates keys for ticket
1443 protection in servers so again support should be possible
1444 with no application modification.
1446 If a client or server wishes to disable RFC4507 support then the option
1447 SSL_OP_NO_TICKET can be set.
1449 Add a TLS extension debugging callback to allow the contents of any client
1450 or server extensions to be examined.
1452 This work was sponsored by Google.
1455 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
1456 OpenSSL should now compile cleanly on gcc 4.2
1457 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
1459 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
1460 support including streaming MAC support: this is required for GOST
1461 ciphersuite support.
1462 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
1464 *) Add option -stream to use PKCS#7 streaming in smime utility. New
1465 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
1466 to output in BER and PEM format.
1469 *) Experimental support for use of HMAC via EVP_PKEY interface. This
1470 allows HMAC to be handled via the EVP_DigestSign*() interface. The
1471 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
1472 ENGINE support for HMAC keys which are unextractable. New -mac and
1473 -macopt options to dgst utility.
1476 *) New option -sigopt to dgst utility. Update dgst to use
1477 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
1478 alternative signing paramaters such as X9.31 or PSS in the dgst
1482 *) Change ssl_cipher_apply_rule(), the internal function that does
1483 the work each time a ciphersuite string requests enabling
1484 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
1485 removing ("!foo+bar") a class of ciphersuites: Now it maintains
1486 the order of disabled ciphersuites such that those ciphersuites
1487 that most recently went from enabled to disabled not only stay
1488 in order with respect to each other, but also have higher priority
1489 than other disabled ciphersuites the next time ciphersuites are
1492 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
1493 the same ciphersuites as with "HIGH" alone, but in a specific
1494 order where the PSK ciphersuites come first (since they are the
1495 most recently disabled ciphersuites when "HIGH" is parsed).
1497 Also, change ssl_create_cipher_list() (using this new
1498 funcionality) such that between otherwise identical
1499 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
1503 *) Change ssl_create_cipher_list() so that it automatically
1504 arranges the ciphersuites in reasonable order before starting
1505 to process the rule string. Thus, the definition for "DEFAULT"
1506 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
1507 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
1508 This makes it much easier to arrive at a reasonable default order
1509 in applications for which anonymous ciphers are OK (meaning
1510 that you can't actually use DEFAULT).
1511 [Bodo Moeller; suggested by Victor Duchovni]
1513 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
1514 processing) into multiple integers instead of setting
1515 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
1516 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
1517 (These masks as well as the individual bit definitions are hidden
1518 away into the non-exported interface ssl/ssl_locl.h, so this
1519 change to the definition of the SSL_CIPHER structure shouldn't
1520 affect applications.) This give us more bits for each of these
1521 categories, so there is no longer a need to coagulate AES128 and
1522 AES256 into a single algorithm bit, and to coagulate Camellia128
1523 and Camellia256 into a single algorithm bit, which has led to all
1526 Thus, among other things, the kludge introduced in 0.9.7m and
1527 0.9.8e for masking out AES256 independently of AES128 or masking
1528 out Camellia256 independently of AES256 is not needed here in 0.9.9.
1530 With the change, we also introduce new ciphersuite aliases that
1531 so far were missing: "AES128", "AES256", "CAMELLIA128", and
1535 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
1536 Use the leftmost N bytes of the signature input if the input is
1537 larger than the prime q (with N being the size in bytes of q).
1540 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
1541 it yet and it is largely untested.
1544 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
1547 *) Initial incomplete changes to avoid need for function casts in OpenSSL
1548 some compilers (gcc 4.2 and later) reject their use. Safestack is
1549 reimplemented. Update ASN1 to avoid use of legacy functions.
1552 *) Win32/64 targets are linked with Winsock2.
1555 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
1556 to external functions. This can be used to increase CRL handling
1557 efficiency especially when CRLs are very large by (for example) storing
1558 the CRL revoked certificates in a database.
1561 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
1562 new CRLs added to a directory can be used. New command line option
1563 -verify_return_error to s_client and s_server. This causes real errors
1564 to be returned by the verify callback instead of carrying on no matter
1565 what. This reflects the way a "real world" verify callback would behave.
1568 *) GOST engine, supporting several GOST algorithms and public key formats.
1569 Kindly donated by Cryptocom.
1572 *) Partial support for Issuing Distribution Point CRL extension. CRLs
1573 partitioned by DP are handled but no indirect CRL or reason partitioning
1574 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
1575 selected via a scoring technique which handles IDP and AKID in CRLs.
1578 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
1579 will ultimately be used for all verify operations: this will remove the
1580 X509_STORE dependency on certificate verification and allow alternative
1581 lookup methods. X509_STORE based implementations of these two callbacks.
1584 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
1585 Modify get_crl() to find a valid (unexpired) CRL if possible.
1588 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
1589 this would be called X509_CRL_cmp() but that name is already used by
1590 a function that just compares CRL issuer names. Cache several CRL
1591 extensions in X509_CRL structure and cache CRLDP in X509.
1594 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
1595 this maps equivalent X509_NAME structures into a consistent structure.
1596 Name comparison can then be performed rapidly using memcmp().
1599 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
1603 *) Allow digests to supply their own micalg string for S/MIME type using
1604 the ctrl EVP_MD_CTRL_MICALG.
1607 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
1608 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
1609 ctrl. It can then customise the structure before and/or after signing
1613 *) New function OBJ_add_sigid() to allow application defined signature OIDs
1614 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
1615 to free up any added signature OIDs.
1618 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
1619 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
1620 digest and cipher tables. New options added to openssl utility:
1621 list-message-digest-algorithms and list-cipher-algorithms.
1624 *) Change the array representation of binary polynomials: the list
1625 of degrees of non-zero coefficients is now terminated with -1.
1626 Previously it was terminated with 0, which was also part of the
1627 value; thus, the array representation was not applicable to
1628 polynomials where t^0 has coefficient zero. This change makes
1629 the array representation useful in a more general context.
1632 *) Various modifications and fixes to SSL/TLS cipher string
1633 handling. For ECC, the code now distinguishes between fixed ECDH
1634 with RSA certificates on the one hand and with ECDSA certificates
1635 on the other hand, since these are separate ciphersuites. The
1636 unused code for Fortezza ciphersuites has been removed.
1638 For consistency with EDH, ephemeral ECDH is now called "EECDH"
1639 (not "ECDHE"). For consistency with the code for DH
1640 certificates, use of ECDH certificates is now considered ECDH
1641 authentication, not RSA or ECDSA authentication (the latter is
1642 merely the CA's signing algorithm and not actively used in the
1645 The temporary ciphersuite alias "ECCdraft" is no longer
1646 available, and ECC ciphersuites are no longer excluded from "ALL"
1647 and "DEFAULT". The following aliases now exist for RFC 4492
1648 ciphersuites, most of these by analogy with the DH case:
1650 kECDHr - ECDH cert, signed with RSA
1651 kECDHe - ECDH cert, signed with ECDSA
1652 kECDH - ECDH cert (signed with either RSA or ECDSA)
1653 kEECDH - ephemeral ECDH
1654 ECDH - ECDH cert or ephemeral ECDH
1660 AECDH - anonymous ECDH
1661 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
1665 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
1666 Use correct micalg parameters depending on digest(s) in signed message.
1669 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
1670 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
1673 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
1674 an engine to register a method. Add ENGINE lookups for methods and
1675 functional reference processing.
1678 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
1679 EVP_{Sign,Verify}* which allow an application to customise the signature
1683 *) New -resign option to smime utility. This adds one or more signers
1684 to an existing PKCS#7 signedData structure. Also -md option to use an
1685 alternative message digest algorithm for signing.
1688 *) Tidy up PKCS#7 routines and add new functions to make it easier to
1689 create PKCS7 structures containing multiple signers. Update smime
1690 application to support multiple signers.
1693 *) New -macalg option to pkcs12 utility to allow setting of an alternative
1697 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
1698 Reorganize PBE internals to lookup from a static table using NIDs,
1699 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
1700 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
1701 PRF which will be automatically used with PBES2.
1704 *) Replace the algorithm specific calls to generate keys in "req" with the
1708 *) Update PKCS#7 enveloped data routines to use new API. This is now
1709 supported by any public key method supporting the encrypt operation. A
1710 ctrl is added to allow the public key algorithm to examine or modify
1711 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
1715 *) Add a ctrl to asn1 method to allow a public key algorithm to express
1716 a default digest type to use. In most cases this will be SHA1 but some
1717 algorithms (such as GOST) need to specify an alternative digest. The
1718 return value indicates how strong the prefernce is 1 means optional and
1719 2 is mandatory (that is it is the only supported type). Modify
1720 ASN1_item_sign() to accept a NULL digest argument to indicate it should
1721 use the default md. Update openssl utilities to use the default digest
1722 type for signing if it is not explicitly indicated.
1725 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
1726 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
1727 signing method from the key type. This effectively removes the link
1728 between digests and public key types.
1731 *) Add an OID cross reference table and utility functions. Its purpose is to
1732 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
1733 rsaEncryption. This will allow some of the algorithm specific hackery
1734 needed to use the correct OID to be removed.
1737 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
1738 structures for PKCS7_sign(). They are now set up by the relevant public
1742 *) Add provisional EC pkey method with support for ECDSA and ECDH.
1745 *) Add support for key derivation (agreement) in the API, DH method and
1749 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
1750 public and private key formats. As a side effect these add additional
1751 command line functionality not previously available: DSA signatures can be
1752 generated and verified using pkeyutl and DH key support and generation in
1757 [Oliver Tappe <zooey@hirschkaefer.de>]
1759 *) New make target "install_html_docs" installs HTML renditions of the
1761 [Oliver Tappe <zooey@hirschkaefer.de>]
1763 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
1764 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
1765 support key and parameter generation and add initial key generation
1766 functionality for RSA.
1769 *) Add functions for main EVP_PKEY_method operations. The undocumented
1770 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
1771 EVP_PKEY_{encrypt,decrypt}_old.
1774 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
1775 key API, doesn't do much yet.
1778 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
1779 public key algorithms. New option to openssl utility:
1780 "list-public-key-algorithms" to print out info.
1783 *) Implement the Supported Elliptic Curves Extension for
1784 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1787 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
1788 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
1791 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
1792 utilities such as rsa, dsa, dsaparam etc except they process any key
1796 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1797 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1798 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
1802 *) Initial support for pluggable public key ASN1.
1803 De-spaghettify the public key ASN1 handling. Move public and private
1804 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
1805 algorithm specific handling to a single module within the relevant
1806 algorithm directory. Add functions to allow (near) opaque processing
1807 of public and private key structures.
1810 *) Implement the Supported Point Formats Extension for
1811 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1814 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
1815 for the psk identity [hint] and the psk callback functions to the
1816 SSL_SESSION, SSL and SSL_CTX structure.
1819 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
1823 SSL_CTX_use_psk_identity_hint
1824 SSL_get_psk_identity_hint
1825 SSL_get_psk_identity
1826 SSL_use_psk_identity_hint
1828 [Mika Kousa and Pasi Eronen of Nokia Corporation]
1830 *) Add RFC 3161 compliant time stamp request creation, response generation
1831 and response verification functionality.
1832 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
1834 *) Add initial support for TLS extensions, specifically for the server_name
1835 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1836 have new members for a host name. The SSL data structure has an
1837 additional member SSL_CTX *initial_ctx so that new sessions can be
1838 stored in that context to allow for session resumption, even after the
1839 SSL has been switched to a new SSL_CTX in reaction to a client's
1840 server_name extension.
1842 New functions (subject to change):
1844 SSL_get_servername()
1845 SSL_get_servername_type()
1848 New CTRL codes and macros (subject to change):
1850 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1851 - SSL_CTX_set_tlsext_servername_callback()
1852 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1853 - SSL_CTX_set_tlsext_servername_arg()
1854 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1856 openssl s_client has a new '-servername ...' option.
1858 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1859 '-key2 ...', '-servername_fatal' (subject to change). This allows
1860 testing the HostName extension for a specific single host name ('-cert'
1861 and '-key' remain fallbacks for handshakes without HostName
1862 negotiation). If the unrecogninzed_name alert has to be sent, this by
1863 default is a warning; it becomes fatal with the '-servername_fatal'
1866 [Peter Sylvester, Remy Allais, Christophe Renou]
1868 *) Whirlpool hash implementation is added.
1871 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1872 bn(64,32). Because of instruction set limitations it doesn't have
1873 any negative impact on performance. This was done mostly in order
1874 to make it possible to share assembler modules, such as bn_mul_mont
1875 implementations, between 32- and 64-bit builds without hassle.
1878 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1879 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1883 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1884 dedicated Montgomery multiplication procedure, is introduced.
1885 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1886 "64-bit" performance on certain 32-bit targets.
1889 *) New option SSL_OP_NO_COMP to disable use of compression selectively
1890 in SSL structures. New SSL ctrl to set maximum send fragment size.
1891 Save memory by seeting the I/O buffer sizes dynamically instead of
1892 using the maximum available value.
1895 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
1896 in addition to the text details.
1899 *) Very, very preliminary EXPERIMENTAL support for printing of general
1900 ASN1 structures. This currently produces rather ugly output and doesn't
1901 handle several customised structures at all.
1904 *) Integrated support for PVK file format and some related formats such
1905 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
1906 these in the 'rsa' and 'dsa' utilities.
1909 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
1912 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
1913 place for the (very old) "NETSCAPE" format certificates which are now
1914 handled using new ASN1 code equivalents.
1917 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
1918 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
1919 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
1922 *) Modify CRL distribution points extension code to print out previously
1923 unsupported fields. Enhance extension setting code to allow setting of
1927 *) Add print and set support for Issuing Distribution Point CRL extension.
1930 *) Change 'Configure' script to enable Camellia by default.
1933 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
1935 *) When rejecting SSL/TLS records due to an incorrect version number, never
1936 update s->server with a new major version number. As of
1937 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
1938 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
1939 the previous behavior could result in a read attempt at NULL when
1940 receiving specific incorrect SSL/TLS records once record payload
1941 protection is active. (CVE-2010-0740)
1942 [Bodo Moeller, Adam Langley <agl@chromium.org>]
1944 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
1945 could be crashed if the relevant tables were not present (e.g. chrooted).
1946 [Tomas Hoger <thoger@redhat.com>]
1948 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
1950 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
1951 [Martin Olsson, Neel Mehta]
1953 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
1954 accommodate for stack sorting, always a write lock!).
1957 *) On some versions of WIN32 Heap32Next is very slow. This can cause
1958 excessive delays in the RAND_poll(): over a minute. As a workaround
1959 include a time check in the inner Heap32Next loop too.
1962 *) The code that handled flushing of data in SSL/TLS originally used the
1963 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
1964 the problem outlined in PR#1949. The fix suggested there however can
1965 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
1966 of Apache). So instead simplify the code to flush unconditionally.
1967 This should be fine since flushing with no data to flush is a no op.
1970 *) Handle TLS versions 2.0 and later properly and correctly use the
1971 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
1972 off ancient servers have a habit of sticking around for a while...
1975 *) Modify compression code so it frees up structures without using the
1976 ex_data callbacks. This works around a problem where some applications
1977 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
1978 restarting) then use compression (e.g. SSL with compression) later.
1979 This results in significant per-connection memory leaks and
1980 has caused some security issues including CVE-2008-1678 and
1984 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
1985 change when encrypting or decrypting.
1988 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
1989 connect and renegotiate with servers which do not support RI.
1990 Until RI is more widely deployed this option is enabled by default.
1993 *) Add "missing" ssl ctrls to clear options and mode.
1996 *) If client attempts to renegotiate and doesn't support RI respond with
1997 a no_renegotiation alert as required by RFC5746. Some renegotiating
1998 TLS clients will continue a connection gracefully when they receive
1999 the alert. Unfortunately OpenSSL mishandled this alert and would hang
2000 waiting for a server hello which it will never receive. Now we treat a
2001 received no_renegotiation alert as a fatal error. This is because
2002 applications requesting a renegotiation might well expect it to succeed
2003 and would have no code in place to handle the server denying it so the
2004 only safe thing to do is to terminate the connection.
2007 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
2008 peer supports secure renegotiation and 0 otherwise. Print out peer
2009 renegotiation support in s_client/s_server.
2012 *) Replace the highly broken and deprecated SPKAC certification method with
2013 the updated NID creation version. This should correctly handle UTF8.
2016 *) Implement RFC5746. Re-enable renegotiation but require the extension
2017 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
2018 turns out to be a bad idea. It has been replaced by
2019 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
2020 SSL_CTX_set_options(). This is really not recommended unless you
2021 know what you are doing.
2022 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
2024 *) Fixes to stateless session resumption handling. Use initial_ctx when
2025 issuing and attempting to decrypt tickets in case it has changed during
2026 servername handling. Use a non-zero length session ID when attempting
2027 stateless session resumption: this makes it possible to determine if
2028 a resumption has occurred immediately after receiving server hello
2029 (several places in OpenSSL subtly assume this) instead of later in
2033 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
2034 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
2035 fixes for a few places where the return code is not checked
2037 [Julia Lawall <julia@diku.dk>]
2039 *) Add --strict-warnings option to Configure script to include devteam
2040 warnings in other configurations.
2043 *) Add support for --libdir option and LIBDIR variable in makefiles. This
2044 makes it possible to install openssl libraries in locations which
2045 have names other than "lib", for example "/usr/lib64" which some
2047 [Steve Henson, based on patch from Jeremy Utley]
2049 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
2050 X690 8.9.12 and can produce some misleading textual output of OIDs.
2051 [Steve Henson, reported by Dan Kaminsky]
2053 *) Delete MD2 from algorithm tables. This follows the recommendation in
2054 several standards that it is not used in new applications due to
2055 several cryptographic weaknesses. For binary compatibility reasons
2056 the MD2 API is still compiled in by default.
2059 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
2063 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
2064 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
2066 [Guenter <lists@gknw.net>]
2068 *) Fix the server certificate chain building code to use X509_verify_cert(),
2069 it used to have an ad-hoc builder which was unable to cope with anything
2070 other than a simple chain.
2071 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
2073 *) Don't check self signed certificate signatures in X509_verify_cert()
2074 by default (a flag can override this): it just wastes time without
2075 adding any security. As a useful side effect self signed root CAs
2076 with non-FIPS digests are now usable in FIPS mode.
2079 *) In dtls1_process_out_of_seq_message() the check if the current message
2080 is already buffered was missing. For every new message was memory
2081 allocated, allowing an attacker to perform an denial of service attack
2082 with sending out of seq handshake messages until there is no memory
2083 left. Additionally every future messege was buffered, even if the
2084 sequence number made no sense and would be part of another handshake.
2085 So only messages with sequence numbers less than 10 in advance will be
2086 buffered. (CVE-2009-1378)
2087 [Robin Seggelmann, discovered by Daniel Mentz]
2089 *) Records are buffered if they arrive with a future epoch to be
2090 processed after finishing the corresponding handshake. There is
2091 currently no limitation to this buffer allowing an attacker to perform
2092 a DOS attack with sending records with future epochs until there is no
2093 memory left. This patch adds the pqueue_size() function to detemine
2094 the size of a buffer and limits the record buffer to 100 entries.
2096 [Robin Seggelmann, discovered by Daniel Mentz]
2098 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
2099 parent structure is freed. (CVE-2009-1379)
2102 *) Handle non-blocking I/O properly in SSL_shutdown() call.
2103 [Darryl Miles <darryl-mailinglists@netbauds.net>]
2106 [Ilya O. <vrghost@gmail.com>]
2108 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
2110 *) Disable renegotiation completely - this fixes a severe security
2111 problem (CVE-2009-3555) at the cost of breaking all
2112 renegotiation. Renegotiation can be re-enabled by setting
2113 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
2114 run-time. This is really not recommended unless you know what
2118 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
2120 *) Don't set val to NULL when freeing up structures, it is freed up by
2121 underlying code. If sizeof(void *) > sizeof(long) this can result in
2122 zeroing past the valid field. (CVE-2009-0789)
2123 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
2125 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
2126 checked correctly. This would allow some invalid signed attributes to
2127 appear to verify correctly. (CVE-2009-0591)
2128 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2130 *) Reject UniversalString and BMPString types with invalid lengths. This
2131 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
2132 a legal length. (CVE-2009-0590)
2135 *) Set S/MIME signing as the default purpose rather than setting it
2136 unconditionally. This allows applications to override it at the store
2140 *) Permit restricted recursion of ASN1 strings. This is needed in practice
2141 to handle some structures.
2144 *) Improve efficiency of mem_gets: don't search whole buffer each time
2146 [Jeremy Shapiro <jnshapir@us.ibm.com>]
2148 *) New -hex option for openssl rand.
2151 *) Print out UTF8String and NumericString when parsing ASN1.
2154 *) Support NumericString type for name components.
2157 *) Allow CC in the environment to override the automatically chosen
2158 compiler. Note that nothing is done to ensure flags work with the
2162 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
2164 *) Properly check EVP_VerifyFinal() and similar return values
2166 [Ben Laurie, Bodo Moeller, Google Security Team]
2168 *) Enable TLS extensions by default.
2171 *) Allow the CHIL engine to be loaded, whether the application is
2172 multithreaded or not. (This does not release the developer from the
2173 obligation to set up the dynamic locking callbacks.)
2174 [Sander Temme <sander@temme.net>]
2176 *) Use correct exit code if there is an error in dgst command.
2177 [Steve Henson; problem pointed out by Roland Dirlewanger]
2179 *) Tweak Configure so that you need to say "experimental-jpake" to enable
2180 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
2183 *) Add experimental JPAKE support, including demo authentication in
2184 s_client and s_server.
2187 *) Set the comparison function in v3_addr_canonize().
2188 [Rob Austein <sra@hactrn.net>]
2190 *) Add support for XMPP STARTTLS in s_client.
2191 [Philip Paeps <philip@freebsd.org>]
2193 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
2194 to ensure that even with this option, only ciphersuites in the
2195 server's preference list will be accepted. (Note that the option
2196 applies only when resuming a session, so the earlier behavior was
2197 just about the algorithm choice for symmetric cryptography.)
2200 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
2202 *) Fix NULL pointer dereference if a DTLS server received
2203 ChangeCipherSpec as first record (CVE-2009-1386).
2206 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
2207 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
2210 *) The fix in 0.9.8c that supposedly got rid of unsafe
2211 double-checked locking was incomplete for RSA blinding,
2212 addressing just one layer of what turns out to have been
2213 doubly unsafe triple-checked locking.
2215 So now fix this for real by retiring the MONT_HELPER macro
2216 in crypto/rsa/rsa_eay.c.
2218 [Bodo Moeller; problem pointed out by Marius Schilder]
2220 *) Various precautionary measures:
2222 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
2224 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
2225 (NB: This would require knowledge of the secret session ticket key
2226 to exploit, in which case you'd be SOL either way.)
2228 - Change bn_nist.c so that it will properly handle input BIGNUMs
2229 outside the expected range.
2231 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
2234 [Neel Mehta, Bodo Moeller]
2236 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
2237 the load fails. Useful for distros.
2238 [Ben Laurie and the FreeBSD team]
2240 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
2243 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
2246 *) Expand ENGINE to support engine supplied SSL client certificate functions.
2248 This work was sponsored by Logica.
2251 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
2252 keystores. Support for SSL/TLS client authentication too.
2253 Not compiled unless enable-capieng specified to Configure.
2255 This work was sponsored by Logica.
2258 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
2259 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
2260 attribute creation routines such as certifcate requests and PKCS#12
2264 Changes between 0.9.8g and 0.9.8h [28 May 2008]
2266 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
2267 handshake which could lead to a cilent crash as found using the
2268 Codenomicon TLS test suite (CVE-2008-1672)
2269 [Steve Henson, Mark Cox]
2271 *) Fix double free in TLS server name extensions which could lead to
2272 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
2275 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
2277 Clear the error queue to ensure that error entries left from
2278 older function calls do not interfere with the correct operation.
2279 [Lutz Jaenicke, Erik de Castro Lopo]
2281 *) Remove root CA certificates of commercial CAs:
2283 The OpenSSL project does not recommend any specific CA and does not
2284 have any policy with respect to including or excluding any CA.
2285 Therefore it does not make any sense to ship an arbitrary selection
2286 of root CA certificates with the OpenSSL software.
2289 *) RSA OAEP patches to fix two separate invalid memory reads.
2290 The first one involves inputs when 'lzero' is greater than
2291 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
2292 before the beginning of from). The second one involves inputs where
2293 the 'db' section contains nothing but zeroes (there is a one-byte
2294 invalid read after the end of 'db').
2295 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2297 *) Partial backport from 0.9.9-dev:
2299 Introduce bn_mul_mont (dedicated Montgomery multiplication
2300 procedure) as a candidate for BIGNUM assembler implementation.
2301 While 0.9.9-dev uses assembler for various architectures, only
2302 x86_64 is available by default here in the 0.9.8 branch, and
2303 32-bit x86 is available through a compile-time setting.
2305 To try the 32-bit x86 assembler implementation, use Configure
2306 option "enable-montasm" (which exists only for this backport).
2308 As "enable-montasm" for 32-bit x86 disclaims code stability
2309 anyway, in this constellation we activate additional code
2310 backported from 0.9.9-dev for further performance improvements,
2311 namely BN_from_montgomery_word. (To enable this otherwise,
2312 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
2314 [Andy Polyakov (backport partially by Bodo Moeller)]
2316 *) Add TLS session ticket callback. This allows an application to set
2317 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
2318 values. This is useful for key rollover for example where several key
2319 sets may exist with different names.
2322 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
2323 This was broken until now in 0.9.8 releases, such that the only way
2324 a registered ENGINE could be used (assuming it initialises
2325 successfully on the host) was to explicitly set it as the default
2326 for the relevant algorithms. This is in contradiction with 0.9.7
2327 behaviour and the documentation. With this fix, when an ENGINE is
2328 registered into a given algorithm's table of implementations, the
2329 'uptodate' flag is reset so that auto-discovery will be used next
2330 time a new context for that algorithm attempts to select an
2332 [Ian Lister (tweaked by Geoff Thorpe)]
2334 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
2335 implemention in the following ways:
2337 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
2340 Lack of BER streaming support means one pass streaming processing is
2341 only supported if data is detached: setting the streaming flag is
2342 ignored for embedded content.
2344 CMS support is disabled by default and must be explicitly enabled
2345 with the enable-cms configuration option.
2348 *) Update the GMP engine glue to do direct copies between BIGNUM and
2349 mpz_t when openssl and GMP use the same limb size. Otherwise the
2350 existing "conversion via a text string export" trick is still used.
2351 [Paul Sheer <paulsheer@gmail.com>]
2353 *) Zlib compression BIO. This is a filter BIO which compressed and
2354 uncompresses any data passed through it.
2357 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
2358 RFC3394 compatible AES key wrapping.
2361 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
2362 sets string data without copying. X509_ALGOR_set0() and
2363 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
2364 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
2365 from an X509_ATTRIBUTE structure optionally checking it occurs only
2366 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
2370 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
2371 to get the expected BN_FLG_CONSTTIME behavior.
2372 [Bodo Moeller (Google)]
2376 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
2377 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
2378 - added some more tests to do_tests.pl
2379 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
2380 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
2381 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
2382 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
2383 - various changes to netware.pl to enable gcc-cross builds on Win32
2385 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
2386 - various changes to fix missing prototype warnings
2387 - fixed x86nasm.pl to create correct asm files for NASM COFF output
2388 - added AES, WHIRLPOOL and CPUID assembler code to build files
2389 - added missing AES assembler make rules to mk1mf.pl
2390 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
2391 [Guenter Knauf <eflash@gmx.net>]
2393 *) Implement certificate status request TLS extension defined in RFC3546.
2394 A client can set the appropriate parameters and receive the encoded
2395 OCSP response via a callback. A server can query the supplied parameters
2396 and set the encoded OCSP response in the callback. Add simplified examples
2397 to s_client and s_server.
2400 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
2402 *) Fix various bugs:
2403 + Binary incompatibility of ssl_ctx_st structure
2404 + DTLS interoperation with non-compliant servers
2405 + Don't call get_session_cb() without proposed session
2406 + Fix ia64 assembler code
2407 [Andy Polyakov, Steve Henson]
2409 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
2411 *) DTLS Handshake overhaul. There were longstanding issues with
2412 OpenSSL DTLS implementation, which were making it impossible for
2413 RFC 4347 compliant client to communicate with OpenSSL server.
2414 Unfortunately just fixing these incompatibilities would "cut off"
2415 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
2416 server keeps tolerating non RFC compliant syntax. The opposite is
2417 not true, 0.9.8f client can not communicate with earlier server.
2418 This update even addresses CVE-2007-4995.
2421 *) Changes to avoid need for function casts in OpenSSL: some compilers
2422 (gcc 4.2 and later) reject their use.
2423 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
2426 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2427 RFC4507bis. The encrypted ticket format is an encrypted encoded
2428 SSL_SESSION structure, that way new session features are automatically
2431 If a client application caches session in an SSL_SESSION structure
2432 support is transparent because tickets are now stored in the encoded
2435 The SSL_CTX structure automatically generates keys for ticket
2436 protection in servers so again support should be possible
2437 with no application modification.
2439 If a client or server wishes to disable RFC4507 support then the option
2440 SSL_OP_NO_TICKET can be set.
2442 Add a TLS extension debugging callback to allow the contents of any client
2443 or server extensions to be examined.
2445 This work was sponsored by Google.
2448 *) Add initial support for TLS extensions, specifically for the server_name
2449 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2450 have new members for a host name. The SSL data structure has an
2451 additional member SSL_CTX *initial_ctx so that new sessions can be
2452 stored in that context to allow for session resumption, even after the
2453 SSL has been switched to a new SSL_CTX in reaction to a client's
2454 server_name extension.
2456 New functions (subject to change):
2458 SSL_get_servername()
2459 SSL_get_servername_type()
2462 New CTRL codes and macros (subject to change):
2464 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2465 - SSL_CTX_set_tlsext_servername_callback()
2466 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2467 - SSL_CTX_set_tlsext_servername_arg()
2468 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2470 openssl s_client has a new '-servername ...' option.
2472 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2473 '-key2 ...', '-servername_fatal' (subject to change). This allows
2474 testing the HostName extension for a specific single host name ('-cert'
2475 and '-key' remain fallbacks for handshakes without HostName
2476 negotiation). If the unrecogninzed_name alert has to be sent, this by
2477 default is a warning; it becomes fatal with the '-servername_fatal'
2480 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
2482 *) Add AES and SSE2 assembly language support to VC++ build.
2485 *) Mitigate attack on final subtraction in Montgomery reduction.
2488 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
2489 (which previously caused an internal error).
2492 *) Squeeze another 10% out of IGE mode when in != out.
2495 *) AES IGE mode speedup.
2496 [Dean Gaudet (Google)]
2498 *) Add the Korean symmetric 128-bit cipher SEED (see
2499 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
2500 add SEED ciphersuites from RFC 4162:
2502 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
2503 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
2504 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
2505 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
2507 To minimize changes between patchlevels in the OpenSSL 0.9.8
2508 series, SEED remains excluded from compilation unless OpenSSL
2509 is configured with 'enable-seed'.
2510 [KISA, Bodo Moeller]
2512 *) Mitigate branch prediction attacks, which can be practical if a
2513 single processor is shared, allowing a spy process to extract
2514 information. For detailed background information, see
2515 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
2516 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
2517 and Necessary Software Countermeasures"). The core of the change
2518 are new versions BN_div_no_branch() and
2519 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
2520 respectively, which are slower, but avoid the security-relevant
2521 conditional branches. These are automatically called by BN_div()
2522 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
2523 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
2524 remove a conditional branch.
2526 BN_FLG_CONSTTIME is the new name for the previous
2527 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
2528 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
2529 in the exponent causes BN_mod_exp_mont() to use the alternative
2530 implementation in BN_mod_exp_mont_consttime().) The old name
2531 remains as a deprecated alias.
2533 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
2534 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
2535 constant-time implementations for more than just exponentiation.
2536 Here too the old name is kept as a deprecated alias.
2538 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2539 the BN_BLINDING structure gets an independent copy of the
2540 modulus. This means that the previous "BIGNUM *m" argument to
2541 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2542 essentially becomes "const BIGNUM *m", although we can't actually
2543 change this in the header file before 0.9.9. It allows
2544 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2545 enable BN_FLG_CONSTTIME.
2547 [Matthew D Wood (Intel Corp)]
2549 *) In the SSL/TLS server implementation, be strict about session ID
2550 context matching (which matters if an application uses a single
2551 external cache for different purposes). Previously,
2552 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2553 set. This did ensure strict client verification, but meant that,
2554 with applications using a single external cache for quite
2555 different requirements, clients could circumvent ciphersuite
2556 restrictions for a given session ID context by starting a session
2557 in a different context.
2560 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2561 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2562 authentication-only ciphersuites.
2565 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
2566 not complete and could lead to a possible single byte overflow
2567 (CVE-2007-5135) [Ben Laurie]
2569 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
2571 *) Since AES128 and AES256 (and similarly Camellia128 and
2572 Camellia256) share a single mask bit in the logic of
2573 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2574 kludge to work properly if AES128 is available and AES256 isn't
2575 (or if Camellia128 is available and Camellia256 isn't).
2578 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
2579 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
2580 When a point or a seed is encoded in a BIT STRING, we need to
2581 prevent the removal of trailing zero bits to get the proper DER
2582 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
2583 of a NamedBitList, for which trailing 0 bits need to be removed.)
2586 *) Have SSL/TLS server implementation tolerate "mismatched" record
2587 protocol version while receiving ClientHello even if the
2588 ClientHello is fragmented. (The server can't insist on the
2589 particular protocol version it has chosen before the ServerHello
2590 message has informed the client about his choice.)
2593 *) Add RFC 3779 support.
2594 [Rob Austein for ARIN, Ben Laurie]
2596 *) Load error codes if they are not already present instead of using a
2597 static variable. This allows them to be cleanly unloaded and reloaded.
2598 Improve header file function name parsing.
2601 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
2602 or CAPABILITY handshake as required by RFCs.
2605 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
2607 *) Introduce limits to prevent malicious keys being able to
2608 cause a denial of service. (CVE-2006-2940)
2609 [Steve Henson, Bodo Moeller]
2611 *) Fix ASN.1 parsing of certain invalid structures that can result
2612 in a denial of service. (CVE-2006-2937) [Steve Henson]
2614 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2615 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2617 *) Fix SSL client code which could crash if connecting to a
2618 malicious SSLv2 server. (CVE-2006-4343)
2619 [Tavis Ormandy and Will Drewry, Google Security Team]
2621 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
2622 match only those. Before that, "AES256-SHA" would be interpreted
2623 as a pattern and match "AES128-SHA" too (since AES128-SHA got
2624 the same strength classification in 0.9.7h) as we currently only
2625 have a single AES bit in the ciphersuite description bitmap.
2626 That change, however, also applied to ciphersuite strings such as
2627 "RC4-MD5" that intentionally matched multiple ciphersuites --
2628 namely, SSL 2.0 ciphersuites in addition to the more common ones
2629 from SSL 3.0/TLS 1.0.
2631 So we change the selection algorithm again: Naming an explicit
2632 ciphersuite selects this one ciphersuite, and any other similar
2633 ciphersuite (same bitmap) from *other* protocol versions.
2634 Thus, "RC4-MD5" again will properly select both the SSL 2.0
2635 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
2637 Since SSL 2.0 does not have any ciphersuites for which the
2638 128/256 bit distinction would be relevant, this works for now.
2639 The proper fix will be to use different bits for AES128 and
2640 AES256, which would have avoided the problems from the beginning;
2641 however, bits are scarce, so we can only do this in a new release
2642 (not just a patchlevel) when we can change the SSL_CIPHER
2643 definition to split the single 'unsigned long mask' bitmap into
2644 multiple values to extend the available space.
2648 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2650 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2651 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2653 *) Add AES IGE and biIGE modes.
2656 *) Change the Unix randomness entropy gathering to use poll() when
2657 possible instead of select(), since the latter has some
2658 undesirable limitations.
2659 [Darryl Miles via Richard Levitte and Bodo Moeller]
2661 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2662 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2663 cannot be implicitly activated as part of, e.g., the "AES" alias.
2664 However, please upgrade to OpenSSL 0.9.9[-dev] for
2665 non-experimental use of the ECC ciphersuites to get TLS extension
2666 support, which is required for curve and point format negotiation
2667 to avoid potential handshake problems.
2670 *) Disable rogue ciphersuites:
2672 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2673 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2674 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2676 The latter two were purportedly from
2677 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2680 Also deactivate the remaining ciphersuites from
2681 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2682 unofficial, and the ID has long expired.
2685 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2686 dual-core machines) and other potential thread-safety issues.
2689 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
2690 versions), which is now available for royalty-free use
2691 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
2692 Also, add Camellia TLS ciphersuites from RFC 4132.
2694 To minimize changes between patchlevels in the OpenSSL 0.9.8
2695 series, Camellia remains excluded from compilation unless OpenSSL
2696 is configured with 'enable-camellia'.
2699 *) Disable the padding bug check when compression is in use. The padding
2700 bug check assumes the first packet is of even length, this is not
2701 necessarily true if compresssion is enabled and can result in false
2702 positives causing handshake failure. The actual bug test is ancient
2703 code so it is hoped that implementations will either have fixed it by
2704 now or any which still have the bug do not support compression.
2707 Changes between 0.9.8a and 0.9.8b [04 May 2006]
2709 *) When applying a cipher rule check to see if string match is an explicit
2710 cipher suite and only match that one cipher suite if it is.
2713 *) Link in manifests for VC++ if needed.
2714 [Austin Ziegler <halostatue@gmail.com>]
2716 *) Update support for ECC-based TLS ciphersuites according to
2717 draft-ietf-tls-ecc-12.txt with proposed changes (but without
2718 TLS extensions, which are supported starting with the 0.9.9
2719 branch, not in the OpenSSL 0.9.8 branch).
2722 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
2723 opaque EVP_CIPHER_CTX handling.
2726 *) Fixes and enhancements to zlib compression code. We now only use
2727 "zlib1.dll" and use the default __cdecl calling convention on Win32
2728 to conform with the standards mentioned here:
2729 http://www.zlib.net/DLL_FAQ.txt
2730 Static zlib linking now works on Windows and the new --with-zlib-include
2731 --with-zlib-lib options to Configure can be used to supply the location
2732 of the headers and library. Gracefully handle case where zlib library
2736 *) Several fixes and enhancements to the OID generation code. The old code
2737 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
2738 handle numbers larger than ULONG_MAX, truncated printing and had a
2739 non standard OBJ_obj2txt() behaviour.
2742 *) Add support for building of engines under engine/ as shared libraries
2743 under VC++ build system.
2746 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
2747 Hopefully, we will not see any false combination of paths any more.
2750 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
2752 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2753 (part of SSL_OP_ALL). This option used to disable the
2754 countermeasure against man-in-the-middle protocol-version
2755 rollback in the SSL 2.0 server implementation, which is a bad
2756 idea. (CVE-2005-2969)
2758 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2759 for Information Security, National Institute of Advanced Industrial
2760 Science and Technology [AIST], Japan)]
2762 *) Add two function to clear and return the verify parameter flags.
2765 *) Keep cipherlists sorted in the source instead of sorting them at
2766 runtime, thus removing the need for a lock.
2769 *) Avoid some small subgroup attacks in Diffie-Hellman.
2770 [Nick Mathewson and Ben Laurie]
2772 *) Add functions for well-known primes.
2775 *) Extended Windows CE support.
2776 [Satoshi Nakamura and Andy Polyakov]
2778 *) Initialize SSL_METHOD structures at compile time instead of during
2779 runtime, thus removing the need for a lock.
2782 *) Make PKCS7_decrypt() work even if no certificate is supplied by
2783 attempting to decrypt each encrypted key in turn. Add support to
2787 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
2789 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
2792 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
2795 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
2796 key into the same file any more.
2799 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
2802 *) Add -utf8 command line and config file option to 'ca'.
2803 [Stefan <stf@udoma.org]
2805 *) Removed the macro des_crypt(), as it seems to conflict with some
2806 libraries. Use DES_crypt().
2809 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
2810 involves renaming the source and generated shared-libs for
2811 both. The engines will accept the corrected or legacy ids
2812 ('ncipher' and '4758_cca' respectively) when binding. NB,
2813 this only applies when building 'shared'.
2814 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
2816 *) Add attribute functions to EVP_PKEY structure. Modify
2817 PKCS12_create() to recognize a CSP name attribute and
2818 use it. Make -CSP option work again in pkcs12 utility.
2821 *) Add new functionality to the bn blinding code:
2822 - automatic re-creation of the BN_BLINDING parameters after
2823 a fixed number of uses (currently 32)
2824 - add new function for parameter creation
2825 - introduce flags to control the update behaviour of the
2826 BN_BLINDING parameters
2827 - hide BN_BLINDING structure
2828 Add a second BN_BLINDING slot to the RSA structure to improve
2829 performance when a single RSA object is shared among several
2833 *) Add support for DTLS.
2834 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
2836 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
2837 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
2840 *) Remove buggy and incompletet DH cert support from
2841 ssl/ssl_rsa.c and ssl/s3_both.c
2844 *) Use SHA-1 instead of MD5 as the default digest algorithm for
2845 the apps/openssl applications.
2848 *) Compile clean with "-Wall -Wmissing-prototypes
2849 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
2850 DEBUG_SAFESTACK must also be set.
2853 *) Change ./Configure so that certain algorithms can be disabled by default.
2854 The new counterpiece to "no-xxx" is "enable-xxx".
2856 The patented RC5 and MDC2 algorithms will now be disabled unless
2857 "enable-rc5" and "enable-mdc2", respectively, are specified.
2859 (IDEA remains enabled despite being patented. This is because IDEA
2860 is frequently required for interoperability, and there is no license
2861 fee for non-commercial use. As before, "no-idea" can be used to
2862 avoid this algorithm.)
2866 *) Add processing of proxy certificates (see RFC 3820). This work was
2867 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
2868 EGEE (Enabling Grids for E-science in Europe).
2871 *) RC4 performance overhaul on modern architectures/implementations, such
2872 as Intel P4, IA-64 and AMD64.
2875 *) New utility extract-section.pl. This can be used specify an alternative
2876 section number in a pod file instead of having to treat each file as
2877 a separate case in Makefile. This can be done by adding two lines to the
2880 =for comment openssl_section:XXX
2882 The blank line is mandatory.
2886 *) New arguments -certform, -keyform and -pass for s_client and s_server
2887 to allow alternative format key and certificate files and passphrase
2891 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
2892 update associated structures and add various utility functions.
2894 Add new policy related verify parameters, include policy checking in
2895 standard verify code. Enhance 'smime' application with extra parameters
2896 to support policy checking and print out.
2899 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
2900 Nehemiah processors. These extensions support AES encryption in hardware
2901 as well as RNG (though RNG support is currently disabled).
2902 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
2904 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
2907 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
2908 [Andy Polyakov and a number of other people]
2910 *) Improved PowerPC platform support. Most notably BIGNUM assembler
2911 implementation contributed by IBM.
2912 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
2914 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
2915 exponent rather than 'unsigned long'. There is a corresponding change to
2916 the new 'rsa_keygen' element of the RSA_METHOD structure.
2917 [Jelte Jansen, Geoff Thorpe]
2919 *) Functionality for creating the initial serial number file is now
2920 moved from CA.pl to the 'ca' utility with a new option -create_serial.
2922 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
2923 number file to 1, which is bound to cause problems. To avoid
2924 the problems while respecting compatibility between different 0.9.7
2925 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
2926 CA.pl for serial number initialization. With the new release 0.9.8,
2927 we can fix the problem directly in the 'ca' utility.)
2930 *) Reduced header interdepencies by declaring more opaque objects in
2931 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
2932 give fewer recursive includes, which could break lazy source code - so
2933 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
2934 developers should define this symbol when building and using openssl to
2935 ensure they track the recommended behaviour, interfaces, [etc], but
2936 backwards-compatible behaviour prevails when this isn't defined.
2939 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
2942 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
2943 This will generate a random key of the appropriate length based on the
2944 cipher context. The EVP_CIPHER can provide its own random key generation
2945 routine to support keys of a specific form. This is used in the des and
2946 3des routines to generate a key of the correct parity. Update S/MIME
2947 code to use new functions and hence generate correct parity DES keys.
2948 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
2949 valid (weak or incorrect parity).
2952 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
2953 as looking them up. This is useful when the verified structure may contain
2954 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
2955 present unless the new PKCS7_NO_CRL flag is asserted.
2958 *) Extend ASN1 oid configuration module. It now additionally accepts the
2961 shortName = some long name, 1.2.3.4
2964 *) Reimplemented the BN_CTX implementation. There is now no more static
2965 limitation on the number of variables it can handle nor the depth of the
2966 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
2967 information can now expand as required, and rather than having a single
2968 static array of bignums, BN_CTX now uses a linked-list of such arrays
2969 allowing it to expand on demand whilst maintaining the usefulness of
2970 BN_CTX's "bundling".
2973 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
2974 to allow all RSA operations to function using a single BN_CTX.
2977 *) Preliminary support for certificate policy evaluation and checking. This
2978 is initially intended to pass the tests outlined in "Conformance Testing
2979 of Relying Party Client Certificate Path Processing Logic" v1.07.
2982 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
2983 remained unused and not that useful. A variety of other little bignum
2984 tweaks and fixes have also been made continuing on from the audit (see
2988 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
2989 associated ASN1, EVP and SSL functions and old ASN1 macros.
2992 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
2993 and this should never fail. So the return value from the use of
2994 BN_set_word() (which can fail due to needless expansion) is now deprecated;
2995 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
2998 *) BN_CTX_get() should return zero-valued bignums, providing the same
2999 initialised value as BN_new().
3000 [Geoff Thorpe, suggested by Ulf Möller]
3002 *) Support for inhibitAnyPolicy certificate extension.
3005 *) An audit of the BIGNUM code is underway, for which debugging code is
3006 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
3007 is considered valid when processing BIGNUMs, and causes execution to
3008 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
3009 further steps are taken to deliberately pollute unused data in BIGNUM
3010 structures to try and expose faulty code further on. For now, openssl will
3011 (in its default mode of operation) continue to tolerate the inconsistent
3012 forms that it has tolerated in the past, but authors and packagers should
3013 consider trying openssl and their own applications when compiled with
3014 these debugging symbols defined. It will help highlight potential bugs in
3015 their own code, and will improve the test coverage for OpenSSL itself. At
3016 some point, these tighter rules will become openssl's default to improve
3017 maintainability, though the assert()s and other overheads will remain only
3018 in debugging configurations. See bn.h for more details.
3019 [Geoff Thorpe, Nils Larsch, Ulf Möller]
3021 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
3022 that can only be obtained through BN_CTX_new() (which implicitly
3023 initialises it). The presence of this function only made it possible
3024 to overwrite an existing structure (and cause memory leaks).
3027 *) Because of the callback-based approach for implementing LHASH as a
3028 template type, lh_insert() adds opaque objects to hash-tables and
3029 lh_doall() or lh_doall_arg() are typically used with a destructor callback
3030 to clean up those corresponding objects before destroying the hash table
3031 (and losing the object pointers). So some over-zealous constifications in
3032 LHASH have been relaxed so that lh_insert() does not take (nor store) the
3033 objects as "const" and the lh_doall[_arg] callback wrappers are not
3034 prototyped to have "const" restrictions on the object pointers they are
3035 given (and so aren't required to cast them away any more).
3038 *) The tmdiff.h API was so ugly and minimal that our own timing utility
3039 (speed) prefers to use its own implementation. The two implementations
3040 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
3041 its object type properly exposed (MS_TM) instead of casting to/from "char
3042 *". This may still change yet if someone realises MS_TM and "ms_time_***"
3043 aren't necessarily the greatest nomenclatures - but this is what was used
3044 internally to the implementation so I've used that for now.
3047 *) Ensure that deprecated functions do not get compiled when
3048 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
3049 the self-tests were still using deprecated key-generation functions so
3050 these have been updated also.
3053 *) Reorganise PKCS#7 code to separate the digest location functionality
3054 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
3055 New function PKCS7_set_digest() to set the digest type for PKCS#7
3056 digestedData type. Add additional code to correctly generate the
3057 digestedData type and add support for this type in PKCS7 initialization
3061 *) New function PKCS7_set0_type_other() this initializes a PKCS7
3062 structure of type "other".
3065 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
3066 sure the loop does correctly stop and breaking ("division by zero")
3067 modulus operations are not performed. The (pre-generated) prime
3068 table crypto/bn/bn_prime.h was already correct, but it could not be
3069 re-generated on some platforms because of the "division by zero"
3070 situation in the script.
3071 [Ralf S. Engelschall]
3073 *) Update support for ECC-based TLS ciphersuites according to
3074 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
3075 SHA-1 now is only used for "small" curves (where the
3076 representation of a field element takes up to 24 bytes); for
3077 larger curves, the field element resulting from ECDH is directly
3078 used as premaster secret.
3079 [Douglas Stebila (Sun Microsystems Laboratories)]
3081 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
3082 curve secp160r1 to the tests.
3083 [Douglas Stebila (Sun Microsystems Laboratories)]
3085 *) Add the possibility to load symbols globally with DSO.
3086 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
3088 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
3089 control of the error stack.
3092 *) Add support for STORE in ENGINE.
3095 *) Add the STORE type. The intention is to provide a common interface
3096 to certificate and key stores, be they simple file-based stores, or
3097 HSM-type store, or LDAP stores, or...
3098 NOTE: The code is currently UNTESTED and isn't really used anywhere.
3101 *) Add a generic structure called OPENSSL_ITEM. This can be used to
3102 pass a list of arguments to any function as well as provide a way
3103 for a function to pass data back to the caller.
3106 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
3107 works like BUF_strdup() but can be used to duplicate a portion of
3108 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
3112 *) Add the function sk_find_ex() which works like sk_find(), but will
3113 return an index to an element even if an exact match couldn't be
3114 found. The index is guaranteed to point at the element where the
3115 searched-for key would be inserted to preserve sorting order.
3118 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
3119 takes an extra flags argument for optional functionality. Currently,
3120 the following flags are defined:
3122 OBJ_BSEARCH_VALUE_ON_NOMATCH
3123 This one gets OBJ_bsearch_ex() to return a pointer to the first
3124 element where the comparing function returns a negative or zero
3127 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
3128 This one gets OBJ_bsearch_ex() to return a pointer to the first
3129 element where the comparing function returns zero. This is useful
3130 if there are more than one element where the comparing function
3134 *) Make it possible to create self-signed certificates with 'openssl ca'
3135 in such a way that the self-signed certificate becomes part of the
3136 CA database and uses the same mechanisms for serial number generation
3137 as all other certificate signing. The new flag '-selfsign' enables
3138 this functionality. Adapt CA.sh and CA.pl.in.
3141 *) Add functionality to check the public key of a certificate request
3142 against a given private. This is useful to check that a certificate
3143 request can be signed by that key (self-signing).
3146 *) Make it possible to have multiple active certificates with the same
3147 subject in the CA index file. This is done only if the keyword
3148 'unique_subject' is set to 'no' in the main CA section (default
3149 if 'CA_default') of the configuration file. The value is saved
3150 with the database itself in a separate index attribute file,
3151 named like the index file with '.attr' appended to the name.
3154 *) Generate muti valued AVAs using '+' notation in config files for
3158 *) Support for nameConstraints certificate extension.
3161 *) Support for policyConstraints certificate extension.
3164 *) Support for policyMappings certificate extension.
3167 *) Make sure the default DSA_METHOD implementation only uses its
3168 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
3169 and change its own handlers to be NULL so as to remove unnecessary
3170 indirection. This lets alternative implementations fallback to the
3171 default implementation more easily.
3174 *) Support for directoryName in GeneralName related extensions
3178 *) Make it possible to link applications using Makefile.shared.
3179 Make that possible even when linking against static libraries!
3182 *) Support for single pass processing for S/MIME signing. This now
3183 means that S/MIME signing can be done from a pipe, in addition
3184 cleartext signing (multipart/signed type) is effectively streaming
3185 and the signed data does not need to be all held in memory.
3187 This is done with a new flag PKCS7_STREAM. When this flag is set
3188 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
3189 is done after the data is output (and digests calculated) in
3190 SMIME_write_PKCS7().
3193 *) Add full support for -rpath/-R, both in shared libraries and
3194 applications, at least on the platforms where it's known how
3198 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
3199 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
3200 will now compute a table of multiples of the generator that
3201 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
3202 faster (notably in the case of a single point multiplication,
3203 scalar * generator).
3204 [Nils Larsch, Bodo Moeller]
3206 *) IPv6 support for certificate extensions. The various extensions
3207 which use the IP:a.b.c.d can now take IPv6 addresses using the
3208 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
3212 *) Added an ENGINE that implements RSA by performing private key
3213 exponentiations with the GMP library. The conversions to and from
3214 GMP's mpz_t format aren't optimised nor are any montgomery forms
3215 cached, and on x86 it appears OpenSSL's own performance has caught up.
3216 However there are likely to be other architectures where GMP could
3217 provide a boost. This ENGINE is not built in by default, but it can be
3218 specified at Configure time and should be accompanied by the necessary
3219 linker additions, eg;
3220 ./config -DOPENSSL_USE_GMP -lgmp
3223 *) "openssl engine" will not display ENGINE/DSO load failure errors when
3224 testing availability of engines with "-t" - the old behaviour is
3225 produced by increasing the feature's verbosity with "-tt".
3228 *) ECDSA routines: under certain error conditions uninitialized BN objects
3229 could be freed. Solution: make sure initialization is performed early
3230 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
3234 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
3235 and DH_METHOD (eg. by ENGINE implementations) to override the normal
3236 software implementations. For DSA and DH, parameter generation can
3237 also be overriden by providing the appropriate method callbacks.
3240 *) Change the "progress" mechanism used in key-generation and
3241 primality testing to functions that take a new BN_GENCB pointer in
3242 place of callback/argument pairs. The new API functions have "_ex"
3243 postfixes and the older functions are reimplemented as wrappers for
3244 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
3245 declarations of the old functions to help (graceful) attempts to
3246 migrate to the new functions. Also, the new key-generation API
3247 functions operate on a caller-supplied key-structure and return
3248 success/failure rather than returning a key or NULL - this is to
3249 help make "keygen" another member function of RSA_METHOD etc.
3251 Example for using the new callback interface:
3253 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
3257 BN_GENCB_set(&my_cb, my_callback, my_arg);
3259 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
3260 /* For the meaning of a, b in calls to my_callback(), see the
3261 * documentation of the function that calls the callback.
3262 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
3263 * my_callback should return 1 if it wants BN_is_prime_ex()
3264 * to continue, or 0 to stop.
3269 *) Change the ZLIB compression method to be stateful, and make it
3270 available to TLS with the number defined in
3271 draft-ietf-tls-compression-04.txt.
3274 *) Add the ASN.1 structures and functions for CertificatePair, which
3275 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
3277 CertificatePair ::= SEQUENCE {
3278 forward [0] Certificate OPTIONAL,
3279 reverse [1] Certificate OPTIONAL,
3280 -- at least one of the pair shall be present -- }
3282 Also implement the PEM functions to read and write certificate
3283 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
3285 This needed to be defined, mostly for the sake of the LDAP
3286 attribute crossCertificatePair, but may prove useful elsewhere as
3290 *) Make it possible to inhibit symlinking of shared libraries in
3291 Makefile.shared, for Cygwin's sake.
3294 *) Extend the BIGNUM API by creating a function
3295 void BN_set_negative(BIGNUM *a, int neg);
3296 and a macro that behave like
3297 int BN_is_negative(const BIGNUM *a);
3299 to avoid the need to access 'a->neg' directly in applications.
3302 *) Implement fast modular reduction for pseudo-Mersenne primes
3303 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
3304 EC_GROUP_new_curve_GFp() will now automatically use this
3306 [Nils Larsch <nla@trustcenter.de>]
3308 *) Add new lock type (CRYPTO_LOCK_BN).
3311 *) Change the ENGINE framework to automatically load engines
3312 dynamically from specific directories unless they could be
3313 found to already be built in or loaded. Move all the
3314 current engines except for the cryptodev one to a new
3316 The engines in engines/ are built as shared libraries if
3317 the "shared" options was given to ./Configure or ./config.
3318 Otherwise, they are inserted in libcrypto.a.
3319 /usr/local/ssl/engines is the default directory for dynamic
3320 engines, but that can be overriden at configure time through
3321 the usual use of --prefix and/or --openssldir, and at run
3322 time with the environment variable OPENSSL_ENGINES.
3323 [Geoff Thorpe and Richard Levitte]
3325 *) Add Makefile.shared, a helper makefile to build shared
3326 libraries. Addapt Makefile.org.
3329 *) Add version info to Win32 DLLs.
3330 [Peter 'Luna' Runestig" <peter@runestig.com>]
3332 *) Add new 'medium level' PKCS#12 API. Certificates and keys
3333 can be added using this API to created arbitrary PKCS#12
3334 files while avoiding the low level API.
3336 New options to PKCS12_create(), key or cert can be NULL and
3337 will then be omitted from the output file. The encryption
3338 algorithm NIDs can be set to -1 for no encryption, the mac
3339 iteration count can be set to 0 to omit the mac.
3341 Enhance pkcs12 utility by making the -nokeys and -nocerts
3342 options work when creating a PKCS#12 file. New option -nomac
3343 to omit the mac, NONE can be set for an encryption algorithm.
3344 New code is modified to use the enhanced PKCS12_create()
3345 instead of the low level API.
3348 *) Extend ASN1 encoder to support indefinite length constructed
3349 encoding. This can output sequences tags and octet strings in
3350 this form. Modify pk7_asn1.c to support indefinite length
3351 encoding. This is experimental and needs additional code to
3352 be useful, such as an ASN1 bio and some enhanced streaming
3355 Extend template encode functionality so that tagging is passed
3356 down to the template encoder.
3359 *) Let 'openssl req' fail if an argument to '-newkey' is not
3360 recognized instead of using RSA as a default.
3363 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
3364 As these are not official, they are not included in "ALL";
3365 the "ECCdraft" ciphersuite group alias can be used to select them.
3366 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
3368 *) Add ECDH engine support.
3369 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
3371 *) Add ECDH in new directory crypto/ecdh/.
3372 [Douglas Stebila (Sun Microsystems Laboratories)]
3374 *) Let BN_rand_range() abort with an error after 100 iterations
3375 without success (which indicates a broken PRNG).
3378 *) Change BN_mod_sqrt() so that it verifies that the input value
3379 is really the square of the return value. (Previously,
3380 BN_mod_sqrt would show GIGO behaviour.)
3383 *) Add named elliptic curves over binary fields from X9.62, SECG,
3384 and WAP/WTLS; add OIDs that were still missing.
3386 [Sheueling Chang Shantz and Douglas Stebila
3387 (Sun Microsystems Laboratories)]
3389 *) Extend the EC library for elliptic curves over binary fields
3390 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
3393 EC_GF2m_simple_method
3397 EC_GROUP_new_curve_GF2m
3398 EC_GROUP_set_curve_GF2m
3399 EC_GROUP_get_curve_GF2m
3400 EC_POINT_set_affine_coordinates_GF2m
3401 EC_POINT_get_affine_coordinates_GF2m
3402 EC_POINT_set_compressed_coordinates_GF2m
3404 Point compression for binary fields is disabled by default for
3405 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
3408 As binary polynomials are represented as BIGNUMs, various members
3409 of the EC_GROUP and EC_POINT data structures can be shared
3410 between the implementations for prime fields and binary fields;
3411 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
3412 are essentially identical to their ..._GFp counterparts.
3413 (For simplicity, the '..._GFp' prefix has been dropped from
3414 various internal method names.)
3416 An internal 'field_div' method (similar to 'field_mul' and
3417 'field_sqr') has been added; this is used only for binary fields.
3419 [Sheueling Chang Shantz and Douglas Stebila
3420 (Sun Microsystems Laboratories)]
3422 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
3423 through methods ('mul', 'precompute_mult').
3425 The generic implementations (now internally called 'ec_wNAF_mul'
3426 and 'ec_wNAF_precomputed_mult') remain the default if these
3427 methods are undefined.
3429 [Sheueling Chang Shantz and Douglas Stebila
3430 (Sun Microsystems Laboratories)]
3432 *) New function EC_GROUP_get_degree, which is defined through
3433 EC_METHOD. For curves over prime fields, this returns the bit
3434 length of the modulus.
3436 [Sheueling Chang Shantz and Douglas Stebila
3437 (Sun Microsystems Laboratories)]
3439 *) New functions EC_GROUP_dup, EC_POINT_dup.
3440 (These simply call ..._new and ..._copy).
3442 [Sheueling Chang Shantz and Douglas Stebila
3443 (Sun Microsystems Laboratories)]
3445 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
3446 Polynomials are represented as BIGNUMs (where the sign bit is not
3447 used) in the following functions [macros]:
3450 BN_GF2m_sub [= BN_GF2m_add]
3451 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
3452 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
3453 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
3455 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
3456 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
3457 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
3458 BN_GF2m_cmp [= BN_ucmp]
3460 (Note that only the 'mod' functions are actually for fields GF(2^m).
3461 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
3463 For some functions, an the irreducible polynomial defining a
3464 field can be given as an 'unsigned int[]' with strictly
3465 decreasing elements giving the indices of those bits that are set;
3466 i.e., p[] represents the polynomial
3467 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
3469 p[0] > p[1] > ... > p[k] = 0.
3470 This applies to the following functions:
3475 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
3476 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
3478 BN_GF2m_mod_sqrt_arr
3479 BN_GF2m_mod_solve_quad_arr
3483 Conversion can be performed by the following functions:
3488 bntest.c has additional tests for binary polynomial arithmetic.
3490 Two implementations for BN_GF2m_mod_div() are available.
3491 The default algorithm simply uses BN_GF2m_mod_inv() and
3492 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
3493 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
3494 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
3496 [Sheueling Chang Shantz and Douglas Stebila
3497 (Sun Microsystems Laboratories)]
3499 *) Add new error code 'ERR_R_DISABLED' that can be used when some
3500 functionality is disabled at compile-time.
3501 [Douglas Stebila <douglas.stebila@sun.com>]
3503 *) Change default behaviour of 'openssl asn1parse' so that more
3504 information is visible when viewing, e.g., a certificate:
3506 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
3507 mode the content of non-printable OCTET STRINGs is output in a
3508 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
3509 avoid the appearance of a printable string.
3510 [Nils Larsch <nla@trustcenter.de>]
3512 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
3514 EC_GROUP_set_asn1_flag()
3515 EC_GROUP_get_asn1_flag()
3516 EC_GROUP_set_point_conversion_form()
3517 EC_GROUP_get_point_conversion_form()
3518 These control ASN1 encoding details:
3519 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
3520 has been set to OPENSSL_EC_NAMED_CURVE.
3521 - Points are encoded in uncompressed form by default; options for
3522 asn1_for are as for point2oct, namely
3523 POINT_CONVERSION_COMPRESSED
3524 POINT_CONVERSION_UNCOMPRESSED
3525 POINT_CONVERSION_HYBRID
3527 Also add 'seed' and 'seed_len' members to EC_GROUP with access
3530 EC_GROUP_get0_seed()
3531 EC_GROUP_get_seed_len()
3532 This is used only for ASN1 purposes (so far).
3533 [Nils Larsch <nla@trustcenter.de>]
3535 *) Add 'field_type' member to EC_METHOD, which holds the NID
3536 of the appropriate field type OID. The new function
3537 EC_METHOD_get_field_type() returns this value.
3538 [Nils Larsch <nla@trustcenter.de>]
3543 EC_POINT_point2hex()
3544 EC_POINT_hex2point()
3545 providing useful interfaces to EC_POINT_point2oct() and
3546 EC_POINT_oct2point().
3547 [Nils Larsch <nla@trustcenter.de>]
3549 *) Change internals of the EC library so that the functions
3550 EC_GROUP_set_generator()
3551 EC_GROUP_get_generator()
3552 EC_GROUP_get_order()
3553 EC_GROUP_get_cofactor()
3554 are implemented directly in crypto/ec/ec_lib.c and not dispatched
3555 to methods, which would lead to unnecessary code duplication when
3556 adding different types of curves.
3557 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
3559 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
3560 arithmetic, and such that modified wNAFs are generated
3561 (which avoid length expansion in many cases).
3564 *) Add a function EC_GROUP_check_discriminant() (defined via
3565 EC_METHOD) that verifies that the curve discriminant is non-zero.
3567 Add a function EC_GROUP_check() that makes some sanity tests
3568 on a EC_GROUP, its generator and order. This includes
3569 EC_GROUP_check_discriminant().
3570 [Nils Larsch <nla@trustcenter.de>]
3572 *) Add ECDSA in new directory crypto/ecdsa/.
3574 Add applications 'openssl ecparam' and 'openssl ecdsa'
3575 (these are based on 'openssl dsaparam' and 'openssl dsa').
3577 ECDSA support is also included in various other files across the
3578 library. Most notably,
3579 - 'openssl req' now has a '-newkey ecdsa:file' option;
3580 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
3581 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
3582 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
3583 them suitable for ECDSA where domain parameters must be
3584 extracted before the specific public key;
3585 - ECDSA engine support has been added.
3586 [Nils Larsch <nla@trustcenter.de>]
3588 *) Include some named elliptic curves, and add OIDs from X9.62,
3589 SECG, and WAP/WTLS. Each curve can be obtained from the new
3591 EC_GROUP_new_by_curve_name(),
3592 and the list of available named curves can be obtained with
3593 EC_get_builtin_curves().
3594 Also add a 'curve_name' member to EC_GROUP objects, which can be
3596 EC_GROUP_set_curve_name()
3597 EC_GROUP_get_curve_name()
3598 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
3600 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3601 was actually never needed) and in BN_mul(). The removal in BN_mul()
3602 required a small change in bn_mul_part_recursive() and the addition
3603 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3604 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3605 bn_sub_words() and bn_add_words() except they take arrays with
3609 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
3611 *) Cleanse PEM buffers before freeing them since they may contain
3613 [Benjamin Bennett <ben@psc.edu>]
3615 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3616 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3617 authentication-only ciphersuites.
3620 *) Since AES128 and AES256 share a single mask bit in the logic of
3621 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3622 kludge to work properly if AES128 is available and AES256 isn't.
3625 *) Expand security boundary to match 1.1.1 module.
3628 *) Remove redundant features: hash file source, editing of test vectors
3629 modify fipsld to use external fips_premain.c signature.
3632 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
3633 run algorithm test programs.
3636 *) Make algorithm test programs more tolerant of whitespace.
3639 *) Have SSL/TLS server implementation tolerate "mismatched" record
3640 protocol version while receiving ClientHello even if the
3641 ClientHello is fragmented. (The server can't insist on the
3642 particular protocol version it has chosen before the ServerHello
3643 message has informed the client about his choice.)
3646 *) Load error codes if they are not already present instead of using a
3647 static variable. This allows them to be cleanly unloaded and reloaded.
3650 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3652 *) Introduce limits to prevent malicious keys being able to
3653 cause a denial of service. (CVE-2006-2940)
3654 [Steve Henson, Bodo Moeller]
3656 *) Fix ASN.1 parsing of certain invalid structures that can result
3657 in a denial of service. (CVE-2006-2937) [Steve Henson]
3659 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3660 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3662 *) Fix SSL client code which could crash if connecting to a
3663 malicious SSLv2 server. (CVE-2006-4343)
3664 [Tavis Ormandy and Will Drewry, Google Security Team]
3666 *) Change ciphersuite string processing so that an explicit
3667 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3668 will no longer include "AES128-SHA"), and any other similar
3669 ciphersuite (same bitmap) from *other* protocol versions (so that
3670 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3671 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3672 changes from 0.9.8b and 0.9.8d.
3675 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
3677 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3678 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3680 *) Change the Unix randomness entropy gathering to use poll() when
3681 possible instead of select(), since the latter has some
3682 undesirable limitations.
3683 [Darryl Miles via Richard Levitte and Bodo Moeller]
3685 *) Disable rogue ciphersuites:
3687 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3688 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3689 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3691 The latter two were purportedly from
3692 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3695 Also deactive the remaining ciphersuites from
3696 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3697 unofficial, and the ID has long expired.
3700 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3701 dual-core machines) and other potential thread-safety issues.
3704 Changes between 0.9.7i and 0.9.7j [04 May 2006]
3706 *) Adapt fipsld and the build system to link against the validated FIPS
3707 module in FIPS mode.
3710 *) Fixes for VC++ 2005 build under Windows.
3713 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
3714 from a Windows bash shell such as MSYS. It is autodetected from the
3715 "config" script when run from a VC++ environment. Modify standard VC++
3716 build to use fipscanister.o from the GNU make build.
3719 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
3721 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
3722 The value now differs depending on if you build for FIPS or not.
3723 BEWARE! A program linked with a shared FIPSed libcrypto can't be
3724 safely run with a non-FIPSed libcrypto, as it may crash because of
3725 the difference induced by this change.
3728 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
3730 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3731 (part of SSL_OP_ALL). This option used to disable the
3732 countermeasure against man-in-the-middle protocol-version
3733 rollback in the SSL 2.0 server implementation, which is a bad
3734 idea. (CVE-2005-2969)
3736 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3737 for Information Security, National Institute of Advanced Industrial
3738 Science and Technology [AIST], Japan)]
3740 *) Minimal support for X9.31 signatures and PSS padding modes. This is
3741 mainly for FIPS compliance and not fully integrated at this stage.
3744 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
3745 the exponentiation using a fixed-length exponent. (Otherwise,
3746 the information leaked through timing could expose the secret key
3747 after many signatures; cf. Bleichenbacher's attack on DSA with
3751 *) Make a new fixed-window mod_exp implementation the default for
3752 RSA, DSA, and DH private-key operations so that the sequence of
3753 squares and multiplies and the memory access pattern are
3754 independent of the particular secret key. This will mitigate
3755 cache-timing and potential related attacks.
3757 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
3758 and this is automatically used by BN_mod_exp_mont() if the new flag
3759 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
3760 will use this BN flag for private exponents unless the flag
3761 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
3762 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
3764 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
3766 *) Change the client implementation for SSLv23_method() and
3767 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
3768 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
3769 (Previously, the SSL 2.0 backwards compatible Client Hello
3770 message format would be used even with SSL_OP_NO_SSLv2.)
3773 *) Add support for smime-type MIME parameter in S/MIME messages which some
3777 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
3778 a threadsafe manner. Modify rsa code to use new function and add calls
3779 to dsa and dh code (which had race conditions before).
3782 *) Include the fixed error library code in the C error file definitions
3783 instead of fixing them up at runtime. This keeps the error code
3784 structures constant.
3787 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
3789 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
3792 *) Fixes for newer kerberos headers. NB: the casts are needed because
3793 the 'length' field is signed on one version and unsigned on another
3794 with no (?) obvious way to tell the difference, without these VC++
3795 complains. Also the "definition" of FAR (blank) is no longer included
3796 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
3797 some needed definitions.
3800 *) Undo Cygwin change.
3803 *) Added support for proxy certificates according to RFC 3820.
3804 Because they may be a security thread to unaware applications,
3805 they must be explicitely allowed in run-time. See
3806 docs/HOWTO/proxy_certificates.txt for further information.
3809 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
3811 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
3812 server and client random values. Previously
3813 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
3814 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
3816 This change has negligible security impact because:
3818 1. Server and client random values still have 24 bytes of pseudo random
3821 2. Server and client random values are sent in the clear in the initial
3824 3. The master secret is derived using the premaster secret (48 bytes in
3825 size for static RSA ciphersuites) as well as client server and random
3828 The OpenSSL team would like to thank the UK NISCC for bringing this issue
3831 [Stephen Henson, reported by UK NISCC]
3833 *) Use Windows randomness collection on Cygwin.
3836 *) Fix hang in EGD/PRNGD query when communication socket is closed
3837 prematurely by EGD/PRNGD.
3838 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
3840 *) Prompt for pass phrases when appropriate for PKCS12 input format.
3843 *) Back-port of selected performance improvements from development
3844 branch, as well as improved support for PowerPC platforms.
3847 *) Add lots of checks for memory allocation failure, error codes to indicate
3848 failure and freeing up memory if a failure occurs.
3849 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
3851 *) Add new -passin argument to dgst.
3854 *) Perform some character comparisons of different types in X509_NAME_cmp:
3855 this is needed for some certificates that reencode DNs into UTF8Strings
3856 (in violation of RFC3280) and can't or wont issue name rollover
3860 *) Make an explicit check during certificate validation to see that
3861 the CA setting in each certificate on the chain is correct. As a
3862 side effect always do the following basic checks on extensions,
3863 not just when there's an associated purpose to the check:
3865 - if there is an unhandled critical extension (unless the user
3866 has chosen to ignore this fault)
3867 - if the path length has been exceeded (if one is set at all)
3868 - that certain extensions fit the associated purpose (if one has
3872 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
3874 *) Avoid a race condition when CRLs are checked in a multi threaded
3875 environment. This would happen due to the reordering of the revoked
3876 entries during signature checking and serial number lookup. Now the
3877 encoding is cached and the serial number sort performed under a lock.
3878 Add new STACK function sk_is_sorted().
3881 *) Add Delta CRL to the extension code.
3884 *) Various fixes to s3_pkt.c so alerts are sent properly.
3885 [David Holmes <d.holmes@f5.com>]
3887 *) Reduce the chances of duplicate issuer name and serial numbers (in
3888 violation of RFC3280) using the OpenSSL certificate creation utilities.
3889 This is done by creating a random 64 bit value for the initial serial
3890 number when a serial number file is created or when a self signed
3891 certificate is created using 'openssl req -x509'. The initial serial
3892 number file is created using 'openssl x509 -next_serial' in CA.pl
3893 rather than being initialized to 1.
3896 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
3898 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3899 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3900 [Joe Orton, Steve Henson]
3902 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
3904 [Joe Orton, Steve Henson]
3906 *) Make it possible to have multiple active certificates with the same
3907 subject in the CA index file. This is done only if the keyword
3908 'unique_subject' is set to 'no' in the main CA section (default
3909 if 'CA_default') of the configuration file. The value is saved
3910 with the database itself in a separate index attribute file,
3911 named like the index file with '.attr' appended to the name.
3914 *) X509 verify fixes. Disable broken certificate workarounds when
3915 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
3916 keyUsage extension present. Don't accept CRLs with unhandled critical
3917 extensions: since verify currently doesn't process CRL extensions this
3918 rejects a CRL with *any* critical extensions. Add new verify error codes
3922 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
3923 A clarification of RFC2560 will require the use of OCTET STRINGs and
3924 some implementations cannot handle the current raw format. Since OpenSSL
3925 copies and compares OCSP nonces as opaque blobs without any attempt at
3926 parsing them this should not create any compatibility issues.
3929 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
3930 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
3931 this HMAC (and other) operations are several times slower than OpenSSL
3935 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
3936 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3938 *) Use the correct content when signing type "other".
3941 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
3943 *) Fix various bugs revealed by running the NISCC test suite:
3945 Stop out of bounds reads in the ASN1 code when presented with
3946 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3948 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
3950 If verify callback ignores invalid public key errors don't try to check
3951 certificate signature with the NULL public key.
3955 *) New -ignore_err option in ocsp application to stop the server
3956 exiting on the first error in a request.
3959 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3960 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3964 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3965 extra data after the compression methods not only for TLS 1.0
3966 but also for SSL 3.0 (as required by the specification).
3967 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3969 *) Change X509_certificate_type() to mark the key as exported/exportable
3970 when it's 512 *bits* long, not 512 bytes.
3973 *) Change AES_cbc_encrypt() so it outputs exact multiple of
3974 blocks during encryption.
3977 *) Various fixes to base64 BIO and non blocking I/O. On write
3978 flushes were not handled properly if the BIO retried. On read
3979 data was not being buffered properly and had various logic bugs.
3980 This also affects blocking I/O when the data being decoded is a
3984 *) Various S/MIME bugfixes and compatibility changes:
3985 output correct application/pkcs7 MIME type if
3986 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
3987 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
3988 of files as .eml work). Correctly handle very long lines in MIME
3992 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
3994 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3995 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3996 a protocol version number mismatch like a decryption error
3997 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4000 *) Turn on RSA blinding by default in the default implementation
4001 to avoid a timing attack. Applications that don't want it can call
4002 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4003 They would be ill-advised to do so in most cases.
4004 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4006 *) Change RSA blinding code so that it works when the PRNG is not
4007 seeded (in this case, the secret RSA exponent is abused as
4008 an unpredictable seed -- if it is not unpredictable, there
4009 is no point in blinding anyway). Make RSA blinding thread-safe
4010 by remembering the creator's thread ID in rsa->blinding and
4011 having all other threads use local one-time blinding factors
4012 (this requires more computation than sharing rsa->blinding, but
4013 avoids excessive locking; and if an RSA object is not shared
4014 between threads, blinding will still be very fast).
4017 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
4018 ENGINE as defaults for all supported algorithms irrespective of
4019 the 'flags' parameter. 'flags' is now honoured, so applications
4020 should make sure they are passing it correctly.
4023 *) Target "mingw" now allows native Windows code to be generated in
4024 the Cygwin environment as well as with the MinGW compiler.
4027 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
4029 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4030 via timing by performing a MAC computation even if incorrrect
4031 block cipher padding has been found. This is a countermeasure
4032 against active attacks where the attacker has to distinguish
4033 between bad padding and a MAC verification error. (CVE-2003-0078)
4035 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4036 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4037 Martin Vuagnoux (EPFL, Ilion)]
4039 *) Make the no-err option work as intended. The intention with no-err
4040 is not to have the whole error stack handling routines removed from
4041 libcrypto, it's only intended to remove all the function name and
4042 reason texts, thereby removing some of the footprint that may not
4043 be interesting if those errors aren't displayed anyway.
4045 NOTE: it's still possible for any application or module to have it's
4046 own set of error texts inserted. The routines are there, just not
4047 used by default when no-err is given.
4050 *) Add support for FreeBSD on IA64.
4051 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
4053 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
4054 Kerberos function mit_des_cbc_cksum(). Before this change,
4055 the value returned by DES_cbc_cksum() was like the one from
4056 mit_des_cbc_cksum(), except the bytes were swapped.
4057 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
4059 *) Allow an application to disable the automatic SSL chain building.
4060 Before this a rather primitive chain build was always performed in
4061 ssl3_output_cert_chain(): an application had no way to send the
4062 correct chain if the automatic operation produced an incorrect result.
4064 Now the chain builder is disabled if either:
4066 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
4068 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
4070 The reasoning behind this is that an application would not want the
4071 auto chain building to take place if extra chain certificates are
4072 present and it might also want a means of sending no additional
4073 certificates (for example the chain has two certificates and the
4077 *) Add the possibility to build without the ENGINE framework.
4078 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4080 *) Under Win32 gmtime() can return NULL: check return value in
4081 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
4084 *) DSA routines: under certain error conditions uninitialized BN objects
4085 could be freed. Solution: make sure initialization is performed early
4086 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
4087 Nils Larsch <nla@trustcenter.de> via PR#459)
4090 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
4091 checked on reconnect on the client side, therefore session resumption
4092 could still fail with a "ssl session id is different" error. This
4093 behaviour is masked when SSL_OP_ALL is used due to
4094 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
4095 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4096 followup to PR #377.
4099 *) IA-32 assembler support enhancements: unified ELF targets, support
4100 for SCO/Caldera platforms, fix for Cygwin shared build.
4103 *) Add support for FreeBSD on sparc64. As a consequence, support for
4104 FreeBSD on non-x86 processors is separate from x86 processors on
4105 the config script, much like the NetBSD support.
4106 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
4108 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
4110 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
4113 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
4114 code (06) was taken as the first octet of the session ID and the last
4115 octet was ignored consequently. As a result SSLv2 client side session
4116 caching could not have worked due to the session ID mismatch between
4118 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4122 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
4123 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
4127 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
4128 seems that in spite of existing for more than a year, many application
4129 author have done nothing to provide the necessary callbacks, which
4130 means that this particular engine will not work properly anywhere.
4131 This is a very unfortunate situation which forces us, in the name
4132 of usability, to give the hw_ncipher.c a static lock, which is part
4134 NOTE: This is for the 0.9.7 series ONLY. This hack will never
4135 appear in 0.9.8 or later. We EXPECT application authors to have
4136 dealt properly with this when 0.9.8 is released (unless we actually
4137 make such changes in the libcrypto locking code that changes will
4138 have to be made anyway).
4141 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
4142 octets have been read, EOF or an error occurs. Without this change
4143 some truncated ASN1 structures will not produce an error.
4146 *) Disable Heimdal support, since it hasn't been fully implemented.
4147 Still give the possibility to force the use of Heimdal, but with
4148 warnings and a request that patches get sent to openssl-dev.
4151 *) Add the VC-CE target, introduce the WINCE sysname, and add
4152 INSTALL.WCE and appropriate conditionals to make it build.
4153 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4155 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
4156 cygssl-x.y.z.dll, where x, y and z are the major, minor and
4157 edit numbers of the version.
4158 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4160 *) Introduce safe string copy and catenation functions
4161 (BUF_strlcpy() and BUF_strlcat()).
4162 [Ben Laurie (CHATS) and Richard Levitte]
4164 *) Avoid using fixed-size buffers for one-line DNs.
4165 [Ben Laurie (CHATS)]
4167 *) Add BUF_MEM_grow_clean() to avoid information leakage when
4168 resizing buffers containing secrets, and use where appropriate.
4169 [Ben Laurie (CHATS)]
4171 *) Avoid using fixed size buffers for configuration file location.
4172 [Ben Laurie (CHATS)]
4174 *) Avoid filename truncation for various CA files.
4175 [Ben Laurie (CHATS)]
4177 *) Use sizeof in preference to magic numbers.
4178 [Ben Laurie (CHATS)]
4180 *) Avoid filename truncation in cert requests.
4181 [Ben Laurie (CHATS)]
4183 *) Add assertions to check for (supposedly impossible) buffer
4185 [Ben Laurie (CHATS)]
4187 *) Don't cache truncated DNS entries in the local cache (this could
4188 potentially lead to a spoofing attack).
4189 [Ben Laurie (CHATS)]
4191 *) Fix various buffers to be large enough for hex/decimal
4192 representations in a platform independent manner.
4193 [Ben Laurie (CHATS)]
4195 *) Add CRYPTO_realloc_clean() to avoid information leakage when
4196 resizing buffers containing secrets, and use where appropriate.
4197 [Ben Laurie (CHATS)]
4199 *) Add BIO_indent() to avoid much slightly worrying code to do
4201 [Ben Laurie (CHATS)]
4203 *) Convert sprintf()/BIO_puts() to BIO_printf().
4204 [Ben Laurie (CHATS)]
4206 *) buffer_gets() could terminate with the buffer only half
4208 [Ben Laurie (CHATS)]
4210 *) Add assertions to prevent user-supplied crypto functions from
4211 overflowing internal buffers by having large block sizes, etc.
4212 [Ben Laurie (CHATS)]
4214 *) New OPENSSL_assert() macro (similar to assert(), but enabled
4216 [Ben Laurie (CHATS)]
4218 *) Eliminate unused copy of key in RC4.
4219 [Ben Laurie (CHATS)]
4221 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
4222 [Ben Laurie (CHATS)]
4224 *) Fix off-by-one error in EGD path.
4225 [Ben Laurie (CHATS)]
4227 *) If RANDFILE path is too long, ignore instead of truncating.
4228 [Ben Laurie (CHATS)]
4230 *) Eliminate unused and incorrectly sized X.509 structure
4232 [Ben Laurie (CHATS)]
4234 *) Eliminate unused and dangerous function knumber().
4235 [Ben Laurie (CHATS)]
4237 *) Eliminate unused and dangerous structure, KSSL_ERR.
4238 [Ben Laurie (CHATS)]
4240 *) Protect against overlong session ID context length in an encoded
4241 session object. Since these are local, this does not appear to be
4243 [Ben Laurie (CHATS)]
4245 *) Change from security patch (see 0.9.6e below) that did not affect
4246 the 0.9.6 release series:
4248 Remote buffer overflow in SSL3 protocol - an attacker could
4249 supply an oversized master key in Kerberos-enabled versions.
4251 [Ben Laurie (CHATS)]
4253 *) Change the SSL kerb5 codes to match RFC 2712.
4256 *) Make -nameopt work fully for req and add -reqopt switch.
4257 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
4259 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
4260 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
4262 *) Make sure tests can be performed even if the corresponding algorithms
4263 have been removed entirely. This was also the last step to make
4264 OpenSSL compilable with DJGPP under all reasonable conditions.
4265 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
4267 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
4268 to allow version independent disabling of normally unselected ciphers,
4269 which may be activated as a side-effect of selecting a single cipher.
4271 (E.g., cipher list string "RSA" enables ciphersuites that are left
4272 out of "ALL" because they do not provide symmetric encryption.
4273 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
4274 [Lutz Jaenicke, Bodo Moeller]
4276 *) Add appropriate support for separate platform-dependent build
4277 directories. The recommended way to make a platform-dependent
4278 build directory is the following (tested on Linux), maybe with
4281 # Place yourself outside of the OpenSSL source tree. In
4282 # this example, the environment variable OPENSSL_SOURCE
4283 # is assumed to contain the absolute OpenSSL source directory.
4284 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
4285 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
4286 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
4287 mkdir -p `dirname $F`
4288 ln -s $OPENSSL_SOURCE/$F $F
4291 To be absolutely sure not to disturb the source tree, a "make clean"
4292 is a good thing. If it isn't successfull, don't worry about it,
4293 it probably means the source directory is very clean.
4296 *) Make sure any ENGINE control commands make local copies of string
4297 pointers passed to them whenever necessary. Otherwise it is possible
4298 the caller may have overwritten (or deallocated) the original string
4299 data when a later ENGINE operation tries to use the stored values.
4300 [Götz Babin-Ebell <babinebell@trustcenter.de>]
4302 *) Improve diagnostics in file reading and command-line digests.
4303 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
4305 *) Add AES modes CFB and OFB to the object database. Correct an
4306 error in AES-CFB decryption.
4309 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
4310 allows existing EVP_CIPHER_CTX structures to be reused after
4311 calling EVP_*Final(). This behaviour is used by encryption
4312 BIOs and some applications. This has the side effect that
4313 applications must explicitly clean up cipher contexts with
4314 EVP_CIPHER_CTX_cleanup() or they will leak memory.
4317 *) Check the values of dna and dnb in bn_mul_recursive before calling
4318 bn_mul_comba (a non zero value means the a or b arrays do not contain
4319 n2 elements) and fallback to bn_mul_normal if either is not zero.
4322 *) Fix escaping of non-ASCII characters when using the -subj option
4323 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
4326 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
4327 form for "surname", serialNumber has no short form.
4328 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
4329 therefore remove "mail" short name for "internet 7".
4330 The OID for unique identifiers in X509 certificates is
4331 x500UniqueIdentifier, not uniqueIdentifier.
4332 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
4335 *) Add an "init" command to the ENGINE config module and auto initialize
4336 ENGINEs. Without any "init" command the ENGINE will be initialized
4337 after all ctrl commands have been executed on it. If init=1 the
4338 ENGINE is initailized at that point (ctrls before that point are run
4339 on the uninitialized ENGINE and after on the initialized one). If
4340 init=0 then the ENGINE will not be iniatialized at all.
4343 *) Fix the 'app_verify_callback' interface so that the user-defined
4344 argument is actually passed to the callback: In the
4345 SSL_CTX_set_cert_verify_callback() prototype, the callback
4346 declaration has been changed from
4349 int (*cb)(X509_STORE_CTX *,void *);
4350 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
4351 i=s->ctx->app_verify_callback(&ctx)
4352 has been changed into
4353 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
4355 To update applications using SSL_CTX_set_cert_verify_callback(),
4356 a dummy argument can be added to their callback functions.
4357 [D. K. Smetters <smetters@parc.xerox.com>]
4359 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
4360 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
4362 *) Add and OPENSSL_LOAD_CONF define which will cause
4363 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
4364 This allows older applications to transparently support certain
4365 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
4366 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
4367 load the config file and OPENSSL_add_all_algorithms_conf() which will
4368 always load it have also been added.
4371 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
4372 Adjust NIDs and EVP layer.
4373 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4375 *) Config modules support in openssl utility.
4377 Most commands now load modules from the config file,
4378 though in a few (such as version) this isn't done
4379 because it couldn't be used for anything.
4381 In the case of ca and req the config file used is
4382 the same as the utility itself: that is the -config
4383 command line option can be used to specify an
4387 *) Move default behaviour from OPENSSL_config(). If appname is NULL
4388 use "openssl_conf" if filename is NULL use default openssl config file.
4391 *) Add an argument to OPENSSL_config() to allow the use of an alternative
4392 config section name. Add a new flag to tolerate a missing config file
4393 and move code to CONF_modules_load_file().
4396 *) Support for crypto accelerator cards from Accelerated Encryption
4397 Processing, www.aep.ie. (Use engine 'aep')
4398 The support was copied from 0.9.6c [engine] and adapted/corrected
4399 to work with the new engine framework.
4400 [AEP Inc. and Richard Levitte]
4402 *) Support for SureWare crypto accelerator cards from Baltimore
4403 Technologies. (Use engine 'sureware')
4404 The support was copied from 0.9.6c [engine] and adapted
4405 to work with the new engine framework.
4408 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
4409 make the newer ENGINE framework commands for the CHIL engine work.
4410 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
4412 *) Make it possible to produce shared libraries on ReliantUNIX.
4413 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
4415 *) Add the configuration target debug-linux-ppro.
4416 Make 'openssl rsa' use the general key loading routines
4417 implemented in apps.c, and make those routines able to
4418 handle the key format FORMAT_NETSCAPE and the variant
4420 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4422 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4423 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4425 *) Add -keyform to rsautl, and document -engine.
4426 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
4428 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
4429 BIO_R_NO_SUCH_FILE error code rather than the generic
4430 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
4433 *) Add new functions
4435 ERR_peek_last_error_line
4436 ERR_peek_last_error_line_data.
4437 These are similar to
4440 ERR_peek_error_line_data,
4441 but report on the latest error recorded rather than the first one
4442 still in the error queue.
4443 [Ben Laurie, Bodo Moeller]
4445 *) default_algorithms option in ENGINE config module. This allows things
4447 default_algorithms = ALL
4448 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
4451 *) Prelminary ENGINE config module.
4454 *) New experimental application configuration code.
4457 *) Change the AES code to follow the same name structure as all other
4458 symmetric ciphers, and behave the same way. Move everything to
4459 the directory crypto/aes, thereby obsoleting crypto/rijndael.
4460 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4462 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
4463 [Ben Laurie and Theo de Raadt]
4465 *) Add option to output public keys in req command.
4466 [Massimiliano Pala madwolf@openca.org]
4468 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
4469 (up to about 10% better than before for P-192 and P-224).
4472 *) New functions/macros
4474 SSL_CTX_set_msg_callback(ctx, cb)
4475 SSL_CTX_set_msg_callback_arg(ctx, arg)
4476 SSL_set_msg_callback(ssl, cb)
4477 SSL_set_msg_callback_arg(ssl, arg)
4479 to request calling a callback function
4481 void cb(int write_p, int version, int content_type,
4482 const void *buf, size_t len, SSL *ssl, void *arg)
4484 whenever a protocol message has been completely received
4485 (write_p == 0) or sent (write_p == 1). Here 'version' is the
4486 protocol version according to which the SSL library interprets
4487 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
4488 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
4489 the content type as defined in the SSL 3.0/TLS 1.0 protocol
4490 specification (change_cipher_spec(20), alert(21), handshake(22)).
4491 'buf' and 'len' point to the actual message, 'ssl' to the
4492 SSL object, and 'arg' is the application-defined value set by
4493 SSL[_CTX]_set_msg_callback_arg().
4495 'openssl s_client' and 'openssl s_server' have new '-msg' options
4496 to enable a callback that displays all protocol messages.
4499 *) Change the shared library support so shared libraries are built as
4500 soon as the corresponding static library is finished, and thereby get
4501 openssl and the test programs linked against the shared library.
4502 This still only happens when the keyword "shard" has been given to
4503 the configuration scripts.
4505 NOTE: shared library support is still an experimental thing, and
4506 backward binary compatibility is still not guaranteed.
4507 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
4509 *) Add support for Subject Information Access extension.
4510 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4512 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
4513 additional bytes when new memory had to be allocated, not just
4514 when reusing an existing buffer.
4517 *) New command line and configuration option 'utf8' for the req command.
4518 This allows field values to be specified as UTF8 strings.
4521 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
4522 runs for the former and machine-readable output for the latter.
4525 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
4526 of the e-mail address in the DN (i.e., it will go into a certificate
4527 extension only). The new configuration file option 'email_in_dn = no'
4528 has the same effect.
4529 [Massimiliano Pala madwolf@openca.org]
4531 *) Change all functions with names starting with des_ to be starting
4532 with DES_ instead. Add wrappers that are compatible with libdes,
4533 but are named _ossl_old_des_*. Finally, add macros that map the
4534 des_* symbols to the corresponding _ossl_old_des_* if libdes
4535 compatibility is desired. If OpenSSL 0.9.6c compatibility is
4536 desired, the des_* symbols will be mapped to DES_*, with one
4539 Since we provide two compatibility mappings, the user needs to
4540 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
4541 compatibility is desired. The default (i.e., when that macro
4542 isn't defined) is OpenSSL 0.9.6c compatibility.
4544 There are also macros that enable and disable the support of old
4545 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
4546 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
4547 are defined, the default will apply: to support the old des routines.
4549 In either case, one must include openssl/des.h to get the correct
4550 definitions. Do not try to just include openssl/des_old.h, that
4553 NOTE: This is a major break of an old API into a new one. Software
4554 authors are encouraged to switch to the DES_ style functions. Some
4555 time in the future, des_old.h and the libdes compatibility functions
4556 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
4557 default), and then completely removed.
4560 *) Test for certificates which contain unsupported critical extensions.
4561 If such a certificate is found during a verify operation it is
4562 rejected by default: this behaviour can be overridden by either
4563 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
4564 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
4565 X509_supported_extension() has also been added which returns 1 if a
4566 particular extension is supported.
4569 *) Modify the behaviour of EVP cipher functions in similar way to digests
4570 to retain compatibility with existing code.
4573 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
4574 compatibility with existing code. In particular the 'ctx' parameter does
4575 not have to be to be initialized before the call to EVP_DigestInit() and
4576 it is tidied up after a call to EVP_DigestFinal(). New function
4577 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
4578 EVP_MD_CTX_copy() changed to not require the destination to be
4579 initialized valid and new function EVP_MD_CTX_copy_ex() added which
4580 requires the destination to be valid.
4582 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
4583 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
4586 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
4587 so that complete 'Handshake' protocol structures are kept in memory
4588 instead of overwriting 'msg_type' and 'length' with 'body' data.
4591 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
4592 [Massimo Santin via Richard Levitte]
4594 *) Major restructuring to the underlying ENGINE code. This includes
4595 reduction of linker bloat, separation of pure "ENGINE" manipulation
4596 (initialisation, etc) from functionality dealing with implementations
4597 of specific crypto iterfaces. This change also introduces integrated
4598 support for symmetric ciphers and digest implementations - so ENGINEs
4599 can now accelerate these by providing EVP_CIPHER and EVP_MD
4600 implementations of their own. This is detailed in crypto/engine/README
4601 as it couldn't be adequately described here. However, there are a few
4602 API changes worth noting - some RSA, DSA, DH, and RAND functions that
4603 were changed in the original introduction of ENGINE code have now
4604 reverted back - the hooking from this code to ENGINE is now a good
4605 deal more passive and at run-time, operations deal directly with
4606 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
4607 dereferencing through an ENGINE pointer any more. Also, the ENGINE
4608 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
4609 they were not being used by the framework as there is no concept of a
4610 BIGNUM_METHOD and they could not be generalised to the new
4611 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
4612 ENGINE_cpy() has been removed as it cannot be consistently defined in
4616 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
4619 *) Change mkdef.pl to sort symbols that get the same entry number,
4620 and make sure the automatically generated functions ERR_load_*
4621 become part of libeay.num as well.
4624 *) New function SSL_renegotiate_pending(). This returns true once
4625 renegotiation has been requested (either SSL_renegotiate() call
4626 or HelloRequest/ClientHello receveived from the peer) and becomes
4627 false once a handshake has been completed.
4628 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
4629 sends a HelloRequest, but does not ensure that a handshake takes
4630 place. SSL_renegotiate_pending() is useful for checking if the
4631 client has followed the request.)
4634 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
4635 By default, clients may request session resumption even during
4636 renegotiation (if session ID contexts permit); with this option,
4637 session resumption is possible only in the first handshake.
4639 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4640 more bits available for options that should not be part of
4641 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4644 *) Add some demos for certificate and certificate request creation.
4647 *) Make maximum certificate chain size accepted from the peer application
4648 settable (SSL*_get/set_max_cert_list()), as proposed by
4649 "Douglas E. Engert" <deengert@anl.gov>.
4652 *) Add support for shared libraries for Unixware-7
4653 (Boyd Lynn Gerber <gerberb@zenez.com>).
4656 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4657 be done prior to destruction. Use this to unload error strings from
4658 ENGINEs that load their own error strings. NB: This adds two new API
4659 functions to "get" and "set" this destroy handler in an ENGINE.
4662 *) Alter all existing ENGINE implementations (except "openssl" and
4663 "openbsd") to dynamically instantiate their own error strings. This
4664 makes them more flexible to be built both as statically-linked ENGINEs
4665 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4666 Also, add stub code to each that makes building them as self-contained
4667 shared-libraries easier (see README.ENGINE).
4670 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4671 implementations into applications that are completely implemented in
4672 self-contained shared-libraries. The "dynamic" ENGINE exposes control
4673 commands that can be used to configure what shared-library to load and
4674 to control aspects of the way it is handled. Also, made an update to
4675 the README.ENGINE file that brings its information up-to-date and
4676 provides some information and instructions on the "dynamic" ENGINE
4677 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
4680 *) Make it possible to unload ranges of ERR strings with a new
4681 "ERR_unload_strings" function.
4684 *) Add a copy() function to EVP_MD.
4687 *) Make EVP_MD routines take a context pointer instead of just the
4688 md_data void pointer.
4691 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
4692 that the digest can only process a single chunk of data
4693 (typically because it is provided by a piece of
4694 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
4695 is only going to provide a single chunk of data, and hence the
4696 framework needn't accumulate the data for oneshot drivers.
4699 *) As with "ERR", make it possible to replace the underlying "ex_data"
4700 functions. This change also alters the storage and management of global
4701 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
4702 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
4703 index counters. The API functions that use this state have been changed
4704 to take a "class_index" rather than pointers to the class's local STACK
4705 and counter, and there is now an API function to dynamically create new
4706 classes. This centralisation allows us to (a) plug a lot of the
4707 thread-safety problems that existed, and (b) makes it possible to clean
4708 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
4709 such data would previously have always leaked in application code and
4710 workarounds were in place to make the memory debugging turn a blind eye
4711 to it. Application code that doesn't use this new function will still
4712 leak as before, but their memory debugging output will announce it now
4713 rather than letting it slide.
4715 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
4716 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
4717 has a return value to indicate success or failure.
4720 *) Make it possible to replace the underlying "ERR" functions such that the
4721 global state (2 LHASH tables and 2 locks) is only used by the "default"
4722 implementation. This change also adds two functions to "get" and "set"
4723 the implementation prior to it being automatically set the first time
4724 any other ERR function takes place. Ie. an application can call "get",
4725 pass the return value to a module it has just loaded, and that module
4726 can call its own "set" function using that value. This means the
4727 module's "ERR" operations will use (and modify) the error state in the
4728 application and not in its own statically linked copy of OpenSSL code.
4731 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
4732 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
4733 the operation, and provides a more encapsulated way for external code
4734 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
4735 to use these functions rather than manually incrementing the counts.
4737 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
4740 *) Add EVP test program.
4743 *) Add symmetric cipher support to ENGINE. Expect the API to change!
4746 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
4747 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
4748 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
4749 These allow a CRL to be built without having to access X509_CRL fields
4750 directly. Modify 'ca' application to use new functions.
4753 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
4754 bug workarounds. Rollback attack detection is a security feature.
4755 The problem will only arise on OpenSSL servers when TLSv1 is not
4756 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
4757 Software authors not wanting to support TLSv1 will have special reasons
4758 for their choice and can explicitly enable this option.
4759 [Bodo Moeller, Lutz Jaenicke]
4761 *) Rationalise EVP so it can be extended: don't include a union of
4762 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
4763 (similar to those existing for EVP_CIPHER_CTX).
4768 EVP_MD_CTX_init(&md); /* new function call */
4769 EVP_DigestInit(&md, EVP_sha1());
4770 EVP_DigestUpdate(&md, in, len);
4771 EVP_DigestFinal(&md, out, NULL);
4772 EVP_MD_CTX_cleanup(&md); /* new function call */
4776 *) Make DES key schedule conform to the usual scheme, as well as
4777 correcting its structure. This means that calls to DES functions
4778 now have to pass a pointer to a des_key_schedule instead of a
4779 plain des_key_schedule (which was actually always a pointer
4782 des_key_schedule ks;
4784 des_set_key_checked(..., &ks);
4785 des_ncbc_encrypt(..., &ks, ...);
4787 (Note that a later change renames 'des_...' into 'DES_...'.)
4790 *) Initial reduction of linker bloat: the use of some functions, such as
4791 PEM causes large amounts of unused functions to be linked in due to
4792 poor organisation. For example pem_all.c contains every PEM function
4793 which has a knock on effect of linking in large amounts of (unused)
4794 ASN1 code. Grouping together similar functions and splitting unrelated
4795 functions prevents this.
4798 *) Cleanup of EVP macros.
4801 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
4802 correct _ecb suffix.
4805 *) Add initial OCSP responder support to ocsp application. The
4806 revocation information is handled using the text based index
4807 use by the ca application. The responder can either handle
4808 requests generated internally, supplied in files (for example
4809 via a CGI script) or using an internal minimal server.
4812 *) Add configuration choices to get zlib compression for TLS.
4815 *) Changes to Kerberos SSL for RFC 2712 compliance:
4816 1. Implemented real KerberosWrapper, instead of just using
4817 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
4818 2. Implemented optional authenticator field of KerberosWrapper.
4820 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
4821 and authenticator structs; see crypto/krb5/.
4823 Generalized Kerberos calls to support multiple Kerberos libraries.
4824 [Vern Staats <staatsvr@asc.hpc.mil>,
4825 Jeffrey Altman <jaltman@columbia.edu>
4826 via Richard Levitte]
4828 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
4829 already does with RSA. testdsa.h now has 'priv_key/pub_key'
4830 values for each of the key sizes rather than having just
4831 parameters (and 'speed' generating keys each time).
4834 *) Speed up EVP routines.
4837 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
4838 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
4839 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
4840 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
4842 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
4843 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
4844 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
4847 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
4849 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
4852 *) Added the OS2-EMX target.
4853 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
4855 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
4856 to support NCONF routines in extension code. New function CONF_set_nconf()
4857 to allow functions which take an NCONF to also handle the old LHASH
4858 structure: this means that the old CONF compatible routines can be
4859 retained (in particular wrt extensions) without having to duplicate the
4860 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
4863 *) Enhance the general user interface with mechanisms for inner control
4864 and with possibilities to have yes/no kind of prompts.
4867 *) Change all calls to low level digest routines in the library and
4868 applications to use EVP. Add missing calls to HMAC_cleanup() and
4869 don't assume HMAC_CTX can be copied using memcpy().
4870 [Verdon Walker <VWalker@novell.com>, Steve Henson]
4872 *) Add the possibility to control engines through control names but with
4873 arbitrary arguments instead of just a string.
4874 Change the key loaders to take a UI_METHOD instead of a callback
4875 function pointer. NOTE: this breaks binary compatibility with earlier
4876 versions of OpenSSL [engine].
4877 Adapt the nCipher code for these new conditions and add a card insertion
4881 *) Enhance the general user interface with mechanisms to better support
4882 dialog box interfaces, application-defined prompts, the possibility
4883 to use defaults (for example default passwords from somewhere else)
4884 and interrupts/cancellations.
4887 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
4888 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
4891 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
4892 tidy up some unnecessarily weird code in 'sk_new()').
4893 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
4895 *) Change the key loading routines for ENGINEs to use the same kind
4896 callback (pem_password_cb) as all other routines that need this
4900 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
4901 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
4902 than this minimum value is recommended.
4905 *) New random seeder for OpenVMS, using the system process statistics
4906 that are easily reachable.
4909 *) Windows apparently can't transparently handle global
4910 variables defined in DLLs. Initialisations such as:
4912 const ASN1_ITEM *it = &ASN1_INTEGER_it;
4914 wont compile. This is used by the any applications that need to
4915 declare their own ASN1 modules. This was fixed by adding the option
4916 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
4917 needed for static libraries under Win32.
4920 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
4921 setting of purpose and trust fields. New X509_STORE trust and
4922 purpose functions and tidy up setting in other SSL functions.
4925 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
4926 structure. These are inherited by X509_STORE_CTX when it is
4927 initialised. This allows various defaults to be set in the
4928 X509_STORE structure (such as flags for CRL checking and custom
4929 purpose or trust settings) for functions which only use X509_STORE_CTX
4930 internally such as S/MIME.
4932 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
4933 trust settings if they are not set in X509_STORE. This allows X509_STORE
4934 purposes and trust (in S/MIME for example) to override any set by default.
4936 Add command line options for CRL checking to smime, s_client and s_server
4940 *) Initial CRL based revocation checking. If the CRL checking flag(s)
4941 are set then the CRL is looked up in the X509_STORE structure and
4942 its validity and signature checked, then if the certificate is found
4943 in the CRL the verify fails with a revoked error.
4945 Various new CRL related callbacks added to X509_STORE_CTX structure.
4947 Command line options added to 'verify' application to support this.
4949 This needs some additional work, such as being able to handle multiple
4950 CRLs with different times, extension based lookup (rather than just
4951 by subject name) and ultimately more complete V2 CRL extension
4955 *) Add a general user interface API (crypto/ui/). This is designed
4956 to replace things like des_read_password and friends (backward
4957 compatibility functions using this new API are provided).
4958 The purpose is to remove prompting functions from the DES code
4959 section as well as provide for prompting through dialog boxes in
4960 a window system and the like.
4963 *) Add "ex_data" support to ENGINE so implementations can add state at a
4964 per-structure level rather than having to store it globally.
4967 *) Make it possible for ENGINE structures to be copied when retrieved by
4968 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
4969 This causes the "original" ENGINE structure to act like a template,
4970 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
4971 operational state can be localised to each ENGINE structure, despite the
4972 fact they all share the same "methods". New ENGINE structures returned in
4973 this case have no functional references and the return value is the single
4974 structural reference. This matches the single structural reference returned
4975 by ENGINE_by_id() normally, when it is incremented on the pre-existing
4979 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
4980 needs to match any other type at all we need to manually clear the
4984 *) Changes to the "openssl engine" utility to include;
4985 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
4986 about an ENGINE's available control commands.
4987 - executing control commands from command line arguments using the
4988 '-pre' and '-post' switches. '-post' is only used if '-t' is
4989 specified and the ENGINE is successfully initialised. The syntax for
4990 the individual commands are colon-separated, for example;
4991 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
4994 *) New dynamic control command support for ENGINEs. ENGINEs can now
4995 declare their own commands (numbers), names (strings), descriptions,
4996 and input types for run-time discovery by calling applications. A
4997 subset of these commands are implicitly classed as "executable"
4998 depending on their input type, and only these can be invoked through
4999 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
5000 can be based on user input, config files, etc). The distinction is
5001 that "executable" commands cannot return anything other than a boolean
5002 result and can only support numeric or string input, whereas some
5003 discoverable commands may only be for direct use through
5004 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
5005 pointers, or other custom uses. The "executable" commands are to
5006 support parameterisations of ENGINE behaviour that can be
5007 unambiguously defined by ENGINEs and used consistently across any
5008 OpenSSL-based application. Commands have been added to all the
5009 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
5010 control over shared-library paths without source code alterations.
5013 *) Changed all ENGINE implementations to dynamically allocate their
5014 ENGINEs rather than declaring them statically. Apart from this being
5015 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
5016 this also allows the implementations to compile without using the
5017 internal engine_int.h header.
5020 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
5021 'const' value. Any code that should be able to modify a RAND_METHOD
5022 should already have non-const pointers to it (ie. they should only
5023 modify their own ones).
5026 *) Made a variety of little tweaks to the ENGINE code.
5027 - "atalla" and "ubsec" string definitions were moved from header files
5028 to C code. "nuron" string definitions were placed in variables
5029 rather than hard-coded - allowing parameterisation of these values
5030 later on via ctrl() commands.
5031 - Removed unused "#if 0"'d code.
5032 - Fixed engine list iteration code so it uses ENGINE_free() to release
5033 structural references.
5034 - Constified the RAND_METHOD element of ENGINE structures.
5035 - Constified various get/set functions as appropriate and added
5036 missing functions (including a catch-all ENGINE_cpy that duplicates
5037 all ENGINE values onto a new ENGINE except reference counts/state).
5038 - Removed NULL parameter checks in get/set functions. Setting a method
5039 or function to NULL is a way of cancelling out a previously set
5040 value. Passing a NULL ENGINE parameter is just plain stupid anyway
5041 and doesn't justify the extra error symbols and code.
5042 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
5043 flags from engine_int.h to engine.h.
5044 - Changed prototypes for ENGINE handler functions (init(), finish(),
5045 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
5048 *) Implement binary inversion algorithm for BN_mod_inverse in addition
5049 to the algorithm using long division. The binary algorithm can be
5050 used only if the modulus is odd. On 32-bit systems, it is faster
5051 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
5052 roughly 5-15% for 256-bit moduli), so we use it only for moduli
5053 up to 450 bits. In 64-bit environments, the binary algorithm
5054 appears to be advantageous for much longer moduli; here we use it
5055 for moduli up to 2048 bits.
5058 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
5059 could not support the combine flag in choice fields.
5062 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
5063 extensions from a certificate request to the certificate.
5066 *) Allow multiple 'certopt' and 'nameopt' options to be separated
5067 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
5068 file: this allows the display of the certificate about to be
5069 signed to be customised, to allow certain fields to be included
5070 or excluded and extension details. The old system didn't display
5071 multicharacter strings properly, omitted fields not in the policy
5072 and couldn't display additional details such as extensions.
5075 *) Function EC_POINTs_mul for multiple scalar multiplication
5076 of an arbitrary number of elliptic curve points
5077 \sum scalars[i]*points[i],
5078 optionally including the generator defined for the EC_GROUP:
5079 scalar*generator + \sum scalars[i]*points[i].
5081 EC_POINT_mul is a simple wrapper function for the typical case
5082 that the point list has just one item (besides the optional
5086 *) First EC_METHODs for curves over GF(p):
5088 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
5089 operations and provides various method functions that can also
5090 operate with faster implementations of modular arithmetic.
5092 EC_GFp_mont_method() reuses most functions that are part of
5093 EC_GFp_simple_method, but uses Montgomery arithmetic.
5095 [Bodo Moeller; point addition and point doubling
5096 implementation directly derived from source code provided by
5097 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
5099 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
5100 crypto/ec/ec_lib.c):
5102 Curves are EC_GROUP objects (with an optional group generator)
5103 based on EC_METHODs that are built into the library.
5105 Points are EC_POINT objects based on EC_GROUP objects.
5107 Most of the framework would be able to handle curves over arbitrary
5108 finite fields, but as there are no obvious types for fields other
5109 than GF(p), some functions are limited to that for now.
5112 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
5113 that the file contains a complete HTTP response.
5116 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
5117 change the def and num file printf format specifier from "%-40sXXX"
5118 to "%-39s XXX". The latter will always guarantee a space after the
5119 field while the former will cause them to run together if the field
5120 is 40 of more characters long.
5123 *) Constify the cipher and digest 'method' functions and structures
5124 and modify related functions to take constant EVP_MD and EVP_CIPHER
5128 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
5129 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
5132 *) Modify EVP_Digest*() routines so they now return values. Although the
5133 internal software routines can never fail additional hardware versions
5137 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
5139 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
5140 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
5143 ERR_R_NESTED_ASN1_ERROR
5145 ERR_R_MISSING_ASN1_EOS
5146 were 4 .. 9, conflicting with
5147 ERR_LIB_RSA (= ERR_R_RSA_LIB)
5149 ERR_LIB_PEM (= ERR_R_PEM_LIB).
5150 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
5152 Add new error code 'ERR_R_INTERNAL_ERROR'.
5155 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
5159 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
5160 sets the subject name for a new request or supersedes the
5161 subject name in a given request. Formats that can be parsed are
5162 'CN=Some Name, OU=myOU, C=IT'
5164 'CN=Some Name/OU=myOU/C=IT'.
5166 Add options '-batch' and '-verbose' to 'openssl req'.
5167 [Massimiliano Pala <madwolf@hackmasters.net>]
5169 *) Introduce the possibility to access global variables through
5170 functions on platform were that's the best way to handle exporting
5171 global variables in shared libraries. To enable this functionality,
5172 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
5173 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
5174 is normally done by Configure or something similar).
5176 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
5177 in the source file (foo.c) like this:
5179 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
5180 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
5182 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
5183 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
5185 OPENSSL_DECLARE_GLOBAL(int,foo);
5186 #define foo OPENSSL_GLOBAL_REF(foo)
5187 OPENSSL_DECLARE_GLOBAL(double,bar);
5188 #define bar OPENSSL_GLOBAL_REF(bar)
5190 The #defines are very important, and therefore so is including the
5191 header file everywhere where the defined globals are used.
5193 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
5194 of ASN.1 items, but that structure is a bit different.
5196 The largest change is in util/mkdef.pl which has been enhanced with
5197 better and easier to understand logic to choose which symbols should
5198 go into the Windows .def files as well as a number of fixes and code
5199 cleanup (among others, algorithm keywords are now sorted
5200 lexicographically to avoid constant rewrites).
5203 *) In BN_div() keep a copy of the sign of 'num' before writing the
5204 result to 'rm' because if rm==num the value will be overwritten
5205 and produce the wrong result if 'num' is negative: this caused
5206 problems with BN_mod() and BN_nnmod().
5209 *) Function OCSP_request_verify(). This checks the signature on an
5210 OCSP request and verifies the signer certificate. The signer
5211 certificate is just checked for a generic purpose and OCSP request
5215 *) Add OCSP_check_validity() function to check the validity of OCSP
5216 responses. OCSP responses are prepared in real time and may only
5217 be a few seconds old. Simply checking that the current time lies
5218 between thisUpdate and nextUpdate max reject otherwise valid responses
5219 caused by either OCSP responder or client clock inaccuracy. Instead
5220 we allow thisUpdate and nextUpdate to fall within a certain period of
5221 the current time. The age of the response can also optionally be
5222 checked. Two new options -validity_period and -status_age added to
5226 *) If signature or public key algorithm is unrecognized print out its
5227 OID rather that just UNKNOWN.
5230 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
5231 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
5232 ID to be generated from the issuer certificate alone which can then be
5233 passed to OCSP_id_issuer_cmp().
5236 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
5237 ASN1 modules to export functions returning ASN1_ITEM pointers
5238 instead of the ASN1_ITEM structures themselves. This adds several
5239 new macros which allow the underlying ASN1 function/structure to
5240 be accessed transparently. As a result code should not use ASN1_ITEM
5241 references directly (such as &X509_it) but instead use the relevant
5242 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
5243 use of the new ASN1 code on platforms where exporting structures
5244 is problematical (for example in shared libraries) but exporting
5245 functions returning pointers to structures is not.
5248 *) Add support for overriding the generation of SSL/TLS session IDs.
5249 These callbacks can be registered either in an SSL_CTX or per SSL.
5250 The purpose of this is to allow applications to control, if they wish,
5251 the arbitrary values chosen for use as session IDs, particularly as it
5252 can be useful for session caching in multiple-server environments. A
5253 command-line switch for testing this (and any client code that wishes
5254 to use such a feature) has been added to "s_server".
5255 [Geoff Thorpe, Lutz Jaenicke]
5257 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
5258 of the form '#if defined(...) || defined(...) || ...' and
5259 '#if !defined(...) && !defined(...) && ...'. This also avoids
5260 the growing number of special cases it was previously handling.
5263 *) Make all configuration macros available for application by making
5264 sure they are available in opensslconf.h, by giving them names starting
5265 with "OPENSSL_" to avoid conflicts with other packages and by making
5266 sure e_os2.h will cover all platform-specific cases together with
5268 Additionally, it is now possible to define configuration/platform-
5269 specific names (called "system identities"). In the C code, these
5270 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
5271 macro with the name beginning with "OPENSSL_SYS_", which is determined
5272 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
5276 *) New option -set_serial to 'req' and 'x509' this allows the serial
5277 number to use to be specified on the command line. Previously self
5278 signed certificates were hard coded with serial number 0 and the
5279 CA options of 'x509' had to use a serial number in a file which was
5283 *) New options to 'ca' utility to support V2 CRL entry extensions.
5284 Currently CRL reason, invalidity date and hold instruction are
5285 supported. Add new CRL extensions to V3 code and some new objects.
5288 *) New function EVP_CIPHER_CTX_set_padding() this is used to
5289 disable standard block padding (aka PKCS#5 padding) in the EVP
5290 API, which was previously mandatory. This means that the data is
5291 not padded in any way and so the total length much be a multiple
5292 of the block size, otherwise an error occurs.
5295 *) Initial (incomplete) OCSP SSL support.
5298 *) New function OCSP_parse_url(). This splits up a URL into its host,
5299 port and path components: primarily to parse OCSP URLs. New -url
5300 option to ocsp utility.
5303 *) New nonce behavior. The return value of OCSP_check_nonce() now
5304 reflects the various checks performed. Applications can decide
5305 whether to tolerate certain situations such as an absent nonce
5306 in a response when one was present in a request: the ocsp application
5307 just prints out a warning. New function OCSP_add1_basic_nonce()
5308 this is to allow responders to include a nonce in a response even if
5309 the request is nonce-less.
5312 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
5313 skipped when using openssl x509 multiple times on a single input file,
5314 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
5317 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
5318 set string type: to handle setting ASN1_TIME structures. Fix ca
5319 utility to correctly initialize revocation date of CRLs.
5322 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
5323 the clients preferred ciphersuites and rather use its own preferences.
5324 Should help to work around M$ SGC (Server Gated Cryptography) bug in
5325 Internet Explorer by ensuring unchanged hash method during stepup.
5326 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
5329 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
5330 to aes and add a new 'exist' option to print out symbols that don't
5334 *) Additional options to ocsp utility to allow flags to be set and
5335 additional certificates supplied.
5338 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
5339 OCSP client a number of certificate to only verify the response
5343 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
5344 handle the new API. Currently only ECB, CBC modes supported. Add new
5347 Add TLS AES ciphersuites as described in RFC3268, "Advanced
5348 Encryption Standard (AES) Ciphersuites for Transport Layer
5349 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
5350 not enabled by default and were not part of the "ALL" ciphersuite
5351 alias because they were not yet official; they could be
5352 explicitly requested by specifying the "AESdraft" ciphersuite
5353 group alias. In the final release of OpenSSL 0.9.7, the group
5354 alias is called "AES" and is part of "ALL".)
5355 [Ben Laurie, Steve Henson, Bodo Moeller]
5357 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
5358 request to response.
5361 *) Functions for OCSP responders. OCSP_request_onereq_count(),
5362 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
5363 extract information from a certificate request. OCSP_response_create()
5364 creates a response and optionally adds a basic response structure.
5365 OCSP_basic_add1_status() adds a complete single response to a basic
5366 response and returns the OCSP_SINGLERESP structure just added (to allow
5367 extensions to be included for example). OCSP_basic_add1_cert() adds a
5368 certificate to a basic response and OCSP_basic_sign() signs a basic
5369 response with various flags. New helper functions ASN1_TIME_check()
5370 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
5371 (converts ASN1_TIME to GeneralizedTime).
5374 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
5375 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
5376 structure from a certificate. X509_pubkey_digest() digests the public_key
5377 contents: this is used in various key identifiers.
5380 *) Make sk_sort() tolerate a NULL argument.
5381 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
5383 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
5384 passed by the function are trusted implicitly. If any of them signed the
5385 response then it is assumed to be valid and is not verified.
5388 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
5389 to data. This was previously part of the PKCS7 ASN1 code. This
5390 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
5391 [Steve Henson, reported by Kenneth R. Robinette
5392 <support@securenetterm.com>]
5394 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
5395 routines: without these tracing memory leaks is very painful.
5396 Fix leaks in PKCS12 and PKCS7 routines.
5399 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
5400 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
5401 effectively meant GeneralizedTime would never be used. Now it
5402 is initialised to -1 but X509_time_adj() now has to check the value
5403 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
5404 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
5405 [Steve Henson, reported by Kenneth R. Robinette
5406 <support@securenetterm.com>]
5408 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
5409 result in a zero length in the ASN1_INTEGER structure which was
5410 not consistent with the structure when d2i_ASN1_INTEGER() was used
5411 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
5412 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
5413 where it did not print out a minus for negative ASN1_INTEGER.
5416 *) Add summary printout to ocsp utility. The various functions which
5417 convert status values to strings have been renamed to:
5418 OCSP_response_status_str(), OCSP_cert_status_str() and
5419 OCSP_crl_reason_str() and are no longer static. New options
5420 to verify nonce values and to disable verification. OCSP response
5421 printout format cleaned up.
5424 *) Add additional OCSP certificate checks. These are those specified
5425 in RFC2560. This consists of two separate checks: the CA of the
5426 certificate being checked must either be the OCSP signer certificate
5427 or the issuer of the OCSP signer certificate. In the latter case the
5428 OCSP signer certificate must contain the OCSP signing extended key
5429 usage. This check is performed by attempting to match the OCSP
5430 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
5431 in the OCSP_CERTID structures of the response.
5434 *) Initial OCSP certificate verification added to OCSP_basic_verify()
5435 and related routines. This uses the standard OpenSSL certificate
5436 verify routines to perform initial checks (just CA validity) and
5437 to obtain the certificate chain. Then additional checks will be
5438 performed on the chain. Currently the root CA is checked to see
5439 if it is explicitly trusted for OCSP signing. This is used to set
5440 a root CA as a global signing root: that is any certificate that
5441 chains to that CA is an acceptable OCSP signing certificate.
5444 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
5445 extensions from a separate configuration file.
5446 As when reading extensions from the main configuration file,
5447 the '-extensions ...' option may be used for specifying the
5449 [Massimiliano Pala <madwolf@comune.modena.it>]
5451 *) New OCSP utility. Allows OCSP requests to be generated or
5452 read. The request can be sent to a responder and the output
5453 parsed, outputed or printed in text form. Not complete yet:
5454 still needs to check the OCSP response validity.
5457 *) New subcommands for 'openssl ca':
5458 'openssl ca -status <serial>' prints the status of the cert with
5459 the given serial number (according to the index file).
5460 'openssl ca -updatedb' updates the expiry status of certificates
5462 [Massimiliano Pala <madwolf@comune.modena.it>]
5464 *) New '-newreq-nodes' command option to CA.pl. This is like
5465 '-newreq', but calls 'openssl req' with the '-nodes' option
5466 so that the resulting key is not encrypted.
5467 [Damien Miller <djm@mindrot.org>]
5469 *) New configuration for the GNU Hurd.
5470 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
5472 *) Initial code to implement OCSP basic response verify. This
5473 is currently incomplete. Currently just finds the signer's
5474 certificate and verifies the signature on the response.
5477 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
5478 value of OPENSSLDIR. This is available via the new '-d' option
5479 to 'openssl version', and is also included in 'openssl version -a'.
5482 *) Allowing defining memory allocation callbacks that will be given
5483 file name and line number information in additional arguments
5484 (a const char* and an int). The basic functionality remains, as
5485 well as the original possibility to just replace malloc(),
5486 realloc() and free() by functions that do not know about these
5487 additional arguments. To register and find out the current
5488 settings for extended allocation functions, the following
5489 functions are provided:
5491 CRYPTO_set_mem_ex_functions
5492 CRYPTO_set_locked_mem_ex_functions
5493 CRYPTO_get_mem_ex_functions
5494 CRYPTO_get_locked_mem_ex_functions
5496 These work the same way as CRYPTO_set_mem_functions and friends.
5497 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
5498 extended allocation function is enabled.
5499 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
5500 a conventional allocation function is enabled.
5501 [Richard Levitte, Bodo Moeller]
5503 *) Finish off removing the remaining LHASH function pointer casts.
5504 There should no longer be any prototype-casting required when using
5505 the LHASH abstraction, and any casts that remain are "bugs". See
5506 the callback types and macros at the head of lhash.h for details
5507 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
5510 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
5511 If /dev/[u]random devices are not available or do not return enough
5512 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
5514 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
5515 /etc/entropy will be queried once each in this sequence, quering stops
5516 when enough entropy was collected without querying more sockets.
5519 *) Change the Unix RAND_poll() variant to be able to poll several
5520 random devices, as specified by DEVRANDOM, until a sufficient amount
5521 of data has been collected. We spend at most 10 ms on each file
5522 (select timeout) and read in non-blocking mode. DEVRANDOM now
5523 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
5524 (previously it was just the string "/dev/urandom"), so on typical
5525 platforms the 10 ms delay will never occur.
5526 Also separate out the Unix variant to its own file, rand_unix.c.
5527 For VMS, there's a currently-empty rand_vms.c.
5530 *) Move OCSP client related routines to ocsp_cl.c. These
5531 provide utility functions which an application needing
5532 to issue a request to an OCSP responder and analyse the
5533 response will typically need: as opposed to those which an
5534 OCSP responder itself would need which will be added later.
5536 OCSP_request_sign() signs an OCSP request with an API similar
5537 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
5538 response. OCSP_response_get1_basic() extracts basic response
5539 from response. OCSP_resp_find_status(): finds and extracts status
5540 information from an OCSP_CERTID structure (which will be created
5541 when the request structure is built). These are built from lower
5542 level functions which work on OCSP_SINGLERESP structures but
5543 wont normally be used unless the application wishes to examine
5544 extensions in the OCSP response for example.
5546 Replace nonce routines with a pair of functions.
5547 OCSP_request_add1_nonce() adds a nonce value and optionally
5548 generates a random value. OCSP_check_nonce() checks the
5549 validity of the nonce in an OCSP response.
5552 *) Change function OCSP_request_add() to OCSP_request_add0_id().
5553 This doesn't copy the supplied OCSP_CERTID and avoids the
5554 need to free up the newly created id. Change return type
5555 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
5556 This can then be used to add extensions to the request.
5557 Deleted OCSP_request_new(), since most of its functionality
5558 is now in OCSP_REQUEST_new() (and the case insensitive name
5559 clash) apart from the ability to set the request name which
5560 will be added elsewhere.
5563 *) Update OCSP API. Remove obsolete extensions argument from
5564 various functions. Extensions are now handled using the new
5565 OCSP extension code. New simple OCSP HTTP function which
5566 can be used to send requests and parse the response.
5569 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
5570 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
5571 uses the special reorder version of SET OF to sort the attributes
5572 and reorder them to match the encoded order. This resolves a long
5573 standing problem: a verify on a PKCS7 structure just after signing
5574 it used to fail because the attribute order did not match the
5575 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
5576 it uses the received order. This is necessary to tolerate some broken
5577 software that does not order SET OF. This is handled by encoding
5578 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
5579 to produce the required SET OF.
5582 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
5583 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
5584 files to get correct declarations of the ASN.1 item variables.
5587 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
5588 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
5589 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
5590 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
5591 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
5592 ASN1_ITEM and no wrapper functions.
5595 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
5596 replace the old function pointer based I/O routines. Change most of
5597 the *_d2i_bio() and *_d2i_fp() functions to use these.
5600 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
5601 lines, recognice more "algorithms" that can be deselected, and make
5602 it complain about algorithm deselection that isn't recognised.
5605 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
5606 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
5607 to use new functions. Add NO_ASN1_OLD which can be set to remove
5608 some old style ASN1 functions: this can be used to determine if old
5609 code will still work when these eventually go away.
5612 *) New extension functions for OCSP structures, these follow the
5613 same conventions as certificates and CRLs.
5616 *) New function X509V3_add1_i2d(). This automatically encodes and
5617 adds an extension. Its behaviour can be customised with various
5618 flags to append, replace or delete. Various wrappers added for
5619 certifcates and CRLs.
5622 *) Fix to avoid calling the underlying ASN1 print routine when
5623 an extension cannot be parsed. Correct a typo in the
5624 OCSP_SERVICELOC extension. Tidy up print OCSP format.
5627 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
5628 entries for variables.
5631 *) Add functionality to apps/openssl.c for detecting locking
5632 problems: As the program is single-threaded, all we have
5633 to do is register a locking callback using an array for
5634 storing which locks are currently held by the program.
5637 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
5638 SSL_get_ex_data_X509_STORE_idx(), which is used in
5639 ssl_verify_cert_chain() and thus can be called at any time
5640 during TLS/SSL handshakes so that thread-safety is essential.
5641 Unfortunately, the ex_data design is not at all suited
5642 for multi-threaded use, so it probably should be abolished.
5645 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5646 [Broadcom, tweaked and integrated by Geoff Thorpe]
5648 *) Move common extension printing code to new function
5649 X509V3_print_extensions(). Reorganise OCSP print routines and
5650 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5653 *) New function X509_signature_print() to remove duplication in some
5657 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5658 set (this was treated exactly the same as SET OF previously). This
5659 is used to reorder the STACK representing the structure to match the
5660 encoding. This will be used to get round a problem where a PKCS7
5661 structure which was signed could not be verified because the STACK
5662 order did not reflect the encoded order.
5665 *) Reimplement the OCSP ASN1 module using the new code.
5668 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5669 for its ASN1 operations. The old style function pointers still exist
5670 for now but they will eventually go away.
5673 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
5674 completely replaces the old ASN1 functionality with a table driven
5675 encoder and decoder which interprets an ASN1_ITEM structure describing
5676 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
5677 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
5678 has also been converted to the new form.
5681 *) Change BN_mod_exp_recp so that negative moduli are tolerated
5682 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
5683 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
5684 for negative moduli.
5687 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
5688 of not touching the result's sign bit.
5691 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
5695 *) Changed the LHASH code to use prototypes for callbacks, and created
5696 macros to declare and implement thin (optionally static) functions
5697 that provide type-safety and avoid function pointer casting for the
5698 type-specific callbacks.
5701 *) Added Kerberos Cipher Suites to be used with TLS, as written in
5703 [Veers Staats <staatsvr@asc.hpc.mil>,
5704 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
5706 *) Reformat the FAQ so the different questions and answers can be divided
5707 in sections depending on the subject.
5710 *) Have the zlib compression code load ZLIB.DLL dynamically under
5714 *) New function BN_mod_sqrt for computing square roots modulo a prime
5715 (using the probabilistic Tonelli-Shanks algorithm unless
5716 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
5717 be handled deterministically).
5718 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5720 *) Make BN_mod_inverse faster by explicitly handling small quotients
5721 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
5722 512 bits], about 30% for larger ones [1024 or 2048 bits].)
5725 *) New function BN_kronecker.
5728 *) Fix BN_gcd so that it works on negative inputs; the result is
5729 positive unless both parameters are zero.
5730 Previously something reasonably close to an infinite loop was
5731 possible because numbers could be growing instead of shrinking
5732 in the implementation of Euclid's algorithm.
5735 *) Fix BN_is_word() and BN_is_one() macros to take into account the
5736 sign of the number in question.
5738 Fix BN_is_word(a,w) to work correctly for w == 0.
5740 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
5741 because its test if the absolute value of 'a' equals 'w'.
5742 Note that BN_abs_is_word does *not* handle w == 0 reliably;
5743 it exists mostly for use in the implementations of BN_is_zero(),
5744 BN_is_one(), and BN_is_word().
5747 *) New function BN_swap.
5750 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
5751 the exponentiation functions are more likely to produce reasonable
5752 results on negative inputs.
5755 *) Change BN_mod_mul so that the result is always non-negative.
5756 Previously, it could be negative if one of the factors was negative;
5757 I don't think anyone really wanted that behaviour.
5760 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
5761 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
5762 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
5763 and add new functions:
5772 BN_mod_lshift1_quick
5776 These functions always generate non-negative results.
5778 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
5779 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
5781 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
5782 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
5783 be reduced modulo m.
5784 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5787 The following entry accidentily appeared in the CHANGES file
5788 distributed with OpenSSL 0.9.7. The modifications described in
5789 it do *not* apply to OpenSSL 0.9.7.
5791 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5792 was actually never needed) and in BN_mul(). The removal in BN_mul()
5793 required a small change in bn_mul_part_recursive() and the addition
5794 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5795 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5796 bn_sub_words() and bn_add_words() except they take arrays with
5801 *) In 'openssl passwd', verify passwords read from the terminal
5802 unless the '-salt' option is used (which usually means that
5803 verification would just waste user's time since the resulting
5804 hash is going to be compared with some given password hash)
5805 or the new '-noverify' option is used.
5807 This is an incompatible change, but it does not affect
5808 non-interactive use of 'openssl passwd' (passwords on the command
5809 line, '-stdin' option, '-in ...' option) and thus should not
5813 *) Remove all references to RSAref, since there's no more need for it.
5816 *) Make DSO load along a path given through an environment variable
5817 (SHLIB_PATH) with shl_load().
5820 *) Constify the ENGINE code as a result of BIGNUM constification.
5821 Also constify the RSA code and most things related to it. In a
5822 few places, most notable in the depth of the ASN.1 code, ugly
5823 casts back to non-const were required (to be solved at a later
5827 *) Make it so the openssl application has all engines loaded by default.
5830 *) Constify the BIGNUM routines a little more.
5833 *) Add the following functions:
5835 ENGINE_load_cswift()
5837 ENGINE_load_atalla()
5839 ENGINE_load_builtin_engines()
5841 That way, an application can itself choose if external engines that
5842 are built-in in OpenSSL shall ever be used or not. The benefit is
5843 that applications won't have to be linked with libdl or other dso
5844 libraries unless it's really needed.
5846 Changed 'openssl engine' to load all engines on demand.
5847 Changed the engine header files to avoid the duplication of some
5848 declarations (they differed!).
5851 *) 'openssl engine' can now list capabilities.
5854 *) Better error reporting in 'openssl engine'.
5857 *) Never call load_dh_param(NULL) in s_server.
5860 *) Add engine application. It can currently list engines by name and
5861 identity, and test if they are actually available.
5864 *) Improve RPM specification file by forcing symbolic linking and making
5865 sure the installed documentation is also owned by root.root.
5866 [Damien Miller <djm@mindrot.org>]
5868 *) Give the OpenSSL applications more possibilities to make use of
5869 keys (public as well as private) handled by engines.
5872 *) Add OCSP code that comes from CertCo.
5875 *) Add VMS support for the Rijndael code.
5878 *) Added untested support for Nuron crypto accelerator.
5881 *) Add support for external cryptographic devices. This code was
5882 previously distributed separately as the "engine" branch.
5883 [Geoff Thorpe, Richard Levitte]
5885 *) Rework the filename-translation in the DSO code. It is now possible to
5886 have far greater control over how a "name" is turned into a filename
5887 depending on the operating environment and any oddities about the
5888 different shared library filenames on each system.
5891 *) Support threads on FreeBSD-elf in Configure.
5894 *) Fix for SHA1 assembly problem with MASM: it produces
5895 warnings about corrupt line number information when assembling
5896 with debugging information. This is caused by the overlapping
5898 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
5901 NCONF_get_number() has no error checking at all. As a replacement,
5902 NCONF_get_number_e() is defined (_e for "error checking") and is
5903 promoted strongly. The old NCONF_get_number is kept around for
5904 binary backward compatibility.
5905 Make it possible for methods to load from something other than a BIO,
5906 by providing a function pointer that is given a name instead of a BIO.
5907 For example, this could be used to load configuration data from an
5911 *) Fix for non blocking accept BIOs. Added new I/O special reason
5912 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
5913 with non blocking I/O was not possible because no retry code was
5914 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
5918 *) Added the beginnings of Rijndael support.
5921 *) Fix for bug in DirectoryString mask setting. Add support for
5922 X509_NAME_print_ex() in 'req' and X509_print_ex() function
5923 to allow certificate printing to more controllable, additional
5924 'certopt' option to 'x509' to allow new printing options to be
5928 *) Clean old EAY MD5 hack from e_os.h.
5931 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
5933 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5934 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5935 [Joe Orton, Steve Henson]
5937 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
5939 *) Fix additional bug revealed by the NISCC test suite:
5941 Stop bug triggering large recursion when presented with
5942 certain ASN.1 tags (CVE-2003-0851)
5945 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
5947 *) Fix various bugs revealed by running the NISCC test suite:
5949 Stop out of bounds reads in the ASN1 code when presented with
5950 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5952 If verify callback ignores invalid public key errors don't try to check
5953 certificate signature with the NULL public key.
5957 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5958 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5962 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5963 extra data after the compression methods not only for TLS 1.0
5964 but also for SSL 3.0 (as required by the specification).
5965 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5967 *) Change X509_certificate_type() to mark the key as exported/exportable
5968 when it's 512 *bits* long, not 512 bytes.
5971 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
5973 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5974 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5975 a protocol version number mismatch like a decryption error
5976 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5979 *) Turn on RSA blinding by default in the default implementation
5980 to avoid a timing attack. Applications that don't want it can call
5981 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5982 They would be ill-advised to do so in most cases.
5983 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5985 *) Change RSA blinding code so that it works when the PRNG is not
5986 seeded (in this case, the secret RSA exponent is abused as
5987 an unpredictable seed -- if it is not unpredictable, there
5988 is no point in blinding anyway). Make RSA blinding thread-safe
5989 by remembering the creator's thread ID in rsa->blinding and
5990 having all other threads use local one-time blinding factors
5991 (this requires more computation than sharing rsa->blinding, but
5992 avoids excessive locking; and if an RSA object is not shared
5993 between threads, blinding will still be very fast).
5996 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
5998 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5999 via timing by performing a MAC computation even if incorrrect
6000 block cipher padding has been found. This is a countermeasure
6001 against active attacks where the attacker has to distinguish
6002 between bad padding and a MAC verification error. (CVE-2003-0078)
6004 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6005 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6006 Martin Vuagnoux (EPFL, Ilion)]
6008 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
6010 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
6011 memory from it's contents. This is done with a counter that will
6012 place alternating values in each byte. This can be used to solve
6013 two issues: 1) the removal of calls to memset() by highly optimizing
6014 compilers, and 2) cleansing with other values than 0, since those can
6015 be read through on certain media, for example a swap space on disk.
6018 *) Bugfix: client side session caching did not work with external caching,
6019 because the session->cipher setting was not restored when reloading
6020 from the external cache. This problem was masked, when
6021 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
6022 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
6025 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
6026 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
6027 [Zeev Lieber <zeev-l@yahoo.com>]
6029 *) Undo an undocumented change introduced in 0.9.6e which caused
6030 repeated calls to OpenSSL_add_all_ciphers() and
6031 OpenSSL_add_all_digests() to be ignored, even after calling
6035 *) Change the default configuration reader to deal with last line not
6036 being properly terminated.
6039 *) Change X509_NAME_cmp() so it applies the special rules on handling
6040 DN values that are of type PrintableString, as well as RDNs of type
6041 emailAddress where the value has the type ia5String.
6042 [stefank@valicert.com via Richard Levitte]
6044 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
6045 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
6046 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
6047 the bitwise-OR of the two for use by the majority of applications
6048 wanting this behaviour, and update the docs. The documented
6049 behaviour and actual behaviour were inconsistent and had been
6050 changing anyway, so this is more a bug-fix than a behavioural
6052 [Geoff Thorpe, diagnosed by Nadav Har'El]
6054 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
6055 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
6058 *) Fix initialization code race conditions in
6059 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
6060 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
6061 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
6062 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
6063 ssl2_get_cipher_by_char(),
6064 ssl3_get_cipher_by_char().
6065 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
6067 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
6068 the cached sessions are flushed, as the remove_cb() might use ex_data
6069 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
6070 (see [openssl.org #212]).
6071 [Geoff Thorpe, Lutz Jaenicke]
6073 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
6074 length, instead of the encoding length to d2i_ASN1_OBJECT.
6077 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
6079 *) [In 0.9.6g-engine release:]
6080 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
6081 [Lynn Gazis <lgazis@rainbow.com>]
6083 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
6085 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
6086 and get fix the header length calculation.
6087 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
6088 Alon Kantor <alonk@checkpoint.com> (and others),
6091 *) Use proper error handling instead of 'assertions' in buffer
6092 overflow checks added in 0.9.6e. This prevents DoS (the
6093 assertions could call abort()).
6094 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
6096 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
6098 *) Add various sanity checks to asn1_get_length() to reject
6099 the ASN1 length bytes if they exceed sizeof(long), will appear
6100 negative or the content length exceeds the length of the
6102 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6104 *) Fix cipher selection routines: ciphers without encryption had no flags
6105 for the cipher strength set and where therefore not handled correctly
6106 by the selection routines (PR #130).
6109 *) Fix EVP_dsa_sha macro.
6113 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
6114 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
6115 that was added in OpenSSL 0.9.6d.
6117 As the countermeasure turned out to be incompatible with some
6118 broken SSL implementations, the new option is part of SSL_OP_ALL.
6119 SSL_OP_ALL is usually employed when compatibility with weird SSL
6120 implementations is desired (e.g. '-bugs' option to 's_client' and
6121 's_server'), so the new option is automatically set in many
6125 *) Changes in security patch:
6127 Changes marked "(CHATS)" were sponsored by the Defense Advanced
6128 Research Projects Agency (DARPA) and Air Force Research Laboratory,
6129 Air Force Materiel Command, USAF, under agreement number
6132 *) Add various sanity checks to asn1_get_length() to reject
6133 the ASN1 length bytes if they exceed sizeof(long), will appear
6134 negative or the content length exceeds the length of the
6135 supplied buffer. (CVE-2002-0659)
6136 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6138 *) Assertions for various potential buffer overflows, not known to
6140 [Ben Laurie (CHATS)]
6142 *) Various temporary buffers to hold ASCII versions of integers were
6143 too small for 64 bit platforms. (CVE-2002-0655)
6144 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
6146 *) Remote buffer overflow in SSL3 protocol - an attacker could
6147 supply an oversized session ID to a client. (CVE-2002-0656)
6148 [Ben Laurie (CHATS)]
6150 *) Remote buffer overflow in SSL2 protocol - an attacker could
6151 supply an oversized client master key. (CVE-2002-0656)
6152 [Ben Laurie (CHATS)]
6154 Changes between 0.9.6c and 0.9.6d [9 May 2002]
6156 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
6157 encoded as NULL) with id-dsa-with-sha1.
6158 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
6160 *) Check various X509_...() return values in apps/req.c.
6161 [Nils Larsch <nla@trustcenter.de>]
6163 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
6164 an end-of-file condition would erronously be flagged, when the CRLF
6165 was just at the end of a processed block. The bug was discovered when
6166 processing data through a buffering memory BIO handing the data to a
6167 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
6168 <ptsekov@syntrex.com> and Nedelcho Stanev.
6171 *) Implement a countermeasure against a vulnerability recently found
6172 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
6173 before application data chunks to avoid the use of known IVs
6174 with data potentially chosen by the attacker.
6177 *) Fix length checks in ssl3_get_client_hello().
6180 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
6181 to prevent ssl3_read_internal() from incorrectly assuming that
6182 ssl3_read_bytes() found application data while handshake
6183 processing was enabled when in fact s->s3->in_read_app_data was
6184 merely automatically cleared during the initial handshake.
6185 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
6187 *) Fix object definitions for Private and Enterprise: they were not
6188 recognized in their shortname (=lowercase) representation. Extend
6189 obj_dat.pl to issue an error when using undefined keywords instead
6190 of silently ignoring the problem (Svenning Sorensen
6191 <sss@sss.dnsalias.net>).
6194 *) Fix DH_generate_parameters() so that it works for 'non-standard'
6195 generators, i.e. generators other than 2 and 5. (Previously, the
6196 code did not properly initialise the 'add' and 'rem' values to
6197 BN_generate_prime().)
6199 In the new general case, we do not insist that 'generator' is
6200 actually a primitive root: This requirement is rather pointless;
6201 a generator of the order-q subgroup is just as good, if not
6205 *) Map new X509 verification errors to alerts. Discovered and submitted by
6206 Tom Wu <tom@arcot.com>.
6209 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
6210 returning non-zero before the data has been completely received
6211 when using non-blocking I/O.
6212 [Bodo Moeller; problem pointed out by John Hughes]
6214 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
6215 [Ben Laurie, Lutz Jaenicke]
6217 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
6218 Yoram Zahavi <YoramZ@gilian.com>).
6221 *) Add information about CygWin 1.3 and on, and preserve proper
6222 configuration for the versions before that.
6223 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6225 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
6226 check whether we deal with a copy of a session and do not delete from
6227 the cache in this case. Problem reported by "Izhar Shoshani Levi"
6228 <izhar@checkpoint.com>.
6231 *) Do not store session data into the internal session cache, if it
6232 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
6233 flag is set). Proposed by Aslam <aslam@funk.com>.
6236 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
6240 *) [In 0.9.6d-engine release:]
6241 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6242 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6244 *) Add the configuration target linux-s390x.
6245 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
6247 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
6248 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
6249 variable as an indication that a ClientHello message has been
6250 received. As the flag value will be lost between multiple
6251 invocations of ssl3_accept when using non-blocking I/O, the
6252 function may not be aware that a handshake has actually taken
6253 place, thus preventing a new session from being added to the
6256 To avoid this problem, we now set s->new_session to 2 instead of
6257 using a local variable.
6258 [Lutz Jaenicke, Bodo Moeller]
6260 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
6261 if the SSL_R_LENGTH_MISMATCH error is detected.
6262 [Geoff Thorpe, Bodo Moeller]
6264 *) New 'shared_ldflag' column in Configure platform table.
6267 *) Fix EVP_CIPHER_mode macro.
6268 ["Dan S. Camper" <dan@bti.net>]
6270 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
6271 type, we must throw them away by setting rr->length to 0.
6272 [D P Chang <dpc@qualys.com>]
6274 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
6276 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
6277 <Dominikus.Scherkl@biodata.com>. (The previous implementation
6278 worked incorrectly for those cases where range = 10..._2 and
6279 3*range is two bits longer than range.)
6282 *) Only add signing time to PKCS7 structures if it is not already
6286 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
6287 OBJ_ld_ce should be OBJ_id_ce.
6288 Also some ip-pda OIDs in crypto/objects/objects.txt were
6289 incorrect (cf. RFC 3039).
6290 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
6292 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
6293 returns early because it has nothing to do.
6294 [Andy Schneider <andy.schneider@bjss.co.uk>]
6296 *) [In 0.9.6c-engine release:]
6297 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
6298 [Andy Schneider <andy.schneider@bjss.co.uk>]
6300 *) [In 0.9.6c-engine release:]
6301 Add support for Cryptographic Appliance's keyserver technology.
6302 (Use engine 'keyclient')
6303 [Cryptographic Appliances and Geoff Thorpe]
6305 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
6306 is called via tools/c89.sh because arguments have to be
6307 rearranged (all '-L' options must appear before the first object
6309 [Richard Shapiro <rshapiro@abinitio.com>]
6311 *) [In 0.9.6c-engine release:]
6312 Add support for Broadcom crypto accelerator cards, backported
6314 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
6316 *) [In 0.9.6c-engine release:]
6317 Add support for SureWare crypto accelerator cards from
6318 Baltimore Technologies. (Use engine 'sureware')
6319 [Baltimore Technologies and Mark Cox]
6321 *) [In 0.9.6c-engine release:]
6322 Add support for crypto accelerator cards from Accelerated
6323 Encryption Processing, www.aep.ie. (Use engine 'aep')
6324 [AEP Inc. and Mark Cox]
6326 *) Add a configuration entry for gcc on UnixWare.
6327 [Gary Benson <gbenson@redhat.com>]
6329 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
6330 messages are stored in a single piece (fixed-length part and
6331 variable-length part combined) and fix various bugs found on the way.
6334 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
6335 instead. BIO_gethostbyname() does not know what timeouts are
6336 appropriate, so entries would stay in cache even when they have
6338 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
6340 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
6341 faced with a pathologically small ClientHello fragment that does
6342 not contain client_version: Instead of aborting with an error,
6343 simply choose the highest available protocol version (i.e.,
6344 TLS 1.0 unless it is disabled). In practice, ClientHello
6345 messages are never sent like this, but this change gives us
6346 strictly correct behaviour at least for TLS.
6349 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
6350 never resets s->method to s->ctx->method when called from within
6351 one of the SSL handshake functions.
6352 [Bodo Moeller; problem pointed out by Niko Baric]
6354 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
6355 (sent using the client's version number) if client_version is
6356 smaller than the protocol version in use. Also change
6357 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
6358 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
6359 the client will at least see that alert.
6362 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
6366 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
6367 client receives HelloRequest while in a handshake.
6368 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
6370 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
6371 should end in 'break', not 'goto end' which circuments various
6372 cleanups done in state SSL_ST_OK. But session related stuff
6373 must be disabled for SSL_ST_OK in the case that we just sent a
6376 Also avoid some overhead by not calling ssl_init_wbio_buffer()
6377 before just sending a HelloRequest.
6378 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
6380 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
6381 reveal whether illegal block cipher padding was found or a MAC
6382 verification error occured. (Neither SSLerr() codes nor alerts
6383 are directly visible to potential attackers, but the information
6384 may leak via logfiles.)
6386 Similar changes are not required for the SSL 2.0 implementation
6387 because the number of padding bytes is sent in clear for SSL 2.0,
6388 and the extra bytes are just ignored. However ssl/s2_pkt.c
6389 failed to verify that the purported number of padding bytes is in
6393 *) Add OpenUNIX-8 support including shared libraries
6394 (Boyd Lynn Gerber <gerberb@zenez.com>).
6397 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
6398 'wristwatch attack' using huge encoding parameters (cf.
6399 James H. Manger's CRYPTO 2001 paper). Note that the
6400 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
6401 encoding parameters and hence was not vulnerable.
6404 *) BN_sqr() bug fix.
6405 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
6407 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
6408 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
6409 followed by modular reduction.
6410 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
6412 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
6413 equivalent based on BN_pseudo_rand() instead of BN_rand().
6416 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
6417 This function was broken, as the check for a new client hello message
6418 to handle SGC did not allow these large messages.
6419 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
6422 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
6425 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
6426 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
6429 *) Rework the configuration and shared library support for Tru64 Unix.
6430 The configuration part makes use of modern compiler features and
6431 still retains old compiler behavior for those that run older versions
6432 of the OS. The shared library support part includes a variant that
6433 uses the RPATH feature, and is available through the special
6434 configuration target "alpha-cc-rpath", which will never be selected
6436 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
6438 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
6439 with the same message size as in ssl3_get_certificate_request().
6440 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
6441 messages might inadvertently be reject as too long.
6442 [Petr Lampa <lampa@fee.vutbr.cz>]
6444 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
6447 *) Modified SSL library such that the verify_callback that has been set
6448 specificly for an SSL object with SSL_set_verify() is actually being
6449 used. Before the change, a verify_callback set with this function was
6450 ignored and the verify_callback() set in the SSL_CTX at the time of
6451 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
6452 to allow the necessary settings.
6455 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
6456 explicitly to NULL, as at least on Solaris 8 this seems not always to be
6457 done automatically (in contradiction to the requirements of the C
6458 standard). This made problems when used from OpenSSH.
6461 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
6462 dh->length and always used
6464 BN_rand_range(priv_key, dh->p).
6466 BN_rand_range() is not necessary for Diffie-Hellman, and this
6467 specific range makes Diffie-Hellman unnecessarily inefficient if
6468 dh->length (recommended exponent length) is much smaller than the
6469 length of dh->p. We could use BN_rand_range() if the order of
6470 the subgroup was stored in the DH structure, but we only have
6475 BN_rand(priv_key, l, ...)
6477 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
6483 RSA_eay_public_encrypt
6484 RSA_eay_private_decrypt
6485 RSA_eay_private_encrypt (signing)
6486 RSA_eay_public_decrypt (signature verification)
6488 (default implementations for RSA_public_encrypt,
6489 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
6490 always reject numbers >= n.
6493 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
6494 to synchronize access to 'locking_thread'. This is necessary on
6495 systems where access to 'locking_thread' (an 'unsigned long'
6496 variable) is not atomic.
6499 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
6500 *before* setting the 'crypto_lock_rand' flag. The previous code had
6501 a race condition if 0 is a valid thread ID.
6502 [Travis Vitek <vitek@roguewave.com>]
6504 *) Add support for shared libraries under Irix.
6505 [Albert Chin-A-Young <china@thewrittenword.com>]
6507 *) Add configuration option to build on Linux on both big-endian and
6509 [Ralf Baechle <ralf@uni-koblenz.de>]
6511 *) Add the possibility to create shared libraries on HP-UX.
6514 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
6516 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
6517 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
6518 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
6519 PRNG state recovery was possible based on the output of
6520 one PRNG request appropriately sized to gain knowledge on
6521 'md' followed by enough consecutive 1-byte PRNG requests
6522 to traverse all of 'state'.
6524 1. When updating 'md_local' (the current thread's copy of 'md')
6525 during PRNG output generation, hash all of the previous
6526 'md_local' value, not just the half used for PRNG output.
6528 2. Make the number of bytes from 'state' included into the hash
6529 independent from the number of PRNG bytes requested.
6531 The first measure alone would be sufficient to avoid
6532 Markku-Juhani's attack. (Actually it had never occurred
6533 to me that the half of 'md_local' used for chaining was the
6534 half from which PRNG output bytes were taken -- I had always
6535 assumed that the secret half would be used.) The second
6536 measure makes sure that additional data from 'state' is never
6537 mixed into 'md_local' in small portions; this heuristically
6538 further strengthens the PRNG.
6541 *) Fix crypto/bn/asm/mips3.s.
6544 *) When only the key is given to "enc", the IV is undefined. Print out
6545 an error message in this case.
6548 *) Handle special case when X509_NAME is empty in X509 printing routines.
6551 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
6552 positive and less than q.
6555 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
6556 used: it isn't thread safe and the add_lock_callback should handle
6558 [Paul Rose <Paul.Rose@bridge.com>]
6560 *) Verify that incoming data obeys the block size in
6561 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
6565 [Ulf Möller, Bodo Möller]
6567 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
6568 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
6569 when fixing the server behaviour for backwards-compatible 'client
6570 hello' messages. (Note that the attack is impractical against
6571 SSL 3.0 and TLS 1.0 anyway because length and version checking
6572 means that the probability of guessing a valid ciphertext is
6573 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
6576 Before 0.9.5, the countermeasure (hide the error by generating a
6577 random 'decryption result') did not work properly because
6578 ERR_clear_error() was missing, meaning that SSL_get_error() would
6579 detect the supposedly ignored error.
6581 Both problems are now fixed.
6584 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
6585 (previously it was 1024).
6588 *) Fix for compatibility mode trust settings: ignore trust settings
6589 unless some valid trust or reject settings are present.
6592 *) Fix for blowfish EVP: its a variable length cipher.
6595 *) Fix various bugs related to DSA S/MIME verification. Handle missing
6596 parameters in DSA public key structures and return an error in the
6597 DSA routines if parameters are absent.
6600 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
6601 in the current directory if neither $RANDFILE nor $HOME was set.
6602 RAND_file_name() in 0.9.6a returned NULL in this case. This has
6603 caused some confusion to Windows users who haven't defined $HOME.
6604 Thus RAND_file_name() is changed again: e_os.h can define a
6605 DEFAULT_HOME, which will be used if $HOME is not set.
6606 For Windows, we use "C:"; on other platforms, we still require
6607 environment variables.
6609 *) Move 'if (!initialized) RAND_poll()' into regions protected by
6610 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
6611 having multiple threads call RAND_poll() concurrently.
6614 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
6615 combination of a flag and a thread ID variable.
6616 Otherwise while one thread is in ssleay_rand_bytes (which sets the
6617 flag), *other* threads can enter ssleay_add_bytes without obeying
6618 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
6619 that they do not hold after the first thread unsets add_do_not_lock).
6622 *) Change bctest again: '-x' expressions are not available in all
6626 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
6628 *) Fix a couple of memory leaks in PKCS7_dataDecode()
6629 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
6631 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
6632 the default extension for executables, if any. Also, make the perl
6633 scripts that use symlink() to test if it really exists and use "cp"
6634 if it doesn't. All this made OpenSSL compilable and installable in
6638 *) Fix for asn1_GetSequence() for indefinite length constructed data.
6639 If SEQUENCE is length is indefinite just set c->slen to the total
6640 amount of data available.
6641 [Steve Henson, reported by shige@FreeBSD.org]
6642 [This change does not apply to 0.9.7.]
6644 *) Change bctest to avoid here-documents inside command substitution
6645 (workaround for FreeBSD /bin/sh bug).
6646 For compatibility with Ultrix, avoid shell functions (introduced
6647 in the bctest version that searches along $PATH).
6650 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6651 with des_encrypt() defined on some operating systems, like Solaris
6655 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6656 On the Importance of Eliminating Errors in Cryptographic
6657 Computations, J. Cryptology 14 (2001) 2, 101-119,
6658 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6661 *) MIPS assembler BIGNUM division bug fix.
6664 *) Disabled incorrect Alpha assembler code.
6667 *) Fix PKCS#7 decode routines so they correctly update the length
6668 after reading an EOC for the EXPLICIT tag.
6670 [This change does not apply to 0.9.7.]
6672 *) Fix bug in PKCS#12 key generation routines. This was triggered
6673 if a 3DES key was generated with a 0 initial byte. Include
6674 PKCS12_BROKEN_KEYGEN compilation option to retain the old
6675 (but broken) behaviour.
6678 *) Enhance bctest to search for a working bc along $PATH and print
6680 [Tim Rice <tim@multitalents.net> via Richard Levitte]
6682 *) Fix memory leaks in err.c: free err_data string if necessary;
6683 don't write to the wrong index in ERR_set_error_data.
6686 *) Implement ssl23_peek (analogous to ssl23_read), which previously
6690 *) Replace rdtsc with _emit statements for VC++ version 5.
6691 [Jeremy Cooper <jeremy@baymoo.org>]
6693 *) Make it possible to reuse SSLv2 sessions.
6696 *) In copy_email() check for >= 0 as a return value for
6697 X509_NAME_get_index_by_NID() since 0 is a valid index.
6698 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
6700 *) Avoid coredump with unsupported or invalid public keys by checking if
6701 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
6702 PKCS7_verify() fails with non detached data.
6705 *) Don't use getenv in library functions when run as setuid/setgid.
6706 New function OPENSSL_issetugid().
6709 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
6710 due to incorrect handling of multi-threading:
6712 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
6714 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
6716 3. Count how many times MemCheck_off() has been called so that
6717 nested use can be treated correctly. This also avoids
6718 inband-signalling in the previous code (which relied on the
6719 assumption that thread ID 0 is impossible).
6722 *) Add "-rand" option also to s_client and s_server.
6725 *) Fix CPU detection on Irix 6.x.
6726 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
6727 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6729 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
6732 [This change does not apply to 0.9.7.]
6734 *) Use the cached encoding of an X509_NAME structure rather than
6735 copying it. This is apparently the reason for the libsafe "errors"
6736 but the code is actually correct.
6739 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
6740 Bleichenbacher's DSA attack.
6741 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
6742 to be set and top=0 forces the highest bit to be set; top=-1 is new
6743 and leaves the highest bit random.
6744 [Ulf Moeller, Bodo Moeller]
6746 *) In the NCONF_...-based implementations for CONF_... queries
6747 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
6748 a temporary CONF structure with the data component set to NULL
6749 (which gives segmentation faults in lh_retrieve).
6750 Instead, use NULL for the CONF pointer in CONF_get_string and
6751 CONF_get_number (which may use environment variables) and directly
6752 return NULL from CONF_get_section.
6755 *) Fix potential buffer overrun for EBCDIC.
6758 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
6759 keyUsage if basicConstraints absent for a CA.
6762 *) Make SMIME_write_PKCS7() write mail header values with a format that
6763 is more generally accepted (no spaces before the semicolon), since
6764 some programs can't parse those values properly otherwise. Also make
6765 sure BIO's that break lines after each write do not create invalid
6769 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
6770 macros previously used would not encode an empty SEQUENCE OF
6771 and break the signature.
6773 [This change does not apply to 0.9.7.]
6775 *) Zero the premaster secret after deriving the master secret in
6779 *) Add some EVP_add_digest_alias registrations (as found in
6780 OpenSSL_add_all_digests()) to SSL_library_init()
6781 aka OpenSSL_add_ssl_algorithms(). This provides improved
6782 compatibility with peers using X.509 certificates
6783 with unconventional AlgorithmIdentifier OIDs.
6786 *) Fix for Irix with NO_ASM.
6787 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6789 *) ./config script fixes.
6790 [Ulf Moeller, Richard Levitte]
6792 *) Fix 'openssl passwd -1'.
6795 *) Change PKCS12_key_gen_asc() so it can cope with non null
6796 terminated strings whose length is passed in the passlen
6797 parameter, for example from PEM callbacks. This was done
6798 by adding an extra length parameter to asc2uni().
6799 [Steve Henson, reported by <oddissey@samsung.co.kr>]
6801 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
6802 call failed, free the DSA structure.
6805 *) Fix to uni2asc() to cope with zero length Unicode strings.
6806 These are present in some PKCS#12 files.
6809 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
6810 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
6811 when writing a 32767 byte record.
6812 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
6814 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
6815 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
6817 (RSA objects have a reference count access to which is protected
6818 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
6819 so they are meant to be shared between threads.)
6820 [Bodo Moeller, Geoff Thorpe; original patch submitted by
6821 "Reddie, Steven" <Steven.Reddie@ca.com>]
6823 *) Fix a deadlock in CRYPTO_mem_leaks().
6826 *) Use better test patterns in bntest.
6829 *) rand_win.c fix for Borland C.
6832 *) BN_rshift bugfix for n == 0.
6835 *) Add a 'bctest' script that checks for some known 'bc' bugs
6836 so that 'make test' does not abort just because 'bc' is broken.
6839 *) Store verify_result within SSL_SESSION also for client side to
6840 avoid potential security hole. (Re-used sessions on the client side
6841 always resulted in verify_result==X509_V_OK, not using the original
6842 result of the server certificate verification.)
6845 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
6846 SSL3_RT_APPLICATION_DATA, return 0.
6847 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
6851 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
6852 releases, have been re-implemented by renaming the previous
6853 implementations of ssl2_read and ssl3_read to ssl2_read_internal
6854 and ssl3_read_internal, respectively, and adding 'peek' parameters
6855 to them. The new ssl[23]_{read,peek} functions are calls to
6856 ssl[23]_read_internal with the 'peek' flag set appropriately.
6857 A 'peek' parameter has also been added to ssl3_read_bytes, which
6858 does the actual work for ssl3_read_internal.
6861 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
6862 the method-specific "init()" handler. Also clean up ex_data after
6863 calling the method-specific "finish()" handler. Previously, this was
6864 happening the other way round.
6867 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
6868 The previous value, 12, was not always sufficient for BN_mod_exp().
6871 *) Make sure that shared libraries get the internal name engine with
6872 the full version number and not just 0. This should mark the
6873 shared libraries as not backward compatible. Of course, this should
6874 be changed again when we can guarantee backward binary compatibility.
6877 *) Fix typo in get_cert_by_subject() in by_dir.c
6878 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
6880 *) Rework the system to generate shared libraries:
6882 - Make note of the expected extension for the shared libraries and
6883 if there is a need for symbolic links from for example libcrypto.so.0
6884 to libcrypto.so.0.9.7. There is extended info in Configure for
6887 - Make as few rebuilds of the shared libraries as possible.
6889 - Still avoid linking the OpenSSL programs with the shared libraries.
6891 - When installing, install the shared libraries separately from the
6895 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
6897 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
6898 and not in SSL_clear because the latter is also used by the
6899 accept/connect functions; previously, the settings made by
6900 SSL_set_read_ahead would be lost during the handshake.
6901 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
6903 *) Correct util/mkdef.pl to be selective about disabled algorithms.
6904 Previously, it would create entries for disableed algorithms no
6908 *) Added several new manual pages for SSL_* function.
6911 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
6913 *) In ssl23_get_client_hello, generate an error message when faced
6914 with an initial SSL 3.0/TLS record that is too small to contain the
6915 first two bytes of the ClientHello message, i.e. client_version.
6916 (Note that this is a pathologic case that probably has never happened
6917 in real life.) The previous approach was to use the version number
6918 from the record header as a substitute; but our protocol choice
6919 should not depend on that one because it is not authenticated
6920 by the Finished messages.
6923 *) More robust randomness gathering functions for Windows.
6924 [Jeffrey Altman <jaltman@columbia.edu>]
6926 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
6927 not set then we don't setup the error code for issuer check errors
6928 to avoid possibly overwriting other errors which the callback does
6929 handle. If an application does set the flag then we assume it knows
6930 what it is doing and can handle the new informational codes
6934 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
6935 a general "ANY" type, as such it should be able to decode anything
6936 including tagged types. However it didn't check the class so it would
6937 wrongly interpret tagged types in the same way as their universal
6938 counterpart and unknown types were just rejected. Changed so that the
6939 tagged and unknown types are handled in the same way as a SEQUENCE:
6940 that is the encoding is stored intact. There is also a new type
6941 "V_ASN1_OTHER" which is used when the class is not universal, in this
6942 case we have no idea what the actual type is so we just lump them all
6946 *) On VMS, stdout may very well lead to a file that is written to
6947 in a record-oriented fashion. That means that every write() will
6948 write a separate record, which will be read separately by the
6949 programs trying to read from it. This can be very confusing.
6951 The solution is to put a BIO filter in the way that will buffer
6952 text until a linefeed is reached, and then write everything a
6953 line at a time, so every record written will be an actual line,
6954 not chunks of lines and not (usually doesn't happen, but I've
6955 seen it once) several lines in one record. BIO_f_linebuffer() is
6958 Currently, it's a VMS-only method, because that's where it has
6959 been tested well enough.
6962 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
6963 it can return incorrect results.
6964 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
6965 but it was in 0.9.6-beta[12].)
6968 *) Disable the check for content being present when verifying detached
6969 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
6970 include zero length content when signing messages.
6973 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
6974 BIO_ctrl (for BIO pairs).
6977 *) Add DSO method for VMS.
6980 *) Bug fix: Montgomery multiplication could produce results with the
6984 *) Add RPM specification openssl.spec and modify it to build three
6985 packages. The default package contains applications, application
6986 documentation and run-time libraries. The devel package contains
6987 include files, static libraries and function documentation. The
6988 doc package contains the contents of the doc directory. The original
6989 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
6992 *) Add a large number of documentation files for many SSL routines.
6993 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6995 *) Add a configuration entry for Sony News 4.
6996 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
6998 *) Don't set the two most significant bits to one when generating a
6999 random number < q in the DSA library.
7002 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
7003 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
7004 the underlying transport is blocking) if a handshake took place.
7005 (The default behaviour is needed by applications such as s_client
7006 and s_server that use select() to determine when to use SSL_read;
7007 but for applications that know in advance when to expect data, it
7008 just makes things more complicated.)
7011 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
7015 *) Add a few more EBCDIC conditionals that make `req' and `x509'
7016 work better on such systems.
7017 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7019 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
7020 Update PKCS12_parse() so it copies the friendlyName and the
7021 keyid to the certificates aux info.
7024 *) Fix bug in PKCS7_verify() which caused an infinite loop
7025 if there was more than one signature.
7026 [Sven Uszpelkat <su@celocom.de>]
7028 *) Major change in util/mkdef.pl to include extra information
7029 about each symbol, as well as presentig variables as well
7030 as functions. This change means that there's n more need
7031 to rebuild the .num files when some algorithms are excluded.
7034 *) Allow the verify time to be set by an application,
7035 rather than always using the current time.
7038 *) Phase 2 verify code reorganisation. The certificate
7039 verify code now looks up an issuer certificate by a
7040 number of criteria: subject name, authority key id
7041 and key usage. It also verifies self signed certificates
7042 by the same criteria. The main comparison function is
7043 X509_check_issued() which performs these checks.
7045 Lot of changes were necessary in order to support this
7046 without completely rewriting the lookup code.
7048 Authority and subject key identifier are now cached.
7050 The LHASH 'certs' is X509_STORE has now been replaced
7051 by a STACK_OF(X509_OBJECT). This is mainly because an
7052 LHASH can't store or retrieve multiple objects with
7053 the same hash value.
7055 As a result various functions (which were all internal
7056 use only) have changed to handle the new X509_STORE
7057 structure. This will break anything that messed round
7058 with X509_STORE internally.
7060 The functions X509_STORE_add_cert() now checks for an
7061 exact match, rather than just subject name.
7063 The X509_STORE API doesn't directly support the retrieval
7064 of multiple certificates matching a given criteria, however
7065 this can be worked round by performing a lookup first
7066 (which will fill the cache with candidate certificates)
7067 and then examining the cache for matches. This is probably
7068 the best we can do without throwing out X509_LOOKUP
7069 entirely (maybe later...).
7071 The X509_VERIFY_CTX structure has been enhanced considerably.
7073 All certificate lookup operations now go via a get_issuer()
7074 callback. Although this currently uses an X509_STORE it
7075 can be replaced by custom lookups. This is a simple way
7076 to bypass the X509_STORE hackery necessary to make this
7077 work and makes it possible to use more efficient techniques
7078 in future. A very simple version which uses a simple
7079 STACK for its trusted certificate store is also provided
7080 using X509_STORE_CTX_trusted_stack().
7082 The verify_cb() and verify() callbacks now have equivalents
7083 in the X509_STORE_CTX structure.
7085 X509_STORE_CTX also has a 'flags' field which can be used
7086 to customise the verify behaviour.
7089 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
7090 excludes S/MIME capabilities.
7093 *) When a certificate request is read in keep a copy of the
7094 original encoding of the signed data and use it when outputing
7095 again. Signatures then use the original encoding rather than
7096 a decoded, encoded version which may cause problems if the
7097 request is improperly encoded.
7100 *) For consistency with other BIO_puts implementations, call
7101 buffer_write(b, ...) directly in buffer_puts instead of calling
7104 In BIO_puts, increment b->num_write as in BIO_write.
7105 [Peter.Sylvester@EdelWeb.fr]
7107 *) Fix BN_mul_word for the case where the word is 0. (We have to use
7108 BN_zero, we may not return a BIGNUM with an array consisting of
7112 *) Avoid calling abort() from within the library when problems are
7113 detected, except if preprocessor symbols have been defined
7114 (such as REF_CHECK, BN_DEBUG etc.).
7117 *) New openssl application 'rsautl'. This utility can be
7118 used for low level RSA operations. DER public key
7119 BIO/fp routines also added.
7122 *) New Configure entry and patches for compiling on QNX 4.
7123 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
7125 *) A demo state-machine implementation was sponsored by
7126 Nuron (http://www.nuron.com/) and is now available in
7127 demos/state_machine.
7130 *) New options added to the 'dgst' utility for signature
7131 generation and verification.
7134 *) Unrecognized PKCS#7 content types are now handled via a
7135 catch all ASN1_TYPE structure. This allows unsupported
7136 types to be stored as a "blob" and an application can
7137 encode and decode it manually.
7140 *) Fix various signed/unsigned issues to make a_strex.c
7142 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
7144 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
7145 length if passed a buffer. ASN1_INTEGER_to_BN failed
7146 if passed a NULL BN and its argument was negative.
7147 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
7149 *) Modification to PKCS#7 encoding routines to output definite
7150 length encoding. Since currently the whole structures are in
7151 memory there's not real point in using indefinite length
7152 constructed encoding. However if OpenSSL is compiled with
7153 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
7156 *) Added BIO_vprintf() and BIO_vsnprintf().
7159 *) Added more prefixes to parse for in the the strings written
7160 through a logging bio, to cover all the levels that are available
7161 through syslog. The prefixes are now:
7163 PANIC, EMERG, EMR => LOG_EMERG
7164 ALERT, ALR => LOG_ALERT
7165 CRIT, CRI => LOG_CRIT
7166 ERROR, ERR => LOG_ERR
7167 WARNING, WARN, WAR => LOG_WARNING
7168 NOTICE, NOTE, NOT => LOG_NOTICE
7169 INFO, INF => LOG_INFO
7170 DEBUG, DBG => LOG_DEBUG
7172 and as before, if none of those prefixes are present at the
7173 beginning of the string, LOG_ERR is chosen.
7175 On Win32, the LOG_* levels are mapped according to this:
7177 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
7178 LOG_WARNING => EVENTLOG_WARNING_TYPE
7179 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
7183 *) Made it possible to reconfigure with just the configuration
7184 argument "reconf" or "reconfigure". The command line arguments
7185 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
7186 and are retrieved from there when reconfiguring.
7190 [Assar Westerlund <assar@sics.se>, Richard Levitte]
7192 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
7195 *) The obj_dat.pl script was messing up the sorting of object
7196 names. The reason was that it compared the quoted version
7197 of strings as a result "OCSP" > "OCSP Signing" because
7198 " > SPACE. Changed script to store unquoted versions of
7199 names and add quotes on output. It was also omitting some
7200 names from the lookup table if they were given a default
7201 value (that is if SN is missing it is given the same
7202 value as LN and vice versa), these are now added on the
7203 grounds that if an object has a name we should be able to
7204 look it up. Finally added warning output when duplicate
7205 short or long names are found.
7208 *) Changes needed for Tandem NSK.
7209 [Scott Uroff <scott@xypro.com>]
7211 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
7212 RSA_padding_check_SSLv23(), special padding was never detected
7213 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
7214 version rollback attacks was not effective.
7216 In s23_clnt.c, don't use special rollback-attack detection padding
7217 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
7218 client; similarly, in s23_srvr.c, don't do the rollback check if
7219 SSL 2.0 is the only protocol enabled in the server.
7222 *) Make it possible to get hexdumps of unprintable data with 'openssl
7223 asn1parse'. By implication, the functions ASN1_parse_dump() and
7224 BIO_dump_indent() are added.
7227 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
7228 these print out strings and name structures based on various
7229 flags including RFC2253 support and proper handling of
7230 multibyte characters. Added options to the 'x509' utility
7231 to allow the various flags to be set.
7234 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
7235 Also change the functions X509_cmp_current_time() and
7236 X509_gmtime_adj() work with an ASN1_TIME structure,
7237 this will enable certificates using GeneralizedTime in validity
7238 dates to be checked.
7241 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
7242 negative public key encodings) on by default,
7243 NO_NEG_PUBKEY_BUG can be set to disable it.
7246 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
7247 content octets. An i2c_ASN1_OBJECT is unnecessary because
7248 the encoding can be trivially obtained from the structure.
7251 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
7252 not read locks (CRYPTO_r_[un]lock).
7255 *) A first attempt at creating official support for shared
7256 libraries through configuration. I've kept it so the
7257 default is static libraries only, and the OpenSSL programs
7258 are always statically linked for now, but there are
7259 preparations for dynamic linking in place.
7260 This has been tested on Linux and Tru64.
7263 *) Randomness polling function for Win9x, as described in:
7264 Peter Gutmann, Software Generation of Practically Strong
7268 *) Fix so PRNG is seeded in req if using an already existing
7272 *) New options to smime application. -inform and -outform
7273 allow alternative formats for the S/MIME message including
7274 PEM and DER. The -content option allows the content to be
7275 specified separately. This should allow things like Netscape
7276 form signing output easier to verify.
7279 *) Fix the ASN1 encoding of tags using the 'long form'.
7282 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
7283 STRING types. These convert content octets to and from the
7284 underlying type. The actual tag and length octets are
7285 already assumed to have been read in and checked. These
7286 are needed because all other string types have virtually
7287 identical handling apart from the tag. By having versions
7288 of the ASN1 functions that just operate on content octets
7289 IMPLICIT tagging can be handled properly. It also allows
7290 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
7291 and ASN1_INTEGER are identical apart from the tag.
7294 *) Change the handling of OID objects as follows:
7296 - New object identifiers are inserted in objects.txt, following
7297 the syntax given in objects.README.
7298 - objects.pl is used to process obj_mac.num and create a new
7300 - obj_dat.pl is used to create a new obj_dat.h, using the data in
7303 This is currently kind of a hack, and the perl code in objects.pl
7304 isn't very elegant, but it works as I intended. The simplest way
7305 to check that it worked correctly is to look in obj_dat.h and
7306 check the array nid_objs and make sure the objects haven't moved
7307 around (this is important!). Additions are OK, as well as
7308 consistent name changes.
7311 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
7314 *) Addition of the command line parameter '-rand file' to 'openssl req'.
7315 The given file adds to whatever has already been seeded into the
7316 random pool through the RANDFILE configuration file option or
7317 environment variable, or the default random state file.
7320 *) mkstack.pl now sorts each macro group into lexical order.
7321 Previously the output order depended on the order the files
7322 appeared in the directory, resulting in needless rewriting
7326 *) Patches to make OpenSSL compile under Win32 again. Mostly
7327 work arounds for the VC++ problem that it treats func() as
7328 func(void). Also stripped out the parts of mkdef.pl that
7329 added extra typesafe functions: these no longer exist.
7332 *) Reorganisation of the stack code. The macros are now all
7333 collected in safestack.h . Each macro is defined in terms of
7334 a "stack macro" of the form SKM_<name>(type, a, b). The
7335 DEBUG_SAFESTACK is now handled in terms of function casts,
7336 this has the advantage of retaining type safety without the
7337 use of additional functions. If DEBUG_SAFESTACK is not defined
7338 then the non typesafe macros are used instead. Also modified the
7339 mkstack.pl script to handle the new form. Needs testing to see
7340 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
7341 the default if no major problems. Similar behaviour for ASN1_SET_OF
7342 and PKCS12_STACK_OF.
7345 *) When some versions of IIS use the 'NET' form of private key the
7346 key derivation algorithm is different. Normally MD5(password) is
7347 used as a 128 bit RC4 key. In the modified case
7348 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
7349 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
7350 as the old Netscape_RSA functions except they have an additional
7351 'sgckey' parameter which uses the modified algorithm. Also added
7352 an -sgckey command line option to the rsa utility. Thanks to
7353 Adrian Peck <bertie@ncipher.com> for posting details of the modified
7354 algorithm to openssl-dev.
7357 *) The evp_local.h macros were using 'c.##kname' which resulted in
7358 invalid expansion on some systems (SCO 5.0.5 for example).
7359 Corrected to 'c.kname'.
7360 [Phillip Porch <root@theporch.com>]
7362 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
7363 a STACK of email addresses from a certificate or request, these look
7364 in the subject name and the subject alternative name extensions and
7365 omit any duplicate addresses.
7368 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
7369 This makes DSA verification about 2 % faster.
7372 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
7373 (meaning that now 2^5 values will be precomputed, which is only 4 KB
7374 plus overhead for 1024 bit moduli).
7375 This makes exponentiations about 0.5 % faster for 1024 bit
7376 exponents (as measured by "openssl speed rsa2048").
7379 *) Rename memory handling macros to avoid conflicts with other
7381 Malloc => OPENSSL_malloc
7382 Malloc_locked => OPENSSL_malloc_locked
7383 Realloc => OPENSSL_realloc
7384 Free => OPENSSL_free
7387 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
7388 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
7391 *) CygWin32 support.
7392 [John Jarvie <jjarvie@newsguy.com>]
7394 *) The type-safe stack code has been rejigged. It is now only compiled
7395 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
7396 by default all type-specific stack functions are "#define"d back to
7397 standard stack functions. This results in more streamlined output
7398 but retains the type-safety checking possibilities of the original
7402 *) The STACK code has been cleaned up, and certain type declarations
7403 that didn't make a lot of sense have been brought in line. This has
7404 also involved a cleanup of sorts in safestack.h to more correctly
7405 map type-safe stack functions onto their plain stack counterparts.
7406 This work has also resulted in a variety of "const"ifications of
7407 lots of the code, especially "_cmp" operations which should normally
7408 be prototyped with "const" parameters anyway.
7411 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
7412 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
7413 (The PRNG state consists of two parts, the large pool 'state' and 'md',
7414 where all of 'md' is used each time the PRNG is used, but 'state'
7415 is used only indexed by a cyclic counter. As entropy may not be
7416 well distributed from the beginning, 'md' is important as a
7417 chaining variable. However, the output function chains only half
7418 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
7419 all of 'md', and seeding with STATE_SIZE dummy bytes will result
7420 in all of 'state' being rewritten, with the new values depending
7421 on virtually all of 'md'. This overcomes the 80 bit limitation.)
7424 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
7425 the handshake is continued after ssl_verify_cert_chain();
7426 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
7427 can lead to 'unexplainable' connection aborts later.
7428 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
7430 *) Major EVP API cipher revision.
7431 Add hooks for extra EVP features. This allows various cipher
7432 parameters to be set in the EVP interface. Support added for variable
7433 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
7434 setting of RC2 and RC5 parameters.
7436 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
7439 Remove lots of duplicated code from the EVP library. For example *every*
7440 cipher init() function handles the 'iv' in the same way according to the
7441 cipher mode. They also all do nothing if the 'key' parameter is NULL and
7442 for CFB and OFB modes they zero ctx->num.
7444 New functionality allows removal of S/MIME code RC2 hack.
7446 Most of the routines have the same form and so can be declared in terms
7449 By shifting this to the top level EVP_CipherInit() it can be removed from
7450 all individual ciphers. If the cipher wants to handle IVs or keys
7451 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
7454 Change lots of functions like EVP_EncryptUpdate() to now return a
7455 value: although software versions of the algorithms cannot fail
7456 any installed hardware versions can.
7459 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
7460 this option is set, tolerate broken clients that send the negotiated
7461 protocol version number instead of the requested protocol version
7465 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
7466 i.e. non-zero for export ciphersuites, zero otherwise.
7467 Previous versions had this flag inverted, inconsistent with
7468 rsa_tmp_cb (..._TMP_RSA_CB).
7469 [Bodo Moeller; problem reported by Amit Chopra]
7471 *) Add missing DSA library text string. Work around for some IIS
7472 key files with invalid SEQUENCE encoding.
7475 *) Add a document (doc/standards.txt) that list all kinds of standards
7476 and so on that are implemented in OpenSSL.
7479 *) Enhance c_rehash script. Old version would mishandle certificates
7480 with the same subject name hash and wouldn't handle CRLs at all.
7481 Added -fingerprint option to crl utility, to support new c_rehash
7485 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
7488 *) Fix for SSL server purpose checking. Server checking was
7489 rejecting certificates which had extended key usage present
7490 but no ssl client purpose.
7491 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
7493 *) Make PKCS#12 code work with no password. The PKCS#12 spec
7494 is a little unclear about how a blank password is handled.
7495 Since the password in encoded as a BMPString with terminating
7496 double NULL a zero length password would end up as just the
7497 double NULL. However no password at all is different and is
7498 handled differently in the PKCS#12 key generation code. NS
7499 treats a blank password as zero length. MSIE treats it as no
7500 password on export: but it will try both on import. We now do
7501 the same: PKCS12_parse() tries zero length and no password if
7502 the password is set to "" or NULL (NULL is now a valid password:
7503 it wasn't before) as does the pkcs12 application.
7506 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
7507 perror when PEM_read_bio_X509_REQ fails, the error message must
7508 be obtained from the error queue.
7511 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
7512 it in ERR_remove_state if appropriate, and change ERR_get_state
7513 accordingly to avoid race conditions (this is necessary because
7514 thread_hash is no longer constant once set).
7517 *) Bugfix for linux-elf makefile.one.
7520 *) RSA_get_default_method() will now cause a default
7521 RSA_METHOD to be chosen if one doesn't exist already.
7522 Previously this was only set during a call to RSA_new()
7523 or RSA_new_method(NULL) meaning it was possible for
7524 RSA_get_default_method() to return NULL.
7527 *) Added native name translation to the existing DSO code
7528 that will convert (if the flag to do so is set) filenames
7529 that are sufficiently small and have no path information
7530 into a canonical native form. Eg. "blah" converted to
7531 "libblah.so" or "blah.dll" etc.
7534 *) New function ERR_error_string_n(e, buf, len) which is like
7535 ERR_error_string(e, buf), but writes at most 'len' bytes
7536 including the 0 terminator. For ERR_error_string_n, 'buf'
7538 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
7540 *) CONF library reworked to become more general. A new CONF
7541 configuration file reader "class" is implemented as well as a
7542 new functions (NCONF_*, for "New CONF") to handle it. The now
7543 old CONF_* functions are still there, but are reimplemented to
7544 work in terms of the new functions. Also, a set of functions
7545 to handle the internal storage of the configuration data is
7546 provided to make it easier to write new configuration file
7547 reader "classes" (I can definitely see something reading a
7548 configuration file in XML format, for example), called _CONF_*,
7549 or "the configuration storage API"...
7551 The new configuration file reading functions are:
7553 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
7554 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
7556 NCONF_default, NCONF_WIN32
7558 NCONF_dump_fp, NCONF_dump_bio
7560 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
7561 NCONF_new creates a new CONF object. This works in the same way
7562 as other interfaces in OpenSSL, like the BIO interface.
7563 NCONF_dump_* dump the internal storage of the configuration file,
7564 which is useful for debugging. All other functions take the same
7565 arguments as the old CONF_* functions wth the exception of the
7566 first that must be a `CONF *' instead of a `LHASH *'.
7568 To make it easer to use the new classes with the old CONF_* functions,
7569 the function CONF_set_default_method is provided.
7572 *) Add '-tls1' option to 'openssl ciphers', which was already
7573 mentioned in the documentation but had not been implemented.
7574 (This option is not yet really useful because even the additional
7575 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
7578 *) Initial DSO code added into libcrypto for letting OpenSSL (and
7579 OpenSSL-based applications) load shared libraries and bind to
7580 them in a portable way.
7581 [Geoff Thorpe, with contributions from Richard Levitte]
7583 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
7585 *) Make sure _lrotl and _lrotr are only used with MSVC.
7587 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
7588 (the default implementation of RAND_status).
7590 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
7591 to '-clrext' (= clear extensions), as intended and documented.
7592 [Bodo Moeller; inconsistency pointed out by Michael Attili
7595 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
7596 was larger than the MD block size.
7597 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
7599 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
7600 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
7601 using the passed key: if the passed key was a private key the result
7602 of X509_print(), for example, would be to print out all the private key
7606 *) des_quad_cksum() byte order bug fix.
7607 [Ulf Möller, using the problem description in krb4-0.9.7, where
7608 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
7610 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
7612 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
7614 *) For easily testing in shell scripts whether some command
7615 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
7616 returns with exit code 0 iff no command of the given name is available.
7617 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
7618 the output goes to stdout and nothing is printed to stderr.
7619 Additional arguments are always ignored.
7621 Since for each cipher there is a command of the same name,
7622 the 'no-cipher' compilation switches can be tested this way.
7624 ('openssl no-XXX' is not able to detect pseudo-commands such
7625 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
7628 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
7631 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
7632 is set; it will be thrown away anyway because each handshake creates
7634 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
7635 to parameters -- in previous versions (since OpenSSL 0.9.3) the
7636 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
7637 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7640 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7641 'Q' and 'R' lose their special meanings (quit/renegotiate).
7642 This is part of what -quiet does; unlike -quiet, -ign_eof
7643 does not suppress any output.
7646 *) Add compatibility options to the purpose and trust code. The
7647 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7648 accepts a certificate or CA, this was the previous behaviour,
7649 with all the associated security issues.
7651 X509_TRUST_COMPAT is the old trust behaviour: only and
7652 automatically trust self signed roots in certificate store. A
7653 new trust setting X509_TRUST_DEFAULT is used to specify that
7654 a purpose has no associated trust setting and it should instead
7655 use the value in the default purpose.
7658 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7659 and fix a memory leak.
7662 *) In util/mkerr.pl (which implements 'make errors'), preserve
7663 reason strings from the previous version of the .c file, as
7664 the default to have only downcase letters (and digits) in
7665 automatically generated reasons codes is not always appropriate.
7668 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7669 using strerror. Previously, ERR_reason_error_string() returned
7670 library names as reason strings for SYSerr; but SYSerr is a special
7671 case where small numbers are errno values, not library numbers.
7674 *) Add '-dsaparam' option to 'openssl dhparam' application. This
7675 converts DSA parameters into DH parameters. (When creating parameters,
7676 DSA_generate_parameters is used.)
7679 *) Include 'length' (recommended exponent length) in C code generated
7680 by 'openssl dhparam -C'.
7683 *) The second argument to set_label in perlasm was already being used
7684 so couldn't be used as a "file scope" flag. Moved to third argument
7688 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
7689 instead of RAND_bytes for encryption IVs and salts.
7692 *) Include RAND_status() into RAND_METHOD instead of implementing
7693 it only for md_rand.c Otherwise replacing the PRNG by calling
7694 RAND_set_rand_method would be impossible.
7697 *) Don't let DSA_generate_key() enter an infinite loop if the random
7698 number generation fails.
7701 *) New 'rand' application for creating pseudo-random output.
7704 *) Added configuration support for Linux/IA64
7705 [Rolf Haberrecker <rolf@suse.de>]
7707 *) Assembler module support for Mingw32.
7710 *) Shared library support for HPUX (in shlib/).
7711 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
7713 *) Shared library support for Solaris gcc.
7714 [Lutz Behnke <behnke@trustcenter.de>]
7716 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
7718 *) PKCS7_encrypt() was adding text MIME headers twice because they
7719 were added manually and by SMIME_crlf_copy().
7722 *) In bntest.c don't call BN_rand with zero bits argument.
7723 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
7725 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
7726 case was implemented. This caused BN_div_recp() to fail occasionally.
7729 *) Add an optional second argument to the set_label() in the perl
7730 assembly language builder. If this argument exists and is set
7731 to 1 it signals that the assembler should use a symbol whose
7732 scope is the entire file, not just the current function. This
7733 is needed with MASM which uses the format label:: for this scope.
7734 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
7736 *) Change the ASN1 types so they are typedefs by default. Before
7737 almost all types were #define'd to ASN1_STRING which was causing
7738 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
7742 *) Change names of new functions to the new get1/get0 naming
7743 convention: After 'get1', the caller owns a reference count
7744 and has to call ..._free; 'get0' returns a pointer to some
7745 data structure without incrementing reference counters.
7746 (Some of the existing 'get' functions increment a reference
7747 counter, some don't.)
7748 Similarly, 'set1' and 'add1' functions increase reference
7749 counters or duplicate objects.
7752 *) Allow for the possibility of temp RSA key generation failure:
7753 the code used to assume it always worked and crashed on failure.
7756 *) Fix potential buffer overrun problem in BIO_printf().
7757 [Ulf Möller, using public domain code by Patrick Powell; problem
7758 pointed out by David Sacerdote <das33@cornell.edu>]
7760 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
7761 RAND_egd() and RAND_status(). In the command line application,
7762 the EGD socket can be specified like a seed file using RANDFILE
7766 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
7767 Some CAs (e.g. Verisign) distribute certificates in this form.
7770 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
7771 list to exclude them. This means that no special compilation option
7772 is needed to use anonymous DH: it just needs to be included in the
7776 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
7777 EVP_MD_type. The old functionality is available in a new macro called
7778 EVP_MD_md(). Change code that uses it and update docs.
7781 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
7782 where the 'void *' argument is replaced by a function pointer argument.
7783 Previously 'void *' was abused to point to functions, which works on
7784 many platforms, but is not correct. As these functions are usually
7785 called by macros defined in OpenSSL header files, most source code
7786 should work without changes.
7789 *) <openssl/opensslconf.h> (which is created by Configure) now contains
7790 sections with information on -D... compiler switches used for
7791 compiling the library so that applications can see them. To enable
7792 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
7793 must be defined. E.g.,
7794 #define OPENSSL_ALGORITHM_DEFINES
7795 #include <openssl/opensslconf.h>
7796 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
7797 [Richard Levitte, Ulf and Bodo Möller]
7799 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
7803 *) Change the 'other' type in certificate aux info to a STACK_OF
7804 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
7805 the required ASN1 format: arbitrary types determined by an OID.
7808 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
7809 argument to 'req'. This is not because the function is newer or
7810 better than others it just uses the work 'NEW' in the certificate
7811 request header lines. Some software needs this.
7814 *) Reorganise password command line arguments: now passwords can be
7815 obtained from various sources. Delete the PEM_cb function and make
7816 it the default behaviour: i.e. if the callback is NULL and the
7817 usrdata argument is not NULL interpret it as a null terminated pass
7818 phrase. If usrdata and the callback are NULL then the pass phrase
7819 is prompted for as usual.
7822 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
7823 the support is automatically enabled. The resulting binaries will
7824 autodetect the card and use it if present.
7825 [Ben Laurie and Compaq Inc.]
7827 *) Work around for Netscape hang bug. This sends certificate request
7828 and server done in one record. Since this is perfectly legal in the
7829 SSL/TLS protocol it isn't a "bug" option and is on by default. See
7830 the bugs/SSLv3 entry for more info.
7833 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
7836 *) Add -rand argument to smime and pkcs12 applications and read/write
7840 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
7843 *) Add command line password options to the remaining applications.
7846 *) Bug fix for BN_div_recp() for numerators with an even number of
7850 *) More tests in bntest.c, and changed test_bn output.
7853 *) ./config recognizes MacOS X now.
7856 *) Bug fix for BN_div() when the first words of num and divsor are
7857 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
7860 *) Add support for various broken PKCS#8 formats, and command line
7861 options to produce them.
7864 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
7865 get temporary BIGNUMs from a BN_CTX.
7868 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
7872 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
7873 include a #define from the old name to the new. The original intent
7874 was that statically linked binaries could for example just call
7875 SSLeay_add_all_ciphers() to just add ciphers to the table and not
7876 link with digests. This never worked becayse SSLeay_add_all_digests()
7877 and SSLeay_add_all_ciphers() were in the same source file so calling
7878 one would link with the other. They are now in separate source files.
7881 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
7884 *) Use a less unusual form of the Miller-Rabin primality test (it used
7885 a binary algorithm for exponentiation integrated into the Miller-Rabin
7886 loop, our standard modexp algorithms are faster).
7889 *) Support for the EBCDIC character set completed.
7890 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
7892 *) Source code cleanups: use const where appropriate, eliminate casts,
7893 use void * instead of char * in lhash.
7896 *) Bugfix: ssl3_send_server_key_exchange was not restartable
7897 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
7898 this the server could overwrite ephemeral keys that the client
7902 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
7903 using 50 iterations of the Rabin-Miller test.
7905 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
7906 iterations of the Rabin-Miller test as required by the appendix
7907 to FIPS PUB 186[-1]) instead of DSA_is_prime.
7908 As BN_is_prime_fasttest includes trial division, DSA parameter
7909 generation becomes much faster.
7911 This implies a change for the callback functions in DSA_is_prime
7912 and DSA_generate_parameters: The callback function is called once
7913 for each positive witness in the Rabin-Miller test, not just
7914 occasionally in the inner loop; and the parameters to the
7915 callback function now provide an iteration count for the outer
7916 loop rather than for the current invocation of the inner loop.
7917 DSA_generate_parameters additionally can call the callback
7918 function with an 'iteration count' of -1, meaning that a
7919 candidate has passed the trial division test (when q is generated
7920 from an application-provided seed, trial division is skipped).
7923 *) New function BN_is_prime_fasttest that optionally does trial
7924 division before starting the Rabin-Miller test and has
7925 an additional BN_CTX * argument (whereas BN_is_prime always
7926 has to allocate at least one BN_CTX).
7927 'callback(1, -1, cb_arg)' is called when a number has passed the
7928 trial division stage.
7931 *) Fix for bug in CRL encoding. The validity dates weren't being handled
7935 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
7938 *) New function BN_pseudo_rand().
7941 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
7942 bignum version of BN_from_montgomery() with the working code from
7943 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
7947 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
7948 made it impossible to use the same SSL_SESSION data structure in
7949 SSL2 clients in multiple threads.
7952 *) The return value of RAND_load_file() no longer counts bytes obtained
7953 by stat(). RAND_load_file(..., -1) is new and uses the complete file
7954 to seed the PRNG (previously an explicit byte count was required).
7955 [Ulf Möller, Bodo Möller]
7957 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
7958 used (char *) instead of (void *) and had casts all over the place.
7961 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
7964 *) Retain source code compatibility for BN_prime_checks macro:
7965 BN_is_prime(..., BN_prime_checks, ...) now uses
7966 BN_prime_checks_for_size to determine the appropriate number of
7967 Rabin-Miller iterations.
7970 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
7971 DH_CHECK_P_NOT_SAFE_PRIME.
7972 (Check if this is true? OpenPGP calls them "strong".)
7975 *) Merge the functionality of "dh" and "gendh" programs into a new program
7976 "dhparam". The old programs are retained for now but will handle DH keys
7977 (instead of parameters) in future.
7980 *) Make the ciphers, s_server and s_client programs check the return values
7981 when a new cipher list is set.
7984 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
7985 ciphers. Before when the 56bit ciphers were enabled the sorting was
7988 The syntax for the cipher sorting has been extended to support sorting by
7989 cipher-strength (using the strength_bits hard coded in the tables).
7990 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
7992 Fix a bug in the cipher-command parser: when supplying a cipher command
7993 string with an "undefined" symbol (neither command nor alphanumeric
7994 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
7995 an error is flagged.
7997 Due to the strength-sorting extension, the code of the
7998 ssl_create_cipher_list() function was completely rearranged. I hope that
7999 the readability was also increased :-)
8000 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8002 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
8003 for the first serial number and places 2 in the serial number file. This
8004 avoids problems when the root CA is created with serial number zero and
8005 the first user certificate has the same issuer name and serial number
8009 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
8010 the new code. Add documentation for this stuff.
8013 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
8014 X509_*() to X509at_*() on the grounds that they don't handle X509
8015 structures and behave in an analagous way to the X509v3 functions:
8016 they shouldn't be called directly but wrapper functions should be used
8019 So we also now have some wrapper functions that call the X509at functions
8020 when passed certificate requests. (TO DO: similar things can be done with
8021 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
8022 things. Some of these need some d2i or i2d and print functionality
8023 because they handle more complex structures.)
8026 *) Add missing #ifndefs that caused missing symbols when building libssl
8027 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
8028 NO_RSA in ssl/s2*.c.
8029 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
8031 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
8032 has a return value which indicates the quality of the random data
8033 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
8034 error queue. New function RAND_pseudo_bytes() generates output that is
8035 guaranteed to be unique but not unpredictable. RAND_add is like
8036 RAND_seed, but takes an extra argument for an entropy estimate
8037 (RAND_seed always assumes full entropy).
8040 *) Do more iterations of Rabin-Miller probable prime test (specifically,
8041 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
8042 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
8043 in crypto/bn/bn_prime.c for the complete table). This guarantees a
8044 false-positive rate of at most 2^-80 for random input.
8047 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
8050 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
8051 in the 0.9.5 release), this returns the chain
8052 from an X509_CTX structure with a dup of the stack and all
8053 the X509 reference counts upped: so the stack will exist
8054 after X509_CTX_cleanup() has been called. Modify pkcs12.c
8057 Also make SSL_SESSION_print() print out the verify return
8061 *) Add manpage for the pkcs12 command. Also change the default
8062 behaviour so MAC iteration counts are used unless the new
8063 -nomaciter option is used. This improves file security and
8064 only older versions of MSIE (4.0 for example) need it.
8067 *) Honor the no-xxx Configure options when creating .DEF files.
8070 *) Add PKCS#10 attributes to field table: challengePassword,
8071 unstructuredName and unstructuredAddress. These are taken from
8072 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
8073 international characters are used.
8075 More changes to X509_ATTRIBUTE code: allow the setting of types
8076 based on strings. Remove the 'loc' parameter when adding
8077 attributes because these will be a SET OF encoding which is sorted
8081 *) Initial changes to the 'req' utility to allow request generation
8082 automation. This will allow an application to just generate a template
8083 file containing all the field values and have req construct the
8086 Initial support for X509_ATTRIBUTE handling. Stacks of these are
8087 used all over the place including certificate requests and PKCS#7
8088 structures. They are currently handled manually where necessary with
8089 some primitive wrappers for PKCS#7. The new functions behave in a
8090 manner analogous to the X509 extension functions: they allow
8091 attributes to be looked up by NID and added.
8093 Later something similar to the X509V3 code would be desirable to
8094 automatically handle the encoding, decoding and printing of the
8095 more complex types. The string types like challengePassword can
8096 be handled by the string table functions.
8098 Also modified the multi byte string table handling. Now there is
8099 a 'global mask' which masks out certain types. The table itself
8100 can use the flag STABLE_NO_MASK to ignore the mask setting: this
8101 is useful when for example there is only one permissible type
8102 (as in countryName) and using the mask might result in no valid
8106 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
8107 SSL_get_peer_finished to allow applications to obtain the latest
8108 Finished messages sent to the peer or expected from the peer,
8109 respectively. (SSL_get_peer_finished is usually the Finished message
8110 actually received from the peer, otherwise the protocol will be aborted.)
8112 As the Finished message are message digests of the complete handshake
8113 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
8114 be used for external authentication procedures when the authentication
8115 provided by SSL/TLS is not desired or is not enough.
8118 *) Enhanced support for Alpha Linux is added. Now ./config checks if
8119 the host supports BWX extension and if Compaq C is present on the
8120 $PATH. Just exploiting of the BWX extension results in 20-30%
8121 performance kick for some algorithms, e.g. DES and RC4 to mention
8122 a couple. Compaq C in turn generates ~20% faster code for MD5 and
8126 *) Add support for MS "fast SGC". This is arguably a violation of the
8127 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
8128 weak crypto and after checking the certificate is SGC a second one
8129 with strong crypto. MS SGC stops the first handshake after receiving
8130 the server certificate message and sends a second client hello. Since
8131 a server will typically do all the time consuming operations before
8132 expecting any further messages from the client (server key exchange
8133 is the most expensive) there is little difference between the two.
8135 To get OpenSSL to support MS SGC we have to permit a second client
8136 hello message after we have sent server done. In addition we have to
8137 reset the MAC if we do get this second client hello.
8140 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
8141 if a DER encoded private key is RSA or DSA traditional format. Changed
8142 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
8143 format DER encoded private key. Newer code should use PKCS#8 format which
8144 has the key type encoded in the ASN1 structure. Added DER private key
8145 support to pkcs8 application.
8148 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
8149 ciphersuites has been selected (as required by the SSL 3/TLS 1
8150 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
8151 is set, we interpret this as a request to violate the specification
8152 (the worst that can happen is a handshake failure, and 'correct'
8153 behaviour would result in a handshake failure anyway).
8156 *) In SSL_CTX_add_session, take into account that there might be multiple
8157 SSL_SESSION structures with the same session ID (e.g. when two threads
8158 concurrently obtain them from an external cache).
8159 The internal cache can handle only one SSL_SESSION with a given ID,
8160 so if there's a conflict, we now throw out the old one to achieve
8164 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
8165 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
8166 some routines that use cipher OIDs: some ciphers do not have OIDs
8167 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
8171 *) Simplify the trust setting structure and code. Now we just have
8172 two sequences of OIDs for trusted and rejected settings. These will
8173 typically have values the same as the extended key usage extension
8174 and any application specific purposes.
8176 The trust checking code now has a default behaviour: it will just
8177 check for an object with the same NID as the passed id. Functions can
8178 be provided to override either the default behaviour or the behaviour
8179 for a given id. SSL client, server and email already have functions
8180 in place for compatibility: they check the NID and also return "trusted"
8181 if the certificate is self signed.
8184 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
8185 traditional format into an EVP_PKEY structure.
8188 *) Add a password callback function PEM_cb() which either prompts for
8189 a password if usr_data is NULL or otherwise assumes it is a null
8190 terminated password. Allow passwords to be passed on command line
8191 environment or config files in a few more utilities.
8194 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
8195 keys. Add some short names for PKCS#8 PBE algorithms and allow them
8196 to be specified on the command line for the pkcs8 and pkcs12 utilities.
8197 Update documentation.
8200 *) Support for ASN1 "NULL" type. This could be handled before by using
8201 ASN1_TYPE but there wasn't any function that would try to read a NULL
8202 and produce an error if it couldn't. For compatibility we also have
8203 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
8204 don't allocate anything because they don't need to.
8207 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
8209 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
8211 *) Rebuild of the memory allocation routines used by OpenSSL code and
8212 possibly others as well. The purpose is to make an interface that
8213 provide hooks so anyone can build a separate set of allocation and
8214 deallocation routines to be used by OpenSSL, for example memory
8215 pool implementations, or something else, which was previously hard
8216 since Malloc(), Realloc() and Free() were defined as macros having
8217 the values malloc, realloc and free, respectively (except for Win32
8218 compilations). The same is provided for memory debugging code.
8219 OpenSSL already comes with functionality to find memory leaks, but
8220 this gives people a chance to debug other memory problems.
8222 With these changes, a new set of functions and macros have appeared:
8224 CRYPTO_set_mem_debug_functions() [F]
8225 CRYPTO_get_mem_debug_functions() [F]
8226 CRYPTO_dbg_set_options() [F]
8227 CRYPTO_dbg_get_options() [F]
8228 CRYPTO_malloc_debug_init() [M]
8230 The memory debug functions are NULL by default, unless the library
8231 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
8232 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
8233 gives the standard debugging functions that come with OpenSSL) or
8234 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
8235 provided by the library user) must be used. When the standard
8236 debugging functions are used, CRYPTO_dbg_set_options can be used to
8237 request additional information:
8238 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
8239 the CRYPTO_MDEBUG_xxx macro when compiling the library.
8241 Also, things like CRYPTO_set_mem_functions will always give the
8242 expected result (the new set of functions is used for allocation
8243 and deallocation) at all times, regardless of platform and compiler
8246 To finish it up, some functions that were never use in any other
8247 way than through macros have a new API and new semantic:
8250 CRYPTO_dbg_realloc()
8253 All macros of value have retained their old syntax.
8254 [Richard Levitte and Bodo Moeller]
8256 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
8257 ordering of SMIMECapabilities wasn't in "strength order" and there
8258 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
8262 *) Some ASN1 types with illegal zero length encoding (INTEGER,
8263 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
8264 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
8266 *) Merge in my S/MIME library for OpenSSL. This provides a simple
8267 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
8268 functionality to handle multipart/signed properly) and a utility
8269 called 'smime' to call all this stuff. This is based on code I
8270 originally wrote for Celo who have kindly allowed it to be
8271 included in OpenSSL.
8274 *) Add variants des_set_key_checked and des_set_key_unchecked of
8275 des_set_key (aka des_key_sched). Global variable des_check_key
8276 decides which of these is called by des_set_key; this way
8277 des_check_key behaves as it always did, but applications and
8278 the library itself, which was buggy for des_check_key == 1,
8279 have a cleaner way to pick the version they need.
8282 *) New function PKCS12_newpass() which changes the password of a
8286 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
8287 dynamic mix. In both cases the ids can be used as an index into the
8288 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
8289 functions so they accept a list of the field values and the
8290 application doesn't need to directly manipulate the X509_TRUST
8294 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
8298 *) Modify the way the V3 extension code looks up extensions. This now
8299 works in a similar way to the object code: we have some "standard"
8300 extensions in a static table which is searched with OBJ_bsearch()
8301 and the application can add dynamic ones if needed. The file
8302 crypto/x509v3/ext_dat.h now has the info: this file needs to be
8303 updated whenever a new extension is added to the core code and kept
8304 in ext_nid order. There is a simple program 'tabtest.c' which checks
8305 this. New extensions are not added too often so this file can readily
8306 be maintained manually.
8308 There are two big advantages in doing things this way. The extensions
8309 can be looked up immediately and no longer need to be "added" using
8310 X509V3_add_standard_extensions(): this function now does nothing.
8311 [Side note: I get *lots* of email saying the extension code doesn't
8312 work because people forget to call this function]
8313 Also no dynamic allocation is done unless new extensions are added:
8314 so if we don't add custom extensions there is no need to call
8315 X509V3_EXT_cleanup().
8318 *) Modify enc utility's salting as follows: make salting the default. Add a
8319 magic header, so unsalted files fail gracefully instead of just decrypting
8320 to garbage. This is because not salting is a big security hole, so people
8321 should be discouraged from doing it.
8324 *) Fixes and enhancements to the 'x509' utility. It allowed a message
8325 digest to be passed on the command line but it only used this
8326 parameter when signing a certificate. Modified so all relevant
8327 operations are affected by the digest parameter including the
8328 -fingerprint and -x509toreq options. Also -x509toreq choked if a
8329 DSA key was used because it didn't fix the digest.
8332 *) Initial certificate chain verify code. Currently tests the untrusted
8333 certificates for consistency with the verify purpose (which is set
8334 when the X509_STORE_CTX structure is set up) and checks the pathlength.
8336 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
8337 this is because it will reject chains with invalid extensions whereas
8338 every previous version of OpenSSL and SSLeay made no checks at all.
8340 Trust code: checks the root CA for the relevant trust settings. Trust
8341 settings have an initial value consistent with the verify purpose: e.g.
8342 if the verify purpose is for SSL client use it expects the CA to be
8343 trusted for SSL client use. However the default value can be changed to
8344 permit custom trust settings: one example of this would be to only trust
8345 certificates from a specific "secure" set of CAs.
8347 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
8348 which should be used for version portability: especially since the
8349 verify structure is likely to change more often now.
8351 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
8352 to set them. If not set then assume SSL clients will verify SSL servers
8355 Two new options to the verify program: -untrusted allows a set of
8356 untrusted certificates to be passed in and -purpose which sets the
8357 intended purpose of the certificate. If a purpose is set then the
8358 new chain verify code is used to check extension consistency.
8361 *) Support for the authority information access extension.
8364 *) Modify RSA and DSA PEM read routines to transparently handle
8365 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
8366 public keys in a format compatible with certificate
8367 SubjectPublicKeyInfo structures. Unfortunately there were already
8368 functions called *_PublicKey_* which used various odd formats so
8369 these are retained for compatibility: however the DSA variants were
8370 never in a public release so they have been deleted. Changed dsa/rsa
8371 utilities to handle the new format: note no releases ever handled public
8372 keys so we should be OK.
8374 The primary motivation for this change is to avoid the same fiasco
8375 that dogs private keys: there are several incompatible private key
8376 formats some of which are standard and some OpenSSL specific and
8377 require various evil hacks to allow partial transparent handling and
8378 even then it doesn't work with DER formats. Given the option anything
8379 other than PKCS#8 should be dumped: but the other formats have to
8380 stay in the name of compatibility.
8382 With public keys and the benefit of hindsight one standard format
8383 is used which works with EVP_PKEY, RSA or DSA structures: though
8384 it clearly returns an error if you try to read the wrong kind of key.
8386 Added a -pubkey option to the 'x509' utility to output the public key.
8387 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
8388 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
8389 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
8390 that do the same as the EVP_PKEY_assign_*() except they up the
8391 reference count of the added key (they don't "swallow" the
8395 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
8396 CRLs would fail if the file contained no certificates or no CRLs:
8397 added a new function to read in both types and return the number
8398 read: this means that if none are read it will be an error. The
8399 DER versions of the certificate and CRL reader would always fail
8400 because it isn't possible to mix certificates and CRLs in DER format
8401 without choking one or the other routine. Changed this to just read
8402 a certificate: this is the best we can do. Also modified the code
8403 in apps/verify.c to take notice of return codes: it was previously
8404 attempting to read in certificates from NULL pointers and ignoring
8405 any errors: this is one reason why the cert and CRL reader seemed
8406 to work. It doesn't check return codes from the default certificate
8407 routines: these may well fail if the certificates aren't installed.
8410 *) Code to support otherName option in GeneralName.
8413 *) First update to verify code. Change the verify utility
8414 so it warns if it is passed a self signed certificate:
8415 for consistency with the normal behaviour. X509_verify
8416 has been modified to it will now verify a self signed
8417 certificate if *exactly* the same certificate appears
8418 in the store: it was previously impossible to trust a
8419 single self signed certificate. This means that:
8420 openssl verify ss.pem
8421 now gives a warning about a self signed certificate but
8422 openssl verify -CAfile ss.pem ss.pem
8426 *) For servers, store verify_result in SSL_SESSION data structure
8427 (and add it to external session representation).
8428 This is needed when client certificate verifications fails,
8429 but an application-provided verification callback (set by
8430 SSL_CTX_set_cert_verify_callback) allows accepting the session
8431 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
8432 but returns 1): When the session is reused, we have to set
8433 ssl->verify_result to the appropriate error code to avoid
8435 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
8437 *) Fix a bug in the new PKCS#7 code: it didn't consider the
8438 case in PKCS7_dataInit() where the signed PKCS7 structure
8439 didn't contain any existing data because it was being created.
8440 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
8442 *) Add a salt to the key derivation routines in enc.c. This
8443 forms the first 8 bytes of the encrypted file. Also add a
8444 -S option to allow a salt to be input on the command line.
8447 *) New function X509_cmp(). Oddly enough there wasn't a function
8448 to compare two certificates. We do this by working out the SHA1
8449 hash and comparing that. X509_cmp() will be needed by the trust
8453 *) SSL_get1_session() is like SSL_get_session(), but increments
8454 the reference count in the SSL_SESSION returned.
8455 [Geoff Thorpe <geoff@eu.c2.net>]
8457 *) Fix for 'req': it was adding a null to request attributes.
8458 Also change the X509_LOOKUP and X509_INFO code to handle
8459 certificate auxiliary information.
8462 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
8466 *) Add the possibility to add extra information to the memory leak
8467 detecting output, to form tracebacks, showing from where each
8468 allocation was originated: CRYPTO_push_info("constant string") adds
8469 the string plus current file name and line number to a per-thread
8470 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
8471 is like calling CYRPTO_pop_info() until the stack is empty.
8472 Also updated memory leak detection code to be multi-thread-safe.
8475 *) Add options -text and -noout to pkcs7 utility and delete the
8476 encryption options which never did anything. Update docs.
8479 *) Add options to some of the utilities to allow the pass phrase
8480 to be included on either the command line (not recommended on
8481 OSes like Unix) or read from the environment. Update the
8482 manpages and fix a few bugs.
8485 *) Add a few manpages for some of the openssl commands.
8488 *) Fix the -revoke option in ca. It was freeing up memory twice,
8489 leaking and not finding already revoked certificates.
8492 *) Extensive changes to support certificate auxiliary information.
8493 This involves the use of X509_CERT_AUX structure and X509_AUX
8494 functions. An X509_AUX function such as PEM_read_X509_AUX()
8495 can still read in a certificate file in the usual way but it
8496 will also read in any additional "auxiliary information". By
8497 doing things this way a fair degree of compatibility can be
8498 retained: existing certificates can have this information added
8499 using the new 'x509' options.
8501 Current auxiliary information includes an "alias" and some trust
8502 settings. The trust settings will ultimately be used in enhanced
8503 certificate chain verification routines: currently a certificate
8504 can only be trusted if it is self signed and then it is trusted
8508 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
8509 The problem was that one of the replacement routines had not been working
8510 since SSLeay releases. For now the offending routine has been replaced
8511 with non-optimised assembler. Even so, this now gives around 95%
8512 performance improvement for 1024 bit RSA signs.
8515 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
8516 handling. Most clients have the effective key size in bits equal to
8517 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
8518 A few however don't do this and instead use the size of the decrypted key
8519 to determine the RC2 key length and the AlgorithmIdentifier to determine
8520 the effective key length. In this case the effective key length can still
8521 be 40 bits but the key length can be 168 bits for example. This is fixed
8522 by manually forcing an RC2 key into the EVP_PKEY structure because the
8523 EVP code can't currently handle unusual RC2 key sizes: it always assumes
8524 the key length and effective key length are equal.
8527 *) Add a bunch of functions that should simplify the creation of
8528 X509_NAME structures. Now you should be able to do:
8529 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
8530 and have it automatically work out the correct field type and fill in
8531 the structures. The more adventurous can try:
8532 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
8533 and it will (hopefully) work out the correct multibyte encoding.
8536 *) Change the 'req' utility to use the new field handling and multibyte
8537 copy routines. Before the DN field creation was handled in an ad hoc
8538 way in req, ca, and x509 which was rather broken and didn't support
8539 BMPStrings or UTF8Strings. Since some software doesn't implement
8540 BMPStrings or UTF8Strings yet, they can be enabled using the config file
8541 using the dirstring_type option. See the new comment in the default
8542 openssl.cnf for more info.
8545 *) Make crypto/rand/md_rand.c more robust:
8546 - Assure unique random numbers after fork().
8547 - Make sure that concurrent threads access the global counter and
8548 md serializably so that we never lose entropy in them
8549 or use exactly the same state in multiple threads.
8550 Access to the large state is not always serializable because
8551 the additional locking could be a performance killer, and
8552 md should be large enough anyway.
8555 *) New file apps/app_rand.c with commonly needed functionality
8556 for handling the random seed file.
8558 Use the random seed file in some applications that previously did not:
8560 dsaparam -genkey (which also ignored its '-rand' option),
8563 x509 (when signing).
8564 Except on systems with /dev/urandom, it is crucial to have a random
8565 seed file at least for key creation, DSA signing, and for DH exchanges;
8566 for RSA signatures we could do without one.
8568 gendh and gendsa (unlike genrsa) used to read only the first byte
8569 of each file listed in the '-rand' option. The function as previously
8570 found in genrsa is now in app_rand.c and is used by all programs
8571 that support '-rand'.
8574 *) In RAND_write_file, use mode 0600 for creating files;
8575 don't just chmod when it may be too late.
8578 *) Report an error from X509_STORE_load_locations
8579 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
8582 *) New function ASN1_mbstring_copy() this copies a string in either
8583 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
8584 into an ASN1_STRING type. A mask of permissible types is passed
8585 and it chooses the "minimal" type to use or an error if not type
8589 *) Add function equivalents to the various macros in asn1.h. The old
8590 macros are retained with an M_ prefix. Code inside the library can
8591 use the M_ macros. External code (including the openssl utility)
8592 should *NOT* in order to be "shared library friendly".
8595 *) Add various functions that can check a certificate's extensions
8596 to see if it usable for various purposes such as SSL client,
8597 server or S/MIME and CAs of these types. This is currently
8598 VERY EXPERIMENTAL but will ultimately be used for certificate chain
8599 verification. Also added a -purpose flag to x509 utility to
8600 print out all the purposes.
8603 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
8607 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
8608 for, obtain and decode and extension and obtain its critical flag.
8609 This allows all the necessary extension code to be handled in a
8610 single function call.
8613 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
8614 platforms. See crypto/rc4/rc4_enc.c for further details.
8617 *) New -noout option to asn1parse. This causes no output to be produced
8618 its main use is when combined with -strparse and -out to extract data
8619 from a file (which may not be in ASN.1 format).
8622 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
8623 when producing the local key id.
8624 [Richard Levitte <levitte@stacken.kth.se>]
8626 *) New option -dhparam in s_server. This allows a DH parameter file to be
8627 stated explicitly. If it is not stated then it tries the first server
8628 certificate file. The previous behaviour hard coded the filename
8632 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
8633 a public key to be input or output. For example:
8634 openssl rsa -in key.pem -pubout -out pubkey.pem
8635 Also added necessary DSA public key functions to handle this.
8638 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
8639 in the message. This was handled by allowing
8640 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8641 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8643 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8644 to the end of the strings whereas this didn't. This would cause problems
8645 if strings read with d2i_ASN1_bytes() were later modified.
8646 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8648 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8649 data and it contains EOF it will end up returning an error. This is
8650 caused by input 46 bytes long. The cause is due to the way base64
8651 BIOs find the start of base64 encoded data. They do this by trying a
8652 trial decode on each line until they find one that works. When they
8653 do a flag is set and it starts again knowing it can pass all the
8654 data directly through the decoder. Unfortunately it doesn't reset
8655 the context it uses. This means that if EOF is reached an attempt
8656 is made to pass two EOFs through the context and this causes the
8657 resulting error. This can also cause other problems as well. As is
8658 usual with these problems it takes *ages* to find and the fix is
8659 trivial: move one line.
8660 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8662 *) Ugly workaround to get s_client and s_server working under Windows. The
8663 old code wouldn't work because it needed to select() on sockets and the
8664 tty (for keypresses and to see if data could be written). Win32 only
8665 supports select() on sockets so we select() with a 1s timeout on the
8666 sockets and then see if any characters are waiting to be read, if none
8667 are present then we retry, we also assume we can always write data to
8668 the tty. This isn't nice because the code then blocks until we've
8669 received a complete line of data and it is effectively polling the
8670 keyboard at 1s intervals: however it's quite a bit better than not
8671 working at all :-) A dedicated Windows application might handle this
8672 with an event loop for example.
8675 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
8676 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
8677 will be called when RSA_sign() and RSA_verify() are used. This is useful
8678 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
8679 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
8680 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
8681 This necessitated the support of an extra signature type NID_md5_sha1
8682 for SSL signatures and modifications to the SSL library to use it instead
8683 of calling RSA_public_decrypt() and RSA_private_encrypt().
8686 *) Add new -verify -CAfile and -CApath options to the crl program, these
8687 will lookup a CRL issuers certificate and verify the signature in a
8688 similar way to the verify program. Tidy up the crl program so it
8689 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
8690 less strict. It will now permit CRL extensions even if it is not
8691 a V2 CRL: this will allow it to tolerate some broken CRLs.
8694 *) Initialize all non-automatic variables each time one of the openssl
8695 sub-programs is started (this is necessary as they may be started
8696 multiple times from the "OpenSSL>" prompt).
8697 [Lennart Bang, Bodo Moeller]
8699 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
8700 removing all other RSA functionality (this is what NO_RSA does). This
8701 is so (for example) those in the US can disable those operations covered
8702 by the RSA patent while allowing storage and parsing of RSA keys and RSA
8706 *) Non-copying interface to BIO pairs.
8707 (still largely untested)
8710 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
8711 ASCII string. This was handled independently in various places before.
8714 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
8715 UTF8 strings a character at a time.
8718 *) Use client_version from client hello to select the protocol
8719 (s23_srvr.c) and for RSA client key exchange verification
8720 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
8723 *) Add various utility functions to handle SPKACs, these were previously
8724 handled by poking round in the structure internals. Added new function
8725 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
8726 print, verify and generate SPKACs. Based on an original idea from
8727 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
8730 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
8733 *) Allow the config file extension section to be overwritten on the
8734 command line. Based on an original idea from Massimiliano Pala
8735 <madwolf@comune.modena.it>. The new option is called -extensions
8736 and can be applied to ca, req and x509. Also -reqexts to override
8737 the request extensions in req and -crlexts to override the crl extensions
8741 *) Add new feature to the SPKAC handling in ca. Now you can include
8742 the same field multiple times by preceding it by "XXXX." for example:
8745 this is the same syntax as used in the req config file.
8748 *) Allow certificate extensions to be added to certificate requests. These
8749 are specified in a 'req_extensions' option of the req section of the
8750 config file. They can be printed out with the -text option to req but
8751 are otherwise ignored at present.
8754 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
8755 data read consists of only the final block it would not decrypted because
8756 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
8757 A misplaced 'break' also meant the decrypted final block might not be
8758 copied until the next read.
8761 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
8762 a few extra parameters to the DH structure: these will be useful if
8763 for example we want the value of 'q' or implement X9.42 DH.
8766 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
8767 provides hooks that allow the default DSA functions or functions on a
8768 "per key" basis to be replaced. This allows hardware acceleration and
8769 hardware key storage to be handled without major modification to the
8770 library. Also added low level modexp hooks and CRYPTO_EX structure and
8771 associated functions.
8774 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
8775 as "read only": it can't be written to and the buffer it points to will
8776 not be freed. Reading from a read only BIO is much more efficient than
8777 a normal memory BIO. This was added because there are several times when
8778 an area of memory needs to be read from a BIO. The previous method was
8779 to create a memory BIO and write the data to it, this results in two
8780 copies of the data and an O(n^2) reading algorithm. There is a new
8781 function BIO_new_mem_buf() which creates a read only memory BIO from
8782 an area of memory. Also modified the PKCS#7 routines to use read only
8786 *) Bugfix: ssl23_get_client_hello did not work properly when called in
8787 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
8788 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
8789 but a retry condition occured while trying to read the rest.
8792 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
8793 NID_pkcs7_encrypted by default: this was wrong since this should almost
8794 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
8795 the encrypted data type: this is a more sensible place to put it and it
8796 allows the PKCS#12 code to be tidied up that duplicated this
8800 *) Changed obj_dat.pl script so it takes its input and output files on
8801 the command line. This should avoid shell escape redirection problems
8805 *) Initial support for certificate extension requests, these are included
8806 in things like Xenroll certificate requests. Included functions to allow
8807 extensions to be obtained and added.
8810 *) -crlf option to s_client and s_server for sending newlines as
8811 CRLF (as required by many protocols).
8814 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
8816 *) Install libRSAglue.a when OpenSSL is built with RSAref.
8817 [Ralf S. Engelschall]
8819 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
8820 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
8822 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
8826 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
8827 DH parameters/keys (q is lost during that conversion, but the resulting
8828 DH parameters contain its length).
8830 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
8831 much faster than DH_generate_parameters (which creates parameters
8832 where p = 2*q + 1), and also the smaller q makes DH computations
8833 much more efficient (160-bit exponentiation instead of 1024-bit
8834 exponentiation); so this provides a convenient way to support DHE
8835 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
8836 utter importance to use
8837 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8839 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8840 when such DH parameters are used, because otherwise small subgroup
8841 attacks may become possible!
8844 *) Avoid memory leak in i2d_DHparams.
8847 *) Allow the -k option to be used more than once in the enc program:
8848 this allows the same encrypted message to be read by multiple recipients.
8851 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
8852 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
8853 it will always use the numerical form of the OID, even if it has a short
8857 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
8858 method only got called if p,q,dmp1,dmq1,iqmp components were present,
8859 otherwise bn_mod_exp was called. In the case of hardware keys for example
8860 no private key components need be present and it might store extra data
8861 in the RSA structure, which cannot be accessed from bn_mod_exp.
8862 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
8863 private key operations.
8866 *) Added support for SPARC Linux.
8869 *) pem_password_cb function type incompatibly changed from
8870 typedef int pem_password_cb(char *buf, int size, int rwflag);
8872 ....(char *buf, int size, int rwflag, void *userdata);
8873 so that applications can pass data to their callbacks:
8874 The PEM[_ASN1]_{read,write}... functions and macros now take an
8875 additional void * argument, which is just handed through whenever
8876 the password callback is called.
8877 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
8879 New function SSL_CTX_set_default_passwd_cb_userdata.
8881 Compatibility note: As many C implementations push function arguments
8882 onto the stack in reverse order, the new library version is likely to
8883 interoperate with programs that have been compiled with the old
8884 pem_password_cb definition (PEM_whatever takes some data that
8885 happens to be on the stack as its last argument, and the callback
8886 just ignores this garbage); but there is no guarantee whatsoever that
8889 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
8890 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
8891 problems not only on Windows, but also on some Unix platforms.
8892 To avoid problematic command lines, these definitions are now in an
8893 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
8894 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
8897 *) MIPS III/IV assembler module is reimplemented.
8900 *) More DES library cleanups: remove references to srand/rand and
8901 delete an unused file.
8904 *) Add support for the the free Netwide assembler (NASM) under Win32,
8905 since not many people have MASM (ml) and it can be hard to obtain.
8906 This is currently experimental but it seems to work OK and pass all
8907 the tests. Check out INSTALL.W32 for info.
8910 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
8911 without temporary keys kept an extra copy of the server key,
8912 and connections with temporary keys did not free everything in case
8916 *) New function RSA_check_key and new openssl rsa option -check
8917 for verifying the consistency of RSA keys.
8918 [Ulf Moeller, Bodo Moeller]
8920 *) Various changes to make Win32 compile work:
8921 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
8922 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
8923 comparison" warnings.
8924 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
8927 *) Add a debugging option to PKCS#5 v2 key generation function: when
8928 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
8929 derived keys are printed to stderr.
8932 *) Copy the flags in ASN1_STRING_dup().
8933 [Roman E. Pavlov <pre@mo.msk.ru>]
8935 *) The x509 application mishandled signing requests containing DSA
8936 keys when the signing key was also DSA and the parameters didn't match.
8938 It was supposed to omit the parameters when they matched the signing key:
8939 the verifying software was then supposed to automatically use the CA's
8940 parameters if they were absent from the end user certificate.
8942 Omitting parameters is no longer recommended. The test was also
8943 the wrong way round! This was probably due to unusual behaviour in
8944 EVP_cmp_parameters() which returns 1 if the parameters match.
8945 This meant that parameters were omitted when they *didn't* match and
8946 the certificate was useless. Certificates signed with 'ca' didn't have
8948 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
8950 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
8951 The interface is as follows:
8952 Applications can use
8953 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
8954 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
8955 "off" is now the default.
8956 The library internally uses
8957 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
8958 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
8959 to disable memory-checking temporarily.
8961 Some inconsistent states that previously were possible (and were
8962 even the default) are now avoided.
8964 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
8965 with each memory chunk allocated; this is occasionally more helpful
8966 than just having a counter.
8968 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
8970 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
8974 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
8975 which largely parallels "options", but is for changing API behaviour,
8976 whereas "options" are about protocol behaviour.
8977 Initial "mode" flags are:
8979 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
8980 a single record has been written.
8981 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
8982 retries use the same buffer location.
8983 (But all of the contents must be
8987 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
8990 *) Fix problems with no-hmac etc.
8991 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
8993 *) New functions RSA_get_default_method(), RSA_set_method() and
8994 RSA_get_method(). These allows replacement of RSA_METHODs without having
8995 to mess around with the internals of an RSA structure.
8998 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
8999 Also really enable memory leak checks in openssl.c and in some
9001 [Chad C. Mulligan, Bodo Moeller]
9003 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
9004 up the length of negative integers. This has now been simplified to just
9005 store the length when it is first determined and use it later, rather
9006 than trying to keep track of where data is copied and updating it to
9008 [Steve Henson, reported by Brien Wheeler
9009 <bwheeler@authentica-security.com>]
9011 *) Add a new function PKCS7_signatureVerify. This allows the verification
9012 of a PKCS#7 signature but with the signing certificate passed to the
9013 function itself. This contrasts with PKCS7_dataVerify which assumes the
9014 certificate is present in the PKCS#7 structure. This isn't always the
9015 case: certificates can be omitted from a PKCS#7 structure and be
9016 distributed by "out of band" means (such as a certificate database).
9019 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
9020 function prototypes in pem.h, also change util/mkdef.pl to add the
9021 necessary function names.
9024 *) mk1mf.pl (used by Windows builds) did not properly read the
9025 options set by Configure in the top level Makefile, and Configure
9026 was not even able to write more than one option correctly.
9027 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
9030 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
9031 file to be loaded from a BIO or FILE pointer. The BIO version will
9032 for example allow memory BIOs to contain config info.
9035 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
9036 Whoever hopes to achieve shared-library compatibility across versions
9037 must use this, not the compile-time macro.
9038 (Exercise 0.9.4: Which is the minimum library version required by
9040 Note: All this applies only to multi-threaded programs, others don't
9044 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
9045 through a BIO pair triggered the default case, i.e.
9046 SSLerr(...,SSL_R_UNKNOWN_STATE).
9049 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
9050 can use the SSL library even if none of the specific BIOs is
9054 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
9055 for the encoded length.
9056 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
9058 *) Add initial documentation of the X509V3 functions.
9061 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
9062 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
9063 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
9064 secure PKCS#8 private key format with a high iteration count.
9067 *) Fix determination of Perl interpreter: A perl or perl5
9068 _directory_ in $PATH was also accepted as the interpreter.
9069 [Ralf S. Engelschall]
9071 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
9072 wrong with it but it was very old and did things like calling
9073 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
9077 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
9078 to use the new extension code.
9081 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
9082 with macros. This should make it easier to change their form, add extra
9083 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
9087 *) Add to configuration table a new entry that can specify an alternative
9088 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
9089 according to Mark Crispin <MRC@Panda.COM>.
9093 *) DES CBC did not update the IV. Weird.
9096 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
9097 Changing the behaviour of the former might break existing programs --
9098 where IV updating is needed, des_ncbc_encrypt can be used.
9101 *) When bntest is run from "make test" it drives bc to check its
9102 calculations, as well as internally checking them. If an internal check
9103 fails, it needs to cause bc to give a non-zero result or make test carries
9104 on without noticing the failure. Fixed.
9107 *) DES library cleanups.
9110 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
9111 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
9112 ciphers. NOTE: although the key derivation function has been verified
9113 against some published test vectors it has not been extensively tested
9114 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
9118 *) Instead of "mkdir -p", which is not fully portable, use new
9119 Perl script "util/mkdir-p.pl".
9122 *) Rewrite the way password based encryption (PBE) is handled. It used to
9123 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
9124 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
9125 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
9126 the 'parameter' field of the AlgorithmIdentifier is passed to the
9127 underlying key generation function so it must do its own ASN1 parsing.
9128 This has also changed the EVP_PBE_CipherInit() function which now has a
9129 'parameter' argument instead of literal salt and iteration count values
9130 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
9133 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
9134 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
9135 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
9136 KEY" because this clashed with PKCS#8 unencrypted string. Since this
9137 value was just used as a "magic string" and not used directly its
9138 value doesn't matter.
9141 *) Introduce some semblance of const correctness to BN. Shame C doesn't
9145 *) "linux-sparc64" configuration (ultrapenguin).
9146 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
9147 "linux-sparc" configuration.
9148 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
9150 *) config now generates no-xxx options for missing ciphers.
9153 *) Support the EBCDIC character set (work in progress).
9154 File ebcdic.c not yet included because it has a different license.
9155 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9157 *) Support BS2000/OSD-POSIX.
9158 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9160 *) Make callbacks for key generation use void * instead of char *.
9163 *) Make S/MIME samples compile (not yet tested).
9166 *) Additional typesafe stacks.
9169 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
9173 Changes between 0.9.3 and 0.9.3a [29 May 1999]
9175 *) New configuration variant "sco5-gcc".
9177 *) Updated some demos.
9178 [Sean O Riordain, Wade Scholine]
9180 *) Add missing BIO_free at exit of pkcs12 application.
9183 *) Fix memory leak in conf.c.
9186 *) Updates for Win32 to assembler version of MD5.
9189 *) Set #! path to perl in apps/der_chop to where we found it
9190 instead of using a fixed path.
9193 *) SHA library changes for irix64-mips4-cc.
9196 *) Improvements for VMS support.
9200 Changes between 0.9.2b and 0.9.3 [24 May 1999]
9202 *) Bignum library bug fix. IRIX 6 passes "make test" now!
9203 This also avoids the problems with SC4.2 and unpatched SC5.
9204 [Andy Polyakov <appro@fy.chalmers.se>]
9206 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
9207 These are required because of the typesafe stack would otherwise break
9208 existing code. If old code used a structure member which used to be STACK
9209 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
9210 sk_num or sk_value it would produce an error because the num, data members
9211 are not present in STACK_OF. Now it just produces a warning. sk_set
9212 replaces the old method of assigning a value to sk_value
9213 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
9214 that does this will no longer work (and should use sk_set instead) but
9215 this could be regarded as a "questionable" behaviour anyway.
9218 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
9219 correctly handle encrypted S/MIME data.
9222 *) Change type of various DES function arguments from des_cblock
9223 (which means, in function argument declarations, pointer to char)
9224 to des_cblock * (meaning pointer to array with 8 char elements),
9225 which allows the compiler to do more typechecking; it was like
9226 that back in SSLeay, but with lots of ugly casts.
9228 Introduce new type const_des_cblock.
9231 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
9232 problems: find RecipientInfo structure that matches recipient certificate
9233 and initialise the ASN1 structures properly based on passed cipher.
9236 *) Belatedly make the BN tests actually check the results.
9239 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
9240 to and from BNs: it was completely broken. New compilation option
9241 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
9242 key elements as negative integers.
9245 *) Reorganize and speed up MD5.
9246 [Andy Polyakov <appro@fy.chalmers.se>]
9249 [Richard Levitte <richard@levitte.org>]
9251 *) New option -out to asn1parse to allow the parsed structure to be
9252 output to a file. This is most useful when combined with the -strparse
9253 option to examine the output of things like OCTET STRINGS.
9256 *) Make SSL library a little more fool-proof by not requiring any longer
9257 that SSL_set_{accept,connect}_state be called before
9258 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
9259 in many applications because usually everything *appeared* to work as
9260 intended anyway -- now it really works as intended).
9263 *) Move openssl.cnf out of lib/.
9266 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
9267 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
9268 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
9269 [Ralf S. Engelschall]
9271 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
9272 handle PKCS#7 enveloped data properly.
9273 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
9275 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
9276 copying pointers. The cert_st handling is changed by this in
9277 various ways (and thus what used to be known as ctx->default_cert
9278 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
9279 any longer when s->cert does not give us what we need).
9280 ssl_cert_instantiate becomes obsolete by this change.
9281 As soon as we've got the new code right (possibly it already is?),
9282 we have solved a couple of bugs of the earlier code where s->cert
9283 was used as if it could not have been shared with other SSL structures.
9285 Note that using the SSL API in certain dirty ways now will result
9286 in different behaviour than observed with earlier library versions:
9287 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
9288 does not influence s as it used to.
9290 In order to clean up things more thoroughly, inside SSL_SESSION
9291 we don't use CERT any longer, but a new structure SESS_CERT
9292 that holds per-session data (if available); currently, this is
9293 the peer's certificate chain and, for clients, the server's certificate
9294 and temporary key. CERT holds only those values that can have
9295 meaningful defaults in an SSL_CTX.
9298 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
9299 from the internal representation. Various PKCS#7 fixes: remove some
9300 evil casts and set the enc_dig_alg field properly based on the signing
9304 *) Allow PKCS#12 password to be set from the command line or the
9305 environment. Let 'ca' get its config file name from the environment
9306 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
9310 *) Allow certificate policies extension to use an IA5STRING for the
9311 organization field. This is contrary to the PKIX definition but
9312 VeriSign uses it and IE5 only recognises this form. Document 'x509'
9316 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
9317 without disallowing inline assembler and the like for non-pedantic builds.
9320 *) Support Borland C++ builder.
9321 [Janez Jere <jj@void.si>, modified by Ulf Möller]
9326 *) SHA-1 cleanups and performance enhancements.
9327 [Andy Polyakov <appro@fy.chalmers.se>]
9329 *) Sparc v8plus assembler for the bignum library.
9330 [Andy Polyakov <appro@fy.chalmers.se>]
9332 *) Accept any -xxx and +xxx compiler options in Configure.
9335 *) Update HPUX configuration.
9338 *) Add missing sk_<type>_unshift() function to safestack.h
9339 [Ralf S. Engelschall]
9341 *) New function SSL_CTX_use_certificate_chain_file that sets the
9342 "extra_cert"s in addition to the certificate. (This makes sense
9343 only for "PEM" format files, as chains as a whole are not
9347 *) Support verify_depth from the SSL API.
9348 x509_vfy.c had what can be considered an off-by-one-error:
9349 Its depth (which was not part of the external interface)
9350 was actually counting the number of certificates in a chain;
9351 now it really counts the depth.
9354 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
9355 instead of X509err, which often resulted in confusing error
9356 messages since the error codes are not globally unique
9357 (e.g. an alleged error in ssl3_accept when a certificate
9358 didn't match the private key).
9360 *) New function SSL_CTX_set_session_id_context that allows to set a default
9361 value (so that you don't need SSL_set_session_id_context for each
9362 connection using the SSL_CTX).
9365 *) OAEP decoding bug fix.
9368 *) Support INSTALL_PREFIX for package builders, as proposed by
9372 *) New Configure options "threads" and "no-threads". For systems
9373 where the proper compiler options are known (currently Solaris
9374 and Linux), "threads" is the default.
9377 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
9380 *) Install various scripts to $(OPENSSLDIR)/misc, not to
9381 $(INSTALLTOP)/bin -- they shouldn't clutter directories
9382 such as /usr/local/bin.
9385 *) "make linux-shared" to build shared libraries.
9386 [Niels Poppe <niels@netbox.org>]
9388 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
9391 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
9392 extension adding in x509 utility.
9395 *) Remove NOPROTO sections and error code comments.
9398 *) Partial rewrite of the DEF file generator to now parse the ANSI
9402 *) New Configure options --prefix=DIR and --openssldir=DIR.
9405 *) Complete rewrite of the error code script(s). It is all now handled
9406 by one script at the top level which handles error code gathering,
9407 header rewriting and C source file generation. It should be much better
9408 than the old method: it now uses a modified version of Ulf's parser to
9409 read the ANSI prototypes in all header files (thus the old K&R definitions
9410 aren't needed for error creation any more) and do a better job of
9411 translating function codes into names. The old 'ASN1 error code imbedded
9412 in a comment' is no longer necessary and it doesn't use .err files which
9413 have now been deleted. Also the error code call doesn't have to appear all
9414 on one line (which resulted in some large lines...).
9417 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
9420 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
9421 0 (which usually indicates a closed connection), but continue reading.
9424 *) Fix some race conditions.
9427 *) Add support for CRL distribution points extension. Add Certificate
9428 Policies and CRL distribution points documentation.
9431 *) Move the autogenerated header file parts to crypto/opensslconf.h.
9434 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
9435 8 of keying material. Merlin has also confirmed interop with this fix
9436 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
9437 [Merlin Hughes <merlin@baltimore.ie>]
9439 *) Fix lots of warnings.
9440 [Richard Levitte <levitte@stacken.kth.se>]
9442 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
9443 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
9444 [Richard Levitte <levitte@stacken.kth.se>]
9446 *) Fix problems with sizeof(long) == 8.
9447 [Andy Polyakov <appro@fy.chalmers.se>]
9449 *) Change functions to ANSI C.
9452 *) Fix typos in error codes.
9453 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
9455 *) Remove defunct assembler files from Configure.
9458 *) SPARC v8 assembler BIGNUM implementation.
9459 [Andy Polyakov <appro@fy.chalmers.se>]
9461 *) Support for Certificate Policies extension: both print and set.
9462 Various additions to support the r2i method this uses.
9465 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
9466 return a const string when you are expecting an allocated buffer.
9469 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
9470 types DirectoryString and DisplayText.
9473 *) Add code to allow r2i extensions to access the configuration database,
9474 add an LHASH database driver and add several ctx helper functions.
9477 *) Fix an evil bug in bn_expand2() which caused various BN functions to
9478 fail when they extended the size of a BIGNUM.
9481 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
9482 support typesafe stack.
9485 *) Fix typo in SSL_[gs]et_options().
9486 [Nils Frostberg <nils@medcom.se>]
9488 *) Delete various functions and files that belonged to the (now obsolete)
9489 old X509V3 handling code.
9492 *) New Configure option "rsaref".
9495 *) Don't auto-generate pem.h.
9498 *) Introduce type-safe ASN.1 SETs.
9501 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
9502 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
9504 *) Introduce type-safe STACKs. This will almost certainly break lots of code
9505 that links with OpenSSL (well at least cause lots of warnings), but fear
9506 not: the conversion is trivial, and it eliminates loads of evil casts. A
9507 few STACKed things have been converted already. Feel free to convert more.
9508 In the fullness of time, I'll do away with the STACK type altogether.
9511 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
9512 specified in <certfile> by updating the entry in the index.txt file.
9513 This way one no longer has to edit the index.txt file manually for
9514 revoking a certificate. The -revoke option does the gory details now.
9515 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
9517 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
9518 `-text' option at all and this way the `-noout -text' combination was
9519 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
9520 [Ralf S. Engelschall]
9522 *) Make sure a corresponding plain text error message exists for the
9523 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
9524 verify callback function determined that a certificate was revoked.
9525 [Ralf S. Engelschall]
9527 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
9528 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
9529 all available cipers including rc5, which was forgotten until now.
9530 In order to let the testing shell script know which algorithms
9531 are available, a new (up to now undocumented) command
9532 "openssl list-cipher-commands" is used.
9535 *) Bugfix: s_client occasionally would sleep in select() when
9536 it should have checked SSL_pending() first.
9539 *) New functions DSA_do_sign and DSA_do_verify to provide access to
9540 the raw DSA values prior to ASN.1 encoding.
9543 *) Tweaks to Configure
9544 [Niels Poppe <niels@netbox.org>]
9546 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
9550 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
9553 *) New config option to avoid instructions that are illegal on the 80386.
9554 The default code is faster, but requires at least a 486.
9557 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
9558 SSL2_SERVER_VERSION (not used at all) macros, which are now the
9559 same as SSL2_VERSION anyway.
9562 *) New "-showcerts" option for s_client.
9565 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
9566 application. Various cleanups and fixes.
9569 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
9570 modify error routines to work internally. Add error codes and PBE init
9571 to library startup routines.
9574 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
9575 packing functions to asn1 and evp. Changed function names and error
9576 codes along the way.
9579 *) PKCS12 integration: and so it begins... First of several patches to
9580 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
9581 objects to objects.h
9584 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
9585 and display support for Thawte strong extranet extension.
9588 *) Add LinuxPPC support.
9589 [Jeff Dubrule <igor@pobox.org>]
9591 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
9592 bn_div_words in alpha.s.
9593 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
9595 *) Make sure the RSA OAEP test is skipped under -DRSAref because
9596 OAEP isn't supported when OpenSSL is built with RSAref.
9597 [Ulf Moeller <ulf@fitug.de>]
9599 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
9600 so they no longer are missing under -DNOPROTO.
9601 [Soren S. Jorvang <soren@t.dk>]
9604 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
9606 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
9607 doesn't work when the session is reused. Coming soon!
9610 *) Fix a security hole, that allows sessions to be reused in the wrong
9611 context thus bypassing client cert protection! All software that uses
9612 client certs and session caches in multiple contexts NEEDS PATCHING to
9613 allow session reuse! A fuller solution is in the works.
9614 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
9616 *) Some more source tree cleanups (removed obsolete files
9617 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
9618 permission on "config" script to be executable) and a fix for the INSTALL
9620 [Ulf Moeller <ulf@fitug.de>]
9622 *) Remove some legacy and erroneous uses of malloc, free instead of
9624 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
9626 *) Make rsa_oaep_test return non-zero on error.
9627 [Ulf Moeller <ulf@fitug.de>]
9629 *) Add support for native Solaris shared libraries. Configure
9630 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
9631 if someone would make that last step automatic.
9632 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
9634 *) ctx_size was not built with the right compiler during "make links". Fixed.
9637 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
9638 except NULL ciphers". This means the default cipher list will no longer
9639 enable NULL ciphers. They need to be specifically enabled e.g. with
9640 the string "DEFAULT:eNULL".
9643 *) Fix to RSA private encryption routines: if p < q then it would
9644 occasionally produce an invalid result. This will only happen with
9645 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9648 *) Be less restrictive and allow also `perl util/perlpath.pl
9649 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9650 because this way one can also use an interpreter named `perl5' (which is
9651 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9652 installed as `perl').
9653 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9655 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9656 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9658 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9659 advapi32.lib to Win32 build and change the pem test comparision
9660 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9661 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9662 and crypto/des/ede_cbcm_enc.c.
9665 *) DES quad checksum was broken on big-endian architectures. Fixed.
9668 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9669 Win32 test batch file so it (might) work again. The Win32 test batch file
9670 is horrible: I feel ill....
9673 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
9674 in e_os.h. Audit of header files to check ANSI and non ANSI
9675 sections: 10 functions were absent from non ANSI section and not exported
9676 from Windows DLLs. Fixed up libeay.num for new functions.
9679 *) Make `openssl version' output lines consistent.
9680 [Ralf S. Engelschall]
9682 *) Fix Win32 symbol export lists for BIO functions: Added
9683 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
9684 to ms/libeay{16,32}.def.
9685 [Ralf S. Engelschall]
9687 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
9688 fine under Unix and passes some trivial tests I've now added. But the
9689 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
9690 added to make sure no one expects that this stuff really works in the
9691 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
9692 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
9694 [Ralf S. Engelschall]
9696 *) Fix the generation of two part addresses in perl.
9697 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
9699 *) Add config entry for Linux on MIPS.
9700 [John Tobey <jtobey@channel1.com>]
9702 *) Make links whenever Configure is run, unless we are on Windoze.
9705 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
9706 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
9710 *) Add a useful kludge to allow package maintainers to specify compiler and
9711 other platforms details on the command line without having to patch the
9712 Configure script everytime: One now can use ``perl Configure
9713 <id>:<details>'', i.e. platform ids are allowed to have details appended
9714 to them (seperated by colons). This is treated as there would be a static
9715 pre-configured entry in Configure's %table under key <id> with value
9716 <details> and ``perl Configure <id>'' is called. So, when you want to
9717 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
9718 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
9719 now, which overrides the FreeBSD-elf entry on-the-fly.
9720 [Ralf S. Engelschall]
9722 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
9725 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
9726 on the `perl Configure ...' command line. This way one can compile
9727 OpenSSL libraries with Position Independent Code (PIC) which is needed
9728 for linking it into DSOs.
9729 [Ralf S. Engelschall]
9731 *) Remarkably, export ciphers were totally broken and no-one had noticed!
9735 *) Cleaned up the LICENSE document: The official contact for any license
9736 questions now is the OpenSSL core team under openssl-core@openssl.org.
9737 And add a paragraph about the dual-license situation to make sure people
9738 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
9739 to the OpenSSL toolkit.
9740 [Ralf S. Engelschall]
9742 *) General source tree makefile cleanups: Made `making xxx in yyy...'
9743 display consistent in the source tree and replaced `/bin/rm' by `rm'.
9744 Additonally cleaned up the `make links' target: Remove unnecessary
9745 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
9746 to speed processing and no longer clutter the display with confusing
9747 stuff. Instead only the actually done links are displayed.
9748 [Ralf S. Engelschall]
9750 *) Permit null encryption ciphersuites, used for authentication only. It used
9751 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
9752 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
9756 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
9757 signed attributes when verifying signatures (this would break them),
9758 the detached data encoding was wrong and public keys obtained using
9759 X509_get_pubkey() weren't freed.
9762 *) Add text documentation for the BUFFER functions. Also added a work around
9763 to a Win95 console bug. This was triggered by the password read stuff: the
9764 last character typed gets carried over to the next fread(). If you were
9765 generating a new cert request using 'req' for example then the last
9766 character of the passphrase would be CR which would then enter the first
9770 *) Added the new `Includes OpenSSL Cryptography Software' button as
9771 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
9772 button and can be used by applications based on OpenSSL to show the
9773 relationship to the OpenSSL project.
9774 [Ralf S. Engelschall]
9776 *) Remove confusing variables in function signatures in files
9777 ssl/ssl_lib.c and ssl/ssl.h.
9778 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9780 *) Don't install bss_file.c under PREFIX/include/
9781 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9783 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
9784 functions that return function pointers and has support for NT specific
9785 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
9786 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
9787 unsigned to signed types: this was killing the Win32 compile.
9790 *) Add new certificate file to stack functions,
9791 SSL_add_dir_cert_subjects_to_stack() and
9792 SSL_add_file_cert_subjects_to_stack(). These largely supplant
9793 SSL_load_client_CA_file(), and can be used to add multiple certs easily
9794 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
9795 This means that Apache-SSL and similar packages don't have to mess around
9796 to add as many CAs as they want to the preferred list.
9799 *) Experiment with doxygen documentation. Currently only partially applied to
9801 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
9802 openssl.doxy as the configuration file.
9805 *) Get rid of remaining C++-style comments which strict C compilers hate.
9806 [Ralf S. Engelschall, pointed out by Carlos Amengual]
9808 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
9809 compiled in by default: it has problems with large keys.
9812 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
9813 DH private keys and/or callback functions which directly correspond to
9814 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
9815 is needed for applications which have to configure certificates on a
9816 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
9818 For the RSA certificate situation is makes no difference, but
9819 for the DSA certificate situation this fixes the "no shared cipher"
9820 problem where the OpenSSL cipher selection procedure failed because the
9821 temporary keys were not overtaken from the context and the API provided
9822 no way to reconfigure them.
9823 The new functions now let applications reconfigure the stuff and they
9824 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
9825 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
9826 non-public-API function ssl_cert_instantiate() is used as a helper
9827 function and also to reduce code redundancy inside ssl_rsa.c.
9828 [Ralf S. Engelschall]
9830 *) Move s_server -dcert and -dkey options out of the undocumented feature
9831 area because they are useful for the DSA situation and should be
9832 recognized by the users.
9833 [Ralf S. Engelschall]
9835 *) Fix the cipher decision scheme for export ciphers: the export bits are
9836 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
9837 SSL_EXP_MASK. So, the original variable has to be used instead of the
9838 already masked variable.
9839 [Richard Levitte <levitte@stacken.kth.se>]
9841 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
9842 [Richard Levitte <levitte@stacken.kth.se>]
9844 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
9845 from `int' to `unsigned int' because it's a length and initialized by
9846 EVP_DigestFinal() which expects an `unsigned int *'.
9847 [Richard Levitte <levitte@stacken.kth.se>]
9849 *) Don't hard-code path to Perl interpreter on shebang line of Configure
9850 script. Instead use the usual Shell->Perl transition trick.
9851 [Ralf S. Engelschall]
9853 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
9854 (in addition to RSA certificates) to match the behaviour of `openssl dsa
9855 -noout -modulus' as it's already the case for `openssl rsa -noout
9856 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
9857 currently the public key is printed (a decision which was already done by
9858 `openssl dsa -modulus' in the past) which serves a similar purpose.
9859 Additionally the NO_RSA no longer completely removes the whole -modulus
9860 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
9862 [Ralf S. Engelschall]
9864 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
9865 BIO. See the source (crypto/evp/bio_ok.c) for more info.
9866 [Arne Ansper <arne@ats.cyber.ee>]
9868 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
9869 to be added. Now both 'req' and 'ca' can use new objects defined in the
9873 *) Add cool BIO that does syslog (or event log on NT).
9874 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
9876 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
9877 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
9878 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
9879 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
9882 *) Add preliminary config info for new extension code.
9885 *) Make RSA_NO_PADDING really use no padding.
9886 [Ulf Moeller <ulf@fitug.de>]
9888 *) Generate errors when private/public key check is done.
9891 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
9892 for some CRL extensions and new objects added.
9895 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
9896 key usage extension and fuller support for authority key id.
9899 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
9900 padding method for RSA, which is recommended for new applications in PKCS
9901 #1 v2.0 (RFC 2437, October 1998).
9902 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
9903 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
9904 against Bleichbacher's attack on RSA.
9905 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
9908 *) Updates to the new SSL compression code
9909 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9911 *) Fix so that the version number in the master secret, when passed
9912 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
9913 (because the server will not accept higher), that the version number
9914 is 0x03,0x01, not 0x03,0x00
9915 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9917 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
9918 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
9919 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
9922 *) Support for RAW extensions where an arbitrary extension can be
9923 created by including its DER encoding. See apps/openssl.cnf for
9927 *) Make sure latest Perl versions don't interpret some generated C array
9928 code as Perl array code in the crypto/err/err_genc.pl script.
9929 [Lars Weber <3weber@informatik.uni-hamburg.de>]
9931 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
9932 not many people have the assembler. Various Win32 compilation fixes and
9933 update to the INSTALL.W32 file with (hopefully) more accurate Win32
9937 *) Modify configure script 'Configure' to automatically create crypto/date.h
9938 file under Win32 and also build pem.h from pem.org. New script
9939 util/mkfiles.pl to create the MINFO file on environments that can't do a
9940 'make files': perl util/mkfiles.pl >MINFO should work.
9943 *) Major rework of DES function declarations, in the pursuit of correctness
9944 and purity. As a result, many evil casts evaporated, and some weirdness,
9945 too. You may find this causes warnings in your code. Zapping your evil
9946 casts will probably fix them. Mostly.
9949 *) Fix for a typo in asn1.h. Bug fix to object creation script
9950 obj_dat.pl. It considered a zero in an object definition to mean
9951 "end of object": none of the objects in objects.h have any zeros
9952 so it wasn't spotted.
9953 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
9955 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
9956 Masking (CBCM). In the absence of test vectors, the best I have been able
9957 to do is check that the decrypt undoes the encrypt, so far. Send me test
9958 vectors if you have them.
9961 *) Correct calculation of key length for export ciphers (too much space was
9962 allocated for null ciphers). This has not been tested!
9965 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
9966 message is now correct (it understands "crypto" and "ssl" on its
9967 command line). There is also now an "update" option. This will update
9968 the util/ssleay.num and util/libeay.num files with any new functions.
9970 perl util/mkdef.pl crypto ssl update
9971 it will update them.
9974 *) Overhauled the Perl interface (perl/*):
9975 - ported BN stuff to OpenSSL's different BN library
9976 - made the perl/ source tree CVS-aware
9977 - renamed the package from SSLeay to OpenSSL (the files still contain
9978 their history because I've copied them in the repository)
9979 - removed obsolete files (the test scripts will be replaced
9980 by better Test::Harness variants in the future)
9981 [Ralf S. Engelschall]
9983 *) First cut for a very conservative source tree cleanup:
9984 1. merge various obsolete readme texts into doc/ssleay.txt
9985 where we collect the old documents and readme texts.
9986 2. remove the first part of files where I'm already sure that we no
9987 longer need them because of three reasons: either they are just temporary
9988 files which were left by Eric or they are preserved original files where
9989 I've verified that the diff is also available in the CVS via "cvs diff
9990 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
9991 the crypto/md/ stuff).
9992 [Ralf S. Engelschall]
9994 *) More extension code. Incomplete support for subject and issuer alt
9995 name, issuer and authority key id. Change the i2v function parameters
9996 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
9997 what that's for :-) Fix to ASN1 macro which messed up
9998 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
10001 *) Preliminary support for ENUMERATED type. This is largely copied from the
10005 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
10006 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10008 *) Make sure `make rehash' target really finds the `openssl' program.
10009 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10011 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
10012 like to hear about it if this slows down other processors.
10015 *) Add CygWin32 platform information to Configure script.
10016 [Alan Batie <batie@aahz.jf.intel.com>]
10018 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
10019 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
10021 *) New program nseq to manipulate netscape certificate sequences
10024 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
10028 *) Fixes to BN code. Previously the default was to define BN_RECURSION
10029 but the BN code had some problems that would cause failures when
10030 doing certificate verification and some other functions.
10031 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10033 *) Add ASN1 and PEM code to support netscape certificate sequences.
10036 *) Add ASN1 and PEM code to support netscape certificate sequences.
10039 *) Add several PKIX and private extended key usage OIDs.
10042 *) Modify the 'ca' program to handle the new extension code. Modify
10043 openssl.cnf for new extension format, add comments.
10046 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
10047 and add a sample to openssl.cnf so req -x509 now adds appropriate
10051 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
10052 error code, add initial support to X509_print() and x509 application.
10055 *) Takes a deep breath and start addding X509 V3 extension support code. Add
10056 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
10057 stuff is currently isolated and isn't even compiled yet.
10060 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
10061 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
10062 Removed the versions check from X509 routines when loading extensions:
10063 this allows certain broken certificates that don't set the version
10064 properly to be processed.
10067 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
10068 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
10069 can still be regenerated with "make depend".
10072 *) Spelling mistake in C version of CAST-128.
10073 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
10075 *) Changes to the error generation code. The perl script err-code.pl
10076 now reads in the old error codes and retains the old numbers, only
10077 adding new ones if necessary. It also only changes the .err files if new
10078 codes are added. The makefiles have been modified to only insert errors
10079 when needed (to avoid needlessly modifying header files). This is done
10080 by only inserting errors if the .err file is newer than the auto generated
10081 C file. To rebuild all the error codes from scratch (the old behaviour)
10082 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
10083 or delete all the .err files.
10086 *) CAST-128 was incorrectly implemented for short keys. The C version has
10087 been fixed, but is untested. The assembler versions are also fixed, but
10088 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
10089 to regenerate it if needed.
10090 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
10091 Hagino <itojun@kame.net>]
10093 *) File was opened incorrectly in randfile.c.
10094 [Ulf Möller <ulf@fitug.de>]
10096 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
10097 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
10098 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
10099 al: it's just almost always a UTCTime. Note this patch adds new error
10100 codes so do a "make errors" if there are problems.
10103 *) Correct Linux 1 recognition in config.
10104 [Ulf Möller <ulf@fitug.de>]
10106 *) Remove pointless MD5 hash when using DSA keys in ca.
10107 [Anonymous <nobody@replay.com>]
10109 *) Generate an error if given an empty string as a cert directory. Also
10110 generate an error if handed NULL (previously returned 0 to indicate an
10111 error, but didn't set one).
10112 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
10114 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
10117 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
10118 parameters. This was causing a warning which killed off the Win32 compile.
10121 *) Remove C++ style comments from crypto/bn/bn_local.h.
10122 [Neil Costigan <neil.costigan@celocom.com>]
10124 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
10125 based on a text string, looking up short and long names and finally
10126 "dot" format. The "dot" format stuff didn't work. Added new function
10127 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
10128 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
10129 OID is not part of the table.
10132 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
10133 X509_LOOKUP_by_alias().
10136 *) Sort openssl functions by name.
10139 *) Get the gendsa program working (hopefully) and add it to app list. Remove
10140 encryption from sample DSA keys (in case anyone is interested the password
10144 *) Make _all_ *_free functions accept a NULL pointer.
10145 [Frans Heymans <fheymans@isaserver.be>]
10147 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
10149 [Anonymous <nobody@replay.com>]
10151 *) s_server should send the CAfile as acceptable CAs, not its own cert.
10152 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10154 *) Don't blow it for numeric -newkey arguments to apps/req.
10155 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10157 *) Temp key "for export" tests were wrong in s3_srvr.c.
10158 [Anonymous <nobody@replay.com>]
10160 *) Add prototype for temp key callback functions
10161 SSL_CTX_set_tmp_{rsa,dh}_callback().
10164 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
10165 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
10168 *) X509_name_add_entry() freed the wrong thing after an error.
10169 [Arne Ansper <arne@ats.cyber.ee>]
10171 *) rsa_eay.c would attempt to free a NULL context.
10172 [Arne Ansper <arne@ats.cyber.ee>]
10174 *) BIO_s_socket() had a broken should_retry() on Windoze.
10175 [Arne Ansper <arne@ats.cyber.ee>]
10177 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
10178 [Arne Ansper <arne@ats.cyber.ee>]
10180 *) Make sure the already existing X509_STORE->depth variable is initialized
10181 in X509_STORE_new(), but document the fact that this variable is still
10182 unused in the certificate verification process.
10183 [Ralf S. Engelschall]
10185 *) Fix the various library and apps files to free up pkeys obtained from
10186 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
10189 *) Fix reference counting in X509_PUBKEY_get(). This makes
10190 demos/maurice/example2.c work, amongst others, probably.
10191 [Steve Henson and Ben Laurie]
10193 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
10194 `openssl' and second, the shortcut symlinks for the `openssl <command>'
10195 are no longer created. This way we have a single and consistent command
10196 line interface `openssl <command>', similar to `cvs <command>'.
10197 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
10199 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
10200 BIT STRING wrapper always have zero unused bits.
10203 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
10206 *) Make the top-level INSTALL documentation easier to understand.
10209 *) Makefiles updated to exit if an error occurs in a sub-directory
10210 make (including if user presses ^C) [Paul Sutton]
10212 *) Make Montgomery context stuff explicit in RSA data structure.
10215 *) Fix build order of pem and err to allow for generated pem.h.
10218 *) Fix renumbering bug in X509_NAME_delete_entry().
10221 *) Enhanced the err-ins.pl script so it makes the error library number
10222 global and can add a library name. This is needed for external ASN1 and
10223 other error libraries.
10226 *) Fixed sk_insert which never worked properly.
10229 *) Fix ASN1 macros so they can handle indefinite length construted
10230 EXPLICIT tags. Some non standard certificates use these: they can now
10234 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
10235 into a single doc/ssleay.txt bundle. This way the information is still
10236 preserved but no longer messes up this directory. Now it's new room for
10237 the new set of documenation files.
10238 [Ralf S. Engelschall]
10240 *) SETs were incorrectly DER encoded. This was a major pain, because they
10241 shared code with SEQUENCEs, which aren't coded the same. This means that
10242 almost everything to do with SETs or SEQUENCEs has either changed name or
10243 number of arguments.
10244 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
10246 *) Fix test data to work with the above.
10249 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
10250 was already fixed by Eric for 0.9.1 it seems.
10251 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
10253 *) Autodetect FreeBSD3.
10256 *) Fix various bugs in Configure. This affects the following platforms:
10260 unixware-2.0-pentium
10264 *) Eliminate generated files from CVS. Reorder tests to regenerate files
10265 before they are needed.
10268 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
10272 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
10274 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
10275 changed SSLeay to OpenSSL in version strings.
10276 [Ralf S. Engelschall]
10278 *) Some fixups to the top-level documents.
10281 *) Fixed the nasty bug where rsaref.h was not found under compile-time
10282 because the symlink to include/ was missing.
10283 [Ralf S. Engelschall]
10285 *) Incorporated the popular no-RSA/DSA-only patches
10286 which allow to compile a RSA-free SSLeay.
10287 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
10289 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
10290 when "ssleay" is still not found.
10291 [Ralf S. Engelschall]
10293 *) Added more platforms to Configure: Cray T3E, HPUX 11,
10294 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
10296 *) Updated the README file.
10297 [Ralf S. Engelschall]
10299 *) Added various .cvsignore files in the CVS repository subdirs
10300 to make a "cvs update" really silent.
10301 [Ralf S. Engelschall]
10303 *) Recompiled the error-definition header files and added
10304 missing symbols to the Win32 linker tables.
10305 [Ralf S. Engelschall]
10307 *) Cleaned up the top-level documents;
10308 o new files: CHANGES and LICENSE
10309 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
10310 o merged COPYRIGHT into LICENSE
10311 o removed obsolete TODO file
10312 o renamed MICROSOFT to INSTALL.W32
10313 [Ralf S. Engelschall]
10315 *) Removed dummy files from the 0.9.1b source tree:
10316 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
10317 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
10318 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
10319 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
10320 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
10321 [Ralf S. Engelschall]
10323 *) Added various platform portability fixes.
10326 *) The Genesis of the OpenSSL rpject:
10327 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
10328 Young and Tim J. Hudson created while they were working for C2Net until
10330 [The OpenSSL Project]
10333 Changes between 0.9.0b and 0.9.1b [not released]
10335 *) Updated a few CA certificates under certs/
10338 *) Changed some BIGNUM api stuff.
10341 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
10342 DGUX x86, Linux Alpha, etc.
10345 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
10346 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
10350 *) Add -strparse option to asn1pars program which parses nested
10352 [Dr Stephen Henson <shenson@bigfoot.com>]
10354 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
10357 *) DSA fix for "ca" program.
10360 *) Added "-genkey" option to "dsaparam" program.
10363 *) Added RIPE MD160 (rmd160) message digest.
10366 *) Added -a (all) option to "ssleay version" command.
10369 *) Added PLATFORM define which is the id given to Configure.
10372 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
10375 *) Extended the ASN.1 parser routines.
10378 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
10381 *) Added a BN_CTX to the BN library.
10384 *) Fixed the weak key values in DES library
10387 *) Changed API in EVP library for cipher aliases.
10390 *) Added support for RC2/64bit cipher.
10393 *) Converted the lhash library to the crypto/mem.c functions.
10396 *) Added more recognized ASN.1 object ids.
10399 *) Added more RSA padding checks for SSL/TLS.
10402 *) Added BIO proxy/filter functionality.
10405 *) Added extra_certs to SSL_CTX which can be used
10406 send extra CA certificates to the client in the CA cert chain sending
10407 process. It can be configured with SSL_CTX_add_extra_chain_cert().
10410 *) Now Fortezza is denied in the authentication phase because
10411 this is key exchange mechanism is not supported by SSLeay at all.
10414 *) Additional PKCS1 checks.
10417 *) Support the string "TLSv1" for all TLS v1 ciphers.
10420 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
10421 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
10424 *) Fixed a few memory leaks.
10427 *) Fixed various code and comment typos.
10430 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
10431 bytes sent in the client random.
10432 [Edward Bishop <ebishop@spyglass.com>]