5 Changes between 1.0.1r and 1.0.1s [xx XXX xxxx]
7 *) Disable SRP fake user seed to address a server memory leak.
9 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
11 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
12 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
13 was changed to ignore the "fake user" SRP seed, even if the seed
16 Users should use SRP_VBASE_get1_by_user instead. Note that in
17 SRP_VBASE_get1_by_user, caller must free the returned value. Note
18 also that even though configuring the SRP seed attempts to hide
19 invalid usernames by continuing the handshake with fake
20 credentials, this behaviour is not constant time and no strong
21 guarantees are made that the handshake is indistinguishable from
26 Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
28 *) Protection for DH small subgroup attacks
30 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
31 switched on by default and cannot be disabled. This could have some
35 *) SSLv2 doesn't block disabled ciphers
37 A malicious client can negotiate SSLv2 ciphers that have been disabled on
38 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
39 been disabled, provided that the SSLv2 protocol was not also disabled via
42 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
43 and Sebastian Schinzel.
47 *) Reject DH handshakes with parameters shorter than 1024 bits.
50 Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
52 *) Certificate verify crash with missing PSS parameter
54 The signature verification routines will crash with a NULL pointer
55 dereference if presented with an ASN.1 signature using the RSA PSS
56 algorithm and absent mask generation function parameter. Since these
57 routines are used to verify certificate signature algorithms this can be
58 used to crash any certificate verification operation and exploited in a
59 DoS attack. Any application which performs certificate verification is
60 vulnerable including OpenSSL clients and servers which enable client
63 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
67 *) X509_ATTRIBUTE memory leak
69 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
70 memory. This structure is used by the PKCS#7 and CMS routines so any
71 application which reads PKCS#7 or CMS data from untrusted sources is
72 affected. SSL/TLS is not affected.
74 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
79 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
80 This changes the decoding behaviour for some invalid messages,
81 though the change is mostly in the more lenient direction, and
82 legacy behaviour is preserved as much as possible.
85 *) In DSA_generate_parameters_ex, if the provided seed is too short,
86 use a random seed, as already documented.
87 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
89 Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
91 *) Alternate chains certificate forgery
93 During certificate verfification, OpenSSL will attempt to find an
94 alternative certificate chain if the first attempt to build such a chain
95 fails. An error in the implementation of this logic can mean that an
96 attacker could cause certain checks on untrusted certificates to be
97 bypassed, such as the CA flag, enabling them to use a valid leaf
98 certificate to act as a CA and "issue" an invalid certificate.
100 This issue was reported to OpenSSL by Adam Langley/David Benjamin
105 *) Race condition handling PSK identify hint
107 If PSK identity hints are received by a multi-threaded client then
108 the values are wrongly updated in the parent SSL_CTX structure. This can
109 result in a race condition potentially leading to a double free of the
114 Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
115 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
116 incompatibility in the handling of HMAC. The previous ABI has now been
119 Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
121 *) Malformed ECParameters causes infinite loop
123 When processing an ECParameters structure OpenSSL enters an infinite loop
124 if the curve specified is over a specially malformed binary polynomial
127 This can be used to perform denial of service against any
128 system which processes public keys, certificate requests or
129 certificates. This includes TLS clients and TLS servers with
130 client authentication enabled.
132 This issue was reported to OpenSSL by Joseph Barr-Pixton.
136 *) Exploitable out-of-bounds read in X509_cmp_time
138 X509_cmp_time does not properly check the length of the ASN1_TIME
139 string and can read a few bytes out of bounds. In addition,
140 X509_cmp_time accepts an arbitrary number of fractional seconds in the
143 An attacker can use this to craft malformed certificates and CRLs of
144 various sizes and potentially cause a segmentation fault, resulting in
145 a DoS on applications that verify certificates or CRLs. TLS clients
146 that verify CRLs are affected. TLS clients and servers with client
147 authentication enabled may be affected if they use custom verification
150 This issue was reported to OpenSSL by Robert Swiecki (Google), and
151 independently by Hanno Böck.
155 *) PKCS7 crash with missing EnvelopedContent
157 The PKCS#7 parsing code does not handle missing inner EncryptedContent
158 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
159 with missing content and trigger a NULL pointer dereference on parsing.
161 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
162 structures from untrusted sources are affected. OpenSSL clients and
163 servers are not affected.
165 This issue was reported to OpenSSL by Michal Zalewski (Google).
169 *) CMS verify infinite loop with unknown hash function
171 When verifying a signedData message the CMS code can enter an infinite loop
172 if presented with an unknown hash function OID. This can be used to perform
173 denial of service against any system which verifies signedData messages using
175 This issue was reported to OpenSSL by Johannes Bauer.
179 *) Race condition handling NewSessionTicket
181 If a NewSessionTicket is received by a multi-threaded client when attempting to
182 reuse a previous ticket then a race condition can occur potentially leading to
183 a double free of the ticket data.
187 *) Reject DH handshakes with parameters shorter than 768 bits.
188 [Kurt Roeckx and Emilia Kasper]
190 *) dhparam: generate 2048-bit parameters by default.
191 [Kurt Roeckx and Emilia Kasper]
193 Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
195 *) Segmentation fault in ASN1_TYPE_cmp fix
197 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
198 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
199 certificate signature algorithm consistency this can be used to crash any
200 certificate verification operation and exploited in a DoS attack. Any
201 application which performs certificate verification is vulnerable including
202 OpenSSL clients and servers which enable client authentication.
206 *) ASN.1 structure reuse memory corruption fix
208 Reusing a structure in ASN.1 parsing may allow an attacker to cause
209 memory corruption via an invalid write. Such reuse is and has been
210 strongly discouraged and is believed to be rare.
212 Applications that parse structures containing CHOICE or ANY DEFINED BY
213 components may be affected. Certificate parsing (d2i_X509 and related
214 functions) are however not affected. OpenSSL clients and servers are
219 *) PKCS7 NULL pointer dereferences fix
221 The PKCS#7 parsing code does not handle missing outer ContentInfo
222 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
223 missing content and trigger a NULL pointer dereference on parsing.
225 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
226 otherwise parse PKCS#7 structures from untrusted sources are
227 affected. OpenSSL clients and servers are not affected.
229 This issue was reported to OpenSSL by Michal Zalewski (Google).
233 *) DoS via reachable assert in SSLv2 servers fix
235 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
236 servers that both support SSLv2 and enable export cipher suites by sending
237 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
239 This issue was discovered by Sean Burford (Google) and Emilia Käsper
240 (OpenSSL development team).
244 *) Use After Free following d2i_ECPrivatekey error fix
246 A malformed EC private key file consumed via the d2i_ECPrivateKey function
247 could cause a use after free condition. This, in turn, could cause a double
248 free in several private key parsing functions (such as d2i_PrivateKey
249 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
250 for applications that receive EC private keys from untrusted
251 sources. This scenario is considered rare.
253 This issue was discovered by the BoringSSL project and fixed in their
258 *) X509_to_X509_REQ NULL pointer deref fix
260 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
261 the certificate key is invalid. This function is rarely used in practice.
263 This issue was discovered by Brian Carpenter.
267 *) Removed the export ciphers from the DEFAULT ciphers
270 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
272 *) Build fixes for the Windows and OpenVMS platforms
273 [Matt Caswell and Richard Levitte]
275 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
277 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
278 message can cause a segmentation fault in OpenSSL due to a NULL pointer
279 dereference. This could lead to a Denial Of Service attack. Thanks to
280 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
284 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
285 dtls1_buffer_record function under certain conditions. In particular this
286 could occur if an attacker sent repeated DTLS records with the same
287 sequence number but for the next epoch. The memory leak could be exploited
288 by an attacker in a Denial of Service attack through memory exhaustion.
289 Thanks to Chris Mueller for reporting this issue.
293 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
294 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
295 method would be set to NULL which could later result in a NULL pointer
296 dereference. Thanks to Frank Schmirler for reporting this issue.
300 *) Abort handshake if server key exchange message is omitted for ephemeral
303 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
304 reporting this issue.
308 *) Remove non-export ephemeral RSA code on client and server. This code
309 violated the TLS standard by allowing the use of temporary RSA keys in
310 non-export ciphersuites and could be used by a server to effectively
311 downgrade the RSA key length used to a value smaller than the server
312 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
313 INRIA or reporting this issue.
317 *) Fixed issue where DH client certificates are accepted without verification.
318 An OpenSSL server will accept a DH certificate for client authentication
319 without the certificate verify message. This effectively allows a client to
320 authenticate without the use of a private key. This only affects servers
321 which trust a client certificate authority which issues certificates
322 containing DH keys: these are extremely rare and hardly ever encountered.
323 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
328 *) Ensure that the session ID context of an SSL is updated when its
329 SSL_CTX is updated via SSL_set_SSL_CTX.
331 The session ID context is typically set from the parent SSL_CTX,
332 and can vary with the CTX.
335 *) Fix various certificate fingerprint issues.
337 By using non-DER or invalid encodings outside the signed portion of a
338 certificate the fingerprint can be changed without breaking the signature.
339 Although no details of the signed portion of the certificate can be changed
340 this can cause problems with some applications: e.g. those using the
341 certificate fingerprint for blacklists.
343 1. Reject signatures with non zero unused bits.
345 If the BIT STRING containing the signature has non zero unused bits reject
346 the signature. All current signature algorithms require zero unused bits.
348 2. Check certificate algorithm consistency.
350 Check the AlgorithmIdentifier inside TBS matches the one in the
351 certificate signature. NB: this will result in signature failure
352 errors for some broken certificates.
354 Thanks to Konrad Kraszewski from Google for reporting this issue.
356 3. Check DSA/ECDSA signatures use DER.
358 Reencode DSA/ECDSA signatures and compare with the original received
359 signature. Return an error if there is a mismatch.
361 This will reject various cases including garbage after signature
362 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
363 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
364 (negative or with leading zeroes).
366 Further analysis was conducted and fixes were developed by Stephen Henson
367 of the OpenSSL core team.
372 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
373 results on some platforms, including x86_64. This bug occurs at random
374 with a very low probability, and is not known to be exploitable in any
375 way, though its exact impact is difficult to determine. Thanks to Pieter
376 Wuille (Blockstream) who reported this issue and also suggested an initial
377 fix. Further analysis was conducted by the OpenSSL development team and
378 Adam Langley of Google. The final fix was developed by Andy Polyakov of
379 the OpenSSL core team.
383 *) Do not resume sessions on the server if the negotiated protocol
384 version does not match the session's version. Resuming with a different
385 version, while not strictly forbidden by the RFC, is of questionable
386 sanity and breaks all known clients.
387 [David Benjamin, Emilia Käsper]
389 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
390 early CCS messages during renegotiation. (Note that because
391 renegotiation is encrypted, this early CCS was not exploitable.)
394 *) Tighten client-side session ticket handling during renegotiation:
395 ensure that the client only accepts a session ticket if the server sends
396 the extension anew in the ServerHello. Previously, a TLS client would
397 reuse the old extension state and thus accept a session ticket if one was
398 announced in the initial ServerHello.
400 Similarly, ensure that the client requires a session ticket if one
401 was advertised in the ServerHello. Previously, a TLS client would
402 ignore a missing NewSessionTicket message.
405 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
409 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
410 sends a carefully crafted handshake message, to cause OpenSSL to fail
411 to free up to 64k of memory causing a memory leak. This could be
412 exploited in a Denial Of Service attack. This issue affects OpenSSL
413 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
414 whether SRTP is used or configured. Implementations of OpenSSL that
415 have been compiled with OPENSSL_NO_SRTP defined are not affected.
417 The fix was developed by the OpenSSL team.
421 *) Session Ticket Memory Leak.
423 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
424 integrity of that ticket is first verified. In the event of a session
425 ticket integrity check failing, OpenSSL will fail to free memory
426 causing a memory leak. By sending a large number of invalid session
427 tickets an attacker could exploit this issue in a Denial Of Service
432 *) Build option no-ssl3 is incomplete.
434 When OpenSSL is configured with "no-ssl3" as a build option, servers
435 could accept and complete a SSL 3.0 handshake, and clients could be
436 configured to send them.
438 [Akamai and the OpenSSL team]
440 *) Add support for TLS_FALLBACK_SCSV.
441 Client applications doing fallback retries should call
442 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
444 [Adam Langley, Bodo Moeller]
446 *) Add additional DigestInfo checks.
448 Reencode DigestInto in DER and check against the original when
449 verifying RSA signature: this will reject any improperly encoded
450 DigestInfo structures.
452 Note: this is a precautionary measure and no attacks are currently known.
456 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
458 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
459 SRP code can be overrun an internal buffer. Add sanity check that
460 g, A, B < N to SRP code.
462 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
463 Group for discovering this issue.
467 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
468 TLS 1.0 instead of higher protocol versions when the ClientHello message
469 is badly fragmented. This allows a man-in-the-middle attacker to force a
470 downgrade to TLS 1.0 even if both the server and the client support a
471 higher protocol version, by modifying the client's TLS records.
473 Thanks to David Benjamin and Adam Langley (Google) for discovering and
474 researching this issue.
478 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
479 to a denial of service attack. A malicious server can crash the client
480 with a null pointer dereference (read) by specifying an anonymous (EC)DH
481 ciphersuite and sending carefully crafted handshake messages.
483 Thanks to Felix Gröbert (Google) for discovering and researching this
488 *) By sending carefully crafted DTLS packets an attacker could cause openssl
489 to leak memory. This can be exploited through a Denial of Service attack.
490 Thanks to Adam Langley for discovering and researching this issue.
494 *) An attacker can force openssl to consume large amounts of memory whilst
495 processing DTLS handshake messages. This can be exploited through a
496 Denial of Service attack.
497 Thanks to Adam Langley for discovering and researching this issue.
501 *) An attacker can force an error condition which causes openssl to crash
502 whilst processing DTLS packets due to memory being freed twice. This
503 can be exploited through a Denial of Service attack.
504 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
509 *) If a multithreaded client connects to a malicious server using a resumed
510 session and the server sends an ec point format extension it could write
511 up to 255 bytes to freed memory.
513 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
518 *) A malicious server can crash an OpenSSL client with a null pointer
519 dereference (read) by specifying an SRP ciphersuite even though it was not
520 properly negotiated with the client. This can be exploited through a
521 Denial of Service attack.
523 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
524 discovering and researching this issue.
528 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
529 X509_name_oneline, X509_name_print_ex et al. to leak some information
530 from the stack. Applications may be affected if they echo pretty printing
531 output to the attacker.
533 Thanks to Ivan Fratric (Google) for discovering this issue.
535 [Emilia Käsper, and Steve Henson]
537 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
538 for corner cases. (Certain input points at infinity could lead to
539 bogus results, with non-infinity inputs mapped to infinity too.)
542 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
544 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
545 handshake can force the use of weak keying material in OpenSSL
546 SSL/TLS clients and servers.
548 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
549 researching this issue. (CVE-2014-0224)
550 [KIKUCHI Masashi, Steve Henson]
552 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
553 OpenSSL DTLS client the code can be made to recurse eventually crashing
556 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
558 [Imre Rad, Steve Henson]
560 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
561 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
562 client or server. This is potentially exploitable to run arbitrary
563 code on a vulnerable client or server.
565 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
566 [Jüri Aedla, Steve Henson]
568 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
569 are subject to a denial of service attack.
571 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
572 this issue. (CVE-2014-3470)
573 [Felix Gröbert, Ivan Fratric, Steve Henson]
575 *) Harmonize version and its documentation. -f flag is used to display
577 [mancha <mancha1@zoho.com>]
579 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
581 [mancha <mancha1@zoho.com>]
583 *) Fix some double frees. These are not thought to be exploitable.
584 [mancha <mancha1@zoho.com>]
586 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
588 *) A missing bounds check in the handling of the TLS heartbeat extension
589 can be used to reveal up to 64k of memory to a connected client or
592 Thanks for Neel Mehta of Google Security for discovering this bug and to
593 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
594 preparing the fix (CVE-2014-0160)
595 [Adam Langley, Bodo Moeller]
597 *) Fix for the attack described in the paper "Recovering OpenSSL
598 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
599 by Yuval Yarom and Naomi Benger. Details can be obtained from:
600 http://eprint.iacr.org/2014/140
602 Thanks to Yuval Yarom and Naomi Benger for discovering this
603 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
604 [Yuval Yarom and Naomi Benger]
606 *) TLS pad extension: draft-agl-tls-padding-03
608 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
609 TLS client Hello record length value would otherwise be > 255 and
610 less that 512 pad with a dummy extension containing zeroes so it
611 is at least 512 bytes long.
613 [Adam Langley, Steve Henson]
615 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
617 *) Fix for TLS record tampering bug. A carefully crafted invalid
618 handshake could crash OpenSSL with a NULL pointer exception.
619 Thanks to Anton Johansson for reporting this issues.
622 *) Keep original DTLS digest and encryption contexts in retransmission
623 structures so we can use the previous session parameters if they need
624 to be resent. (CVE-2013-6450)
627 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
628 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
629 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
630 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
631 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
632 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
633 [Rob Stradling, Adam Langley]
635 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
637 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
638 supporting platforms or when small records were transferred.
639 [Andy Polyakov, Steve Henson]
641 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
643 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
645 This addresses the flaw in CBC record processing discovered by
646 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
647 at: http://www.isg.rhul.ac.uk/tls/
649 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
650 Security Group at Royal Holloway, University of London
651 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
652 Emilia Käsper for the initial patch.
654 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
656 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
657 ciphersuites which can be exploited in a denial of service attack.
658 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
659 and detecting this bug and to Wolfgang Ettlinger
660 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
664 *) Return an error when checking OCSP signatures when key is NULL.
665 This fixes a DoS attack. (CVE-2013-0166)
668 *) Make openssl verify return errors.
669 [Chris Palmer <palmer@google.com> and Ben Laurie]
671 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
672 the right response is stapled. Also change SSL_get_certificate()
673 so it returns the certificate actually sent.
674 See http://rt.openssl.org/Ticket/Display.html?id=2836.
675 [Rob Stradling <rob.stradling@comodo.com>]
677 *) Fix possible deadlock when decoding public keys.
680 *) Don't use TLS 1.0 record version number in initial client hello
684 Changes between 1.0.1b and 1.0.1c [10 May 2012]
686 *) Sanity check record length before skipping explicit IV in TLS
687 1.2, 1.1 and DTLS to fix DoS attack.
689 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
690 fuzzing as a service testing platform.
694 *) Initialise tkeylen properly when encrypting CMS messages.
695 Thanks to Solar Designer of Openwall for reporting this issue.
698 *) In FIPS mode don't try to use composite ciphers as they are not
702 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
704 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
705 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
706 mean any application compiled against OpenSSL 1.0.0 headers setting
707 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
708 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
709 0x10000000L Any application which was previously compiled against
710 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
711 will need to be recompiled as a result. Letting be results in
712 inability to disable specifically TLS 1.1 and in client context,
713 in unlike event, limit maximum offered version to TLS 1.0 [see below].
716 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
717 disable just protocol X, but all protocols above X *if* there are
718 protocols *below* X still enabled. In more practical terms it means
719 that if application wants to disable TLS1.0 in favor of TLS1.1 and
720 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
721 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
725 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
727 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
728 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
729 in CRYPTO_realloc_clean.
731 Thanks to Tavis Ormandy, Google Security Team, for discovering this
732 issue and to Adam Langley <agl@chromium.org> for fixing it.
734 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
736 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
739 *) Workarounds for some broken servers that "hang" if a client hello
740 record length exceeds 255 bytes.
742 1. Do not use record version number > TLS 1.0 in initial client
743 hello: some (but not all) hanging servers will now work.
744 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
745 the number of ciphers sent in the client hello. This should be
746 set to an even number, such as 50, for example by passing:
747 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
748 Most broken servers should now work.
749 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
750 TLS 1.2 client support entirely.
753 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
756 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
758 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
759 STRING form instead of a DigestInfo.
762 *) The format used for MDC2 RSA signatures is inconsistent between EVP
763 and the RSA_sign/RSA_verify functions. This was made more apparent when
764 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
765 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
766 the correct format in RSA_verify so both forms transparently work.
769 *) Some servers which support TLS 1.0 can choke if we initially indicate
770 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
771 encrypted premaster secret. As a workaround use the maximum pemitted
772 client version in client hello, this should keep such servers happy
773 and still work with previous versions of OpenSSL.
776 *) Add support for TLS/DTLS heartbeats.
777 [Robin Seggelmann <seggelmann@fh-muenster.de>]
779 *) Add support for SCTP.
780 [Robin Seggelmann <seggelmann@fh-muenster.de>]
782 *) Improved PRNG seeding for VOS.
783 [Paul Green <Paul.Green@stratus.com>]
785 *) Extensive assembler packs updates, most notably:
787 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
788 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
789 - x86_64: bit-sliced AES implementation;
790 - ARM: NEON support, contemporary platforms optimizations;
791 - s390x: z196 support;
792 - *: GHASH and GF(2^m) multiplication implementations;
796 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
797 (removal of unnecessary code)
798 [Peter Sylvester <peter.sylvester@edelweb.fr>]
800 *) Add TLS key material exporter from RFC 5705.
803 *) Add DTLS-SRTP negotiation from RFC 5764.
806 *) Add Next Protocol Negotiation,
807 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
808 disabled with a no-npn flag to config or Configure. Code donated
810 [Adam Langley <agl@google.com> and Ben Laurie]
812 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
813 NIST-P256, NIST-P521, with constant-time single point multiplication on
814 typical inputs. Compiler support for the nonstandard type __uint128_t is
815 required to use this (present in gcc 4.4 and later, for 64-bit builds).
816 Code made available under Apache License version 2.0.
818 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
819 line to include this in your build of OpenSSL, and run "make depend" (or
820 "make update"). This enables the following EC_METHODs:
822 EC_GFp_nistp224_method()
823 EC_GFp_nistp256_method()
824 EC_GFp_nistp521_method()
826 EC_GROUP_new_by_curve_name() will automatically use these (while
827 EC_GROUP_new_curve_GFp() currently prefers the more flexible
829 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
831 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
832 all platforms. Move ssize_t definition from e_os.h to the public
833 header file e_os2.h as it now appears in public header file cms.h
836 *) New -sigopt option to the ca, req and x509 utilities. Additional
837 signature parameters can be passed using this option and in
841 *) Add RSA PSS signing function. This will generate and set the
842 appropriate AlgorithmIdentifiers for PSS based on those in the
843 corresponding EVP_MD_CTX structure. No application support yet.
846 *) Support for companion algorithm specific ASN1 signing routines.
847 New function ASN1_item_sign_ctx() signs a pre-initialised
848 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
849 the appropriate parameters.
852 *) Add new algorithm specific ASN1 verification initialisation function
853 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
854 handling will be the same no matter what EVP_PKEY_METHOD is used.
855 Add a PSS handler to support verification of PSS signatures: checked
856 against a number of sample certificates.
859 *) Add signature printing for PSS. Add PSS OIDs.
860 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
862 *) Add algorithm specific signature printing. An individual ASN1 method
863 can now print out signatures instead of the standard hex dump.
865 More complex signatures (e.g. PSS) can print out more meaningful
866 information. Include DSA version that prints out the signature
870 *) Password based recipient info support for CMS library: implementing
874 *) Split password based encryption into PBES2 and PBKDF2 functions. This
875 neatly separates the code into cipher and PBE sections and is required
876 for some algorithms that split PBES2 into separate pieces (such as
880 *) Session-handling fixes:
881 - Fix handling of connections that are resuming with a session ID,
882 but also support Session Tickets.
883 - Fix a bug that suppressed issuing of a new ticket if the client
884 presented a ticket with an expired session.
885 - Try to set the ticket lifetime hint to something reasonable.
886 - Make tickets shorter by excluding irrelevant information.
887 - On the client side, don't ignore renewed tickets.
888 [Adam Langley, Bodo Moeller (Google)]
890 *) Fix PSK session representation.
893 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
895 This work was sponsored by Intel.
898 *) Add GCM support to TLS library. Some custom code is needed to split
899 the IV between the fixed (from PRF) and explicit (from TLS record)
900 portions. This adds all GCM ciphersuites supported by RFC5288 and
901 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
902 add a special AESGCM string for GCM only.
905 *) Expand range of ctrls for AES GCM. Permit setting invocation
906 field on decrypt and retrieval of invocation field only on encrypt.
909 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
910 As required by RFC5289 these ciphersuites cannot be used if for
911 versions of TLS earlier than 1.2.
914 *) For FIPS capable OpenSSL interpret a NULL default public key method
915 as unset and return the appopriate default but do *not* set the default.
916 This means we can return the appopriate method in applications that
917 swicth between FIPS and non-FIPS modes.
920 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
921 ENGINE is used then we cannot handle that in the FIPS module so we
922 keep original code iff non-FIPS operations are allowed.
925 *) Add -attime option to openssl utilities.
926 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
928 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
931 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
932 FIPS EC methods unconditionally for now.
935 *) New build option no-ec2m to disable characteristic 2 code.
938 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
939 all cases can be covered as some introduce binary incompatibilities.
942 *) Redirect RSA operations to FIPS module including keygen,
943 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
946 *) Add similar low level API blocking to ciphers.
949 *) Low level digest APIs are not approved in FIPS mode: any attempt
950 to use these will cause a fatal error. Applications that *really* want
951 to use them can use the private_* version instead.
954 *) Redirect cipher operations to FIPS module for FIPS builds.
957 *) Redirect digest operations to FIPS module for FIPS builds.
960 *) Update build system to add "fips" flag which will link in fipscanister.o
961 for static and shared library builds embedding a signature if needed.
964 *) Output TLS supported curves in preference order instead of numerical
965 order. This is currently hardcoded for the highest order curves first.
966 This should be configurable so applications can judge speed vs strength.
969 *) Add TLS v1.2 server support for client authentication.
972 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
976 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
977 FIPS modules versions.
980 *) Add TLS v1.2 client side support for client authentication. Keep cache
981 of handshake records longer as we don't know the hash algorithm to use
982 until after the certificate request message is received.
985 *) Initial TLS v1.2 client support. Add a default signature algorithms
986 extension including all the algorithms we support. Parse new signature
987 format in client key exchange. Relax some ECC signing restrictions for
988 TLS v1.2 as indicated in RFC5246.
991 *) Add server support for TLS v1.2 signature algorithms extension. Switch
992 to new signature format when needed using client digest preference.
993 All server ciphersuites should now work correctly in TLS v1.2. No client
994 support yet and no support for client certificates.
997 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
998 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
999 ciphersuites. At present only RSA key exchange ciphersuites work with
1000 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
1001 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
1002 and version checking.
1005 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
1006 with this defined it will not be affected by any changes to ssl internal
1007 structures. Add several utility functions to allow openssl application
1008 to work with OPENSSL_NO_SSL_INTERN defined.
1012 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
1014 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
1017 *) Permit abbreviated handshakes when renegotiating using the function
1018 SSL_renegotiate_abbreviated().
1019 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1021 *) Add call to ENGINE_register_all_complete() to
1022 ENGINE_load_builtin_engines(), so some implementations get used
1023 automatically instead of needing explicit application support.
1026 *) Add support for TLS key exporter as described in RFC5705.
1027 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
1029 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
1030 a few changes are required:
1032 Add SSL_OP_NO_TLSv1_1 flag.
1033 Add TLSv1_1 methods.
1034 Update version checking logic to handle version 1.1.
1035 Add explicit IV handling (ported from DTLS code).
1036 Add command line options to s_client/s_server.
1039 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
1041 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1042 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1043 content decryption and always return the same error. Note: this attack
1044 needs on average 2^20 messages so it only affects automated senders. The
1045 old behaviour can be reenabled in the CMS code by setting the
1046 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1047 an MMA defence is not necessary.
1048 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1049 this issue. (CVE-2012-0884)
1052 *) Fix CVE-2011-4619: make sure we really are receiving a
1053 client hello before rejecting multiple SGC restarts. Thanks to
1054 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1057 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
1059 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1060 Thanks to Antonio Martin, Enterprise Secure Access Research and
1061 Development, Cisco Systems, Inc. for discovering this bug and
1062 preparing a fix. (CVE-2012-0050)
1065 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
1067 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1068 of the Vaudenay padding oracle attack on CBC mode encryption
1069 which enables an efficient plaintext recovery attack against
1070 the OpenSSL implementation of DTLS. Their attack exploits timing
1071 differences arising during decryption processing. A research
1072 paper describing this attack can be found at:
1073 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1074 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1075 Security Group at Royal Holloway, University of London
1076 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1077 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1078 for preparing the fix. (CVE-2011-4108)
1079 [Robin Seggelmann, Michael Tuexen]
1081 *) Clear bytes used for block padding of SSL 3.0 records.
1083 [Adam Langley (Google)]
1085 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1086 Kadianakis <desnacked@gmail.com> for discovering this issue and
1087 Adam Langley for preparing the fix. (CVE-2011-4619)
1088 [Adam Langley (Google)]
1090 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
1091 [Andrey Kulikov <amdeich@gmail.com>]
1093 *) Prevent malformed RFC3779 data triggering an assertion failure.
1094 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1095 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1096 [Rob Austein <sra@hactrn.net>]
1098 *) Improved PRNG seeding for VOS.
1099 [Paul Green <Paul.Green@stratus.com>]
1101 *) Fix ssl_ciph.c set-up race.
1102 [Adam Langley (Google)]
1104 *) Fix spurious failures in ecdsatest.c.
1105 [Emilia Käsper (Google)]
1107 *) Fix the BIO_f_buffer() implementation (which was mixing different
1108 interpretations of the '..._len' fields).
1109 [Adam Langley (Google)]
1111 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1112 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1113 threads won't reuse the same blinding coefficients.
1115 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1116 lock to call BN_BLINDING_invert_ex, and avoids one use of
1117 BN_BLINDING_update for each BN_BLINDING structure (previously,
1118 the last update always remained unused).
1119 [Emilia Käsper (Google)]
1121 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
1122 [Bob Buckholz (Google)]
1124 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
1126 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
1127 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
1128 [Kaspar Brand <ossl@velox.ch>]
1130 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1131 for multi-threaded use of ECDH. (CVE-2011-3210)
1132 [Adam Langley (Google)]
1134 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1137 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
1138 signature public key algorithm by using OID xref utilities instead.
1139 Before this you could only use some ECC ciphersuites with SHA1 only.
1142 *) Add protection against ECDSA timing attacks as mentioned in the paper
1143 by Billy Bob Brumley and Nicola Tuveri, see:
1145 http://eprint.iacr.org/2011/232.pdf
1147 [Billy Bob Brumley and Nicola Tuveri]
1149 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
1151 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1152 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1154 *) Fix bug in string printing code: if *any* escaping is enabled we must
1155 escape the escape character (backslash) or the resulting string is
1159 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
1161 *) Disable code workaround for ancient and obsolete Netscape browsers
1162 and servers: an attacker can use it in a ciphersuite downgrade attack.
1163 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1166 *) Fixed J-PAKE implementation error, originally discovered by
1167 Sebastien Martini, further info and confirmation from Stefan
1168 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1171 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
1173 *) Fix extension code to avoid race conditions which can result in a buffer
1174 overrun vulnerability: resumed sessions must not be modified as they can
1175 be shared by multiple threads. CVE-2010-3864
1178 *) Fix WIN32 build system to correctly link an ENGINE directory into
1182 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
1184 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
1186 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
1188 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
1190 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
1191 context. The operation can be customised via the ctrl mechanism in
1192 case ENGINEs want to include additional functionality.
1195 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
1198 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
1199 output hashes compatible with older versions of OpenSSL.
1200 [Willy Weisz <weisz@vcpc.univie.ac.at>]
1202 *) Fix compression algorithm handling: if resuming a session use the
1203 compression algorithm of the resumed session instead of determining
1204 it from client hello again. Don't allow server to change algorithm.
1207 *) Add load_crls() function to apps tidying load_certs() too. Add option
1208 to verify utility to allow additional CRLs to be included.
1211 *) Update OCSP request code to permit adding custom headers to the request:
1212 some responders need this.
1215 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
1217 [Julia Lawall <julia@diku.dk>]
1219 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
1220 needlessly dereferenced structures, used obsolete functions and
1221 didn't handle all updated verify codes correctly.
1224 *) Disable MD2 in the default configuration.
1227 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
1228 indicate the initial BIO being pushed or popped. This makes it possible
1229 to determine whether the BIO is the one explicitly called or as a result
1230 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
1231 it handles reference counts correctly and doesn't zero out the I/O bio
1232 when it is not being explicitly popped. WARNING: applications which
1233 included workarounds for the old buggy behaviour will need to be modified
1234 or they could free up already freed BIOs.
1237 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
1238 renaming to all platforms (within the 0.9.8 branch, this was
1239 done conditionally on Netware platforms to avoid a name clash).
1240 [Guenter <lists@gknw.net>]
1242 *) Add ECDHE and PSK support to DTLS.
1243 [Michael Tuexen <tuexen@fh-muenster.de>]
1245 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
1249 *) Add "missing" function EVP_MD_flags() (without this the only way to
1250 retrieve a digest flags is by accessing the structure directly. Update
1251 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
1252 or cipher is registered as in the "from" argument. Print out all
1253 registered digests in the dgst usage message instead of manually
1254 attempting to work them out.
1257 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
1258 this allows the use of compression and extensions. Change default cipher
1259 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
1260 by default unless an application cipher string requests it.
1263 *) Alter match criteria in PKCS12_parse(). It used to try to use local
1264 key ids to find matching certificates and keys but some PKCS#12 files
1265 don't follow the (somewhat unwritten) rules and this strategy fails.
1266 Now just gather all certificates together and the first private key
1267 then look for the first certificate that matches the key.
1270 *) Support use of registered digest and cipher names for dgst and cipher
1271 commands instead of having to add each one as a special case. So now
1278 openssl dgst -sha256 foo
1280 and this works for ENGINE based algorithms too.
1284 *) Update Gost ENGINE to support parameter files.
1285 [Victor B. Wagner <vitus@cryptocom.ru>]
1287 *) Support GeneralizedTime in ca utility.
1288 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
1290 *) Enhance the hash format used for certificate directory links. The new
1291 form uses the canonical encoding (meaning equivalent names will work
1292 even if they aren't identical) and uses SHA1 instead of MD5. This form
1293 is incompatible with the older format and as a result c_rehash should
1294 be used to rebuild symbolic links.
1297 *) Make PKCS#8 the default write format for private keys, replacing the
1298 traditional format. This form is standardised, more secure and doesn't
1299 include an implicit MD5 dependency.
1302 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
1303 committed to OpenSSL should pass this lot as a minimum.
1306 *) Add session ticket override functionality for use by EAP-FAST.
1307 [Jouni Malinen <j@w1.fi>]
1309 *) Modify HMAC functions to return a value. Since these can be implemented
1310 in an ENGINE errors can occur.
1313 *) Type-checked OBJ_bsearch_ex.
1316 *) Type-checked OBJ_bsearch. Also some constification necessitated
1317 by type-checking. Still to come: TXT_DB, bsearch(?),
1318 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
1322 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
1323 seconds to a tm structure directly, instead of going through OS
1324 specific date routines. This avoids any issues with OS routines such
1325 as the year 2038 bug. New *_adj() functions for ASN1 time structures
1326 and X509_time_adj_ex() to cover the extended range. The existing
1327 X509_time_adj() is still usable and will no longer have any date issues.
1330 *) Delta CRL support. New use deltas option which will attempt to locate
1331 and search any appropriate delta CRLs available.
1333 This work was sponsored by Google.
1336 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
1337 code and add additional score elements. Validate alternate CRL paths
1338 as part of the CRL checking and indicate a new error "CRL path validation
1339 error" in this case. Applications wanting additional details can use
1340 the verify callback and check the new "parent" field. If this is not
1341 NULL CRL path validation is taking place. Existing applications wont
1342 see this because it requires extended CRL support which is off by
1345 This work was sponsored by Google.
1348 *) Support for freshest CRL extension.
1350 This work was sponsored by Google.
1353 *) Initial indirect CRL support. Currently only supported in the CRLs
1354 passed directly and not via lookup. Process certificate issuer
1355 CRL entry extension and lookup CRL entries by bother issuer name
1356 and serial number. Check and process CRL issuer entry in IDP extension.
1358 This work was sponsored by Google.
1361 *) Add support for distinct certificate and CRL paths. The CRL issuer
1362 certificate is validated separately in this case. Only enabled if
1363 an extended CRL support flag is set: this flag will enable additional
1364 CRL functionality in future.
1366 This work was sponsored by Google.
1369 *) Add support for policy mappings extension.
1371 This work was sponsored by Google.
1374 *) Fixes to pathlength constraint, self issued certificate handling,
1375 policy processing to align with RFC3280 and PKITS tests.
1377 This work was sponsored by Google.
1380 *) Support for name constraints certificate extension. DN, email, DNS
1381 and URI types are currently supported.
1383 This work was sponsored by Google.
1386 *) To cater for systems that provide a pointer-based thread ID rather
1387 than numeric, deprecate the current numeric thread ID mechanism and
1388 replace it with a structure and associated callback type. This
1389 mechanism allows a numeric "hash" to be extracted from a thread ID in
1390 either case, and on platforms where pointers are larger than 'long',
1391 mixing is done to help ensure the numeric 'hash' is usable even if it
1392 can't be guaranteed unique. The default mechanism is to use "&errno"
1393 as a pointer-based thread ID to distinguish between threads.
1395 Applications that want to provide their own thread IDs should now use
1396 CRYPTO_THREADID_set_callback() to register a callback that will call
1397 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
1399 Note that ERR_remove_state() is now deprecated, because it is tied
1400 to the assumption that thread IDs are numeric. ERR_remove_state(0)
1401 to free the current thread's error state should be replaced by
1402 ERR_remove_thread_state(NULL).
1404 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
1405 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
1406 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
1407 application was previously providing a numeric thread callback that
1408 was inappropriate for distinguishing threads, then uniqueness might
1409 have been obtained with &errno that happened immediately in the
1410 intermediate development versions of OpenSSL; this is no longer the
1411 case, the numeric thread callback will now override the automatic use
1413 [Geoff Thorpe, with help from Bodo Moeller]
1415 *) Initial support for different CRL issuing certificates. This covers a
1416 simple case where the self issued certificates in the chain exist and
1417 the real CRL issuer is higher in the existing chain.
1419 This work was sponsored by Google.
1422 *) Removed effectively defunct crypto/store from the build.
1425 *) Revamp of STACK to provide stronger type-checking. Still to come:
1426 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
1427 ASN1_STRING, CONF_VALUE.
1430 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
1431 RAM on SSL connections. This option can save about 34k per idle SSL.
1434 *) Revamp of LHASH to provide stronger type-checking. Still to come:
1435 STACK, TXT_DB, bsearch, qsort.
1438 *) Initial support for Cryptographic Message Syntax (aka CMS) based
1439 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
1440 support for data, signedData, compressedData, digestedData and
1441 encryptedData, envelopedData types included. Scripts to check against
1442 RFC4134 examples draft and interop and consistency checks of many
1443 content types and variants.
1446 *) Add options to enc utility to support use of zlib compression BIO.
1449 *) Extend mk1mf to support importing of options and assembly language
1450 files from Configure script, currently only included in VC-WIN32.
1451 The assembly language rules can now optionally generate the source
1452 files from the associated perl scripts.
1455 *) Implement remaining functionality needed to support GOST ciphersuites.
1456 Interop testing has been performed using CryptoPro implementations.
1457 [Victor B. Wagner <vitus@cryptocom.ru>]
1459 *) s390x assembler pack.
1462 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
1466 *) Implement Opaque PRF Input TLS extension as specified in
1467 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
1468 official specification yet and no extension type assignment by
1469 IANA exists, this extension (for now) will have to be explicitly
1470 enabled when building OpenSSL by providing the extension number
1471 to use. For example, specify an option
1473 -DTLSEXT_TYPE_opaque_prf_input=0x9527
1475 to the "config" or "Configure" script to enable the extension,
1476 assuming extension number 0x9527 (which is a completely arbitrary
1477 and unofficial assignment based on the MD5 hash of the Internet
1478 Draft). Note that by doing so, you potentially lose
1479 interoperability with other TLS implementations since these might
1480 be using the same extension number for other purposes.
1482 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
1483 opaque PRF input value to use in the handshake. This will create
1484 an interal copy of the length-'len' string at 'src', and will
1485 return non-zero for success.
1487 To get more control and flexibility, provide a callback function
1490 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
1491 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
1495 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
1498 Callback function 'cb' will be called in handshakes, and is
1499 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
1500 Argument 'arg' is for application purposes (the value as given to
1501 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
1502 be provided to the callback function). The callback function
1503 has to return non-zero to report success: usually 1 to use opaque
1504 PRF input just if possible, or 2 to enforce use of the opaque PRF
1505 input. In the latter case, the library will abort the handshake
1506 if opaque PRF input is not successfully negotiated.
1508 Arguments 'peerinput' and 'len' given to the callback function
1509 will always be NULL and 0 in the case of a client. A server will
1510 see the client's opaque PRF input through these variables if
1511 available (NULL and 0 otherwise). Note that if the server
1512 provides an opaque PRF input, the length must be the same as the
1513 length of the client's opaque PRF input.
1515 Note that the callback function will only be called when creating
1516 a new session (session resumption can resume whatever was
1517 previously negotiated), and will not be called in SSL 2.0
1518 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
1519 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
1520 for applications that need to enforce opaque PRF input.
1524 *) Update ssl code to support digests other than SHA1+MD5 for handshake
1527 [Victor B. Wagner <vitus@cryptocom.ru>]
1529 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1530 RFC4507bis. The encrypted ticket format is an encrypted encoded
1531 SSL_SESSION structure, that way new session features are automatically
1534 If a client application caches session in an SSL_SESSION structure
1535 support is transparent because tickets are now stored in the encoded
1538 The SSL_CTX structure automatically generates keys for ticket
1539 protection in servers so again support should be possible
1540 with no application modification.
1542 If a client or server wishes to disable RFC4507 support then the option
1543 SSL_OP_NO_TICKET can be set.
1545 Add a TLS extension debugging callback to allow the contents of any client
1546 or server extensions to be examined.
1548 This work was sponsored by Google.
1551 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
1552 OpenSSL should now compile cleanly on gcc 4.2
1553 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
1555 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
1556 support including streaming MAC support: this is required for GOST
1557 ciphersuite support.
1558 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
1560 *) Add option -stream to use PKCS#7 streaming in smime utility. New
1561 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
1562 to output in BER and PEM format.
1565 *) Experimental support for use of HMAC via EVP_PKEY interface. This
1566 allows HMAC to be handled via the EVP_DigestSign*() interface. The
1567 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
1568 ENGINE support for HMAC keys which are unextractable. New -mac and
1569 -macopt options to dgst utility.
1572 *) New option -sigopt to dgst utility. Update dgst to use
1573 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
1574 alternative signing paramaters such as X9.31 or PSS in the dgst
1578 *) Change ssl_cipher_apply_rule(), the internal function that does
1579 the work each time a ciphersuite string requests enabling
1580 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
1581 removing ("!foo+bar") a class of ciphersuites: Now it maintains
1582 the order of disabled ciphersuites such that those ciphersuites
1583 that most recently went from enabled to disabled not only stay
1584 in order with respect to each other, but also have higher priority
1585 than other disabled ciphersuites the next time ciphersuites are
1588 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
1589 the same ciphersuites as with "HIGH" alone, but in a specific
1590 order where the PSK ciphersuites come first (since they are the
1591 most recently disabled ciphersuites when "HIGH" is parsed).
1593 Also, change ssl_create_cipher_list() (using this new
1594 funcionality) such that between otherwise identical
1595 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
1599 *) Change ssl_create_cipher_list() so that it automatically
1600 arranges the ciphersuites in reasonable order before starting
1601 to process the rule string. Thus, the definition for "DEFAULT"
1602 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
1603 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
1604 This makes it much easier to arrive at a reasonable default order
1605 in applications for which anonymous ciphers are OK (meaning
1606 that you can't actually use DEFAULT).
1607 [Bodo Moeller; suggested by Victor Duchovni]
1609 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
1610 processing) into multiple integers instead of setting
1611 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
1612 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
1613 (These masks as well as the individual bit definitions are hidden
1614 away into the non-exported interface ssl/ssl_locl.h, so this
1615 change to the definition of the SSL_CIPHER structure shouldn't
1616 affect applications.) This give us more bits for each of these
1617 categories, so there is no longer a need to coagulate AES128 and
1618 AES256 into a single algorithm bit, and to coagulate Camellia128
1619 and Camellia256 into a single algorithm bit, which has led to all
1622 Thus, among other things, the kludge introduced in 0.9.7m and
1623 0.9.8e for masking out AES256 independently of AES128 or masking
1624 out Camellia256 independently of AES256 is not needed here in 0.9.9.
1626 With the change, we also introduce new ciphersuite aliases that
1627 so far were missing: "AES128", "AES256", "CAMELLIA128", and
1631 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
1632 Use the leftmost N bytes of the signature input if the input is
1633 larger than the prime q (with N being the size in bytes of q).
1636 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
1637 it yet and it is largely untested.
1640 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
1643 *) Initial incomplete changes to avoid need for function casts in OpenSSL
1644 some compilers (gcc 4.2 and later) reject their use. Safestack is
1645 reimplemented. Update ASN1 to avoid use of legacy functions.
1648 *) Win32/64 targets are linked with Winsock2.
1651 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
1652 to external functions. This can be used to increase CRL handling
1653 efficiency especially when CRLs are very large by (for example) storing
1654 the CRL revoked certificates in a database.
1657 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
1658 new CRLs added to a directory can be used. New command line option
1659 -verify_return_error to s_client and s_server. This causes real errors
1660 to be returned by the verify callback instead of carrying on no matter
1661 what. This reflects the way a "real world" verify callback would behave.
1664 *) GOST engine, supporting several GOST algorithms and public key formats.
1665 Kindly donated by Cryptocom.
1668 *) Partial support for Issuing Distribution Point CRL extension. CRLs
1669 partitioned by DP are handled but no indirect CRL or reason partitioning
1670 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
1671 selected via a scoring technique which handles IDP and AKID in CRLs.
1674 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
1675 will ultimately be used for all verify operations: this will remove the
1676 X509_STORE dependency on certificate verification and allow alternative
1677 lookup methods. X509_STORE based implementations of these two callbacks.
1680 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
1681 Modify get_crl() to find a valid (unexpired) CRL if possible.
1684 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
1685 this would be called X509_CRL_cmp() but that name is already used by
1686 a function that just compares CRL issuer names. Cache several CRL
1687 extensions in X509_CRL structure and cache CRLDP in X509.
1690 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
1691 this maps equivalent X509_NAME structures into a consistent structure.
1692 Name comparison can then be performed rapidly using memcmp().
1695 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
1699 *) Allow digests to supply their own micalg string for S/MIME type using
1700 the ctrl EVP_MD_CTRL_MICALG.
1703 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
1704 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
1705 ctrl. It can then customise the structure before and/or after signing
1709 *) New function OBJ_add_sigid() to allow application defined signature OIDs
1710 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
1711 to free up any added signature OIDs.
1714 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
1715 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
1716 digest and cipher tables. New options added to openssl utility:
1717 list-message-digest-algorithms and list-cipher-algorithms.
1720 *) Change the array representation of binary polynomials: the list
1721 of degrees of non-zero coefficients is now terminated with -1.
1722 Previously it was terminated with 0, which was also part of the
1723 value; thus, the array representation was not applicable to
1724 polynomials where t^0 has coefficient zero. This change makes
1725 the array representation useful in a more general context.
1728 *) Various modifications and fixes to SSL/TLS cipher string
1729 handling. For ECC, the code now distinguishes between fixed ECDH
1730 with RSA certificates on the one hand and with ECDSA certificates
1731 on the other hand, since these are separate ciphersuites. The
1732 unused code for Fortezza ciphersuites has been removed.
1734 For consistency with EDH, ephemeral ECDH is now called "EECDH"
1735 (not "ECDHE"). For consistency with the code for DH
1736 certificates, use of ECDH certificates is now considered ECDH
1737 authentication, not RSA or ECDSA authentication (the latter is
1738 merely the CA's signing algorithm and not actively used in the
1741 The temporary ciphersuite alias "ECCdraft" is no longer
1742 available, and ECC ciphersuites are no longer excluded from "ALL"
1743 and "DEFAULT". The following aliases now exist for RFC 4492
1744 ciphersuites, most of these by analogy with the DH case:
1746 kECDHr - ECDH cert, signed with RSA
1747 kECDHe - ECDH cert, signed with ECDSA
1748 kECDH - ECDH cert (signed with either RSA or ECDSA)
1749 kEECDH - ephemeral ECDH
1750 ECDH - ECDH cert or ephemeral ECDH
1756 AECDH - anonymous ECDH
1757 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
1761 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
1762 Use correct micalg parameters depending on digest(s) in signed message.
1765 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
1766 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
1769 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
1770 an engine to register a method. Add ENGINE lookups for methods and
1771 functional reference processing.
1774 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
1775 EVP_{Sign,Verify}* which allow an application to customise the signature
1779 *) New -resign option to smime utility. This adds one or more signers
1780 to an existing PKCS#7 signedData structure. Also -md option to use an
1781 alternative message digest algorithm for signing.
1784 *) Tidy up PKCS#7 routines and add new functions to make it easier to
1785 create PKCS7 structures containing multiple signers. Update smime
1786 application to support multiple signers.
1789 *) New -macalg option to pkcs12 utility to allow setting of an alternative
1793 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
1794 Reorganize PBE internals to lookup from a static table using NIDs,
1795 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
1796 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
1797 PRF which will be automatically used with PBES2.
1800 *) Replace the algorithm specific calls to generate keys in "req" with the
1804 *) Update PKCS#7 enveloped data routines to use new API. This is now
1805 supported by any public key method supporting the encrypt operation. A
1806 ctrl is added to allow the public key algorithm to examine or modify
1807 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
1811 *) Add a ctrl to asn1 method to allow a public key algorithm to express
1812 a default digest type to use. In most cases this will be SHA1 but some
1813 algorithms (such as GOST) need to specify an alternative digest. The
1814 return value indicates how strong the prefernce is 1 means optional and
1815 2 is mandatory (that is it is the only supported type). Modify
1816 ASN1_item_sign() to accept a NULL digest argument to indicate it should
1817 use the default md. Update openssl utilities to use the default digest
1818 type for signing if it is not explicitly indicated.
1821 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
1822 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
1823 signing method from the key type. This effectively removes the link
1824 between digests and public key types.
1827 *) Add an OID cross reference table and utility functions. Its purpose is to
1828 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
1829 rsaEncryption. This will allow some of the algorithm specific hackery
1830 needed to use the correct OID to be removed.
1833 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
1834 structures for PKCS7_sign(). They are now set up by the relevant public
1838 *) Add provisional EC pkey method with support for ECDSA and ECDH.
1841 *) Add support for key derivation (agreement) in the API, DH method and
1845 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
1846 public and private key formats. As a side effect these add additional
1847 command line functionality not previously available: DSA signatures can be
1848 generated and verified using pkeyutl and DH key support and generation in
1853 [Oliver Tappe <zooey@hirschkaefer.de>]
1855 *) New make target "install_html_docs" installs HTML renditions of the
1857 [Oliver Tappe <zooey@hirschkaefer.de>]
1859 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
1860 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
1861 support key and parameter generation and add initial key generation
1862 functionality for RSA.
1865 *) Add functions for main EVP_PKEY_method operations. The undocumented
1866 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
1867 EVP_PKEY_{encrypt,decrypt}_old.
1870 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
1871 key API, doesn't do much yet.
1874 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
1875 public key algorithms. New option to openssl utility:
1876 "list-public-key-algorithms" to print out info.
1879 *) Implement the Supported Elliptic Curves Extension for
1880 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1883 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
1884 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
1887 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
1888 utilities such as rsa, dsa, dsaparam etc except they process any key
1892 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1893 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1894 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
1898 *) Initial support for pluggable public key ASN1.
1899 De-spaghettify the public key ASN1 handling. Move public and private
1900 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
1901 algorithm specific handling to a single module within the relevant
1902 algorithm directory. Add functions to allow (near) opaque processing
1903 of public and private key structures.
1906 *) Implement the Supported Point Formats Extension for
1907 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1910 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
1911 for the psk identity [hint] and the psk callback functions to the
1912 SSL_SESSION, SSL and SSL_CTX structure.
1915 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
1919 SSL_CTX_use_psk_identity_hint
1920 SSL_get_psk_identity_hint
1921 SSL_get_psk_identity
1922 SSL_use_psk_identity_hint
1924 [Mika Kousa and Pasi Eronen of Nokia Corporation]
1926 *) Add RFC 3161 compliant time stamp request creation, response generation
1927 and response verification functionality.
1928 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
1930 *) Add initial support for TLS extensions, specifically for the server_name
1931 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1932 have new members for a host name. The SSL data structure has an
1933 additional member SSL_CTX *initial_ctx so that new sessions can be
1934 stored in that context to allow for session resumption, even after the
1935 SSL has been switched to a new SSL_CTX in reaction to a client's
1936 server_name extension.
1938 New functions (subject to change):
1940 SSL_get_servername()
1941 SSL_get_servername_type()
1944 New CTRL codes and macros (subject to change):
1946 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1947 - SSL_CTX_set_tlsext_servername_callback()
1948 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1949 - SSL_CTX_set_tlsext_servername_arg()
1950 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1952 openssl s_client has a new '-servername ...' option.
1954 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1955 '-key2 ...', '-servername_fatal' (subject to change). This allows
1956 testing the HostName extension for a specific single host name ('-cert'
1957 and '-key' remain fallbacks for handshakes without HostName
1958 negotiation). If the unrecogninzed_name alert has to be sent, this by
1959 default is a warning; it becomes fatal with the '-servername_fatal'
1962 [Peter Sylvester, Remy Allais, Christophe Renou]
1964 *) Whirlpool hash implementation is added.
1967 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1968 bn(64,32). Because of instruction set limitations it doesn't have
1969 any negative impact on performance. This was done mostly in order
1970 to make it possible to share assembler modules, such as bn_mul_mont
1971 implementations, between 32- and 64-bit builds without hassle.
1974 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1975 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1979 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1980 dedicated Montgomery multiplication procedure, is introduced.
1981 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1982 "64-bit" performance on certain 32-bit targets.
1985 *) New option SSL_OP_NO_COMP to disable use of compression selectively
1986 in SSL structures. New SSL ctrl to set maximum send fragment size.
1987 Save memory by seeting the I/O buffer sizes dynamically instead of
1988 using the maximum available value.
1991 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
1992 in addition to the text details.
1995 *) Very, very preliminary EXPERIMENTAL support for printing of general
1996 ASN1 structures. This currently produces rather ugly output and doesn't
1997 handle several customised structures at all.
2000 *) Integrated support for PVK file format and some related formats such
2001 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
2002 these in the 'rsa' and 'dsa' utilities.
2005 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
2008 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
2009 place for the (very old) "NETSCAPE" format certificates which are now
2010 handled using new ASN1 code equivalents.
2013 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
2014 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
2015 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
2018 *) Modify CRL distribution points extension code to print out previously
2019 unsupported fields. Enhance extension setting code to allow setting of
2023 *) Add print and set support for Issuing Distribution Point CRL extension.
2026 *) Change 'Configure' script to enable Camellia by default.
2029 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
2031 *) When rejecting SSL/TLS records due to an incorrect version number, never
2032 update s->server with a new major version number. As of
2033 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
2034 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
2035 the previous behavior could result in a read attempt at NULL when
2036 receiving specific incorrect SSL/TLS records once record payload
2037 protection is active. (CVE-2010-0740)
2038 [Bodo Moeller, Adam Langley <agl@chromium.org>]
2040 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
2041 could be crashed if the relevant tables were not present (e.g. chrooted).
2042 [Tomas Hoger <thoger@redhat.com>]
2044 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
2046 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
2047 [Martin Olsson, Neel Mehta]
2049 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
2050 accommodate for stack sorting, always a write lock!).
2053 *) On some versions of WIN32 Heap32Next is very slow. This can cause
2054 excessive delays in the RAND_poll(): over a minute. As a workaround
2055 include a time check in the inner Heap32Next loop too.
2058 *) The code that handled flushing of data in SSL/TLS originally used the
2059 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
2060 the problem outlined in PR#1949. The fix suggested there however can
2061 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
2062 of Apache). So instead simplify the code to flush unconditionally.
2063 This should be fine since flushing with no data to flush is a no op.
2066 *) Handle TLS versions 2.0 and later properly and correctly use the
2067 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
2068 off ancient servers have a habit of sticking around for a while...
2071 *) Modify compression code so it frees up structures without using the
2072 ex_data callbacks. This works around a problem where some applications
2073 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
2074 restarting) then use compression (e.g. SSL with compression) later.
2075 This results in significant per-connection memory leaks and
2076 has caused some security issues including CVE-2008-1678 and
2080 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
2081 change when encrypting or decrypting.
2084 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
2085 connect and renegotiate with servers which do not support RI.
2086 Until RI is more widely deployed this option is enabled by default.
2089 *) Add "missing" ssl ctrls to clear options and mode.
2092 *) If client attempts to renegotiate and doesn't support RI respond with
2093 a no_renegotiation alert as required by RFC5746. Some renegotiating
2094 TLS clients will continue a connection gracefully when they receive
2095 the alert. Unfortunately OpenSSL mishandled this alert and would hang
2096 waiting for a server hello which it will never receive. Now we treat a
2097 received no_renegotiation alert as a fatal error. This is because
2098 applications requesting a renegotiation might well expect it to succeed
2099 and would have no code in place to handle the server denying it so the
2100 only safe thing to do is to terminate the connection.
2103 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
2104 peer supports secure renegotiation and 0 otherwise. Print out peer
2105 renegotiation support in s_client/s_server.
2108 *) Replace the highly broken and deprecated SPKAC certification method with
2109 the updated NID creation version. This should correctly handle UTF8.
2112 *) Implement RFC5746. Re-enable renegotiation but require the extension
2113 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
2114 turns out to be a bad idea. It has been replaced by
2115 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
2116 SSL_CTX_set_options(). This is really not recommended unless you
2117 know what you are doing.
2118 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
2120 *) Fixes to stateless session resumption handling. Use initial_ctx when
2121 issuing and attempting to decrypt tickets in case it has changed during
2122 servername handling. Use a non-zero length session ID when attempting
2123 stateless session resumption: this makes it possible to determine if
2124 a resumption has occurred immediately after receiving server hello
2125 (several places in OpenSSL subtly assume this) instead of later in
2129 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
2130 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
2131 fixes for a few places where the return code is not checked
2133 [Julia Lawall <julia@diku.dk>]
2135 *) Add --strict-warnings option to Configure script to include devteam
2136 warnings in other configurations.
2139 *) Add support for --libdir option and LIBDIR variable in makefiles. This
2140 makes it possible to install openssl libraries in locations which
2141 have names other than "lib", for example "/usr/lib64" which some
2143 [Steve Henson, based on patch from Jeremy Utley]
2145 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
2146 X690 8.9.12 and can produce some misleading textual output of OIDs.
2147 [Steve Henson, reported by Dan Kaminsky]
2149 *) Delete MD2 from algorithm tables. This follows the recommendation in
2150 several standards that it is not used in new applications due to
2151 several cryptographic weaknesses. For binary compatibility reasons
2152 the MD2 API is still compiled in by default.
2155 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
2159 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
2160 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
2162 [Guenter <lists@gknw.net>]
2164 *) Fix the server certificate chain building code to use X509_verify_cert(),
2165 it used to have an ad-hoc builder which was unable to cope with anything
2166 other than a simple chain.
2167 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
2169 *) Don't check self signed certificate signatures in X509_verify_cert()
2170 by default (a flag can override this): it just wastes time without
2171 adding any security. As a useful side effect self signed root CAs
2172 with non-FIPS digests are now usable in FIPS mode.
2175 *) In dtls1_process_out_of_seq_message() the check if the current message
2176 is already buffered was missing. For every new message was memory
2177 allocated, allowing an attacker to perform an denial of service attack
2178 with sending out of seq handshake messages until there is no memory
2179 left. Additionally every future messege was buffered, even if the
2180 sequence number made no sense and would be part of another handshake.
2181 So only messages with sequence numbers less than 10 in advance will be
2182 buffered. (CVE-2009-1378)
2183 [Robin Seggelmann, discovered by Daniel Mentz]
2185 *) Records are buffered if they arrive with a future epoch to be
2186 processed after finishing the corresponding handshake. There is
2187 currently no limitation to this buffer allowing an attacker to perform
2188 a DOS attack with sending records with future epochs until there is no
2189 memory left. This patch adds the pqueue_size() function to detemine
2190 the size of a buffer and limits the record buffer to 100 entries.
2192 [Robin Seggelmann, discovered by Daniel Mentz]
2194 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
2195 parent structure is freed. (CVE-2009-1379)
2198 *) Handle non-blocking I/O properly in SSL_shutdown() call.
2199 [Darryl Miles <darryl-mailinglists@netbauds.net>]
2202 [Ilya O. <vrghost@gmail.com>]
2204 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
2206 *) Disable renegotiation completely - this fixes a severe security
2207 problem (CVE-2009-3555) at the cost of breaking all
2208 renegotiation. Renegotiation can be re-enabled by setting
2209 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
2210 run-time. This is really not recommended unless you know what
2214 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
2216 *) Don't set val to NULL when freeing up structures, it is freed up by
2217 underlying code. If sizeof(void *) > sizeof(long) this can result in
2218 zeroing past the valid field. (CVE-2009-0789)
2219 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
2221 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
2222 checked correctly. This would allow some invalid signed attributes to
2223 appear to verify correctly. (CVE-2009-0591)
2224 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2226 *) Reject UniversalString and BMPString types with invalid lengths. This
2227 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
2228 a legal length. (CVE-2009-0590)
2231 *) Set S/MIME signing as the default purpose rather than setting it
2232 unconditionally. This allows applications to override it at the store
2236 *) Permit restricted recursion of ASN1 strings. This is needed in practice
2237 to handle some structures.
2240 *) Improve efficiency of mem_gets: don't search whole buffer each time
2242 [Jeremy Shapiro <jnshapir@us.ibm.com>]
2244 *) New -hex option for openssl rand.
2247 *) Print out UTF8String and NumericString when parsing ASN1.
2250 *) Support NumericString type for name components.
2253 *) Allow CC in the environment to override the automatically chosen
2254 compiler. Note that nothing is done to ensure flags work with the
2258 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
2260 *) Properly check EVP_VerifyFinal() and similar return values
2262 [Ben Laurie, Bodo Moeller, Google Security Team]
2264 *) Enable TLS extensions by default.
2267 *) Allow the CHIL engine to be loaded, whether the application is
2268 multithreaded or not. (This does not release the developer from the
2269 obligation to set up the dynamic locking callbacks.)
2270 [Sander Temme <sander@temme.net>]
2272 *) Use correct exit code if there is an error in dgst command.
2273 [Steve Henson; problem pointed out by Roland Dirlewanger]
2275 *) Tweak Configure so that you need to say "experimental-jpake" to enable
2276 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
2279 *) Add experimental JPAKE support, including demo authentication in
2280 s_client and s_server.
2283 *) Set the comparison function in v3_addr_canonize().
2284 [Rob Austein <sra@hactrn.net>]
2286 *) Add support for XMPP STARTTLS in s_client.
2287 [Philip Paeps <philip@freebsd.org>]
2289 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
2290 to ensure that even with this option, only ciphersuites in the
2291 server's preference list will be accepted. (Note that the option
2292 applies only when resuming a session, so the earlier behavior was
2293 just about the algorithm choice for symmetric cryptography.)
2296 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
2298 *) Fix NULL pointer dereference if a DTLS server received
2299 ChangeCipherSpec as first record (CVE-2009-1386).
2302 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
2303 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
2306 *) The fix in 0.9.8c that supposedly got rid of unsafe
2307 double-checked locking was incomplete for RSA blinding,
2308 addressing just one layer of what turns out to have been
2309 doubly unsafe triple-checked locking.
2311 So now fix this for real by retiring the MONT_HELPER macro
2312 in crypto/rsa/rsa_eay.c.
2314 [Bodo Moeller; problem pointed out by Marius Schilder]
2316 *) Various precautionary measures:
2318 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
2320 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
2321 (NB: This would require knowledge of the secret session ticket key
2322 to exploit, in which case you'd be SOL either way.)
2324 - Change bn_nist.c so that it will properly handle input BIGNUMs
2325 outside the expected range.
2327 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
2330 [Neel Mehta, Bodo Moeller]
2332 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
2333 the load fails. Useful for distros.
2334 [Ben Laurie and the FreeBSD team]
2336 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
2339 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
2342 *) Expand ENGINE to support engine supplied SSL client certificate functions.
2344 This work was sponsored by Logica.
2347 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
2348 keystores. Support for SSL/TLS client authentication too.
2349 Not compiled unless enable-capieng specified to Configure.
2351 This work was sponsored by Logica.
2354 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
2355 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
2356 attribute creation routines such as certifcate requests and PKCS#12
2360 Changes between 0.9.8g and 0.9.8h [28 May 2008]
2362 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
2363 handshake which could lead to a cilent crash as found using the
2364 Codenomicon TLS test suite (CVE-2008-1672)
2365 [Steve Henson, Mark Cox]
2367 *) Fix double free in TLS server name extensions which could lead to
2368 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
2371 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
2373 Clear the error queue to ensure that error entries left from
2374 older function calls do not interfere with the correct operation.
2375 [Lutz Jaenicke, Erik de Castro Lopo]
2377 *) Remove root CA certificates of commercial CAs:
2379 The OpenSSL project does not recommend any specific CA and does not
2380 have any policy with respect to including or excluding any CA.
2381 Therefore it does not make any sense to ship an arbitrary selection
2382 of root CA certificates with the OpenSSL software.
2385 *) RSA OAEP patches to fix two separate invalid memory reads.
2386 The first one involves inputs when 'lzero' is greater than
2387 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
2388 before the beginning of from). The second one involves inputs where
2389 the 'db' section contains nothing but zeroes (there is a one-byte
2390 invalid read after the end of 'db').
2391 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2393 *) Partial backport from 0.9.9-dev:
2395 Introduce bn_mul_mont (dedicated Montgomery multiplication
2396 procedure) as a candidate for BIGNUM assembler implementation.
2397 While 0.9.9-dev uses assembler for various architectures, only
2398 x86_64 is available by default here in the 0.9.8 branch, and
2399 32-bit x86 is available through a compile-time setting.
2401 To try the 32-bit x86 assembler implementation, use Configure
2402 option "enable-montasm" (which exists only for this backport).
2404 As "enable-montasm" for 32-bit x86 disclaims code stability
2405 anyway, in this constellation we activate additional code
2406 backported from 0.9.9-dev for further performance improvements,
2407 namely BN_from_montgomery_word. (To enable this otherwise,
2408 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
2410 [Andy Polyakov (backport partially by Bodo Moeller)]
2412 *) Add TLS session ticket callback. This allows an application to set
2413 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
2414 values. This is useful for key rollover for example where several key
2415 sets may exist with different names.
2418 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
2419 This was broken until now in 0.9.8 releases, such that the only way
2420 a registered ENGINE could be used (assuming it initialises
2421 successfully on the host) was to explicitly set it as the default
2422 for the relevant algorithms. This is in contradiction with 0.9.7
2423 behaviour and the documentation. With this fix, when an ENGINE is
2424 registered into a given algorithm's table of implementations, the
2425 'uptodate' flag is reset so that auto-discovery will be used next
2426 time a new context for that algorithm attempts to select an
2428 [Ian Lister (tweaked by Geoff Thorpe)]
2430 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
2431 implemention in the following ways:
2433 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
2436 Lack of BER streaming support means one pass streaming processing is
2437 only supported if data is detached: setting the streaming flag is
2438 ignored for embedded content.
2440 CMS support is disabled by default and must be explicitly enabled
2441 with the enable-cms configuration option.
2444 *) Update the GMP engine glue to do direct copies between BIGNUM and
2445 mpz_t when openssl and GMP use the same limb size. Otherwise the
2446 existing "conversion via a text string export" trick is still used.
2447 [Paul Sheer <paulsheer@gmail.com>]
2449 *) Zlib compression BIO. This is a filter BIO which compressed and
2450 uncompresses any data passed through it.
2453 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
2454 RFC3394 compatible AES key wrapping.
2457 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
2458 sets string data without copying. X509_ALGOR_set0() and
2459 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
2460 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
2461 from an X509_ATTRIBUTE structure optionally checking it occurs only
2462 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
2466 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
2467 to get the expected BN_FLG_CONSTTIME behavior.
2468 [Bodo Moeller (Google)]
2472 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
2473 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
2474 - added some more tests to do_tests.pl
2475 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
2476 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
2477 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
2478 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
2479 - various changes to netware.pl to enable gcc-cross builds on Win32
2481 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
2482 - various changes to fix missing prototype warnings
2483 - fixed x86nasm.pl to create correct asm files for NASM COFF output
2484 - added AES, WHIRLPOOL and CPUID assembler code to build files
2485 - added missing AES assembler make rules to mk1mf.pl
2486 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
2487 [Guenter Knauf <eflash@gmx.net>]
2489 *) Implement certificate status request TLS extension defined in RFC3546.
2490 A client can set the appropriate parameters and receive the encoded
2491 OCSP response via a callback. A server can query the supplied parameters
2492 and set the encoded OCSP response in the callback. Add simplified examples
2493 to s_client and s_server.
2496 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
2498 *) Fix various bugs:
2499 + Binary incompatibility of ssl_ctx_st structure
2500 + DTLS interoperation with non-compliant servers
2501 + Don't call get_session_cb() without proposed session
2502 + Fix ia64 assembler code
2503 [Andy Polyakov, Steve Henson]
2505 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
2507 *) DTLS Handshake overhaul. There were longstanding issues with
2508 OpenSSL DTLS implementation, which were making it impossible for
2509 RFC 4347 compliant client to communicate with OpenSSL server.
2510 Unfortunately just fixing these incompatibilities would "cut off"
2511 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
2512 server keeps tolerating non RFC compliant syntax. The opposite is
2513 not true, 0.9.8f client can not communicate with earlier server.
2514 This update even addresses CVE-2007-4995.
2517 *) Changes to avoid need for function casts in OpenSSL: some compilers
2518 (gcc 4.2 and later) reject their use.
2519 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
2522 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2523 RFC4507bis. The encrypted ticket format is an encrypted encoded
2524 SSL_SESSION structure, that way new session features are automatically
2527 If a client application caches session in an SSL_SESSION structure
2528 support is transparent because tickets are now stored in the encoded
2531 The SSL_CTX structure automatically generates keys for ticket
2532 protection in servers so again support should be possible
2533 with no application modification.
2535 If a client or server wishes to disable RFC4507 support then the option
2536 SSL_OP_NO_TICKET can be set.
2538 Add a TLS extension debugging callback to allow the contents of any client
2539 or server extensions to be examined.
2541 This work was sponsored by Google.
2544 *) Add initial support for TLS extensions, specifically for the server_name
2545 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2546 have new members for a host name. The SSL data structure has an
2547 additional member SSL_CTX *initial_ctx so that new sessions can be
2548 stored in that context to allow for session resumption, even after the
2549 SSL has been switched to a new SSL_CTX in reaction to a client's
2550 server_name extension.
2552 New functions (subject to change):
2554 SSL_get_servername()
2555 SSL_get_servername_type()
2558 New CTRL codes and macros (subject to change):
2560 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2561 - SSL_CTX_set_tlsext_servername_callback()
2562 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2563 - SSL_CTX_set_tlsext_servername_arg()
2564 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2566 openssl s_client has a new '-servername ...' option.
2568 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2569 '-key2 ...', '-servername_fatal' (subject to change). This allows
2570 testing the HostName extension for a specific single host name ('-cert'
2571 and '-key' remain fallbacks for handshakes without HostName
2572 negotiation). If the unrecogninzed_name alert has to be sent, this by
2573 default is a warning; it becomes fatal with the '-servername_fatal'
2576 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
2578 *) Add AES and SSE2 assembly language support to VC++ build.
2581 *) Mitigate attack on final subtraction in Montgomery reduction.
2584 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
2585 (which previously caused an internal error).
2588 *) Squeeze another 10% out of IGE mode when in != out.
2591 *) AES IGE mode speedup.
2592 [Dean Gaudet (Google)]
2594 *) Add the Korean symmetric 128-bit cipher SEED (see
2595 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
2596 add SEED ciphersuites from RFC 4162:
2598 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
2599 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
2600 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
2601 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
2603 To minimize changes between patchlevels in the OpenSSL 0.9.8
2604 series, SEED remains excluded from compilation unless OpenSSL
2605 is configured with 'enable-seed'.
2606 [KISA, Bodo Moeller]
2608 *) Mitigate branch prediction attacks, which can be practical if a
2609 single processor is shared, allowing a spy process to extract
2610 information. For detailed background information, see
2611 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
2612 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
2613 and Necessary Software Countermeasures"). The core of the change
2614 are new versions BN_div_no_branch() and
2615 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
2616 respectively, which are slower, but avoid the security-relevant
2617 conditional branches. These are automatically called by BN_div()
2618 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
2619 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
2620 remove a conditional branch.
2622 BN_FLG_CONSTTIME is the new name for the previous
2623 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
2624 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
2625 in the exponent causes BN_mod_exp_mont() to use the alternative
2626 implementation in BN_mod_exp_mont_consttime().) The old name
2627 remains as a deprecated alias.
2629 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
2630 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
2631 constant-time implementations for more than just exponentiation.
2632 Here too the old name is kept as a deprecated alias.
2634 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2635 the BN_BLINDING structure gets an independent copy of the
2636 modulus. This means that the previous "BIGNUM *m" argument to
2637 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2638 essentially becomes "const BIGNUM *m", although we can't actually
2639 change this in the header file before 0.9.9. It allows
2640 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2641 enable BN_FLG_CONSTTIME.
2643 [Matthew D Wood (Intel Corp)]
2645 *) In the SSL/TLS server implementation, be strict about session ID
2646 context matching (which matters if an application uses a single
2647 external cache for different purposes). Previously,
2648 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2649 set. This did ensure strict client verification, but meant that,
2650 with applications using a single external cache for quite
2651 different requirements, clients could circumvent ciphersuite
2652 restrictions for a given session ID context by starting a session
2653 in a different context.
2656 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2657 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2658 authentication-only ciphersuites.
2661 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
2662 not complete and could lead to a possible single byte overflow
2663 (CVE-2007-5135) [Ben Laurie]
2665 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
2667 *) Since AES128 and AES256 (and similarly Camellia128 and
2668 Camellia256) share a single mask bit in the logic of
2669 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2670 kludge to work properly if AES128 is available and AES256 isn't
2671 (or if Camellia128 is available and Camellia256 isn't).
2674 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
2675 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
2676 When a point or a seed is encoded in a BIT STRING, we need to
2677 prevent the removal of trailing zero bits to get the proper DER
2678 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
2679 of a NamedBitList, for which trailing 0 bits need to be removed.)
2682 *) Have SSL/TLS server implementation tolerate "mismatched" record
2683 protocol version while receiving ClientHello even if the
2684 ClientHello is fragmented. (The server can't insist on the
2685 particular protocol version it has chosen before the ServerHello
2686 message has informed the client about his choice.)
2689 *) Add RFC 3779 support.
2690 [Rob Austein for ARIN, Ben Laurie]
2692 *) Load error codes if they are not already present instead of using a
2693 static variable. This allows them to be cleanly unloaded and reloaded.
2694 Improve header file function name parsing.
2697 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
2698 or CAPABILITY handshake as required by RFCs.
2701 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
2703 *) Introduce limits to prevent malicious keys being able to
2704 cause a denial of service. (CVE-2006-2940)
2705 [Steve Henson, Bodo Moeller]
2707 *) Fix ASN.1 parsing of certain invalid structures that can result
2708 in a denial of service. (CVE-2006-2937) [Steve Henson]
2710 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2711 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2713 *) Fix SSL client code which could crash if connecting to a
2714 malicious SSLv2 server. (CVE-2006-4343)
2715 [Tavis Ormandy and Will Drewry, Google Security Team]
2717 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
2718 match only those. Before that, "AES256-SHA" would be interpreted
2719 as a pattern and match "AES128-SHA" too (since AES128-SHA got
2720 the same strength classification in 0.9.7h) as we currently only
2721 have a single AES bit in the ciphersuite description bitmap.
2722 That change, however, also applied to ciphersuite strings such as
2723 "RC4-MD5" that intentionally matched multiple ciphersuites --
2724 namely, SSL 2.0 ciphersuites in addition to the more common ones
2725 from SSL 3.0/TLS 1.0.
2727 So we change the selection algorithm again: Naming an explicit
2728 ciphersuite selects this one ciphersuite, and any other similar
2729 ciphersuite (same bitmap) from *other* protocol versions.
2730 Thus, "RC4-MD5" again will properly select both the SSL 2.0
2731 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
2733 Since SSL 2.0 does not have any ciphersuites for which the
2734 128/256 bit distinction would be relevant, this works for now.
2735 The proper fix will be to use different bits for AES128 and
2736 AES256, which would have avoided the problems from the beginning;
2737 however, bits are scarce, so we can only do this in a new release
2738 (not just a patchlevel) when we can change the SSL_CIPHER
2739 definition to split the single 'unsigned long mask' bitmap into
2740 multiple values to extend the available space.
2744 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2746 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2747 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2749 *) Add AES IGE and biIGE modes.
2752 *) Change the Unix randomness entropy gathering to use poll() when
2753 possible instead of select(), since the latter has some
2754 undesirable limitations.
2755 [Darryl Miles via Richard Levitte and Bodo Moeller]
2757 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2758 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2759 cannot be implicitly activated as part of, e.g., the "AES" alias.
2760 However, please upgrade to OpenSSL 0.9.9[-dev] for
2761 non-experimental use of the ECC ciphersuites to get TLS extension
2762 support, which is required for curve and point format negotiation
2763 to avoid potential handshake problems.
2766 *) Disable rogue ciphersuites:
2768 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2769 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2770 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2772 The latter two were purportedly from
2773 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2776 Also deactivate the remaining ciphersuites from
2777 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2778 unofficial, and the ID has long expired.
2781 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2782 dual-core machines) and other potential thread-safety issues.
2785 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
2786 versions), which is now available for royalty-free use
2787 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
2788 Also, add Camellia TLS ciphersuites from RFC 4132.
2790 To minimize changes between patchlevels in the OpenSSL 0.9.8
2791 series, Camellia remains excluded from compilation unless OpenSSL
2792 is configured with 'enable-camellia'.
2795 *) Disable the padding bug check when compression is in use. The padding
2796 bug check assumes the first packet is of even length, this is not
2797 necessarily true if compresssion is enabled and can result in false
2798 positives causing handshake failure. The actual bug test is ancient
2799 code so it is hoped that implementations will either have fixed it by
2800 now or any which still have the bug do not support compression.
2803 Changes between 0.9.8a and 0.9.8b [04 May 2006]
2805 *) When applying a cipher rule check to see if string match is an explicit
2806 cipher suite and only match that one cipher suite if it is.
2809 *) Link in manifests for VC++ if needed.
2810 [Austin Ziegler <halostatue@gmail.com>]
2812 *) Update support for ECC-based TLS ciphersuites according to
2813 draft-ietf-tls-ecc-12.txt with proposed changes (but without
2814 TLS extensions, which are supported starting with the 0.9.9
2815 branch, not in the OpenSSL 0.9.8 branch).
2818 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
2819 opaque EVP_CIPHER_CTX handling.
2822 *) Fixes and enhancements to zlib compression code. We now only use
2823 "zlib1.dll" and use the default __cdecl calling convention on Win32
2824 to conform with the standards mentioned here:
2825 http://www.zlib.net/DLL_FAQ.txt
2826 Static zlib linking now works on Windows and the new --with-zlib-include
2827 --with-zlib-lib options to Configure can be used to supply the location
2828 of the headers and library. Gracefully handle case where zlib library
2832 *) Several fixes and enhancements to the OID generation code. The old code
2833 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
2834 handle numbers larger than ULONG_MAX, truncated printing and had a
2835 non standard OBJ_obj2txt() behaviour.
2838 *) Add support for building of engines under engine/ as shared libraries
2839 under VC++ build system.
2842 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
2843 Hopefully, we will not see any false combination of paths any more.
2846 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
2848 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2849 (part of SSL_OP_ALL). This option used to disable the
2850 countermeasure against man-in-the-middle protocol-version
2851 rollback in the SSL 2.0 server implementation, which is a bad
2852 idea. (CVE-2005-2969)
2854 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2855 for Information Security, National Institute of Advanced Industrial
2856 Science and Technology [AIST], Japan)]
2858 *) Add two function to clear and return the verify parameter flags.
2861 *) Keep cipherlists sorted in the source instead of sorting them at
2862 runtime, thus removing the need for a lock.
2865 *) Avoid some small subgroup attacks in Diffie-Hellman.
2866 [Nick Mathewson and Ben Laurie]
2868 *) Add functions for well-known primes.
2871 *) Extended Windows CE support.
2872 [Satoshi Nakamura and Andy Polyakov]
2874 *) Initialize SSL_METHOD structures at compile time instead of during
2875 runtime, thus removing the need for a lock.
2878 *) Make PKCS7_decrypt() work even if no certificate is supplied by
2879 attempting to decrypt each encrypted key in turn. Add support to
2883 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
2885 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
2888 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
2891 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
2892 key into the same file any more.
2895 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
2898 *) Add -utf8 command line and config file option to 'ca'.
2899 [Stefan <stf@udoma.org]
2901 *) Removed the macro des_crypt(), as it seems to conflict with some
2902 libraries. Use DES_crypt().
2905 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
2906 involves renaming the source and generated shared-libs for
2907 both. The engines will accept the corrected or legacy ids
2908 ('ncipher' and '4758_cca' respectively) when binding. NB,
2909 this only applies when building 'shared'.
2910 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
2912 *) Add attribute functions to EVP_PKEY structure. Modify
2913 PKCS12_create() to recognize a CSP name attribute and
2914 use it. Make -CSP option work again in pkcs12 utility.
2917 *) Add new functionality to the bn blinding code:
2918 - automatic re-creation of the BN_BLINDING parameters after
2919 a fixed number of uses (currently 32)
2920 - add new function for parameter creation
2921 - introduce flags to control the update behaviour of the
2922 BN_BLINDING parameters
2923 - hide BN_BLINDING structure
2924 Add a second BN_BLINDING slot to the RSA structure to improve
2925 performance when a single RSA object is shared among several
2929 *) Add support for DTLS.
2930 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
2932 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
2933 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
2936 *) Remove buggy and incompletet DH cert support from
2937 ssl/ssl_rsa.c and ssl/s3_both.c
2940 *) Use SHA-1 instead of MD5 as the default digest algorithm for
2941 the apps/openssl applications.
2944 *) Compile clean with "-Wall -Wmissing-prototypes
2945 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
2946 DEBUG_SAFESTACK must also be set.
2949 *) Change ./Configure so that certain algorithms can be disabled by default.
2950 The new counterpiece to "no-xxx" is "enable-xxx".
2952 The patented RC5 and MDC2 algorithms will now be disabled unless
2953 "enable-rc5" and "enable-mdc2", respectively, are specified.
2955 (IDEA remains enabled despite being patented. This is because IDEA
2956 is frequently required for interoperability, and there is no license
2957 fee for non-commercial use. As before, "no-idea" can be used to
2958 avoid this algorithm.)
2962 *) Add processing of proxy certificates (see RFC 3820). This work was
2963 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
2964 EGEE (Enabling Grids for E-science in Europe).
2967 *) RC4 performance overhaul on modern architectures/implementations, such
2968 as Intel P4, IA-64 and AMD64.
2971 *) New utility extract-section.pl. This can be used specify an alternative
2972 section number in a pod file instead of having to treat each file as
2973 a separate case in Makefile. This can be done by adding two lines to the
2976 =for comment openssl_section:XXX
2978 The blank line is mandatory.
2982 *) New arguments -certform, -keyform and -pass for s_client and s_server
2983 to allow alternative format key and certificate files and passphrase
2987 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
2988 update associated structures and add various utility functions.
2990 Add new policy related verify parameters, include policy checking in
2991 standard verify code. Enhance 'smime' application with extra parameters
2992 to support policy checking and print out.
2995 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
2996 Nehemiah processors. These extensions support AES encryption in hardware
2997 as well as RNG (though RNG support is currently disabled).
2998 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
3000 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
3003 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
3004 [Andy Polyakov and a number of other people]
3006 *) Improved PowerPC platform support. Most notably BIGNUM assembler
3007 implementation contributed by IBM.
3008 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
3010 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
3011 exponent rather than 'unsigned long'. There is a corresponding change to
3012 the new 'rsa_keygen' element of the RSA_METHOD structure.
3013 [Jelte Jansen, Geoff Thorpe]
3015 *) Functionality for creating the initial serial number file is now
3016 moved from CA.pl to the 'ca' utility with a new option -create_serial.
3018 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
3019 number file to 1, which is bound to cause problems. To avoid
3020 the problems while respecting compatibility between different 0.9.7
3021 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
3022 CA.pl for serial number initialization. With the new release 0.9.8,
3023 we can fix the problem directly in the 'ca' utility.)
3026 *) Reduced header interdepencies by declaring more opaque objects in
3027 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
3028 give fewer recursive includes, which could break lazy source code - so
3029 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
3030 developers should define this symbol when building and using openssl to
3031 ensure they track the recommended behaviour, interfaces, [etc], but
3032 backwards-compatible behaviour prevails when this isn't defined.
3035 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
3038 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
3039 This will generate a random key of the appropriate length based on the
3040 cipher context. The EVP_CIPHER can provide its own random key generation
3041 routine to support keys of a specific form. This is used in the des and
3042 3des routines to generate a key of the correct parity. Update S/MIME
3043 code to use new functions and hence generate correct parity DES keys.
3044 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
3045 valid (weak or incorrect parity).
3048 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
3049 as looking them up. This is useful when the verified structure may contain
3050 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
3051 present unless the new PKCS7_NO_CRL flag is asserted.
3054 *) Extend ASN1 oid configuration module. It now additionally accepts the
3057 shortName = some long name, 1.2.3.4
3060 *) Reimplemented the BN_CTX implementation. There is now no more static
3061 limitation on the number of variables it can handle nor the depth of the
3062 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
3063 information can now expand as required, and rather than having a single
3064 static array of bignums, BN_CTX now uses a linked-list of such arrays
3065 allowing it to expand on demand whilst maintaining the usefulness of
3066 BN_CTX's "bundling".
3069 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
3070 to allow all RSA operations to function using a single BN_CTX.
3073 *) Preliminary support for certificate policy evaluation and checking. This
3074 is initially intended to pass the tests outlined in "Conformance Testing
3075 of Relying Party Client Certificate Path Processing Logic" v1.07.
3078 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
3079 remained unused and not that useful. A variety of other little bignum
3080 tweaks and fixes have also been made continuing on from the audit (see
3084 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
3085 associated ASN1, EVP and SSL functions and old ASN1 macros.
3088 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
3089 and this should never fail. So the return value from the use of
3090 BN_set_word() (which can fail due to needless expansion) is now deprecated;
3091 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
3094 *) BN_CTX_get() should return zero-valued bignums, providing the same
3095 initialised value as BN_new().
3096 [Geoff Thorpe, suggested by Ulf Möller]
3098 *) Support for inhibitAnyPolicy certificate extension.
3101 *) An audit of the BIGNUM code is underway, for which debugging code is
3102 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
3103 is considered valid when processing BIGNUMs, and causes execution to
3104 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
3105 further steps are taken to deliberately pollute unused data in BIGNUM
3106 structures to try and expose faulty code further on. For now, openssl will
3107 (in its default mode of operation) continue to tolerate the inconsistent
3108 forms that it has tolerated in the past, but authors and packagers should
3109 consider trying openssl and their own applications when compiled with
3110 these debugging symbols defined. It will help highlight potential bugs in
3111 their own code, and will improve the test coverage for OpenSSL itself. At
3112 some point, these tighter rules will become openssl's default to improve
3113 maintainability, though the assert()s and other overheads will remain only
3114 in debugging configurations. See bn.h for more details.
3115 [Geoff Thorpe, Nils Larsch, Ulf Möller]
3117 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
3118 that can only be obtained through BN_CTX_new() (which implicitly
3119 initialises it). The presence of this function only made it possible
3120 to overwrite an existing structure (and cause memory leaks).
3123 *) Because of the callback-based approach for implementing LHASH as a
3124 template type, lh_insert() adds opaque objects to hash-tables and
3125 lh_doall() or lh_doall_arg() are typically used with a destructor callback
3126 to clean up those corresponding objects before destroying the hash table
3127 (and losing the object pointers). So some over-zealous constifications in
3128 LHASH have been relaxed so that lh_insert() does not take (nor store) the
3129 objects as "const" and the lh_doall[_arg] callback wrappers are not
3130 prototyped to have "const" restrictions on the object pointers they are
3131 given (and so aren't required to cast them away any more).
3134 *) The tmdiff.h API was so ugly and minimal that our own timing utility
3135 (speed) prefers to use its own implementation. The two implementations
3136 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
3137 its object type properly exposed (MS_TM) instead of casting to/from "char
3138 *". This may still change yet if someone realises MS_TM and "ms_time_***"
3139 aren't necessarily the greatest nomenclatures - but this is what was used
3140 internally to the implementation so I've used that for now.
3143 *) Ensure that deprecated functions do not get compiled when
3144 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
3145 the self-tests were still using deprecated key-generation functions so
3146 these have been updated also.
3149 *) Reorganise PKCS#7 code to separate the digest location functionality
3150 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
3151 New function PKCS7_set_digest() to set the digest type for PKCS#7
3152 digestedData type. Add additional code to correctly generate the
3153 digestedData type and add support for this type in PKCS7 initialization
3157 *) New function PKCS7_set0_type_other() this initializes a PKCS7
3158 structure of type "other".
3161 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
3162 sure the loop does correctly stop and breaking ("division by zero")
3163 modulus operations are not performed. The (pre-generated) prime
3164 table crypto/bn/bn_prime.h was already correct, but it could not be
3165 re-generated on some platforms because of the "division by zero"
3166 situation in the script.
3167 [Ralf S. Engelschall]
3169 *) Update support for ECC-based TLS ciphersuites according to
3170 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
3171 SHA-1 now is only used for "small" curves (where the
3172 representation of a field element takes up to 24 bytes); for
3173 larger curves, the field element resulting from ECDH is directly
3174 used as premaster secret.
3175 [Douglas Stebila (Sun Microsystems Laboratories)]
3177 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
3178 curve secp160r1 to the tests.
3179 [Douglas Stebila (Sun Microsystems Laboratories)]
3181 *) Add the possibility to load symbols globally with DSO.
3182 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
3184 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
3185 control of the error stack.
3188 *) Add support for STORE in ENGINE.
3191 *) Add the STORE type. The intention is to provide a common interface
3192 to certificate and key stores, be they simple file-based stores, or
3193 HSM-type store, or LDAP stores, or...
3194 NOTE: The code is currently UNTESTED and isn't really used anywhere.
3197 *) Add a generic structure called OPENSSL_ITEM. This can be used to
3198 pass a list of arguments to any function as well as provide a way
3199 for a function to pass data back to the caller.
3202 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
3203 works like BUF_strdup() but can be used to duplicate a portion of
3204 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
3208 *) Add the function sk_find_ex() which works like sk_find(), but will
3209 return an index to an element even if an exact match couldn't be
3210 found. The index is guaranteed to point at the element where the
3211 searched-for key would be inserted to preserve sorting order.
3214 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
3215 takes an extra flags argument for optional functionality. Currently,
3216 the following flags are defined:
3218 OBJ_BSEARCH_VALUE_ON_NOMATCH
3219 This one gets OBJ_bsearch_ex() to return a pointer to the first
3220 element where the comparing function returns a negative or zero
3223 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
3224 This one gets OBJ_bsearch_ex() to return a pointer to the first
3225 element where the comparing function returns zero. This is useful
3226 if there are more than one element where the comparing function
3230 *) Make it possible to create self-signed certificates with 'openssl ca'
3231 in such a way that the self-signed certificate becomes part of the
3232 CA database and uses the same mechanisms for serial number generation
3233 as all other certificate signing. The new flag '-selfsign' enables
3234 this functionality. Adapt CA.sh and CA.pl.in.
3237 *) Add functionality to check the public key of a certificate request
3238 against a given private. This is useful to check that a certificate
3239 request can be signed by that key (self-signing).
3242 *) Make it possible to have multiple active certificates with the same
3243 subject in the CA index file. This is done only if the keyword
3244 'unique_subject' is set to 'no' in the main CA section (default
3245 if 'CA_default') of the configuration file. The value is saved
3246 with the database itself in a separate index attribute file,
3247 named like the index file with '.attr' appended to the name.
3250 *) Generate muti valued AVAs using '+' notation in config files for
3254 *) Support for nameConstraints certificate extension.
3257 *) Support for policyConstraints certificate extension.
3260 *) Support for policyMappings certificate extension.
3263 *) Make sure the default DSA_METHOD implementation only uses its
3264 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
3265 and change its own handlers to be NULL so as to remove unnecessary
3266 indirection. This lets alternative implementations fallback to the
3267 default implementation more easily.
3270 *) Support for directoryName in GeneralName related extensions
3274 *) Make it possible to link applications using Makefile.shared.
3275 Make that possible even when linking against static libraries!
3278 *) Support for single pass processing for S/MIME signing. This now
3279 means that S/MIME signing can be done from a pipe, in addition
3280 cleartext signing (multipart/signed type) is effectively streaming
3281 and the signed data does not need to be all held in memory.
3283 This is done with a new flag PKCS7_STREAM. When this flag is set
3284 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
3285 is done after the data is output (and digests calculated) in
3286 SMIME_write_PKCS7().
3289 *) Add full support for -rpath/-R, both in shared libraries and
3290 applications, at least on the platforms where it's known how
3294 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
3295 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
3296 will now compute a table of multiples of the generator that
3297 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
3298 faster (notably in the case of a single point multiplication,
3299 scalar * generator).
3300 [Nils Larsch, Bodo Moeller]
3302 *) IPv6 support for certificate extensions. The various extensions
3303 which use the IP:a.b.c.d can now take IPv6 addresses using the
3304 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
3308 *) Added an ENGINE that implements RSA by performing private key
3309 exponentiations with the GMP library. The conversions to and from
3310 GMP's mpz_t format aren't optimised nor are any montgomery forms
3311 cached, and on x86 it appears OpenSSL's own performance has caught up.
3312 However there are likely to be other architectures where GMP could
3313 provide a boost. This ENGINE is not built in by default, but it can be
3314 specified at Configure time and should be accompanied by the necessary
3315 linker additions, eg;
3316 ./config -DOPENSSL_USE_GMP -lgmp
3319 *) "openssl engine" will not display ENGINE/DSO load failure errors when
3320 testing availability of engines with "-t" - the old behaviour is
3321 produced by increasing the feature's verbosity with "-tt".
3324 *) ECDSA routines: under certain error conditions uninitialized BN objects
3325 could be freed. Solution: make sure initialization is performed early
3326 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
3330 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
3331 and DH_METHOD (eg. by ENGINE implementations) to override the normal
3332 software implementations. For DSA and DH, parameter generation can
3333 also be overriden by providing the appropriate method callbacks.
3336 *) Change the "progress" mechanism used in key-generation and
3337 primality testing to functions that take a new BN_GENCB pointer in
3338 place of callback/argument pairs. The new API functions have "_ex"
3339 postfixes and the older functions are reimplemented as wrappers for
3340 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
3341 declarations of the old functions to help (graceful) attempts to
3342 migrate to the new functions. Also, the new key-generation API
3343 functions operate on a caller-supplied key-structure and return
3344 success/failure rather than returning a key or NULL - this is to
3345 help make "keygen" another member function of RSA_METHOD etc.
3347 Example for using the new callback interface:
3349 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
3353 BN_GENCB_set(&my_cb, my_callback, my_arg);
3355 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
3356 /* For the meaning of a, b in calls to my_callback(), see the
3357 * documentation of the function that calls the callback.
3358 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
3359 * my_callback should return 1 if it wants BN_is_prime_ex()
3360 * to continue, or 0 to stop.
3365 *) Change the ZLIB compression method to be stateful, and make it
3366 available to TLS with the number defined in
3367 draft-ietf-tls-compression-04.txt.
3370 *) Add the ASN.1 structures and functions for CertificatePair, which
3371 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
3373 CertificatePair ::= SEQUENCE {
3374 forward [0] Certificate OPTIONAL,
3375 reverse [1] Certificate OPTIONAL,
3376 -- at least one of the pair shall be present -- }
3378 Also implement the PEM functions to read and write certificate
3379 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
3381 This needed to be defined, mostly for the sake of the LDAP
3382 attribute crossCertificatePair, but may prove useful elsewhere as
3386 *) Make it possible to inhibit symlinking of shared libraries in
3387 Makefile.shared, for Cygwin's sake.
3390 *) Extend the BIGNUM API by creating a function
3391 void BN_set_negative(BIGNUM *a, int neg);
3392 and a macro that behave like
3393 int BN_is_negative(const BIGNUM *a);
3395 to avoid the need to access 'a->neg' directly in applications.
3398 *) Implement fast modular reduction for pseudo-Mersenne primes
3399 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
3400 EC_GROUP_new_curve_GFp() will now automatically use this
3402 [Nils Larsch <nla@trustcenter.de>]
3404 *) Add new lock type (CRYPTO_LOCK_BN).
3407 *) Change the ENGINE framework to automatically load engines
3408 dynamically from specific directories unless they could be
3409 found to already be built in or loaded. Move all the
3410 current engines except for the cryptodev one to a new
3412 The engines in engines/ are built as shared libraries if
3413 the "shared" options was given to ./Configure or ./config.
3414 Otherwise, they are inserted in libcrypto.a.
3415 /usr/local/ssl/engines is the default directory for dynamic
3416 engines, but that can be overriden at configure time through
3417 the usual use of --prefix and/or --openssldir, and at run
3418 time with the environment variable OPENSSL_ENGINES.
3419 [Geoff Thorpe and Richard Levitte]
3421 *) Add Makefile.shared, a helper makefile to build shared
3422 libraries. Addapt Makefile.org.
3425 *) Add version info to Win32 DLLs.
3426 [Peter 'Luna' Runestig" <peter@runestig.com>]
3428 *) Add new 'medium level' PKCS#12 API. Certificates and keys
3429 can be added using this API to created arbitrary PKCS#12
3430 files while avoiding the low level API.
3432 New options to PKCS12_create(), key or cert can be NULL and
3433 will then be omitted from the output file. The encryption
3434 algorithm NIDs can be set to -1 for no encryption, the mac
3435 iteration count can be set to 0 to omit the mac.
3437 Enhance pkcs12 utility by making the -nokeys and -nocerts
3438 options work when creating a PKCS#12 file. New option -nomac
3439 to omit the mac, NONE can be set for an encryption algorithm.
3440 New code is modified to use the enhanced PKCS12_create()
3441 instead of the low level API.
3444 *) Extend ASN1 encoder to support indefinite length constructed
3445 encoding. This can output sequences tags and octet strings in
3446 this form. Modify pk7_asn1.c to support indefinite length
3447 encoding. This is experimental and needs additional code to
3448 be useful, such as an ASN1 bio and some enhanced streaming
3451 Extend template encode functionality so that tagging is passed
3452 down to the template encoder.
3455 *) Let 'openssl req' fail if an argument to '-newkey' is not
3456 recognized instead of using RSA as a default.
3459 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
3460 As these are not official, they are not included in "ALL";
3461 the "ECCdraft" ciphersuite group alias can be used to select them.
3462 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
3464 *) Add ECDH engine support.
3465 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
3467 *) Add ECDH in new directory crypto/ecdh/.
3468 [Douglas Stebila (Sun Microsystems Laboratories)]
3470 *) Let BN_rand_range() abort with an error after 100 iterations
3471 without success (which indicates a broken PRNG).
3474 *) Change BN_mod_sqrt() so that it verifies that the input value
3475 is really the square of the return value. (Previously,
3476 BN_mod_sqrt would show GIGO behaviour.)
3479 *) Add named elliptic curves over binary fields from X9.62, SECG,
3480 and WAP/WTLS; add OIDs that were still missing.
3482 [Sheueling Chang Shantz and Douglas Stebila
3483 (Sun Microsystems Laboratories)]
3485 *) Extend the EC library for elliptic curves over binary fields
3486 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
3489 EC_GF2m_simple_method
3493 EC_GROUP_new_curve_GF2m
3494 EC_GROUP_set_curve_GF2m
3495 EC_GROUP_get_curve_GF2m
3496 EC_POINT_set_affine_coordinates_GF2m
3497 EC_POINT_get_affine_coordinates_GF2m
3498 EC_POINT_set_compressed_coordinates_GF2m
3500 Point compression for binary fields is disabled by default for
3501 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
3504 As binary polynomials are represented as BIGNUMs, various members
3505 of the EC_GROUP and EC_POINT data structures can be shared
3506 between the implementations for prime fields and binary fields;
3507 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
3508 are essentially identical to their ..._GFp counterparts.
3509 (For simplicity, the '..._GFp' prefix has been dropped from
3510 various internal method names.)
3512 An internal 'field_div' method (similar to 'field_mul' and
3513 'field_sqr') has been added; this is used only for binary fields.
3515 [Sheueling Chang Shantz and Douglas Stebila
3516 (Sun Microsystems Laboratories)]
3518 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
3519 through methods ('mul', 'precompute_mult').
3521 The generic implementations (now internally called 'ec_wNAF_mul'
3522 and 'ec_wNAF_precomputed_mult') remain the default if these
3523 methods are undefined.
3525 [Sheueling Chang Shantz and Douglas Stebila
3526 (Sun Microsystems Laboratories)]
3528 *) New function EC_GROUP_get_degree, which is defined through
3529 EC_METHOD. For curves over prime fields, this returns the bit
3530 length of the modulus.
3532 [Sheueling Chang Shantz and Douglas Stebila
3533 (Sun Microsystems Laboratories)]
3535 *) New functions EC_GROUP_dup, EC_POINT_dup.
3536 (These simply call ..._new and ..._copy).
3538 [Sheueling Chang Shantz and Douglas Stebila
3539 (Sun Microsystems Laboratories)]
3541 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
3542 Polynomials are represented as BIGNUMs (where the sign bit is not
3543 used) in the following functions [macros]:
3546 BN_GF2m_sub [= BN_GF2m_add]
3547 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
3548 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
3549 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
3551 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
3552 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
3553 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
3554 BN_GF2m_cmp [= BN_ucmp]
3556 (Note that only the 'mod' functions are actually for fields GF(2^m).
3557 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
3559 For some functions, an the irreducible polynomial defining a
3560 field can be given as an 'unsigned int[]' with strictly
3561 decreasing elements giving the indices of those bits that are set;
3562 i.e., p[] represents the polynomial
3563 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
3565 p[0] > p[1] > ... > p[k] = 0.
3566 This applies to the following functions:
3571 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
3572 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
3574 BN_GF2m_mod_sqrt_arr
3575 BN_GF2m_mod_solve_quad_arr
3579 Conversion can be performed by the following functions:
3584 bntest.c has additional tests for binary polynomial arithmetic.
3586 Two implementations for BN_GF2m_mod_div() are available.
3587 The default algorithm simply uses BN_GF2m_mod_inv() and
3588 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
3589 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
3590 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
3592 [Sheueling Chang Shantz and Douglas Stebila
3593 (Sun Microsystems Laboratories)]
3595 *) Add new error code 'ERR_R_DISABLED' that can be used when some
3596 functionality is disabled at compile-time.
3597 [Douglas Stebila <douglas.stebila@sun.com>]
3599 *) Change default behaviour of 'openssl asn1parse' so that more
3600 information is visible when viewing, e.g., a certificate:
3602 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
3603 mode the content of non-printable OCTET STRINGs is output in a
3604 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
3605 avoid the appearance of a printable string.
3606 [Nils Larsch <nla@trustcenter.de>]
3608 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
3610 EC_GROUP_set_asn1_flag()
3611 EC_GROUP_get_asn1_flag()
3612 EC_GROUP_set_point_conversion_form()
3613 EC_GROUP_get_point_conversion_form()
3614 These control ASN1 encoding details:
3615 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
3616 has been set to OPENSSL_EC_NAMED_CURVE.
3617 - Points are encoded in uncompressed form by default; options for
3618 asn1_for are as for point2oct, namely
3619 POINT_CONVERSION_COMPRESSED
3620 POINT_CONVERSION_UNCOMPRESSED
3621 POINT_CONVERSION_HYBRID
3623 Also add 'seed' and 'seed_len' members to EC_GROUP with access
3626 EC_GROUP_get0_seed()
3627 EC_GROUP_get_seed_len()
3628 This is used only for ASN1 purposes (so far).
3629 [Nils Larsch <nla@trustcenter.de>]
3631 *) Add 'field_type' member to EC_METHOD, which holds the NID
3632 of the appropriate field type OID. The new function
3633 EC_METHOD_get_field_type() returns this value.
3634 [Nils Larsch <nla@trustcenter.de>]
3639 EC_POINT_point2hex()
3640 EC_POINT_hex2point()
3641 providing useful interfaces to EC_POINT_point2oct() and
3642 EC_POINT_oct2point().
3643 [Nils Larsch <nla@trustcenter.de>]
3645 *) Change internals of the EC library so that the functions
3646 EC_GROUP_set_generator()
3647 EC_GROUP_get_generator()
3648 EC_GROUP_get_order()
3649 EC_GROUP_get_cofactor()
3650 are implemented directly in crypto/ec/ec_lib.c and not dispatched
3651 to methods, which would lead to unnecessary code duplication when
3652 adding different types of curves.
3653 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
3655 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
3656 arithmetic, and such that modified wNAFs are generated
3657 (which avoid length expansion in many cases).
3660 *) Add a function EC_GROUP_check_discriminant() (defined via
3661 EC_METHOD) that verifies that the curve discriminant is non-zero.
3663 Add a function EC_GROUP_check() that makes some sanity tests
3664 on a EC_GROUP, its generator and order. This includes
3665 EC_GROUP_check_discriminant().
3666 [Nils Larsch <nla@trustcenter.de>]
3668 *) Add ECDSA in new directory crypto/ecdsa/.
3670 Add applications 'openssl ecparam' and 'openssl ecdsa'
3671 (these are based on 'openssl dsaparam' and 'openssl dsa').
3673 ECDSA support is also included in various other files across the
3674 library. Most notably,
3675 - 'openssl req' now has a '-newkey ecdsa:file' option;
3676 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
3677 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
3678 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
3679 them suitable for ECDSA where domain parameters must be
3680 extracted before the specific public key;
3681 - ECDSA engine support has been added.
3682 [Nils Larsch <nla@trustcenter.de>]
3684 *) Include some named elliptic curves, and add OIDs from X9.62,
3685 SECG, and WAP/WTLS. Each curve can be obtained from the new
3687 EC_GROUP_new_by_curve_name(),
3688 and the list of available named curves can be obtained with
3689 EC_get_builtin_curves().
3690 Also add a 'curve_name' member to EC_GROUP objects, which can be
3692 EC_GROUP_set_curve_name()
3693 EC_GROUP_get_curve_name()
3694 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
3696 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3697 was actually never needed) and in BN_mul(). The removal in BN_mul()
3698 required a small change in bn_mul_part_recursive() and the addition
3699 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3700 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3701 bn_sub_words() and bn_add_words() except they take arrays with
3705 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
3707 *) Cleanse PEM buffers before freeing them since they may contain
3709 [Benjamin Bennett <ben@psc.edu>]
3711 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3712 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3713 authentication-only ciphersuites.
3716 *) Since AES128 and AES256 share a single mask bit in the logic of
3717 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3718 kludge to work properly if AES128 is available and AES256 isn't.
3721 *) Expand security boundary to match 1.1.1 module.
3724 *) Remove redundant features: hash file source, editing of test vectors
3725 modify fipsld to use external fips_premain.c signature.
3728 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
3729 run algorithm test programs.
3732 *) Make algorithm test programs more tolerant of whitespace.
3735 *) Have SSL/TLS server implementation tolerate "mismatched" record
3736 protocol version while receiving ClientHello even if the
3737 ClientHello is fragmented. (The server can't insist on the
3738 particular protocol version it has chosen before the ServerHello
3739 message has informed the client about his choice.)
3742 *) Load error codes if they are not already present instead of using a
3743 static variable. This allows them to be cleanly unloaded and reloaded.
3746 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3748 *) Introduce limits to prevent malicious keys being able to
3749 cause a denial of service. (CVE-2006-2940)
3750 [Steve Henson, Bodo Moeller]
3752 *) Fix ASN.1 parsing of certain invalid structures that can result
3753 in a denial of service. (CVE-2006-2937) [Steve Henson]
3755 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3756 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3758 *) Fix SSL client code which could crash if connecting to a
3759 malicious SSLv2 server. (CVE-2006-4343)
3760 [Tavis Ormandy and Will Drewry, Google Security Team]
3762 *) Change ciphersuite string processing so that an explicit
3763 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3764 will no longer include "AES128-SHA"), and any other similar
3765 ciphersuite (same bitmap) from *other* protocol versions (so that
3766 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3767 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3768 changes from 0.9.8b and 0.9.8d.
3771 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
3773 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3774 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3776 *) Change the Unix randomness entropy gathering to use poll() when
3777 possible instead of select(), since the latter has some
3778 undesirable limitations.
3779 [Darryl Miles via Richard Levitte and Bodo Moeller]
3781 *) Disable rogue ciphersuites:
3783 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3784 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3785 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3787 The latter two were purportedly from
3788 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3791 Also deactive the remaining ciphersuites from
3792 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3793 unofficial, and the ID has long expired.
3796 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3797 dual-core machines) and other potential thread-safety issues.
3800 Changes between 0.9.7i and 0.9.7j [04 May 2006]
3802 *) Adapt fipsld and the build system to link against the validated FIPS
3803 module in FIPS mode.
3806 *) Fixes for VC++ 2005 build under Windows.
3809 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
3810 from a Windows bash shell such as MSYS. It is autodetected from the
3811 "config" script when run from a VC++ environment. Modify standard VC++
3812 build to use fipscanister.o from the GNU make build.
3815 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
3817 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
3818 The value now differs depending on if you build for FIPS or not.
3819 BEWARE! A program linked with a shared FIPSed libcrypto can't be
3820 safely run with a non-FIPSed libcrypto, as it may crash because of
3821 the difference induced by this change.
3824 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
3826 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3827 (part of SSL_OP_ALL). This option used to disable the
3828 countermeasure against man-in-the-middle protocol-version
3829 rollback in the SSL 2.0 server implementation, which is a bad
3830 idea. (CVE-2005-2969)
3832 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3833 for Information Security, National Institute of Advanced Industrial
3834 Science and Technology [AIST], Japan)]
3836 *) Minimal support for X9.31 signatures and PSS padding modes. This is
3837 mainly for FIPS compliance and not fully integrated at this stage.
3840 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
3841 the exponentiation using a fixed-length exponent. (Otherwise,
3842 the information leaked through timing could expose the secret key
3843 after many signatures; cf. Bleichenbacher's attack on DSA with
3847 *) Make a new fixed-window mod_exp implementation the default for
3848 RSA, DSA, and DH private-key operations so that the sequence of
3849 squares and multiplies and the memory access pattern are
3850 independent of the particular secret key. This will mitigate
3851 cache-timing and potential related attacks.
3853 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
3854 and this is automatically used by BN_mod_exp_mont() if the new flag
3855 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
3856 will use this BN flag for private exponents unless the flag
3857 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
3858 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
3860 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
3862 *) Change the client implementation for SSLv23_method() and
3863 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
3864 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
3865 (Previously, the SSL 2.0 backwards compatible Client Hello
3866 message format would be used even with SSL_OP_NO_SSLv2.)
3869 *) Add support for smime-type MIME parameter in S/MIME messages which some
3873 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
3874 a threadsafe manner. Modify rsa code to use new function and add calls
3875 to dsa and dh code (which had race conditions before).
3878 *) Include the fixed error library code in the C error file definitions
3879 instead of fixing them up at runtime. This keeps the error code
3880 structures constant.
3883 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
3885 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
3888 *) Fixes for newer kerberos headers. NB: the casts are needed because
3889 the 'length' field is signed on one version and unsigned on another
3890 with no (?) obvious way to tell the difference, without these VC++
3891 complains. Also the "definition" of FAR (blank) is no longer included
3892 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
3893 some needed definitions.
3896 *) Undo Cygwin change.
3899 *) Added support for proxy certificates according to RFC 3820.
3900 Because they may be a security thread to unaware applications,
3901 they must be explicitely allowed in run-time. See
3902 docs/HOWTO/proxy_certificates.txt for further information.
3905 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
3907 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
3908 server and client random values. Previously
3909 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
3910 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
3912 This change has negligible security impact because:
3914 1. Server and client random values still have 24 bytes of pseudo random
3917 2. Server and client random values are sent in the clear in the initial
3920 3. The master secret is derived using the premaster secret (48 bytes in
3921 size for static RSA ciphersuites) as well as client server and random
3924 The OpenSSL team would like to thank the UK NISCC for bringing this issue
3927 [Stephen Henson, reported by UK NISCC]
3929 *) Use Windows randomness collection on Cygwin.
3932 *) Fix hang in EGD/PRNGD query when communication socket is closed
3933 prematurely by EGD/PRNGD.
3934 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
3936 *) Prompt for pass phrases when appropriate for PKCS12 input format.
3939 *) Back-port of selected performance improvements from development
3940 branch, as well as improved support for PowerPC platforms.
3943 *) Add lots of checks for memory allocation failure, error codes to indicate
3944 failure and freeing up memory if a failure occurs.
3945 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
3947 *) Add new -passin argument to dgst.
3950 *) Perform some character comparisons of different types in X509_NAME_cmp:
3951 this is needed for some certificates that reencode DNs into UTF8Strings
3952 (in violation of RFC3280) and can't or wont issue name rollover
3956 *) Make an explicit check during certificate validation to see that
3957 the CA setting in each certificate on the chain is correct. As a
3958 side effect always do the following basic checks on extensions,
3959 not just when there's an associated purpose to the check:
3961 - if there is an unhandled critical extension (unless the user
3962 has chosen to ignore this fault)
3963 - if the path length has been exceeded (if one is set at all)
3964 - that certain extensions fit the associated purpose (if one has
3968 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
3970 *) Avoid a race condition when CRLs are checked in a multi threaded
3971 environment. This would happen due to the reordering of the revoked
3972 entries during signature checking and serial number lookup. Now the
3973 encoding is cached and the serial number sort performed under a lock.
3974 Add new STACK function sk_is_sorted().
3977 *) Add Delta CRL to the extension code.
3980 *) Various fixes to s3_pkt.c so alerts are sent properly.
3981 [David Holmes <d.holmes@f5.com>]
3983 *) Reduce the chances of duplicate issuer name and serial numbers (in
3984 violation of RFC3280) using the OpenSSL certificate creation utilities.
3985 This is done by creating a random 64 bit value for the initial serial
3986 number when a serial number file is created or when a self signed
3987 certificate is created using 'openssl req -x509'. The initial serial
3988 number file is created using 'openssl x509 -next_serial' in CA.pl
3989 rather than being initialized to 1.
3992 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
3994 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3995 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3996 [Joe Orton, Steve Henson]
3998 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
4000 [Joe Orton, Steve Henson]
4002 *) Make it possible to have multiple active certificates with the same
4003 subject in the CA index file. This is done only if the keyword
4004 'unique_subject' is set to 'no' in the main CA section (default
4005 if 'CA_default') of the configuration file. The value is saved
4006 with the database itself in a separate index attribute file,
4007 named like the index file with '.attr' appended to the name.
4010 *) X509 verify fixes. Disable broken certificate workarounds when
4011 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
4012 keyUsage extension present. Don't accept CRLs with unhandled critical
4013 extensions: since verify currently doesn't process CRL extensions this
4014 rejects a CRL with *any* critical extensions. Add new verify error codes
4018 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
4019 A clarification of RFC2560 will require the use of OCTET STRINGs and
4020 some implementations cannot handle the current raw format. Since OpenSSL
4021 copies and compares OCSP nonces as opaque blobs without any attempt at
4022 parsing them this should not create any compatibility issues.
4025 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
4026 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
4027 this HMAC (and other) operations are several times slower than OpenSSL
4031 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
4032 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4034 *) Use the correct content when signing type "other".
4037 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
4039 *) Fix various bugs revealed by running the NISCC test suite:
4041 Stop out of bounds reads in the ASN1 code when presented with
4042 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4044 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
4046 If verify callback ignores invalid public key errors don't try to check
4047 certificate signature with the NULL public key.
4051 *) New -ignore_err option in ocsp application to stop the server
4052 exiting on the first error in a request.
4055 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4056 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4060 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4061 extra data after the compression methods not only for TLS 1.0
4062 but also for SSL 3.0 (as required by the specification).
4063 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4065 *) Change X509_certificate_type() to mark the key as exported/exportable
4066 when it's 512 *bits* long, not 512 bytes.
4069 *) Change AES_cbc_encrypt() so it outputs exact multiple of
4070 blocks during encryption.
4073 *) Various fixes to base64 BIO and non blocking I/O. On write
4074 flushes were not handled properly if the BIO retried. On read
4075 data was not being buffered properly and had various logic bugs.
4076 This also affects blocking I/O when the data being decoded is a
4080 *) Various S/MIME bugfixes and compatibility changes:
4081 output correct application/pkcs7 MIME type if
4082 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
4083 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
4084 of files as .eml work). Correctly handle very long lines in MIME
4088 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
4090 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4091 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4092 a protocol version number mismatch like a decryption error
4093 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4096 *) Turn on RSA blinding by default in the default implementation
4097 to avoid a timing attack. Applications that don't want it can call
4098 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4099 They would be ill-advised to do so in most cases.
4100 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4102 *) Change RSA blinding code so that it works when the PRNG is not
4103 seeded (in this case, the secret RSA exponent is abused as
4104 an unpredictable seed -- if it is not unpredictable, there
4105 is no point in blinding anyway). Make RSA blinding thread-safe
4106 by remembering the creator's thread ID in rsa->blinding and
4107 having all other threads use local one-time blinding factors
4108 (this requires more computation than sharing rsa->blinding, but
4109 avoids excessive locking; and if an RSA object is not shared
4110 between threads, blinding will still be very fast).
4113 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
4114 ENGINE as defaults for all supported algorithms irrespective of
4115 the 'flags' parameter. 'flags' is now honoured, so applications
4116 should make sure they are passing it correctly.
4119 *) Target "mingw" now allows native Windows code to be generated in
4120 the Cygwin environment as well as with the MinGW compiler.
4123 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
4125 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4126 via timing by performing a MAC computation even if incorrrect
4127 block cipher padding has been found. This is a countermeasure
4128 against active attacks where the attacker has to distinguish
4129 between bad padding and a MAC verification error. (CVE-2003-0078)
4131 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4132 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4133 Martin Vuagnoux (EPFL, Ilion)]
4135 *) Make the no-err option work as intended. The intention with no-err
4136 is not to have the whole error stack handling routines removed from
4137 libcrypto, it's only intended to remove all the function name and
4138 reason texts, thereby removing some of the footprint that may not
4139 be interesting if those errors aren't displayed anyway.
4141 NOTE: it's still possible for any application or module to have it's
4142 own set of error texts inserted. The routines are there, just not
4143 used by default when no-err is given.
4146 *) Add support for FreeBSD on IA64.
4147 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
4149 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
4150 Kerberos function mit_des_cbc_cksum(). Before this change,
4151 the value returned by DES_cbc_cksum() was like the one from
4152 mit_des_cbc_cksum(), except the bytes were swapped.
4153 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
4155 *) Allow an application to disable the automatic SSL chain building.
4156 Before this a rather primitive chain build was always performed in
4157 ssl3_output_cert_chain(): an application had no way to send the
4158 correct chain if the automatic operation produced an incorrect result.
4160 Now the chain builder is disabled if either:
4162 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
4164 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
4166 The reasoning behind this is that an application would not want the
4167 auto chain building to take place if extra chain certificates are
4168 present and it might also want a means of sending no additional
4169 certificates (for example the chain has two certificates and the
4173 *) Add the possibility to build without the ENGINE framework.
4174 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4176 *) Under Win32 gmtime() can return NULL: check return value in
4177 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
4180 *) DSA routines: under certain error conditions uninitialized BN objects
4181 could be freed. Solution: make sure initialization is performed early
4182 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
4183 Nils Larsch <nla@trustcenter.de> via PR#459)
4186 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
4187 checked on reconnect on the client side, therefore session resumption
4188 could still fail with a "ssl session id is different" error. This
4189 behaviour is masked when SSL_OP_ALL is used due to
4190 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
4191 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4192 followup to PR #377.
4195 *) IA-32 assembler support enhancements: unified ELF targets, support
4196 for SCO/Caldera platforms, fix for Cygwin shared build.
4199 *) Add support for FreeBSD on sparc64. As a consequence, support for
4200 FreeBSD on non-x86 processors is separate from x86 processors on
4201 the config script, much like the NetBSD support.
4202 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
4204 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
4206 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
4209 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
4210 code (06) was taken as the first octet of the session ID and the last
4211 octet was ignored consequently. As a result SSLv2 client side session
4212 caching could not have worked due to the session ID mismatch between
4214 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4218 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
4219 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
4223 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
4224 seems that in spite of existing for more than a year, many application
4225 author have done nothing to provide the necessary callbacks, which
4226 means that this particular engine will not work properly anywhere.
4227 This is a very unfortunate situation which forces us, in the name
4228 of usability, to give the hw_ncipher.c a static lock, which is part
4230 NOTE: This is for the 0.9.7 series ONLY. This hack will never
4231 appear in 0.9.8 or later. We EXPECT application authors to have
4232 dealt properly with this when 0.9.8 is released (unless we actually
4233 make such changes in the libcrypto locking code that changes will
4234 have to be made anyway).
4237 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
4238 octets have been read, EOF or an error occurs. Without this change
4239 some truncated ASN1 structures will not produce an error.
4242 *) Disable Heimdal support, since it hasn't been fully implemented.
4243 Still give the possibility to force the use of Heimdal, but with
4244 warnings and a request that patches get sent to openssl-dev.
4247 *) Add the VC-CE target, introduce the WINCE sysname, and add
4248 INSTALL.WCE and appropriate conditionals to make it build.
4249 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4251 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
4252 cygssl-x.y.z.dll, where x, y and z are the major, minor and
4253 edit numbers of the version.
4254 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4256 *) Introduce safe string copy and catenation functions
4257 (BUF_strlcpy() and BUF_strlcat()).
4258 [Ben Laurie (CHATS) and Richard Levitte]
4260 *) Avoid using fixed-size buffers for one-line DNs.
4261 [Ben Laurie (CHATS)]
4263 *) Add BUF_MEM_grow_clean() to avoid information leakage when
4264 resizing buffers containing secrets, and use where appropriate.
4265 [Ben Laurie (CHATS)]
4267 *) Avoid using fixed size buffers for configuration file location.
4268 [Ben Laurie (CHATS)]
4270 *) Avoid filename truncation for various CA files.
4271 [Ben Laurie (CHATS)]
4273 *) Use sizeof in preference to magic numbers.
4274 [Ben Laurie (CHATS)]
4276 *) Avoid filename truncation in cert requests.
4277 [Ben Laurie (CHATS)]
4279 *) Add assertions to check for (supposedly impossible) buffer
4281 [Ben Laurie (CHATS)]
4283 *) Don't cache truncated DNS entries in the local cache (this could
4284 potentially lead to a spoofing attack).
4285 [Ben Laurie (CHATS)]
4287 *) Fix various buffers to be large enough for hex/decimal
4288 representations in a platform independent manner.
4289 [Ben Laurie (CHATS)]
4291 *) Add CRYPTO_realloc_clean() to avoid information leakage when
4292 resizing buffers containing secrets, and use where appropriate.
4293 [Ben Laurie (CHATS)]
4295 *) Add BIO_indent() to avoid much slightly worrying code to do
4297 [Ben Laurie (CHATS)]
4299 *) Convert sprintf()/BIO_puts() to BIO_printf().
4300 [Ben Laurie (CHATS)]
4302 *) buffer_gets() could terminate with the buffer only half
4304 [Ben Laurie (CHATS)]
4306 *) Add assertions to prevent user-supplied crypto functions from
4307 overflowing internal buffers by having large block sizes, etc.
4308 [Ben Laurie (CHATS)]
4310 *) New OPENSSL_assert() macro (similar to assert(), but enabled
4312 [Ben Laurie (CHATS)]
4314 *) Eliminate unused copy of key in RC4.
4315 [Ben Laurie (CHATS)]
4317 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
4318 [Ben Laurie (CHATS)]
4320 *) Fix off-by-one error in EGD path.
4321 [Ben Laurie (CHATS)]
4323 *) If RANDFILE path is too long, ignore instead of truncating.
4324 [Ben Laurie (CHATS)]
4326 *) Eliminate unused and incorrectly sized X.509 structure
4328 [Ben Laurie (CHATS)]
4330 *) Eliminate unused and dangerous function knumber().
4331 [Ben Laurie (CHATS)]
4333 *) Eliminate unused and dangerous structure, KSSL_ERR.
4334 [Ben Laurie (CHATS)]
4336 *) Protect against overlong session ID context length in an encoded
4337 session object. Since these are local, this does not appear to be
4339 [Ben Laurie (CHATS)]
4341 *) Change from security patch (see 0.9.6e below) that did not affect
4342 the 0.9.6 release series:
4344 Remote buffer overflow in SSL3 protocol - an attacker could
4345 supply an oversized master key in Kerberos-enabled versions.
4347 [Ben Laurie (CHATS)]
4349 *) Change the SSL kerb5 codes to match RFC 2712.
4352 *) Make -nameopt work fully for req and add -reqopt switch.
4353 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
4355 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
4356 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
4358 *) Make sure tests can be performed even if the corresponding algorithms
4359 have been removed entirely. This was also the last step to make
4360 OpenSSL compilable with DJGPP under all reasonable conditions.
4361 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
4363 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
4364 to allow version independent disabling of normally unselected ciphers,
4365 which may be activated as a side-effect of selecting a single cipher.
4367 (E.g., cipher list string "RSA" enables ciphersuites that are left
4368 out of "ALL" because they do not provide symmetric encryption.
4369 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
4370 [Lutz Jaenicke, Bodo Moeller]
4372 *) Add appropriate support for separate platform-dependent build
4373 directories. The recommended way to make a platform-dependent
4374 build directory is the following (tested on Linux), maybe with
4377 # Place yourself outside of the OpenSSL source tree. In
4378 # this example, the environment variable OPENSSL_SOURCE
4379 # is assumed to contain the absolute OpenSSL source directory.
4380 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
4381 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
4382 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
4383 mkdir -p `dirname $F`
4384 ln -s $OPENSSL_SOURCE/$F $F
4387 To be absolutely sure not to disturb the source tree, a "make clean"
4388 is a good thing. If it isn't successfull, don't worry about it,
4389 it probably means the source directory is very clean.
4392 *) Make sure any ENGINE control commands make local copies of string
4393 pointers passed to them whenever necessary. Otherwise it is possible
4394 the caller may have overwritten (or deallocated) the original string
4395 data when a later ENGINE operation tries to use the stored values.
4396 [Götz Babin-Ebell <babinebell@trustcenter.de>]
4398 *) Improve diagnostics in file reading and command-line digests.
4399 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
4401 *) Add AES modes CFB and OFB to the object database. Correct an
4402 error in AES-CFB decryption.
4405 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
4406 allows existing EVP_CIPHER_CTX structures to be reused after
4407 calling EVP_*Final(). This behaviour is used by encryption
4408 BIOs and some applications. This has the side effect that
4409 applications must explicitly clean up cipher contexts with
4410 EVP_CIPHER_CTX_cleanup() or they will leak memory.
4413 *) Check the values of dna and dnb in bn_mul_recursive before calling
4414 bn_mul_comba (a non zero value means the a or b arrays do not contain
4415 n2 elements) and fallback to bn_mul_normal if either is not zero.
4418 *) Fix escaping of non-ASCII characters when using the -subj option
4419 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
4422 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
4423 form for "surname", serialNumber has no short form.
4424 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
4425 therefore remove "mail" short name for "internet 7".
4426 The OID for unique identifiers in X509 certificates is
4427 x500UniqueIdentifier, not uniqueIdentifier.
4428 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
4431 *) Add an "init" command to the ENGINE config module and auto initialize
4432 ENGINEs. Without any "init" command the ENGINE will be initialized
4433 after all ctrl commands have been executed on it. If init=1 the
4434 ENGINE is initailized at that point (ctrls before that point are run
4435 on the uninitialized ENGINE and after on the initialized one). If
4436 init=0 then the ENGINE will not be iniatialized at all.
4439 *) Fix the 'app_verify_callback' interface so that the user-defined
4440 argument is actually passed to the callback: In the
4441 SSL_CTX_set_cert_verify_callback() prototype, the callback
4442 declaration has been changed from
4445 int (*cb)(X509_STORE_CTX *,void *);
4446 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
4447 i=s->ctx->app_verify_callback(&ctx)
4448 has been changed into
4449 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
4451 To update applications using SSL_CTX_set_cert_verify_callback(),
4452 a dummy argument can be added to their callback functions.
4453 [D. K. Smetters <smetters@parc.xerox.com>]
4455 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
4456 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
4458 *) Add and OPENSSL_LOAD_CONF define which will cause
4459 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
4460 This allows older applications to transparently support certain
4461 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
4462 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
4463 load the config file and OPENSSL_add_all_algorithms_conf() which will
4464 always load it have also been added.
4467 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
4468 Adjust NIDs and EVP layer.
4469 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4471 *) Config modules support in openssl utility.
4473 Most commands now load modules from the config file,
4474 though in a few (such as version) this isn't done
4475 because it couldn't be used for anything.
4477 In the case of ca and req the config file used is
4478 the same as the utility itself: that is the -config
4479 command line option can be used to specify an
4483 *) Move default behaviour from OPENSSL_config(). If appname is NULL
4484 use "openssl_conf" if filename is NULL use default openssl config file.
4487 *) Add an argument to OPENSSL_config() to allow the use of an alternative
4488 config section name. Add a new flag to tolerate a missing config file
4489 and move code to CONF_modules_load_file().
4492 *) Support for crypto accelerator cards from Accelerated Encryption
4493 Processing, www.aep.ie. (Use engine 'aep')
4494 The support was copied from 0.9.6c [engine] and adapted/corrected
4495 to work with the new engine framework.
4496 [AEP Inc. and Richard Levitte]
4498 *) Support for SureWare crypto accelerator cards from Baltimore
4499 Technologies. (Use engine 'sureware')
4500 The support was copied from 0.9.6c [engine] and adapted
4501 to work with the new engine framework.
4504 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
4505 make the newer ENGINE framework commands for the CHIL engine work.
4506 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
4508 *) Make it possible to produce shared libraries on ReliantUNIX.
4509 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
4511 *) Add the configuration target debug-linux-ppro.
4512 Make 'openssl rsa' use the general key loading routines
4513 implemented in apps.c, and make those routines able to
4514 handle the key format FORMAT_NETSCAPE and the variant
4516 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4518 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4519 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4521 *) Add -keyform to rsautl, and document -engine.
4522 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
4524 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
4525 BIO_R_NO_SUCH_FILE error code rather than the generic
4526 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
4529 *) Add new functions
4531 ERR_peek_last_error_line
4532 ERR_peek_last_error_line_data.
4533 These are similar to
4536 ERR_peek_error_line_data,
4537 but report on the latest error recorded rather than the first one
4538 still in the error queue.
4539 [Ben Laurie, Bodo Moeller]
4541 *) default_algorithms option in ENGINE config module. This allows things
4543 default_algorithms = ALL
4544 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
4547 *) Prelminary ENGINE config module.
4550 *) New experimental application configuration code.
4553 *) Change the AES code to follow the same name structure as all other
4554 symmetric ciphers, and behave the same way. Move everything to
4555 the directory crypto/aes, thereby obsoleting crypto/rijndael.
4556 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4558 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
4559 [Ben Laurie and Theo de Raadt]
4561 *) Add option to output public keys in req command.
4562 [Massimiliano Pala madwolf@openca.org]
4564 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
4565 (up to about 10% better than before for P-192 and P-224).
4568 *) New functions/macros
4570 SSL_CTX_set_msg_callback(ctx, cb)
4571 SSL_CTX_set_msg_callback_arg(ctx, arg)
4572 SSL_set_msg_callback(ssl, cb)
4573 SSL_set_msg_callback_arg(ssl, arg)
4575 to request calling a callback function
4577 void cb(int write_p, int version, int content_type,
4578 const void *buf, size_t len, SSL *ssl, void *arg)
4580 whenever a protocol message has been completely received
4581 (write_p == 0) or sent (write_p == 1). Here 'version' is the
4582 protocol version according to which the SSL library interprets
4583 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
4584 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
4585 the content type as defined in the SSL 3.0/TLS 1.0 protocol
4586 specification (change_cipher_spec(20), alert(21), handshake(22)).
4587 'buf' and 'len' point to the actual message, 'ssl' to the
4588 SSL object, and 'arg' is the application-defined value set by
4589 SSL[_CTX]_set_msg_callback_arg().
4591 'openssl s_client' and 'openssl s_server' have new '-msg' options
4592 to enable a callback that displays all protocol messages.
4595 *) Change the shared library support so shared libraries are built as
4596 soon as the corresponding static library is finished, and thereby get
4597 openssl and the test programs linked against the shared library.
4598 This still only happens when the keyword "shard" has been given to
4599 the configuration scripts.
4601 NOTE: shared library support is still an experimental thing, and
4602 backward binary compatibility is still not guaranteed.
4603 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
4605 *) Add support for Subject Information Access extension.
4606 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4608 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
4609 additional bytes when new memory had to be allocated, not just
4610 when reusing an existing buffer.
4613 *) New command line and configuration option 'utf8' for the req command.
4614 This allows field values to be specified as UTF8 strings.
4617 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
4618 runs for the former and machine-readable output for the latter.
4621 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
4622 of the e-mail address in the DN (i.e., it will go into a certificate
4623 extension only). The new configuration file option 'email_in_dn = no'
4624 has the same effect.
4625 [Massimiliano Pala madwolf@openca.org]
4627 *) Change all functions with names starting with des_ to be starting
4628 with DES_ instead. Add wrappers that are compatible with libdes,
4629 but are named _ossl_old_des_*. Finally, add macros that map the
4630 des_* symbols to the corresponding _ossl_old_des_* if libdes
4631 compatibility is desired. If OpenSSL 0.9.6c compatibility is
4632 desired, the des_* symbols will be mapped to DES_*, with one
4635 Since we provide two compatibility mappings, the user needs to
4636 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
4637 compatibility is desired. The default (i.e., when that macro
4638 isn't defined) is OpenSSL 0.9.6c compatibility.
4640 There are also macros that enable and disable the support of old
4641 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
4642 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
4643 are defined, the default will apply: to support the old des routines.
4645 In either case, one must include openssl/des.h to get the correct
4646 definitions. Do not try to just include openssl/des_old.h, that
4649 NOTE: This is a major break of an old API into a new one. Software
4650 authors are encouraged to switch to the DES_ style functions. Some
4651 time in the future, des_old.h and the libdes compatibility functions
4652 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
4653 default), and then completely removed.
4656 *) Test for certificates which contain unsupported critical extensions.
4657 If such a certificate is found during a verify operation it is
4658 rejected by default: this behaviour can be overridden by either
4659 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
4660 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
4661 X509_supported_extension() has also been added which returns 1 if a
4662 particular extension is supported.
4665 *) Modify the behaviour of EVP cipher functions in similar way to digests
4666 to retain compatibility with existing code.
4669 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
4670 compatibility with existing code. In particular the 'ctx' parameter does
4671 not have to be to be initialized before the call to EVP_DigestInit() and
4672 it is tidied up after a call to EVP_DigestFinal(). New function
4673 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
4674 EVP_MD_CTX_copy() changed to not require the destination to be
4675 initialized valid and new function EVP_MD_CTX_copy_ex() added which
4676 requires the destination to be valid.
4678 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
4679 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
4682 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
4683 so that complete 'Handshake' protocol structures are kept in memory
4684 instead of overwriting 'msg_type' and 'length' with 'body' data.
4687 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
4688 [Massimo Santin via Richard Levitte]
4690 *) Major restructuring to the underlying ENGINE code. This includes
4691 reduction of linker bloat, separation of pure "ENGINE" manipulation
4692 (initialisation, etc) from functionality dealing with implementations
4693 of specific crypto iterfaces. This change also introduces integrated
4694 support for symmetric ciphers and digest implementations - so ENGINEs
4695 can now accelerate these by providing EVP_CIPHER and EVP_MD
4696 implementations of their own. This is detailed in crypto/engine/README
4697 as it couldn't be adequately described here. However, there are a few
4698 API changes worth noting - some RSA, DSA, DH, and RAND functions that
4699 were changed in the original introduction of ENGINE code have now
4700 reverted back - the hooking from this code to ENGINE is now a good
4701 deal more passive and at run-time, operations deal directly with
4702 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
4703 dereferencing through an ENGINE pointer any more. Also, the ENGINE
4704 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
4705 they were not being used by the framework as there is no concept of a
4706 BIGNUM_METHOD and they could not be generalised to the new
4707 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
4708 ENGINE_cpy() has been removed as it cannot be consistently defined in
4712 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
4715 *) Change mkdef.pl to sort symbols that get the same entry number,
4716 and make sure the automatically generated functions ERR_load_*
4717 become part of libeay.num as well.
4720 *) New function SSL_renegotiate_pending(). This returns true once
4721 renegotiation has been requested (either SSL_renegotiate() call
4722 or HelloRequest/ClientHello receveived from the peer) and becomes
4723 false once a handshake has been completed.
4724 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
4725 sends a HelloRequest, but does not ensure that a handshake takes
4726 place. SSL_renegotiate_pending() is useful for checking if the
4727 client has followed the request.)
4730 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
4731 By default, clients may request session resumption even during
4732 renegotiation (if session ID contexts permit); with this option,
4733 session resumption is possible only in the first handshake.
4735 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4736 more bits available for options that should not be part of
4737 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4740 *) Add some demos for certificate and certificate request creation.
4743 *) Make maximum certificate chain size accepted from the peer application
4744 settable (SSL*_get/set_max_cert_list()), as proposed by
4745 "Douglas E. Engert" <deengert@anl.gov>.
4748 *) Add support for shared libraries for Unixware-7
4749 (Boyd Lynn Gerber <gerberb@zenez.com>).
4752 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4753 be done prior to destruction. Use this to unload error strings from
4754 ENGINEs that load their own error strings. NB: This adds two new API
4755 functions to "get" and "set" this destroy handler in an ENGINE.
4758 *) Alter all existing ENGINE implementations (except "openssl" and
4759 "openbsd") to dynamically instantiate their own error strings. This
4760 makes them more flexible to be built both as statically-linked ENGINEs
4761 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4762 Also, add stub code to each that makes building them as self-contained
4763 shared-libraries easier (see README.ENGINE).
4766 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4767 implementations into applications that are completely implemented in
4768 self-contained shared-libraries. The "dynamic" ENGINE exposes control
4769 commands that can be used to configure what shared-library to load and
4770 to control aspects of the way it is handled. Also, made an update to
4771 the README.ENGINE file that brings its information up-to-date and
4772 provides some information and instructions on the "dynamic" ENGINE
4773 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
4776 *) Make it possible to unload ranges of ERR strings with a new
4777 "ERR_unload_strings" function.
4780 *) Add a copy() function to EVP_MD.
4783 *) Make EVP_MD routines take a context pointer instead of just the
4784 md_data void pointer.
4787 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
4788 that the digest can only process a single chunk of data
4789 (typically because it is provided by a piece of
4790 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
4791 is only going to provide a single chunk of data, and hence the
4792 framework needn't accumulate the data for oneshot drivers.
4795 *) As with "ERR", make it possible to replace the underlying "ex_data"
4796 functions. This change also alters the storage and management of global
4797 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
4798 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
4799 index counters. The API functions that use this state have been changed
4800 to take a "class_index" rather than pointers to the class's local STACK
4801 and counter, and there is now an API function to dynamically create new
4802 classes. This centralisation allows us to (a) plug a lot of the
4803 thread-safety problems that existed, and (b) makes it possible to clean
4804 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
4805 such data would previously have always leaked in application code and
4806 workarounds were in place to make the memory debugging turn a blind eye
4807 to it. Application code that doesn't use this new function will still
4808 leak as before, but their memory debugging output will announce it now
4809 rather than letting it slide.
4811 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
4812 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
4813 has a return value to indicate success or failure.
4816 *) Make it possible to replace the underlying "ERR" functions such that the
4817 global state (2 LHASH tables and 2 locks) is only used by the "default"
4818 implementation. This change also adds two functions to "get" and "set"
4819 the implementation prior to it being automatically set the first time
4820 any other ERR function takes place. Ie. an application can call "get",
4821 pass the return value to a module it has just loaded, and that module
4822 can call its own "set" function using that value. This means the
4823 module's "ERR" operations will use (and modify) the error state in the
4824 application and not in its own statically linked copy of OpenSSL code.
4827 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
4828 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
4829 the operation, and provides a more encapsulated way for external code
4830 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
4831 to use these functions rather than manually incrementing the counts.
4833 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
4836 *) Add EVP test program.
4839 *) Add symmetric cipher support to ENGINE. Expect the API to change!
4842 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
4843 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
4844 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
4845 These allow a CRL to be built without having to access X509_CRL fields
4846 directly. Modify 'ca' application to use new functions.
4849 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
4850 bug workarounds. Rollback attack detection is a security feature.
4851 The problem will only arise on OpenSSL servers when TLSv1 is not
4852 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
4853 Software authors not wanting to support TLSv1 will have special reasons
4854 for their choice and can explicitly enable this option.
4855 [Bodo Moeller, Lutz Jaenicke]
4857 *) Rationalise EVP so it can be extended: don't include a union of
4858 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
4859 (similar to those existing for EVP_CIPHER_CTX).
4864 EVP_MD_CTX_init(&md); /* new function call */
4865 EVP_DigestInit(&md, EVP_sha1());
4866 EVP_DigestUpdate(&md, in, len);
4867 EVP_DigestFinal(&md, out, NULL);
4868 EVP_MD_CTX_cleanup(&md); /* new function call */
4872 *) Make DES key schedule conform to the usual scheme, as well as
4873 correcting its structure. This means that calls to DES functions
4874 now have to pass a pointer to a des_key_schedule instead of a
4875 plain des_key_schedule (which was actually always a pointer
4878 des_key_schedule ks;
4880 des_set_key_checked(..., &ks);
4881 des_ncbc_encrypt(..., &ks, ...);
4883 (Note that a later change renames 'des_...' into 'DES_...'.)
4886 *) Initial reduction of linker bloat: the use of some functions, such as
4887 PEM causes large amounts of unused functions to be linked in due to
4888 poor organisation. For example pem_all.c contains every PEM function
4889 which has a knock on effect of linking in large amounts of (unused)
4890 ASN1 code. Grouping together similar functions and splitting unrelated
4891 functions prevents this.
4894 *) Cleanup of EVP macros.
4897 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
4898 correct _ecb suffix.
4901 *) Add initial OCSP responder support to ocsp application. The
4902 revocation information is handled using the text based index
4903 use by the ca application. The responder can either handle
4904 requests generated internally, supplied in files (for example
4905 via a CGI script) or using an internal minimal server.
4908 *) Add configuration choices to get zlib compression for TLS.
4911 *) Changes to Kerberos SSL for RFC 2712 compliance:
4912 1. Implemented real KerberosWrapper, instead of just using
4913 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
4914 2. Implemented optional authenticator field of KerberosWrapper.
4916 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
4917 and authenticator structs; see crypto/krb5/.
4919 Generalized Kerberos calls to support multiple Kerberos libraries.
4920 [Vern Staats <staatsvr@asc.hpc.mil>,
4921 Jeffrey Altman <jaltman@columbia.edu>
4922 via Richard Levitte]
4924 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
4925 already does with RSA. testdsa.h now has 'priv_key/pub_key'
4926 values for each of the key sizes rather than having just
4927 parameters (and 'speed' generating keys each time).
4930 *) Speed up EVP routines.
4933 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
4934 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
4935 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
4936 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
4938 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
4939 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
4940 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
4943 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
4945 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
4948 *) Added the OS2-EMX target.
4949 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
4951 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
4952 to support NCONF routines in extension code. New function CONF_set_nconf()
4953 to allow functions which take an NCONF to also handle the old LHASH
4954 structure: this means that the old CONF compatible routines can be
4955 retained (in particular wrt extensions) without having to duplicate the
4956 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
4959 *) Enhance the general user interface with mechanisms for inner control
4960 and with possibilities to have yes/no kind of prompts.
4963 *) Change all calls to low level digest routines in the library and
4964 applications to use EVP. Add missing calls to HMAC_cleanup() and
4965 don't assume HMAC_CTX can be copied using memcpy().
4966 [Verdon Walker <VWalker@novell.com>, Steve Henson]
4968 *) Add the possibility to control engines through control names but with
4969 arbitrary arguments instead of just a string.
4970 Change the key loaders to take a UI_METHOD instead of a callback
4971 function pointer. NOTE: this breaks binary compatibility with earlier
4972 versions of OpenSSL [engine].
4973 Adapt the nCipher code for these new conditions and add a card insertion
4977 *) Enhance the general user interface with mechanisms to better support
4978 dialog box interfaces, application-defined prompts, the possibility
4979 to use defaults (for example default passwords from somewhere else)
4980 and interrupts/cancellations.
4983 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
4984 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
4987 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
4988 tidy up some unnecessarily weird code in 'sk_new()').
4989 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
4991 *) Change the key loading routines for ENGINEs to use the same kind
4992 callback (pem_password_cb) as all other routines that need this
4996 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
4997 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
4998 than this minimum value is recommended.
5001 *) New random seeder for OpenVMS, using the system process statistics
5002 that are easily reachable.
5005 *) Windows apparently can't transparently handle global
5006 variables defined in DLLs. Initialisations such as:
5008 const ASN1_ITEM *it = &ASN1_INTEGER_it;
5010 wont compile. This is used by the any applications that need to
5011 declare their own ASN1 modules. This was fixed by adding the option
5012 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
5013 needed for static libraries under Win32.
5016 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
5017 setting of purpose and trust fields. New X509_STORE trust and
5018 purpose functions and tidy up setting in other SSL functions.
5021 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
5022 structure. These are inherited by X509_STORE_CTX when it is
5023 initialised. This allows various defaults to be set in the
5024 X509_STORE structure (such as flags for CRL checking and custom
5025 purpose or trust settings) for functions which only use X509_STORE_CTX
5026 internally such as S/MIME.
5028 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
5029 trust settings if they are not set in X509_STORE. This allows X509_STORE
5030 purposes and trust (in S/MIME for example) to override any set by default.
5032 Add command line options for CRL checking to smime, s_client and s_server
5036 *) Initial CRL based revocation checking. If the CRL checking flag(s)
5037 are set then the CRL is looked up in the X509_STORE structure and
5038 its validity and signature checked, then if the certificate is found
5039 in the CRL the verify fails with a revoked error.
5041 Various new CRL related callbacks added to X509_STORE_CTX structure.
5043 Command line options added to 'verify' application to support this.
5045 This needs some additional work, such as being able to handle multiple
5046 CRLs with different times, extension based lookup (rather than just
5047 by subject name) and ultimately more complete V2 CRL extension
5051 *) Add a general user interface API (crypto/ui/). This is designed
5052 to replace things like des_read_password and friends (backward
5053 compatibility functions using this new API are provided).
5054 The purpose is to remove prompting functions from the DES code
5055 section as well as provide for prompting through dialog boxes in
5056 a window system and the like.
5059 *) Add "ex_data" support to ENGINE so implementations can add state at a
5060 per-structure level rather than having to store it globally.
5063 *) Make it possible for ENGINE structures to be copied when retrieved by
5064 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
5065 This causes the "original" ENGINE structure to act like a template,
5066 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
5067 operational state can be localised to each ENGINE structure, despite the
5068 fact they all share the same "methods". New ENGINE structures returned in
5069 this case have no functional references and the return value is the single
5070 structural reference. This matches the single structural reference returned
5071 by ENGINE_by_id() normally, when it is incremented on the pre-existing
5075 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
5076 needs to match any other type at all we need to manually clear the
5080 *) Changes to the "openssl engine" utility to include;
5081 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
5082 about an ENGINE's available control commands.
5083 - executing control commands from command line arguments using the
5084 '-pre' and '-post' switches. '-post' is only used if '-t' is
5085 specified and the ENGINE is successfully initialised. The syntax for
5086 the individual commands are colon-separated, for example;
5087 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
5090 *) New dynamic control command support for ENGINEs. ENGINEs can now
5091 declare their own commands (numbers), names (strings), descriptions,
5092 and input types for run-time discovery by calling applications. A
5093 subset of these commands are implicitly classed as "executable"
5094 depending on their input type, and only these can be invoked through
5095 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
5096 can be based on user input, config files, etc). The distinction is
5097 that "executable" commands cannot return anything other than a boolean
5098 result and can only support numeric or string input, whereas some
5099 discoverable commands may only be for direct use through
5100 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
5101 pointers, or other custom uses. The "executable" commands are to
5102 support parameterisations of ENGINE behaviour that can be
5103 unambiguously defined by ENGINEs and used consistently across any
5104 OpenSSL-based application. Commands have been added to all the
5105 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
5106 control over shared-library paths without source code alterations.
5109 *) Changed all ENGINE implementations to dynamically allocate their
5110 ENGINEs rather than declaring them statically. Apart from this being
5111 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
5112 this also allows the implementations to compile without using the
5113 internal engine_int.h header.
5116 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
5117 'const' value. Any code that should be able to modify a RAND_METHOD
5118 should already have non-const pointers to it (ie. they should only
5119 modify their own ones).
5122 *) Made a variety of little tweaks to the ENGINE code.
5123 - "atalla" and "ubsec" string definitions were moved from header files
5124 to C code. "nuron" string definitions were placed in variables
5125 rather than hard-coded - allowing parameterisation of these values
5126 later on via ctrl() commands.
5127 - Removed unused "#if 0"'d code.
5128 - Fixed engine list iteration code so it uses ENGINE_free() to release
5129 structural references.
5130 - Constified the RAND_METHOD element of ENGINE structures.
5131 - Constified various get/set functions as appropriate and added
5132 missing functions (including a catch-all ENGINE_cpy that duplicates
5133 all ENGINE values onto a new ENGINE except reference counts/state).
5134 - Removed NULL parameter checks in get/set functions. Setting a method
5135 or function to NULL is a way of cancelling out a previously set
5136 value. Passing a NULL ENGINE parameter is just plain stupid anyway
5137 and doesn't justify the extra error symbols and code.
5138 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
5139 flags from engine_int.h to engine.h.
5140 - Changed prototypes for ENGINE handler functions (init(), finish(),
5141 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
5144 *) Implement binary inversion algorithm for BN_mod_inverse in addition
5145 to the algorithm using long division. The binary algorithm can be
5146 used only if the modulus is odd. On 32-bit systems, it is faster
5147 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
5148 roughly 5-15% for 256-bit moduli), so we use it only for moduli
5149 up to 450 bits. In 64-bit environments, the binary algorithm
5150 appears to be advantageous for much longer moduli; here we use it
5151 for moduli up to 2048 bits.
5154 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
5155 could not support the combine flag in choice fields.
5158 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
5159 extensions from a certificate request to the certificate.
5162 *) Allow multiple 'certopt' and 'nameopt' options to be separated
5163 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
5164 file: this allows the display of the certificate about to be
5165 signed to be customised, to allow certain fields to be included
5166 or excluded and extension details. The old system didn't display
5167 multicharacter strings properly, omitted fields not in the policy
5168 and couldn't display additional details such as extensions.
5171 *) Function EC_POINTs_mul for multiple scalar multiplication
5172 of an arbitrary number of elliptic curve points
5173 \sum scalars[i]*points[i],
5174 optionally including the generator defined for the EC_GROUP:
5175 scalar*generator + \sum scalars[i]*points[i].
5177 EC_POINT_mul is a simple wrapper function for the typical case
5178 that the point list has just one item (besides the optional
5182 *) First EC_METHODs for curves over GF(p):
5184 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
5185 operations and provides various method functions that can also
5186 operate with faster implementations of modular arithmetic.
5188 EC_GFp_mont_method() reuses most functions that are part of
5189 EC_GFp_simple_method, but uses Montgomery arithmetic.
5191 [Bodo Moeller; point addition and point doubling
5192 implementation directly derived from source code provided by
5193 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
5195 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
5196 crypto/ec/ec_lib.c):
5198 Curves are EC_GROUP objects (with an optional group generator)
5199 based on EC_METHODs that are built into the library.
5201 Points are EC_POINT objects based on EC_GROUP objects.
5203 Most of the framework would be able to handle curves over arbitrary
5204 finite fields, but as there are no obvious types for fields other
5205 than GF(p), some functions are limited to that for now.
5208 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
5209 that the file contains a complete HTTP response.
5212 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
5213 change the def and num file printf format specifier from "%-40sXXX"
5214 to "%-39s XXX". The latter will always guarantee a space after the
5215 field while the former will cause them to run together if the field
5216 is 40 of more characters long.
5219 *) Constify the cipher and digest 'method' functions and structures
5220 and modify related functions to take constant EVP_MD and EVP_CIPHER
5224 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
5225 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
5228 *) Modify EVP_Digest*() routines so they now return values. Although the
5229 internal software routines can never fail additional hardware versions
5233 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
5235 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
5236 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
5239 ERR_R_NESTED_ASN1_ERROR
5241 ERR_R_MISSING_ASN1_EOS
5242 were 4 .. 9, conflicting with
5243 ERR_LIB_RSA (= ERR_R_RSA_LIB)
5245 ERR_LIB_PEM (= ERR_R_PEM_LIB).
5246 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
5248 Add new error code 'ERR_R_INTERNAL_ERROR'.
5251 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
5255 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
5256 sets the subject name for a new request or supersedes the
5257 subject name in a given request. Formats that can be parsed are
5258 'CN=Some Name, OU=myOU, C=IT'
5260 'CN=Some Name/OU=myOU/C=IT'.
5262 Add options '-batch' and '-verbose' to 'openssl req'.
5263 [Massimiliano Pala <madwolf@hackmasters.net>]
5265 *) Introduce the possibility to access global variables through
5266 functions on platform were that's the best way to handle exporting
5267 global variables in shared libraries. To enable this functionality,
5268 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
5269 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
5270 is normally done by Configure or something similar).
5272 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
5273 in the source file (foo.c) like this:
5275 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
5276 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
5278 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
5279 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
5281 OPENSSL_DECLARE_GLOBAL(int,foo);
5282 #define foo OPENSSL_GLOBAL_REF(foo)
5283 OPENSSL_DECLARE_GLOBAL(double,bar);
5284 #define bar OPENSSL_GLOBAL_REF(bar)
5286 The #defines are very important, and therefore so is including the
5287 header file everywhere where the defined globals are used.
5289 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
5290 of ASN.1 items, but that structure is a bit different.
5292 The largest change is in util/mkdef.pl which has been enhanced with
5293 better and easier to understand logic to choose which symbols should
5294 go into the Windows .def files as well as a number of fixes and code
5295 cleanup (among others, algorithm keywords are now sorted
5296 lexicographically to avoid constant rewrites).
5299 *) In BN_div() keep a copy of the sign of 'num' before writing the
5300 result to 'rm' because if rm==num the value will be overwritten
5301 and produce the wrong result if 'num' is negative: this caused
5302 problems with BN_mod() and BN_nnmod().
5305 *) Function OCSP_request_verify(). This checks the signature on an
5306 OCSP request and verifies the signer certificate. The signer
5307 certificate is just checked for a generic purpose and OCSP request
5311 *) Add OCSP_check_validity() function to check the validity of OCSP
5312 responses. OCSP responses are prepared in real time and may only
5313 be a few seconds old. Simply checking that the current time lies
5314 between thisUpdate and nextUpdate max reject otherwise valid responses
5315 caused by either OCSP responder or client clock inaccuracy. Instead
5316 we allow thisUpdate and nextUpdate to fall within a certain period of
5317 the current time. The age of the response can also optionally be
5318 checked. Two new options -validity_period and -status_age added to
5322 *) If signature or public key algorithm is unrecognized print out its
5323 OID rather that just UNKNOWN.
5326 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
5327 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
5328 ID to be generated from the issuer certificate alone which can then be
5329 passed to OCSP_id_issuer_cmp().
5332 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
5333 ASN1 modules to export functions returning ASN1_ITEM pointers
5334 instead of the ASN1_ITEM structures themselves. This adds several
5335 new macros which allow the underlying ASN1 function/structure to
5336 be accessed transparently. As a result code should not use ASN1_ITEM
5337 references directly (such as &X509_it) but instead use the relevant
5338 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
5339 use of the new ASN1 code on platforms where exporting structures
5340 is problematical (for example in shared libraries) but exporting
5341 functions returning pointers to structures is not.
5344 *) Add support for overriding the generation of SSL/TLS session IDs.
5345 These callbacks can be registered either in an SSL_CTX or per SSL.
5346 The purpose of this is to allow applications to control, if they wish,
5347 the arbitrary values chosen for use as session IDs, particularly as it
5348 can be useful for session caching in multiple-server environments. A
5349 command-line switch for testing this (and any client code that wishes
5350 to use such a feature) has been added to "s_server".
5351 [Geoff Thorpe, Lutz Jaenicke]
5353 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
5354 of the form '#if defined(...) || defined(...) || ...' and
5355 '#if !defined(...) && !defined(...) && ...'. This also avoids
5356 the growing number of special cases it was previously handling.
5359 *) Make all configuration macros available for application by making
5360 sure they are available in opensslconf.h, by giving them names starting
5361 with "OPENSSL_" to avoid conflicts with other packages and by making
5362 sure e_os2.h will cover all platform-specific cases together with
5364 Additionally, it is now possible to define configuration/platform-
5365 specific names (called "system identities"). In the C code, these
5366 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
5367 macro with the name beginning with "OPENSSL_SYS_", which is determined
5368 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
5372 *) New option -set_serial to 'req' and 'x509' this allows the serial
5373 number to use to be specified on the command line. Previously self
5374 signed certificates were hard coded with serial number 0 and the
5375 CA options of 'x509' had to use a serial number in a file which was
5379 *) New options to 'ca' utility to support V2 CRL entry extensions.
5380 Currently CRL reason, invalidity date and hold instruction are
5381 supported. Add new CRL extensions to V3 code and some new objects.
5384 *) New function EVP_CIPHER_CTX_set_padding() this is used to
5385 disable standard block padding (aka PKCS#5 padding) in the EVP
5386 API, which was previously mandatory. This means that the data is
5387 not padded in any way and so the total length much be a multiple
5388 of the block size, otherwise an error occurs.
5391 *) Initial (incomplete) OCSP SSL support.
5394 *) New function OCSP_parse_url(). This splits up a URL into its host,
5395 port and path components: primarily to parse OCSP URLs. New -url
5396 option to ocsp utility.
5399 *) New nonce behavior. The return value of OCSP_check_nonce() now
5400 reflects the various checks performed. Applications can decide
5401 whether to tolerate certain situations such as an absent nonce
5402 in a response when one was present in a request: the ocsp application
5403 just prints out a warning. New function OCSP_add1_basic_nonce()
5404 this is to allow responders to include a nonce in a response even if
5405 the request is nonce-less.
5408 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
5409 skipped when using openssl x509 multiple times on a single input file,
5410 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
5413 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
5414 set string type: to handle setting ASN1_TIME structures. Fix ca
5415 utility to correctly initialize revocation date of CRLs.
5418 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
5419 the clients preferred ciphersuites and rather use its own preferences.
5420 Should help to work around M$ SGC (Server Gated Cryptography) bug in
5421 Internet Explorer by ensuring unchanged hash method during stepup.
5422 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
5425 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
5426 to aes and add a new 'exist' option to print out symbols that don't
5430 *) Additional options to ocsp utility to allow flags to be set and
5431 additional certificates supplied.
5434 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
5435 OCSP client a number of certificate to only verify the response
5439 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
5440 handle the new API. Currently only ECB, CBC modes supported. Add new
5443 Add TLS AES ciphersuites as described in RFC3268, "Advanced
5444 Encryption Standard (AES) Ciphersuites for Transport Layer
5445 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
5446 not enabled by default and were not part of the "ALL" ciphersuite
5447 alias because they were not yet official; they could be
5448 explicitly requested by specifying the "AESdraft" ciphersuite
5449 group alias. In the final release of OpenSSL 0.9.7, the group
5450 alias is called "AES" and is part of "ALL".)
5451 [Ben Laurie, Steve Henson, Bodo Moeller]
5453 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
5454 request to response.
5457 *) Functions for OCSP responders. OCSP_request_onereq_count(),
5458 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
5459 extract information from a certificate request. OCSP_response_create()
5460 creates a response and optionally adds a basic response structure.
5461 OCSP_basic_add1_status() adds a complete single response to a basic
5462 response and returns the OCSP_SINGLERESP structure just added (to allow
5463 extensions to be included for example). OCSP_basic_add1_cert() adds a
5464 certificate to a basic response and OCSP_basic_sign() signs a basic
5465 response with various flags. New helper functions ASN1_TIME_check()
5466 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
5467 (converts ASN1_TIME to GeneralizedTime).
5470 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
5471 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
5472 structure from a certificate. X509_pubkey_digest() digests the public_key
5473 contents: this is used in various key identifiers.
5476 *) Make sk_sort() tolerate a NULL argument.
5477 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
5479 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
5480 passed by the function are trusted implicitly. If any of them signed the
5481 response then it is assumed to be valid and is not verified.
5484 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
5485 to data. This was previously part of the PKCS7 ASN1 code. This
5486 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
5487 [Steve Henson, reported by Kenneth R. Robinette
5488 <support@securenetterm.com>]
5490 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
5491 routines: without these tracing memory leaks is very painful.
5492 Fix leaks in PKCS12 and PKCS7 routines.
5495 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
5496 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
5497 effectively meant GeneralizedTime would never be used. Now it
5498 is initialised to -1 but X509_time_adj() now has to check the value
5499 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
5500 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
5501 [Steve Henson, reported by Kenneth R. Robinette
5502 <support@securenetterm.com>]
5504 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
5505 result in a zero length in the ASN1_INTEGER structure which was
5506 not consistent with the structure when d2i_ASN1_INTEGER() was used
5507 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
5508 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
5509 where it did not print out a minus for negative ASN1_INTEGER.
5512 *) Add summary printout to ocsp utility. The various functions which
5513 convert status values to strings have been renamed to:
5514 OCSP_response_status_str(), OCSP_cert_status_str() and
5515 OCSP_crl_reason_str() and are no longer static. New options
5516 to verify nonce values and to disable verification. OCSP response
5517 printout format cleaned up.
5520 *) Add additional OCSP certificate checks. These are those specified
5521 in RFC2560. This consists of two separate checks: the CA of the
5522 certificate being checked must either be the OCSP signer certificate
5523 or the issuer of the OCSP signer certificate. In the latter case the
5524 OCSP signer certificate must contain the OCSP signing extended key
5525 usage. This check is performed by attempting to match the OCSP
5526 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
5527 in the OCSP_CERTID structures of the response.
5530 *) Initial OCSP certificate verification added to OCSP_basic_verify()
5531 and related routines. This uses the standard OpenSSL certificate
5532 verify routines to perform initial checks (just CA validity) and
5533 to obtain the certificate chain. Then additional checks will be
5534 performed on the chain. Currently the root CA is checked to see
5535 if it is explicitly trusted for OCSP signing. This is used to set
5536 a root CA as a global signing root: that is any certificate that
5537 chains to that CA is an acceptable OCSP signing certificate.
5540 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
5541 extensions from a separate configuration file.
5542 As when reading extensions from the main configuration file,
5543 the '-extensions ...' option may be used for specifying the
5545 [Massimiliano Pala <madwolf@comune.modena.it>]
5547 *) New OCSP utility. Allows OCSP requests to be generated or
5548 read. The request can be sent to a responder and the output
5549 parsed, outputed or printed in text form. Not complete yet:
5550 still needs to check the OCSP response validity.
5553 *) New subcommands for 'openssl ca':
5554 'openssl ca -status <serial>' prints the status of the cert with
5555 the given serial number (according to the index file).
5556 'openssl ca -updatedb' updates the expiry status of certificates
5558 [Massimiliano Pala <madwolf@comune.modena.it>]
5560 *) New '-newreq-nodes' command option to CA.pl. This is like
5561 '-newreq', but calls 'openssl req' with the '-nodes' option
5562 so that the resulting key is not encrypted.
5563 [Damien Miller <djm@mindrot.org>]
5565 *) New configuration for the GNU Hurd.
5566 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
5568 *) Initial code to implement OCSP basic response verify. This
5569 is currently incomplete. Currently just finds the signer's
5570 certificate and verifies the signature on the response.
5573 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
5574 value of OPENSSLDIR. This is available via the new '-d' option
5575 to 'openssl version', and is also included in 'openssl version -a'.
5578 *) Allowing defining memory allocation callbacks that will be given
5579 file name and line number information in additional arguments
5580 (a const char* and an int). The basic functionality remains, as
5581 well as the original possibility to just replace malloc(),
5582 realloc() and free() by functions that do not know about these
5583 additional arguments. To register and find out the current
5584 settings for extended allocation functions, the following
5585 functions are provided:
5587 CRYPTO_set_mem_ex_functions
5588 CRYPTO_set_locked_mem_ex_functions
5589 CRYPTO_get_mem_ex_functions
5590 CRYPTO_get_locked_mem_ex_functions
5592 These work the same way as CRYPTO_set_mem_functions and friends.
5593 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
5594 extended allocation function is enabled.
5595 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
5596 a conventional allocation function is enabled.
5597 [Richard Levitte, Bodo Moeller]
5599 *) Finish off removing the remaining LHASH function pointer casts.
5600 There should no longer be any prototype-casting required when using
5601 the LHASH abstraction, and any casts that remain are "bugs". See
5602 the callback types and macros at the head of lhash.h for details
5603 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
5606 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
5607 If /dev/[u]random devices are not available or do not return enough
5608 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
5610 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
5611 /etc/entropy will be queried once each in this sequence, quering stops
5612 when enough entropy was collected without querying more sockets.
5615 *) Change the Unix RAND_poll() variant to be able to poll several
5616 random devices, as specified by DEVRANDOM, until a sufficient amount
5617 of data has been collected. We spend at most 10 ms on each file
5618 (select timeout) and read in non-blocking mode. DEVRANDOM now
5619 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
5620 (previously it was just the string "/dev/urandom"), so on typical
5621 platforms the 10 ms delay will never occur.
5622 Also separate out the Unix variant to its own file, rand_unix.c.
5623 For VMS, there's a currently-empty rand_vms.c.
5626 *) Move OCSP client related routines to ocsp_cl.c. These
5627 provide utility functions which an application needing
5628 to issue a request to an OCSP responder and analyse the
5629 response will typically need: as opposed to those which an
5630 OCSP responder itself would need which will be added later.
5632 OCSP_request_sign() signs an OCSP request with an API similar
5633 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
5634 response. OCSP_response_get1_basic() extracts basic response
5635 from response. OCSP_resp_find_status(): finds and extracts status
5636 information from an OCSP_CERTID structure (which will be created
5637 when the request structure is built). These are built from lower
5638 level functions which work on OCSP_SINGLERESP structures but
5639 wont normally be used unless the application wishes to examine
5640 extensions in the OCSP response for example.
5642 Replace nonce routines with a pair of functions.
5643 OCSP_request_add1_nonce() adds a nonce value and optionally
5644 generates a random value. OCSP_check_nonce() checks the
5645 validity of the nonce in an OCSP response.
5648 *) Change function OCSP_request_add() to OCSP_request_add0_id().
5649 This doesn't copy the supplied OCSP_CERTID and avoids the
5650 need to free up the newly created id. Change return type
5651 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
5652 This can then be used to add extensions to the request.
5653 Deleted OCSP_request_new(), since most of its functionality
5654 is now in OCSP_REQUEST_new() (and the case insensitive name
5655 clash) apart from the ability to set the request name which
5656 will be added elsewhere.
5659 *) Update OCSP API. Remove obsolete extensions argument from
5660 various functions. Extensions are now handled using the new
5661 OCSP extension code. New simple OCSP HTTP function which
5662 can be used to send requests and parse the response.
5665 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
5666 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
5667 uses the special reorder version of SET OF to sort the attributes
5668 and reorder them to match the encoded order. This resolves a long
5669 standing problem: a verify on a PKCS7 structure just after signing
5670 it used to fail because the attribute order did not match the
5671 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
5672 it uses the received order. This is necessary to tolerate some broken
5673 software that does not order SET OF. This is handled by encoding
5674 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
5675 to produce the required SET OF.
5678 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
5679 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
5680 files to get correct declarations of the ASN.1 item variables.
5683 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
5684 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
5685 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
5686 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
5687 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
5688 ASN1_ITEM and no wrapper functions.
5691 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
5692 replace the old function pointer based I/O routines. Change most of
5693 the *_d2i_bio() and *_d2i_fp() functions to use these.
5696 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
5697 lines, recognice more "algorithms" that can be deselected, and make
5698 it complain about algorithm deselection that isn't recognised.
5701 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
5702 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
5703 to use new functions. Add NO_ASN1_OLD which can be set to remove
5704 some old style ASN1 functions: this can be used to determine if old
5705 code will still work when these eventually go away.
5708 *) New extension functions for OCSP structures, these follow the
5709 same conventions as certificates and CRLs.
5712 *) New function X509V3_add1_i2d(). This automatically encodes and
5713 adds an extension. Its behaviour can be customised with various
5714 flags to append, replace or delete. Various wrappers added for
5715 certifcates and CRLs.
5718 *) Fix to avoid calling the underlying ASN1 print routine when
5719 an extension cannot be parsed. Correct a typo in the
5720 OCSP_SERVICELOC extension. Tidy up print OCSP format.
5723 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
5724 entries for variables.
5727 *) Add functionality to apps/openssl.c for detecting locking
5728 problems: As the program is single-threaded, all we have
5729 to do is register a locking callback using an array for
5730 storing which locks are currently held by the program.
5733 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
5734 SSL_get_ex_data_X509_STORE_idx(), which is used in
5735 ssl_verify_cert_chain() and thus can be called at any time
5736 during TLS/SSL handshakes so that thread-safety is essential.
5737 Unfortunately, the ex_data design is not at all suited
5738 for multi-threaded use, so it probably should be abolished.
5741 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5742 [Broadcom, tweaked and integrated by Geoff Thorpe]
5744 *) Move common extension printing code to new function
5745 X509V3_print_extensions(). Reorganise OCSP print routines and
5746 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5749 *) New function X509_signature_print() to remove duplication in some
5753 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5754 set (this was treated exactly the same as SET OF previously). This
5755 is used to reorder the STACK representing the structure to match the
5756 encoding. This will be used to get round a problem where a PKCS7
5757 structure which was signed could not be verified because the STACK
5758 order did not reflect the encoded order.
5761 *) Reimplement the OCSP ASN1 module using the new code.
5764 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5765 for its ASN1 operations. The old style function pointers still exist
5766 for now but they will eventually go away.
5769 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
5770 completely replaces the old ASN1 functionality with a table driven
5771 encoder and decoder which interprets an ASN1_ITEM structure describing
5772 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
5773 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
5774 has also been converted to the new form.
5777 *) Change BN_mod_exp_recp so that negative moduli are tolerated
5778 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
5779 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
5780 for negative moduli.
5783 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
5784 of not touching the result's sign bit.
5787 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
5791 *) Changed the LHASH code to use prototypes for callbacks, and created
5792 macros to declare and implement thin (optionally static) functions
5793 that provide type-safety and avoid function pointer casting for the
5794 type-specific callbacks.
5797 *) Added Kerberos Cipher Suites to be used with TLS, as written in
5799 [Veers Staats <staatsvr@asc.hpc.mil>,
5800 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
5802 *) Reformat the FAQ so the different questions and answers can be divided
5803 in sections depending on the subject.
5806 *) Have the zlib compression code load ZLIB.DLL dynamically under
5810 *) New function BN_mod_sqrt for computing square roots modulo a prime
5811 (using the probabilistic Tonelli-Shanks algorithm unless
5812 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
5813 be handled deterministically).
5814 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5816 *) Make BN_mod_inverse faster by explicitly handling small quotients
5817 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
5818 512 bits], about 30% for larger ones [1024 or 2048 bits].)
5821 *) New function BN_kronecker.
5824 *) Fix BN_gcd so that it works on negative inputs; the result is
5825 positive unless both parameters are zero.
5826 Previously something reasonably close to an infinite loop was
5827 possible because numbers could be growing instead of shrinking
5828 in the implementation of Euclid's algorithm.
5831 *) Fix BN_is_word() and BN_is_one() macros to take into account the
5832 sign of the number in question.
5834 Fix BN_is_word(a,w) to work correctly for w == 0.
5836 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
5837 because its test if the absolute value of 'a' equals 'w'.
5838 Note that BN_abs_is_word does *not* handle w == 0 reliably;
5839 it exists mostly for use in the implementations of BN_is_zero(),
5840 BN_is_one(), and BN_is_word().
5843 *) New function BN_swap.
5846 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
5847 the exponentiation functions are more likely to produce reasonable
5848 results on negative inputs.
5851 *) Change BN_mod_mul so that the result is always non-negative.
5852 Previously, it could be negative if one of the factors was negative;
5853 I don't think anyone really wanted that behaviour.
5856 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
5857 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
5858 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
5859 and add new functions:
5868 BN_mod_lshift1_quick
5872 These functions always generate non-negative results.
5874 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
5875 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
5877 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
5878 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
5879 be reduced modulo m.
5880 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5883 The following entry accidentily appeared in the CHANGES file
5884 distributed with OpenSSL 0.9.7. The modifications described in
5885 it do *not* apply to OpenSSL 0.9.7.
5887 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5888 was actually never needed) and in BN_mul(). The removal in BN_mul()
5889 required a small change in bn_mul_part_recursive() and the addition
5890 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5891 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5892 bn_sub_words() and bn_add_words() except they take arrays with
5897 *) In 'openssl passwd', verify passwords read from the terminal
5898 unless the '-salt' option is used (which usually means that
5899 verification would just waste user's time since the resulting
5900 hash is going to be compared with some given password hash)
5901 or the new '-noverify' option is used.
5903 This is an incompatible change, but it does not affect
5904 non-interactive use of 'openssl passwd' (passwords on the command
5905 line, '-stdin' option, '-in ...' option) and thus should not
5909 *) Remove all references to RSAref, since there's no more need for it.
5912 *) Make DSO load along a path given through an environment variable
5913 (SHLIB_PATH) with shl_load().
5916 *) Constify the ENGINE code as a result of BIGNUM constification.
5917 Also constify the RSA code and most things related to it. In a
5918 few places, most notable in the depth of the ASN.1 code, ugly
5919 casts back to non-const were required (to be solved at a later
5923 *) Make it so the openssl application has all engines loaded by default.
5926 *) Constify the BIGNUM routines a little more.
5929 *) Add the following functions:
5931 ENGINE_load_cswift()
5933 ENGINE_load_atalla()
5935 ENGINE_load_builtin_engines()
5937 That way, an application can itself choose if external engines that
5938 are built-in in OpenSSL shall ever be used or not. The benefit is
5939 that applications won't have to be linked with libdl or other dso
5940 libraries unless it's really needed.
5942 Changed 'openssl engine' to load all engines on demand.
5943 Changed the engine header files to avoid the duplication of some
5944 declarations (they differed!).
5947 *) 'openssl engine' can now list capabilities.
5950 *) Better error reporting in 'openssl engine'.
5953 *) Never call load_dh_param(NULL) in s_server.
5956 *) Add engine application. It can currently list engines by name and
5957 identity, and test if they are actually available.
5960 *) Improve RPM specification file by forcing symbolic linking and making
5961 sure the installed documentation is also owned by root.root.
5962 [Damien Miller <djm@mindrot.org>]
5964 *) Give the OpenSSL applications more possibilities to make use of
5965 keys (public as well as private) handled by engines.
5968 *) Add OCSP code that comes from CertCo.
5971 *) Add VMS support for the Rijndael code.
5974 *) Added untested support for Nuron crypto accelerator.
5977 *) Add support for external cryptographic devices. This code was
5978 previously distributed separately as the "engine" branch.
5979 [Geoff Thorpe, Richard Levitte]
5981 *) Rework the filename-translation in the DSO code. It is now possible to
5982 have far greater control over how a "name" is turned into a filename
5983 depending on the operating environment and any oddities about the
5984 different shared library filenames on each system.
5987 *) Support threads on FreeBSD-elf in Configure.
5990 *) Fix for SHA1 assembly problem with MASM: it produces
5991 warnings about corrupt line number information when assembling
5992 with debugging information. This is caused by the overlapping
5994 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
5997 NCONF_get_number() has no error checking at all. As a replacement,
5998 NCONF_get_number_e() is defined (_e for "error checking") and is
5999 promoted strongly. The old NCONF_get_number is kept around for
6000 binary backward compatibility.
6001 Make it possible for methods to load from something other than a BIO,
6002 by providing a function pointer that is given a name instead of a BIO.
6003 For example, this could be used to load configuration data from an
6007 *) Fix for non blocking accept BIOs. Added new I/O special reason
6008 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
6009 with non blocking I/O was not possible because no retry code was
6010 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
6014 *) Added the beginnings of Rijndael support.
6017 *) Fix for bug in DirectoryString mask setting. Add support for
6018 X509_NAME_print_ex() in 'req' and X509_print_ex() function
6019 to allow certificate printing to more controllable, additional
6020 'certopt' option to 'x509' to allow new printing options to be
6024 *) Clean old EAY MD5 hack from e_os.h.
6027 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
6029 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6030 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6031 [Joe Orton, Steve Henson]
6033 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
6035 *) Fix additional bug revealed by the NISCC test suite:
6037 Stop bug triggering large recursion when presented with
6038 certain ASN.1 tags (CVE-2003-0851)
6041 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
6043 *) Fix various bugs revealed by running the NISCC test suite:
6045 Stop out of bounds reads in the ASN1 code when presented with
6046 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6048 If verify callback ignores invalid public key errors don't try to check
6049 certificate signature with the NULL public key.
6053 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6054 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6058 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6059 extra data after the compression methods not only for TLS 1.0
6060 but also for SSL 3.0 (as required by the specification).
6061 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6063 *) Change X509_certificate_type() to mark the key as exported/exportable
6064 when it's 512 *bits* long, not 512 bytes.
6067 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
6069 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6070 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6071 a protocol version number mismatch like a decryption error
6072 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6075 *) Turn on RSA blinding by default in the default implementation
6076 to avoid a timing attack. Applications that don't want it can call
6077 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6078 They would be ill-advised to do so in most cases.
6079 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6081 *) Change RSA blinding code so that it works when the PRNG is not
6082 seeded (in this case, the secret RSA exponent is abused as
6083 an unpredictable seed -- if it is not unpredictable, there
6084 is no point in blinding anyway). Make RSA blinding thread-safe
6085 by remembering the creator's thread ID in rsa->blinding and
6086 having all other threads use local one-time blinding factors
6087 (this requires more computation than sharing rsa->blinding, but
6088 avoids excessive locking; and if an RSA object is not shared
6089 between threads, blinding will still be very fast).
6092 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
6094 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6095 via timing by performing a MAC computation even if incorrrect
6096 block cipher padding has been found. This is a countermeasure
6097 against active attacks where the attacker has to distinguish
6098 between bad padding and a MAC verification error. (CVE-2003-0078)
6100 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6101 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6102 Martin Vuagnoux (EPFL, Ilion)]
6104 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
6106 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
6107 memory from it's contents. This is done with a counter that will
6108 place alternating values in each byte. This can be used to solve
6109 two issues: 1) the removal of calls to memset() by highly optimizing
6110 compilers, and 2) cleansing with other values than 0, since those can
6111 be read through on certain media, for example a swap space on disk.
6114 *) Bugfix: client side session caching did not work with external caching,
6115 because the session->cipher setting was not restored when reloading
6116 from the external cache. This problem was masked, when
6117 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
6118 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
6121 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
6122 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
6123 [Zeev Lieber <zeev-l@yahoo.com>]
6125 *) Undo an undocumented change introduced in 0.9.6e which caused
6126 repeated calls to OpenSSL_add_all_ciphers() and
6127 OpenSSL_add_all_digests() to be ignored, even after calling
6131 *) Change the default configuration reader to deal with last line not
6132 being properly terminated.
6135 *) Change X509_NAME_cmp() so it applies the special rules on handling
6136 DN values that are of type PrintableString, as well as RDNs of type
6137 emailAddress where the value has the type ia5String.
6138 [stefank@valicert.com via Richard Levitte]
6140 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
6141 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
6142 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
6143 the bitwise-OR of the two for use by the majority of applications
6144 wanting this behaviour, and update the docs. The documented
6145 behaviour and actual behaviour were inconsistent and had been
6146 changing anyway, so this is more a bug-fix than a behavioural
6148 [Geoff Thorpe, diagnosed by Nadav Har'El]
6150 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
6151 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
6154 *) Fix initialization code race conditions in
6155 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
6156 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
6157 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
6158 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
6159 ssl2_get_cipher_by_char(),
6160 ssl3_get_cipher_by_char().
6161 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
6163 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
6164 the cached sessions are flushed, as the remove_cb() might use ex_data
6165 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
6166 (see [openssl.org #212]).
6167 [Geoff Thorpe, Lutz Jaenicke]
6169 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
6170 length, instead of the encoding length to d2i_ASN1_OBJECT.
6173 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
6175 *) [In 0.9.6g-engine release:]
6176 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
6177 [Lynn Gazis <lgazis@rainbow.com>]
6179 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
6181 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
6182 and get fix the header length calculation.
6183 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
6184 Alon Kantor <alonk@checkpoint.com> (and others),
6187 *) Use proper error handling instead of 'assertions' in buffer
6188 overflow checks added in 0.9.6e. This prevents DoS (the
6189 assertions could call abort()).
6190 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
6192 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
6194 *) Add various sanity checks to asn1_get_length() to reject
6195 the ASN1 length bytes if they exceed sizeof(long), will appear
6196 negative or the content length exceeds the length of the
6198 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6200 *) Fix cipher selection routines: ciphers without encryption had no flags
6201 for the cipher strength set and where therefore not handled correctly
6202 by the selection routines (PR #130).
6205 *) Fix EVP_dsa_sha macro.
6209 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
6210 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
6211 that was added in OpenSSL 0.9.6d.
6213 As the countermeasure turned out to be incompatible with some
6214 broken SSL implementations, the new option is part of SSL_OP_ALL.
6215 SSL_OP_ALL is usually employed when compatibility with weird SSL
6216 implementations is desired (e.g. '-bugs' option to 's_client' and
6217 's_server'), so the new option is automatically set in many
6221 *) Changes in security patch:
6223 Changes marked "(CHATS)" were sponsored by the Defense Advanced
6224 Research Projects Agency (DARPA) and Air Force Research Laboratory,
6225 Air Force Materiel Command, USAF, under agreement number
6228 *) Add various sanity checks to asn1_get_length() to reject
6229 the ASN1 length bytes if they exceed sizeof(long), will appear
6230 negative or the content length exceeds the length of the
6231 supplied buffer. (CVE-2002-0659)
6232 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6234 *) Assertions for various potential buffer overflows, not known to
6236 [Ben Laurie (CHATS)]
6238 *) Various temporary buffers to hold ASCII versions of integers were
6239 too small for 64 bit platforms. (CVE-2002-0655)
6240 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
6242 *) Remote buffer overflow in SSL3 protocol - an attacker could
6243 supply an oversized session ID to a client. (CVE-2002-0656)
6244 [Ben Laurie (CHATS)]
6246 *) Remote buffer overflow in SSL2 protocol - an attacker could
6247 supply an oversized client master key. (CVE-2002-0656)
6248 [Ben Laurie (CHATS)]
6250 Changes between 0.9.6c and 0.9.6d [9 May 2002]
6252 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
6253 encoded as NULL) with id-dsa-with-sha1.
6254 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
6256 *) Check various X509_...() return values in apps/req.c.
6257 [Nils Larsch <nla@trustcenter.de>]
6259 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
6260 an end-of-file condition would erronously be flagged, when the CRLF
6261 was just at the end of a processed block. The bug was discovered when
6262 processing data through a buffering memory BIO handing the data to a
6263 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
6264 <ptsekov@syntrex.com> and Nedelcho Stanev.
6267 *) Implement a countermeasure against a vulnerability recently found
6268 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
6269 before application data chunks to avoid the use of known IVs
6270 with data potentially chosen by the attacker.
6273 *) Fix length checks in ssl3_get_client_hello().
6276 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
6277 to prevent ssl3_read_internal() from incorrectly assuming that
6278 ssl3_read_bytes() found application data while handshake
6279 processing was enabled when in fact s->s3->in_read_app_data was
6280 merely automatically cleared during the initial handshake.
6281 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
6283 *) Fix object definitions for Private and Enterprise: they were not
6284 recognized in their shortname (=lowercase) representation. Extend
6285 obj_dat.pl to issue an error when using undefined keywords instead
6286 of silently ignoring the problem (Svenning Sorensen
6287 <sss@sss.dnsalias.net>).
6290 *) Fix DH_generate_parameters() so that it works for 'non-standard'
6291 generators, i.e. generators other than 2 and 5. (Previously, the
6292 code did not properly initialise the 'add' and 'rem' values to
6293 BN_generate_prime().)
6295 In the new general case, we do not insist that 'generator' is
6296 actually a primitive root: This requirement is rather pointless;
6297 a generator of the order-q subgroup is just as good, if not
6301 *) Map new X509 verification errors to alerts. Discovered and submitted by
6302 Tom Wu <tom@arcot.com>.
6305 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
6306 returning non-zero before the data has been completely received
6307 when using non-blocking I/O.
6308 [Bodo Moeller; problem pointed out by John Hughes]
6310 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
6311 [Ben Laurie, Lutz Jaenicke]
6313 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
6314 Yoram Zahavi <YoramZ@gilian.com>).
6317 *) Add information about CygWin 1.3 and on, and preserve proper
6318 configuration for the versions before that.
6319 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6321 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
6322 check whether we deal with a copy of a session and do not delete from
6323 the cache in this case. Problem reported by "Izhar Shoshani Levi"
6324 <izhar@checkpoint.com>.
6327 *) Do not store session data into the internal session cache, if it
6328 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
6329 flag is set). Proposed by Aslam <aslam@funk.com>.
6332 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
6336 *) [In 0.9.6d-engine release:]
6337 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6338 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6340 *) Add the configuration target linux-s390x.
6341 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
6343 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
6344 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
6345 variable as an indication that a ClientHello message has been
6346 received. As the flag value will be lost between multiple
6347 invocations of ssl3_accept when using non-blocking I/O, the
6348 function may not be aware that a handshake has actually taken
6349 place, thus preventing a new session from being added to the
6352 To avoid this problem, we now set s->new_session to 2 instead of
6353 using a local variable.
6354 [Lutz Jaenicke, Bodo Moeller]
6356 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
6357 if the SSL_R_LENGTH_MISMATCH error is detected.
6358 [Geoff Thorpe, Bodo Moeller]
6360 *) New 'shared_ldflag' column in Configure platform table.
6363 *) Fix EVP_CIPHER_mode macro.
6364 ["Dan S. Camper" <dan@bti.net>]
6366 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
6367 type, we must throw them away by setting rr->length to 0.
6368 [D P Chang <dpc@qualys.com>]
6370 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
6372 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
6373 <Dominikus.Scherkl@biodata.com>. (The previous implementation
6374 worked incorrectly for those cases where range = 10..._2 and
6375 3*range is two bits longer than range.)
6378 *) Only add signing time to PKCS7 structures if it is not already
6382 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
6383 OBJ_ld_ce should be OBJ_id_ce.
6384 Also some ip-pda OIDs in crypto/objects/objects.txt were
6385 incorrect (cf. RFC 3039).
6386 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
6388 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
6389 returns early because it has nothing to do.
6390 [Andy Schneider <andy.schneider@bjss.co.uk>]
6392 *) [In 0.9.6c-engine release:]
6393 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
6394 [Andy Schneider <andy.schneider@bjss.co.uk>]
6396 *) [In 0.9.6c-engine release:]
6397 Add support for Cryptographic Appliance's keyserver technology.
6398 (Use engine 'keyclient')
6399 [Cryptographic Appliances and Geoff Thorpe]
6401 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
6402 is called via tools/c89.sh because arguments have to be
6403 rearranged (all '-L' options must appear before the first object
6405 [Richard Shapiro <rshapiro@abinitio.com>]
6407 *) [In 0.9.6c-engine release:]
6408 Add support for Broadcom crypto accelerator cards, backported
6410 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
6412 *) [In 0.9.6c-engine release:]
6413 Add support for SureWare crypto accelerator cards from
6414 Baltimore Technologies. (Use engine 'sureware')
6415 [Baltimore Technologies and Mark Cox]
6417 *) [In 0.9.6c-engine release:]
6418 Add support for crypto accelerator cards from Accelerated
6419 Encryption Processing, www.aep.ie. (Use engine 'aep')
6420 [AEP Inc. and Mark Cox]
6422 *) Add a configuration entry for gcc on UnixWare.
6423 [Gary Benson <gbenson@redhat.com>]
6425 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
6426 messages are stored in a single piece (fixed-length part and
6427 variable-length part combined) and fix various bugs found on the way.
6430 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
6431 instead. BIO_gethostbyname() does not know what timeouts are
6432 appropriate, so entries would stay in cache even when they have
6434 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
6436 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
6437 faced with a pathologically small ClientHello fragment that does
6438 not contain client_version: Instead of aborting with an error,
6439 simply choose the highest available protocol version (i.e.,
6440 TLS 1.0 unless it is disabled). In practice, ClientHello
6441 messages are never sent like this, but this change gives us
6442 strictly correct behaviour at least for TLS.
6445 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
6446 never resets s->method to s->ctx->method when called from within
6447 one of the SSL handshake functions.
6448 [Bodo Moeller; problem pointed out by Niko Baric]
6450 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
6451 (sent using the client's version number) if client_version is
6452 smaller than the protocol version in use. Also change
6453 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
6454 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
6455 the client will at least see that alert.
6458 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
6462 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
6463 client receives HelloRequest while in a handshake.
6464 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
6466 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
6467 should end in 'break', not 'goto end' which circuments various
6468 cleanups done in state SSL_ST_OK. But session related stuff
6469 must be disabled for SSL_ST_OK in the case that we just sent a
6472 Also avoid some overhead by not calling ssl_init_wbio_buffer()
6473 before just sending a HelloRequest.
6474 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
6476 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
6477 reveal whether illegal block cipher padding was found or a MAC
6478 verification error occured. (Neither SSLerr() codes nor alerts
6479 are directly visible to potential attackers, but the information
6480 may leak via logfiles.)
6482 Similar changes are not required for the SSL 2.0 implementation
6483 because the number of padding bytes is sent in clear for SSL 2.0,
6484 and the extra bytes are just ignored. However ssl/s2_pkt.c
6485 failed to verify that the purported number of padding bytes is in
6489 *) Add OpenUNIX-8 support including shared libraries
6490 (Boyd Lynn Gerber <gerberb@zenez.com>).
6493 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
6494 'wristwatch attack' using huge encoding parameters (cf.
6495 James H. Manger's CRYPTO 2001 paper). Note that the
6496 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
6497 encoding parameters and hence was not vulnerable.
6500 *) BN_sqr() bug fix.
6501 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
6503 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
6504 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
6505 followed by modular reduction.
6506 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
6508 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
6509 equivalent based on BN_pseudo_rand() instead of BN_rand().
6512 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
6513 This function was broken, as the check for a new client hello message
6514 to handle SGC did not allow these large messages.
6515 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
6518 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
6521 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
6522 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
6525 *) Rework the configuration and shared library support for Tru64 Unix.
6526 The configuration part makes use of modern compiler features and
6527 still retains old compiler behavior for those that run older versions
6528 of the OS. The shared library support part includes a variant that
6529 uses the RPATH feature, and is available through the special
6530 configuration target "alpha-cc-rpath", which will never be selected
6532 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
6534 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
6535 with the same message size as in ssl3_get_certificate_request().
6536 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
6537 messages might inadvertently be reject as too long.
6538 [Petr Lampa <lampa@fee.vutbr.cz>]
6540 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
6543 *) Modified SSL library such that the verify_callback that has been set
6544 specificly for an SSL object with SSL_set_verify() is actually being
6545 used. Before the change, a verify_callback set with this function was
6546 ignored and the verify_callback() set in the SSL_CTX at the time of
6547 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
6548 to allow the necessary settings.
6551 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
6552 explicitly to NULL, as at least on Solaris 8 this seems not always to be
6553 done automatically (in contradiction to the requirements of the C
6554 standard). This made problems when used from OpenSSH.
6557 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
6558 dh->length and always used
6560 BN_rand_range(priv_key, dh->p).
6562 BN_rand_range() is not necessary for Diffie-Hellman, and this
6563 specific range makes Diffie-Hellman unnecessarily inefficient if
6564 dh->length (recommended exponent length) is much smaller than the
6565 length of dh->p. We could use BN_rand_range() if the order of
6566 the subgroup was stored in the DH structure, but we only have
6571 BN_rand(priv_key, l, ...)
6573 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
6579 RSA_eay_public_encrypt
6580 RSA_eay_private_decrypt
6581 RSA_eay_private_encrypt (signing)
6582 RSA_eay_public_decrypt (signature verification)
6584 (default implementations for RSA_public_encrypt,
6585 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
6586 always reject numbers >= n.
6589 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
6590 to synchronize access to 'locking_thread'. This is necessary on
6591 systems where access to 'locking_thread' (an 'unsigned long'
6592 variable) is not atomic.
6595 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
6596 *before* setting the 'crypto_lock_rand' flag. The previous code had
6597 a race condition if 0 is a valid thread ID.
6598 [Travis Vitek <vitek@roguewave.com>]
6600 *) Add support for shared libraries under Irix.
6601 [Albert Chin-A-Young <china@thewrittenword.com>]
6603 *) Add configuration option to build on Linux on both big-endian and
6605 [Ralf Baechle <ralf@uni-koblenz.de>]
6607 *) Add the possibility to create shared libraries on HP-UX.
6610 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
6612 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
6613 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
6614 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
6615 PRNG state recovery was possible based on the output of
6616 one PRNG request appropriately sized to gain knowledge on
6617 'md' followed by enough consecutive 1-byte PRNG requests
6618 to traverse all of 'state'.
6620 1. When updating 'md_local' (the current thread's copy of 'md')
6621 during PRNG output generation, hash all of the previous
6622 'md_local' value, not just the half used for PRNG output.
6624 2. Make the number of bytes from 'state' included into the hash
6625 independent from the number of PRNG bytes requested.
6627 The first measure alone would be sufficient to avoid
6628 Markku-Juhani's attack. (Actually it had never occurred
6629 to me that the half of 'md_local' used for chaining was the
6630 half from which PRNG output bytes were taken -- I had always
6631 assumed that the secret half would be used.) The second
6632 measure makes sure that additional data from 'state' is never
6633 mixed into 'md_local' in small portions; this heuristically
6634 further strengthens the PRNG.
6637 *) Fix crypto/bn/asm/mips3.s.
6640 *) When only the key is given to "enc", the IV is undefined. Print out
6641 an error message in this case.
6644 *) Handle special case when X509_NAME is empty in X509 printing routines.
6647 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
6648 positive and less than q.
6651 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
6652 used: it isn't thread safe and the add_lock_callback should handle
6654 [Paul Rose <Paul.Rose@bridge.com>]
6656 *) Verify that incoming data obeys the block size in
6657 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
6661 [Ulf Möller, Bodo Möller]
6663 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
6664 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
6665 when fixing the server behaviour for backwards-compatible 'client
6666 hello' messages. (Note that the attack is impractical against
6667 SSL 3.0 and TLS 1.0 anyway because length and version checking
6668 means that the probability of guessing a valid ciphertext is
6669 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
6672 Before 0.9.5, the countermeasure (hide the error by generating a
6673 random 'decryption result') did not work properly because
6674 ERR_clear_error() was missing, meaning that SSL_get_error() would
6675 detect the supposedly ignored error.
6677 Both problems are now fixed.
6680 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
6681 (previously it was 1024).
6684 *) Fix for compatibility mode trust settings: ignore trust settings
6685 unless some valid trust or reject settings are present.
6688 *) Fix for blowfish EVP: its a variable length cipher.
6691 *) Fix various bugs related to DSA S/MIME verification. Handle missing
6692 parameters in DSA public key structures and return an error in the
6693 DSA routines if parameters are absent.
6696 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
6697 in the current directory if neither $RANDFILE nor $HOME was set.
6698 RAND_file_name() in 0.9.6a returned NULL in this case. This has
6699 caused some confusion to Windows users who haven't defined $HOME.
6700 Thus RAND_file_name() is changed again: e_os.h can define a
6701 DEFAULT_HOME, which will be used if $HOME is not set.
6702 For Windows, we use "C:"; on other platforms, we still require
6703 environment variables.
6705 *) Move 'if (!initialized) RAND_poll()' into regions protected by
6706 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
6707 having multiple threads call RAND_poll() concurrently.
6710 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
6711 combination of a flag and a thread ID variable.
6712 Otherwise while one thread is in ssleay_rand_bytes (which sets the
6713 flag), *other* threads can enter ssleay_add_bytes without obeying
6714 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
6715 that they do not hold after the first thread unsets add_do_not_lock).
6718 *) Change bctest again: '-x' expressions are not available in all
6722 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
6724 *) Fix a couple of memory leaks in PKCS7_dataDecode()
6725 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
6727 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
6728 the default extension for executables, if any. Also, make the perl
6729 scripts that use symlink() to test if it really exists and use "cp"
6730 if it doesn't. All this made OpenSSL compilable and installable in
6734 *) Fix for asn1_GetSequence() for indefinite length constructed data.
6735 If SEQUENCE is length is indefinite just set c->slen to the total
6736 amount of data available.
6737 [Steve Henson, reported by shige@FreeBSD.org]
6738 [This change does not apply to 0.9.7.]
6740 *) Change bctest to avoid here-documents inside command substitution
6741 (workaround for FreeBSD /bin/sh bug).
6742 For compatibility with Ultrix, avoid shell functions (introduced
6743 in the bctest version that searches along $PATH).
6746 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6747 with des_encrypt() defined on some operating systems, like Solaris
6751 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6752 On the Importance of Eliminating Errors in Cryptographic
6753 Computations, J. Cryptology 14 (2001) 2, 101-119,
6754 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6757 *) MIPS assembler BIGNUM division bug fix.
6760 *) Disabled incorrect Alpha assembler code.
6763 *) Fix PKCS#7 decode routines so they correctly update the length
6764 after reading an EOC for the EXPLICIT tag.
6766 [This change does not apply to 0.9.7.]
6768 *) Fix bug in PKCS#12 key generation routines. This was triggered
6769 if a 3DES key was generated with a 0 initial byte. Include
6770 PKCS12_BROKEN_KEYGEN compilation option to retain the old
6771 (but broken) behaviour.
6774 *) Enhance bctest to search for a working bc along $PATH and print
6776 [Tim Rice <tim@multitalents.net> via Richard Levitte]
6778 *) Fix memory leaks in err.c: free err_data string if necessary;
6779 don't write to the wrong index in ERR_set_error_data.
6782 *) Implement ssl23_peek (analogous to ssl23_read), which previously
6786 *) Replace rdtsc with _emit statements for VC++ version 5.
6787 [Jeremy Cooper <jeremy@baymoo.org>]
6789 *) Make it possible to reuse SSLv2 sessions.
6792 *) In copy_email() check for >= 0 as a return value for
6793 X509_NAME_get_index_by_NID() since 0 is a valid index.
6794 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
6796 *) Avoid coredump with unsupported or invalid public keys by checking if
6797 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
6798 PKCS7_verify() fails with non detached data.
6801 *) Don't use getenv in library functions when run as setuid/setgid.
6802 New function OPENSSL_issetugid().
6805 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
6806 due to incorrect handling of multi-threading:
6808 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
6810 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
6812 3. Count how many times MemCheck_off() has been called so that
6813 nested use can be treated correctly. This also avoids
6814 inband-signalling in the previous code (which relied on the
6815 assumption that thread ID 0 is impossible).
6818 *) Add "-rand" option also to s_client and s_server.
6821 *) Fix CPU detection on Irix 6.x.
6822 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
6823 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6825 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
6828 [This change does not apply to 0.9.7.]
6830 *) Use the cached encoding of an X509_NAME structure rather than
6831 copying it. This is apparently the reason for the libsafe "errors"
6832 but the code is actually correct.
6835 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
6836 Bleichenbacher's DSA attack.
6837 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
6838 to be set and top=0 forces the highest bit to be set; top=-1 is new
6839 and leaves the highest bit random.
6840 [Ulf Moeller, Bodo Moeller]
6842 *) In the NCONF_...-based implementations for CONF_... queries
6843 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
6844 a temporary CONF structure with the data component set to NULL
6845 (which gives segmentation faults in lh_retrieve).
6846 Instead, use NULL for the CONF pointer in CONF_get_string and
6847 CONF_get_number (which may use environment variables) and directly
6848 return NULL from CONF_get_section.
6851 *) Fix potential buffer overrun for EBCDIC.
6854 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
6855 keyUsage if basicConstraints absent for a CA.
6858 *) Make SMIME_write_PKCS7() write mail header values with a format that
6859 is more generally accepted (no spaces before the semicolon), since
6860 some programs can't parse those values properly otherwise. Also make
6861 sure BIO's that break lines after each write do not create invalid
6865 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
6866 macros previously used would not encode an empty SEQUENCE OF
6867 and break the signature.
6869 [This change does not apply to 0.9.7.]
6871 *) Zero the premaster secret after deriving the master secret in
6875 *) Add some EVP_add_digest_alias registrations (as found in
6876 OpenSSL_add_all_digests()) to SSL_library_init()
6877 aka OpenSSL_add_ssl_algorithms(). This provides improved
6878 compatibility with peers using X.509 certificates
6879 with unconventional AlgorithmIdentifier OIDs.
6882 *) Fix for Irix with NO_ASM.
6883 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6885 *) ./config script fixes.
6886 [Ulf Moeller, Richard Levitte]
6888 *) Fix 'openssl passwd -1'.
6891 *) Change PKCS12_key_gen_asc() so it can cope with non null
6892 terminated strings whose length is passed in the passlen
6893 parameter, for example from PEM callbacks. This was done
6894 by adding an extra length parameter to asc2uni().
6895 [Steve Henson, reported by <oddissey@samsung.co.kr>]
6897 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
6898 call failed, free the DSA structure.
6901 *) Fix to uni2asc() to cope with zero length Unicode strings.
6902 These are present in some PKCS#12 files.
6905 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
6906 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
6907 when writing a 32767 byte record.
6908 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
6910 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
6911 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
6913 (RSA objects have a reference count access to which is protected
6914 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
6915 so they are meant to be shared between threads.)
6916 [Bodo Moeller, Geoff Thorpe; original patch submitted by
6917 "Reddie, Steven" <Steven.Reddie@ca.com>]
6919 *) Fix a deadlock in CRYPTO_mem_leaks().
6922 *) Use better test patterns in bntest.
6925 *) rand_win.c fix for Borland C.
6928 *) BN_rshift bugfix for n == 0.
6931 *) Add a 'bctest' script that checks for some known 'bc' bugs
6932 so that 'make test' does not abort just because 'bc' is broken.
6935 *) Store verify_result within SSL_SESSION also for client side to
6936 avoid potential security hole. (Re-used sessions on the client side
6937 always resulted in verify_result==X509_V_OK, not using the original
6938 result of the server certificate verification.)
6941 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
6942 SSL3_RT_APPLICATION_DATA, return 0.
6943 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
6947 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
6948 releases, have been re-implemented by renaming the previous
6949 implementations of ssl2_read and ssl3_read to ssl2_read_internal
6950 and ssl3_read_internal, respectively, and adding 'peek' parameters
6951 to them. The new ssl[23]_{read,peek} functions are calls to
6952 ssl[23]_read_internal with the 'peek' flag set appropriately.
6953 A 'peek' parameter has also been added to ssl3_read_bytes, which
6954 does the actual work for ssl3_read_internal.
6957 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
6958 the method-specific "init()" handler. Also clean up ex_data after
6959 calling the method-specific "finish()" handler. Previously, this was
6960 happening the other way round.
6963 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
6964 The previous value, 12, was not always sufficient for BN_mod_exp().
6967 *) Make sure that shared libraries get the internal name engine with
6968 the full version number and not just 0. This should mark the
6969 shared libraries as not backward compatible. Of course, this should
6970 be changed again when we can guarantee backward binary compatibility.
6973 *) Fix typo in get_cert_by_subject() in by_dir.c
6974 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
6976 *) Rework the system to generate shared libraries:
6978 - Make note of the expected extension for the shared libraries and
6979 if there is a need for symbolic links from for example libcrypto.so.0
6980 to libcrypto.so.0.9.7. There is extended info in Configure for
6983 - Make as few rebuilds of the shared libraries as possible.
6985 - Still avoid linking the OpenSSL programs with the shared libraries.
6987 - When installing, install the shared libraries separately from the
6991 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
6993 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
6994 and not in SSL_clear because the latter is also used by the
6995 accept/connect functions; previously, the settings made by
6996 SSL_set_read_ahead would be lost during the handshake.
6997 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
6999 *) Correct util/mkdef.pl to be selective about disabled algorithms.
7000 Previously, it would create entries for disableed algorithms no
7004 *) Added several new manual pages for SSL_* function.
7007 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
7009 *) In ssl23_get_client_hello, generate an error message when faced
7010 with an initial SSL 3.0/TLS record that is too small to contain the
7011 first two bytes of the ClientHello message, i.e. client_version.
7012 (Note that this is a pathologic case that probably has never happened
7013 in real life.) The previous approach was to use the version number
7014 from the record header as a substitute; but our protocol choice
7015 should not depend on that one because it is not authenticated
7016 by the Finished messages.
7019 *) More robust randomness gathering functions for Windows.
7020 [Jeffrey Altman <jaltman@columbia.edu>]
7022 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
7023 not set then we don't setup the error code for issuer check errors
7024 to avoid possibly overwriting other errors which the callback does
7025 handle. If an application does set the flag then we assume it knows
7026 what it is doing and can handle the new informational codes
7030 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
7031 a general "ANY" type, as such it should be able to decode anything
7032 including tagged types. However it didn't check the class so it would
7033 wrongly interpret tagged types in the same way as their universal
7034 counterpart and unknown types were just rejected. Changed so that the
7035 tagged and unknown types are handled in the same way as a SEQUENCE:
7036 that is the encoding is stored intact. There is also a new type
7037 "V_ASN1_OTHER" which is used when the class is not universal, in this
7038 case we have no idea what the actual type is so we just lump them all
7042 *) On VMS, stdout may very well lead to a file that is written to
7043 in a record-oriented fashion. That means that every write() will
7044 write a separate record, which will be read separately by the
7045 programs trying to read from it. This can be very confusing.
7047 The solution is to put a BIO filter in the way that will buffer
7048 text until a linefeed is reached, and then write everything a
7049 line at a time, so every record written will be an actual line,
7050 not chunks of lines and not (usually doesn't happen, but I've
7051 seen it once) several lines in one record. BIO_f_linebuffer() is
7054 Currently, it's a VMS-only method, because that's where it has
7055 been tested well enough.
7058 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
7059 it can return incorrect results.
7060 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
7061 but it was in 0.9.6-beta[12].)
7064 *) Disable the check for content being present when verifying detached
7065 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
7066 include zero length content when signing messages.
7069 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
7070 BIO_ctrl (for BIO pairs).
7073 *) Add DSO method for VMS.
7076 *) Bug fix: Montgomery multiplication could produce results with the
7080 *) Add RPM specification openssl.spec and modify it to build three
7081 packages. The default package contains applications, application
7082 documentation and run-time libraries. The devel package contains
7083 include files, static libraries and function documentation. The
7084 doc package contains the contents of the doc directory. The original
7085 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
7088 *) Add a large number of documentation files for many SSL routines.
7089 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7091 *) Add a configuration entry for Sony News 4.
7092 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
7094 *) Don't set the two most significant bits to one when generating a
7095 random number < q in the DSA library.
7098 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
7099 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
7100 the underlying transport is blocking) if a handshake took place.
7101 (The default behaviour is needed by applications such as s_client
7102 and s_server that use select() to determine when to use SSL_read;
7103 but for applications that know in advance when to expect data, it
7104 just makes things more complicated.)
7107 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
7111 *) Add a few more EBCDIC conditionals that make `req' and `x509'
7112 work better on such systems.
7113 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7115 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
7116 Update PKCS12_parse() so it copies the friendlyName and the
7117 keyid to the certificates aux info.
7120 *) Fix bug in PKCS7_verify() which caused an infinite loop
7121 if there was more than one signature.
7122 [Sven Uszpelkat <su@celocom.de>]
7124 *) Major change in util/mkdef.pl to include extra information
7125 about each symbol, as well as presentig variables as well
7126 as functions. This change means that there's n more need
7127 to rebuild the .num files when some algorithms are excluded.
7130 *) Allow the verify time to be set by an application,
7131 rather than always using the current time.
7134 *) Phase 2 verify code reorganisation. The certificate
7135 verify code now looks up an issuer certificate by a
7136 number of criteria: subject name, authority key id
7137 and key usage. It also verifies self signed certificates
7138 by the same criteria. The main comparison function is
7139 X509_check_issued() which performs these checks.
7141 Lot of changes were necessary in order to support this
7142 without completely rewriting the lookup code.
7144 Authority and subject key identifier are now cached.
7146 The LHASH 'certs' is X509_STORE has now been replaced
7147 by a STACK_OF(X509_OBJECT). This is mainly because an
7148 LHASH can't store or retrieve multiple objects with
7149 the same hash value.
7151 As a result various functions (which were all internal
7152 use only) have changed to handle the new X509_STORE
7153 structure. This will break anything that messed round
7154 with X509_STORE internally.
7156 The functions X509_STORE_add_cert() now checks for an
7157 exact match, rather than just subject name.
7159 The X509_STORE API doesn't directly support the retrieval
7160 of multiple certificates matching a given criteria, however
7161 this can be worked round by performing a lookup first
7162 (which will fill the cache with candidate certificates)
7163 and then examining the cache for matches. This is probably
7164 the best we can do without throwing out X509_LOOKUP
7165 entirely (maybe later...).
7167 The X509_VERIFY_CTX structure has been enhanced considerably.
7169 All certificate lookup operations now go via a get_issuer()
7170 callback. Although this currently uses an X509_STORE it
7171 can be replaced by custom lookups. This is a simple way
7172 to bypass the X509_STORE hackery necessary to make this
7173 work and makes it possible to use more efficient techniques
7174 in future. A very simple version which uses a simple
7175 STACK for its trusted certificate store is also provided
7176 using X509_STORE_CTX_trusted_stack().
7178 The verify_cb() and verify() callbacks now have equivalents
7179 in the X509_STORE_CTX structure.
7181 X509_STORE_CTX also has a 'flags' field which can be used
7182 to customise the verify behaviour.
7185 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
7186 excludes S/MIME capabilities.
7189 *) When a certificate request is read in keep a copy of the
7190 original encoding of the signed data and use it when outputing
7191 again. Signatures then use the original encoding rather than
7192 a decoded, encoded version which may cause problems if the
7193 request is improperly encoded.
7196 *) For consistency with other BIO_puts implementations, call
7197 buffer_write(b, ...) directly in buffer_puts instead of calling
7200 In BIO_puts, increment b->num_write as in BIO_write.
7201 [Peter.Sylvester@EdelWeb.fr]
7203 *) Fix BN_mul_word for the case where the word is 0. (We have to use
7204 BN_zero, we may not return a BIGNUM with an array consisting of
7208 *) Avoid calling abort() from within the library when problems are
7209 detected, except if preprocessor symbols have been defined
7210 (such as REF_CHECK, BN_DEBUG etc.).
7213 *) New openssl application 'rsautl'. This utility can be
7214 used for low level RSA operations. DER public key
7215 BIO/fp routines also added.
7218 *) New Configure entry and patches for compiling on QNX 4.
7219 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
7221 *) A demo state-machine implementation was sponsored by
7222 Nuron (http://www.nuron.com/) and is now available in
7223 demos/state_machine.
7226 *) New options added to the 'dgst' utility for signature
7227 generation and verification.
7230 *) Unrecognized PKCS#7 content types are now handled via a
7231 catch all ASN1_TYPE structure. This allows unsupported
7232 types to be stored as a "blob" and an application can
7233 encode and decode it manually.
7236 *) Fix various signed/unsigned issues to make a_strex.c
7238 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
7240 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
7241 length if passed a buffer. ASN1_INTEGER_to_BN failed
7242 if passed a NULL BN and its argument was negative.
7243 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
7245 *) Modification to PKCS#7 encoding routines to output definite
7246 length encoding. Since currently the whole structures are in
7247 memory there's not real point in using indefinite length
7248 constructed encoding. However if OpenSSL is compiled with
7249 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
7252 *) Added BIO_vprintf() and BIO_vsnprintf().
7255 *) Added more prefixes to parse for in the the strings written
7256 through a logging bio, to cover all the levels that are available
7257 through syslog. The prefixes are now:
7259 PANIC, EMERG, EMR => LOG_EMERG
7260 ALERT, ALR => LOG_ALERT
7261 CRIT, CRI => LOG_CRIT
7262 ERROR, ERR => LOG_ERR
7263 WARNING, WARN, WAR => LOG_WARNING
7264 NOTICE, NOTE, NOT => LOG_NOTICE
7265 INFO, INF => LOG_INFO
7266 DEBUG, DBG => LOG_DEBUG
7268 and as before, if none of those prefixes are present at the
7269 beginning of the string, LOG_ERR is chosen.
7271 On Win32, the LOG_* levels are mapped according to this:
7273 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
7274 LOG_WARNING => EVENTLOG_WARNING_TYPE
7275 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
7279 *) Made it possible to reconfigure with just the configuration
7280 argument "reconf" or "reconfigure". The command line arguments
7281 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
7282 and are retrieved from there when reconfiguring.
7286 [Assar Westerlund <assar@sics.se>, Richard Levitte]
7288 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
7291 *) The obj_dat.pl script was messing up the sorting of object
7292 names. The reason was that it compared the quoted version
7293 of strings as a result "OCSP" > "OCSP Signing" because
7294 " > SPACE. Changed script to store unquoted versions of
7295 names and add quotes on output. It was also omitting some
7296 names from the lookup table if they were given a default
7297 value (that is if SN is missing it is given the same
7298 value as LN and vice versa), these are now added on the
7299 grounds that if an object has a name we should be able to
7300 look it up. Finally added warning output when duplicate
7301 short or long names are found.
7304 *) Changes needed for Tandem NSK.
7305 [Scott Uroff <scott@xypro.com>]
7307 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
7308 RSA_padding_check_SSLv23(), special padding was never detected
7309 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
7310 version rollback attacks was not effective.
7312 In s23_clnt.c, don't use special rollback-attack detection padding
7313 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
7314 client; similarly, in s23_srvr.c, don't do the rollback check if
7315 SSL 2.0 is the only protocol enabled in the server.
7318 *) Make it possible to get hexdumps of unprintable data with 'openssl
7319 asn1parse'. By implication, the functions ASN1_parse_dump() and
7320 BIO_dump_indent() are added.
7323 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
7324 these print out strings and name structures based on various
7325 flags including RFC2253 support and proper handling of
7326 multibyte characters. Added options to the 'x509' utility
7327 to allow the various flags to be set.
7330 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
7331 Also change the functions X509_cmp_current_time() and
7332 X509_gmtime_adj() work with an ASN1_TIME structure,
7333 this will enable certificates using GeneralizedTime in validity
7334 dates to be checked.
7337 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
7338 negative public key encodings) on by default,
7339 NO_NEG_PUBKEY_BUG can be set to disable it.
7342 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
7343 content octets. An i2c_ASN1_OBJECT is unnecessary because
7344 the encoding can be trivially obtained from the structure.
7347 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
7348 not read locks (CRYPTO_r_[un]lock).
7351 *) A first attempt at creating official support for shared
7352 libraries through configuration. I've kept it so the
7353 default is static libraries only, and the OpenSSL programs
7354 are always statically linked for now, but there are
7355 preparations for dynamic linking in place.
7356 This has been tested on Linux and Tru64.
7359 *) Randomness polling function for Win9x, as described in:
7360 Peter Gutmann, Software Generation of Practically Strong
7364 *) Fix so PRNG is seeded in req if using an already existing
7368 *) New options to smime application. -inform and -outform
7369 allow alternative formats for the S/MIME message including
7370 PEM and DER. The -content option allows the content to be
7371 specified separately. This should allow things like Netscape
7372 form signing output easier to verify.
7375 *) Fix the ASN1 encoding of tags using the 'long form'.
7378 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
7379 STRING types. These convert content octets to and from the
7380 underlying type. The actual tag and length octets are
7381 already assumed to have been read in and checked. These
7382 are needed because all other string types have virtually
7383 identical handling apart from the tag. By having versions
7384 of the ASN1 functions that just operate on content octets
7385 IMPLICIT tagging can be handled properly. It also allows
7386 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
7387 and ASN1_INTEGER are identical apart from the tag.
7390 *) Change the handling of OID objects as follows:
7392 - New object identifiers are inserted in objects.txt, following
7393 the syntax given in objects.README.
7394 - objects.pl is used to process obj_mac.num and create a new
7396 - obj_dat.pl is used to create a new obj_dat.h, using the data in
7399 This is currently kind of a hack, and the perl code in objects.pl
7400 isn't very elegant, but it works as I intended. The simplest way
7401 to check that it worked correctly is to look in obj_dat.h and
7402 check the array nid_objs and make sure the objects haven't moved
7403 around (this is important!). Additions are OK, as well as
7404 consistent name changes.
7407 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
7410 *) Addition of the command line parameter '-rand file' to 'openssl req'.
7411 The given file adds to whatever has already been seeded into the
7412 random pool through the RANDFILE configuration file option or
7413 environment variable, or the default random state file.
7416 *) mkstack.pl now sorts each macro group into lexical order.
7417 Previously the output order depended on the order the files
7418 appeared in the directory, resulting in needless rewriting
7422 *) Patches to make OpenSSL compile under Win32 again. Mostly
7423 work arounds for the VC++ problem that it treats func() as
7424 func(void). Also stripped out the parts of mkdef.pl that
7425 added extra typesafe functions: these no longer exist.
7428 *) Reorganisation of the stack code. The macros are now all
7429 collected in safestack.h . Each macro is defined in terms of
7430 a "stack macro" of the form SKM_<name>(type, a, b). The
7431 DEBUG_SAFESTACK is now handled in terms of function casts,
7432 this has the advantage of retaining type safety without the
7433 use of additional functions. If DEBUG_SAFESTACK is not defined
7434 then the non typesafe macros are used instead. Also modified the
7435 mkstack.pl script to handle the new form. Needs testing to see
7436 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
7437 the default if no major problems. Similar behaviour for ASN1_SET_OF
7438 and PKCS12_STACK_OF.
7441 *) When some versions of IIS use the 'NET' form of private key the
7442 key derivation algorithm is different. Normally MD5(password) is
7443 used as a 128 bit RC4 key. In the modified case
7444 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
7445 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
7446 as the old Netscape_RSA functions except they have an additional
7447 'sgckey' parameter which uses the modified algorithm. Also added
7448 an -sgckey command line option to the rsa utility. Thanks to
7449 Adrian Peck <bertie@ncipher.com> for posting details of the modified
7450 algorithm to openssl-dev.
7453 *) The evp_local.h macros were using 'c.##kname' which resulted in
7454 invalid expansion on some systems (SCO 5.0.5 for example).
7455 Corrected to 'c.kname'.
7456 [Phillip Porch <root@theporch.com>]
7458 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
7459 a STACK of email addresses from a certificate or request, these look
7460 in the subject name and the subject alternative name extensions and
7461 omit any duplicate addresses.
7464 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
7465 This makes DSA verification about 2 % faster.
7468 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
7469 (meaning that now 2^5 values will be precomputed, which is only 4 KB
7470 plus overhead for 1024 bit moduli).
7471 This makes exponentiations about 0.5 % faster for 1024 bit
7472 exponents (as measured by "openssl speed rsa2048").
7475 *) Rename memory handling macros to avoid conflicts with other
7477 Malloc => OPENSSL_malloc
7478 Malloc_locked => OPENSSL_malloc_locked
7479 Realloc => OPENSSL_realloc
7480 Free => OPENSSL_free
7483 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
7484 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
7487 *) CygWin32 support.
7488 [John Jarvie <jjarvie@newsguy.com>]
7490 *) The type-safe stack code has been rejigged. It is now only compiled
7491 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
7492 by default all type-specific stack functions are "#define"d back to
7493 standard stack functions. This results in more streamlined output
7494 but retains the type-safety checking possibilities of the original
7498 *) The STACK code has been cleaned up, and certain type declarations
7499 that didn't make a lot of sense have been brought in line. This has
7500 also involved a cleanup of sorts in safestack.h to more correctly
7501 map type-safe stack functions onto their plain stack counterparts.
7502 This work has also resulted in a variety of "const"ifications of
7503 lots of the code, especially "_cmp" operations which should normally
7504 be prototyped with "const" parameters anyway.
7507 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
7508 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
7509 (The PRNG state consists of two parts, the large pool 'state' and 'md',
7510 where all of 'md' is used each time the PRNG is used, but 'state'
7511 is used only indexed by a cyclic counter. As entropy may not be
7512 well distributed from the beginning, 'md' is important as a
7513 chaining variable. However, the output function chains only half
7514 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
7515 all of 'md', and seeding with STATE_SIZE dummy bytes will result
7516 in all of 'state' being rewritten, with the new values depending
7517 on virtually all of 'md'. This overcomes the 80 bit limitation.)
7520 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
7521 the handshake is continued after ssl_verify_cert_chain();
7522 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
7523 can lead to 'unexplainable' connection aborts later.
7524 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
7526 *) Major EVP API cipher revision.
7527 Add hooks for extra EVP features. This allows various cipher
7528 parameters to be set in the EVP interface. Support added for variable
7529 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
7530 setting of RC2 and RC5 parameters.
7532 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
7535 Remove lots of duplicated code from the EVP library. For example *every*
7536 cipher init() function handles the 'iv' in the same way according to the
7537 cipher mode. They also all do nothing if the 'key' parameter is NULL and
7538 for CFB and OFB modes they zero ctx->num.
7540 New functionality allows removal of S/MIME code RC2 hack.
7542 Most of the routines have the same form and so can be declared in terms
7545 By shifting this to the top level EVP_CipherInit() it can be removed from
7546 all individual ciphers. If the cipher wants to handle IVs or keys
7547 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
7550 Change lots of functions like EVP_EncryptUpdate() to now return a
7551 value: although software versions of the algorithms cannot fail
7552 any installed hardware versions can.
7555 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
7556 this option is set, tolerate broken clients that send the negotiated
7557 protocol version number instead of the requested protocol version
7561 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
7562 i.e. non-zero for export ciphersuites, zero otherwise.
7563 Previous versions had this flag inverted, inconsistent with
7564 rsa_tmp_cb (..._TMP_RSA_CB).
7565 [Bodo Moeller; problem reported by Amit Chopra]
7567 *) Add missing DSA library text string. Work around for some IIS
7568 key files with invalid SEQUENCE encoding.
7571 *) Add a document (doc/standards.txt) that list all kinds of standards
7572 and so on that are implemented in OpenSSL.
7575 *) Enhance c_rehash script. Old version would mishandle certificates
7576 with the same subject name hash and wouldn't handle CRLs at all.
7577 Added -fingerprint option to crl utility, to support new c_rehash
7581 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
7584 *) Fix for SSL server purpose checking. Server checking was
7585 rejecting certificates which had extended key usage present
7586 but no ssl client purpose.
7587 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
7589 *) Make PKCS#12 code work with no password. The PKCS#12 spec
7590 is a little unclear about how a blank password is handled.
7591 Since the password in encoded as a BMPString with terminating
7592 double NULL a zero length password would end up as just the
7593 double NULL. However no password at all is different and is
7594 handled differently in the PKCS#12 key generation code. NS
7595 treats a blank password as zero length. MSIE treats it as no
7596 password on export: but it will try both on import. We now do
7597 the same: PKCS12_parse() tries zero length and no password if
7598 the password is set to "" or NULL (NULL is now a valid password:
7599 it wasn't before) as does the pkcs12 application.
7602 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
7603 perror when PEM_read_bio_X509_REQ fails, the error message must
7604 be obtained from the error queue.
7607 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
7608 it in ERR_remove_state if appropriate, and change ERR_get_state
7609 accordingly to avoid race conditions (this is necessary because
7610 thread_hash is no longer constant once set).
7613 *) Bugfix for linux-elf makefile.one.
7616 *) RSA_get_default_method() will now cause a default
7617 RSA_METHOD to be chosen if one doesn't exist already.
7618 Previously this was only set during a call to RSA_new()
7619 or RSA_new_method(NULL) meaning it was possible for
7620 RSA_get_default_method() to return NULL.
7623 *) Added native name translation to the existing DSO code
7624 that will convert (if the flag to do so is set) filenames
7625 that are sufficiently small and have no path information
7626 into a canonical native form. Eg. "blah" converted to
7627 "libblah.so" or "blah.dll" etc.
7630 *) New function ERR_error_string_n(e, buf, len) which is like
7631 ERR_error_string(e, buf), but writes at most 'len' bytes
7632 including the 0 terminator. For ERR_error_string_n, 'buf'
7634 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
7636 *) CONF library reworked to become more general. A new CONF
7637 configuration file reader "class" is implemented as well as a
7638 new functions (NCONF_*, for "New CONF") to handle it. The now
7639 old CONF_* functions are still there, but are reimplemented to
7640 work in terms of the new functions. Also, a set of functions
7641 to handle the internal storage of the configuration data is
7642 provided to make it easier to write new configuration file
7643 reader "classes" (I can definitely see something reading a
7644 configuration file in XML format, for example), called _CONF_*,
7645 or "the configuration storage API"...
7647 The new configuration file reading functions are:
7649 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
7650 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
7652 NCONF_default, NCONF_WIN32
7654 NCONF_dump_fp, NCONF_dump_bio
7656 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
7657 NCONF_new creates a new CONF object. This works in the same way
7658 as other interfaces in OpenSSL, like the BIO interface.
7659 NCONF_dump_* dump the internal storage of the configuration file,
7660 which is useful for debugging. All other functions take the same
7661 arguments as the old CONF_* functions wth the exception of the
7662 first that must be a `CONF *' instead of a `LHASH *'.
7664 To make it easer to use the new classes with the old CONF_* functions,
7665 the function CONF_set_default_method is provided.
7668 *) Add '-tls1' option to 'openssl ciphers', which was already
7669 mentioned in the documentation but had not been implemented.
7670 (This option is not yet really useful because even the additional
7671 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
7674 *) Initial DSO code added into libcrypto for letting OpenSSL (and
7675 OpenSSL-based applications) load shared libraries and bind to
7676 them in a portable way.
7677 [Geoff Thorpe, with contributions from Richard Levitte]
7679 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
7681 *) Make sure _lrotl and _lrotr are only used with MSVC.
7683 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
7684 (the default implementation of RAND_status).
7686 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
7687 to '-clrext' (= clear extensions), as intended and documented.
7688 [Bodo Moeller; inconsistency pointed out by Michael Attili
7691 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
7692 was larger than the MD block size.
7693 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
7695 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
7696 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
7697 using the passed key: if the passed key was a private key the result
7698 of X509_print(), for example, would be to print out all the private key
7702 *) des_quad_cksum() byte order bug fix.
7703 [Ulf Möller, using the problem description in krb4-0.9.7, where
7704 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
7706 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
7708 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
7710 *) For easily testing in shell scripts whether some command
7711 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
7712 returns with exit code 0 iff no command of the given name is available.
7713 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
7714 the output goes to stdout and nothing is printed to stderr.
7715 Additional arguments are always ignored.
7717 Since for each cipher there is a command of the same name,
7718 the 'no-cipher' compilation switches can be tested this way.
7720 ('openssl no-XXX' is not able to detect pseudo-commands such
7721 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
7724 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
7727 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
7728 is set; it will be thrown away anyway because each handshake creates
7730 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
7731 to parameters -- in previous versions (since OpenSSL 0.9.3) the
7732 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
7733 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7736 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7737 'Q' and 'R' lose their special meanings (quit/renegotiate).
7738 This is part of what -quiet does; unlike -quiet, -ign_eof
7739 does not suppress any output.
7742 *) Add compatibility options to the purpose and trust code. The
7743 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7744 accepts a certificate or CA, this was the previous behaviour,
7745 with all the associated security issues.
7747 X509_TRUST_COMPAT is the old trust behaviour: only and
7748 automatically trust self signed roots in certificate store. A
7749 new trust setting X509_TRUST_DEFAULT is used to specify that
7750 a purpose has no associated trust setting and it should instead
7751 use the value in the default purpose.
7754 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7755 and fix a memory leak.
7758 *) In util/mkerr.pl (which implements 'make errors'), preserve
7759 reason strings from the previous version of the .c file, as
7760 the default to have only downcase letters (and digits) in
7761 automatically generated reasons codes is not always appropriate.
7764 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7765 using strerror. Previously, ERR_reason_error_string() returned
7766 library names as reason strings for SYSerr; but SYSerr is a special
7767 case where small numbers are errno values, not library numbers.
7770 *) Add '-dsaparam' option to 'openssl dhparam' application. This
7771 converts DSA parameters into DH parameters. (When creating parameters,
7772 DSA_generate_parameters is used.)
7775 *) Include 'length' (recommended exponent length) in C code generated
7776 by 'openssl dhparam -C'.
7779 *) The second argument to set_label in perlasm was already being used
7780 so couldn't be used as a "file scope" flag. Moved to third argument
7784 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
7785 instead of RAND_bytes for encryption IVs and salts.
7788 *) Include RAND_status() into RAND_METHOD instead of implementing
7789 it only for md_rand.c Otherwise replacing the PRNG by calling
7790 RAND_set_rand_method would be impossible.
7793 *) Don't let DSA_generate_key() enter an infinite loop if the random
7794 number generation fails.
7797 *) New 'rand' application for creating pseudo-random output.
7800 *) Added configuration support for Linux/IA64
7801 [Rolf Haberrecker <rolf@suse.de>]
7803 *) Assembler module support for Mingw32.
7806 *) Shared library support for HPUX (in shlib/).
7807 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
7809 *) Shared library support for Solaris gcc.
7810 [Lutz Behnke <behnke@trustcenter.de>]
7812 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
7814 *) PKCS7_encrypt() was adding text MIME headers twice because they
7815 were added manually and by SMIME_crlf_copy().
7818 *) In bntest.c don't call BN_rand with zero bits argument.
7819 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
7821 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
7822 case was implemented. This caused BN_div_recp() to fail occasionally.
7825 *) Add an optional second argument to the set_label() in the perl
7826 assembly language builder. If this argument exists and is set
7827 to 1 it signals that the assembler should use a symbol whose
7828 scope is the entire file, not just the current function. This
7829 is needed with MASM which uses the format label:: for this scope.
7830 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
7832 *) Change the ASN1 types so they are typedefs by default. Before
7833 almost all types were #define'd to ASN1_STRING which was causing
7834 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
7838 *) Change names of new functions to the new get1/get0 naming
7839 convention: After 'get1', the caller owns a reference count
7840 and has to call ..._free; 'get0' returns a pointer to some
7841 data structure without incrementing reference counters.
7842 (Some of the existing 'get' functions increment a reference
7843 counter, some don't.)
7844 Similarly, 'set1' and 'add1' functions increase reference
7845 counters or duplicate objects.
7848 *) Allow for the possibility of temp RSA key generation failure:
7849 the code used to assume it always worked and crashed on failure.
7852 *) Fix potential buffer overrun problem in BIO_printf().
7853 [Ulf Möller, using public domain code by Patrick Powell; problem
7854 pointed out by David Sacerdote <das33@cornell.edu>]
7856 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
7857 RAND_egd() and RAND_status(). In the command line application,
7858 the EGD socket can be specified like a seed file using RANDFILE
7862 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
7863 Some CAs (e.g. Verisign) distribute certificates in this form.
7866 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
7867 list to exclude them. This means that no special compilation option
7868 is needed to use anonymous DH: it just needs to be included in the
7872 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
7873 EVP_MD_type. The old functionality is available in a new macro called
7874 EVP_MD_md(). Change code that uses it and update docs.
7877 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
7878 where the 'void *' argument is replaced by a function pointer argument.
7879 Previously 'void *' was abused to point to functions, which works on
7880 many platforms, but is not correct. As these functions are usually
7881 called by macros defined in OpenSSL header files, most source code
7882 should work without changes.
7885 *) <openssl/opensslconf.h> (which is created by Configure) now contains
7886 sections with information on -D... compiler switches used for
7887 compiling the library so that applications can see them. To enable
7888 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
7889 must be defined. E.g.,
7890 #define OPENSSL_ALGORITHM_DEFINES
7891 #include <openssl/opensslconf.h>
7892 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
7893 [Richard Levitte, Ulf and Bodo Möller]
7895 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
7899 *) Change the 'other' type in certificate aux info to a STACK_OF
7900 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
7901 the required ASN1 format: arbitrary types determined by an OID.
7904 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
7905 argument to 'req'. This is not because the function is newer or
7906 better than others it just uses the work 'NEW' in the certificate
7907 request header lines. Some software needs this.
7910 *) Reorganise password command line arguments: now passwords can be
7911 obtained from various sources. Delete the PEM_cb function and make
7912 it the default behaviour: i.e. if the callback is NULL and the
7913 usrdata argument is not NULL interpret it as a null terminated pass
7914 phrase. If usrdata and the callback are NULL then the pass phrase
7915 is prompted for as usual.
7918 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
7919 the support is automatically enabled. The resulting binaries will
7920 autodetect the card and use it if present.
7921 [Ben Laurie and Compaq Inc.]
7923 *) Work around for Netscape hang bug. This sends certificate request
7924 and server done in one record. Since this is perfectly legal in the
7925 SSL/TLS protocol it isn't a "bug" option and is on by default. See
7926 the bugs/SSLv3 entry for more info.
7929 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
7932 *) Add -rand argument to smime and pkcs12 applications and read/write
7936 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
7939 *) Add command line password options to the remaining applications.
7942 *) Bug fix for BN_div_recp() for numerators with an even number of
7946 *) More tests in bntest.c, and changed test_bn output.
7949 *) ./config recognizes MacOS X now.
7952 *) Bug fix for BN_div() when the first words of num and divsor are
7953 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
7956 *) Add support for various broken PKCS#8 formats, and command line
7957 options to produce them.
7960 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
7961 get temporary BIGNUMs from a BN_CTX.
7964 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
7968 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
7969 include a #define from the old name to the new. The original intent
7970 was that statically linked binaries could for example just call
7971 SSLeay_add_all_ciphers() to just add ciphers to the table and not
7972 link with digests. This never worked becayse SSLeay_add_all_digests()
7973 and SSLeay_add_all_ciphers() were in the same source file so calling
7974 one would link with the other. They are now in separate source files.
7977 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
7980 *) Use a less unusual form of the Miller-Rabin primality test (it used
7981 a binary algorithm for exponentiation integrated into the Miller-Rabin
7982 loop, our standard modexp algorithms are faster).
7985 *) Support for the EBCDIC character set completed.
7986 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
7988 *) Source code cleanups: use const where appropriate, eliminate casts,
7989 use void * instead of char * in lhash.
7992 *) Bugfix: ssl3_send_server_key_exchange was not restartable
7993 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
7994 this the server could overwrite ephemeral keys that the client
7998 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
7999 using 50 iterations of the Rabin-Miller test.
8001 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
8002 iterations of the Rabin-Miller test as required by the appendix
8003 to FIPS PUB 186[-1]) instead of DSA_is_prime.
8004 As BN_is_prime_fasttest includes trial division, DSA parameter
8005 generation becomes much faster.
8007 This implies a change for the callback functions in DSA_is_prime
8008 and DSA_generate_parameters: The callback function is called once
8009 for each positive witness in the Rabin-Miller test, not just
8010 occasionally in the inner loop; and the parameters to the
8011 callback function now provide an iteration count for the outer
8012 loop rather than for the current invocation of the inner loop.
8013 DSA_generate_parameters additionally can call the callback
8014 function with an 'iteration count' of -1, meaning that a
8015 candidate has passed the trial division test (when q is generated
8016 from an application-provided seed, trial division is skipped).
8019 *) New function BN_is_prime_fasttest that optionally does trial
8020 division before starting the Rabin-Miller test and has
8021 an additional BN_CTX * argument (whereas BN_is_prime always
8022 has to allocate at least one BN_CTX).
8023 'callback(1, -1, cb_arg)' is called when a number has passed the
8024 trial division stage.
8027 *) Fix for bug in CRL encoding. The validity dates weren't being handled
8031 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
8034 *) New function BN_pseudo_rand().
8037 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
8038 bignum version of BN_from_montgomery() with the working code from
8039 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
8043 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
8044 made it impossible to use the same SSL_SESSION data structure in
8045 SSL2 clients in multiple threads.
8048 *) The return value of RAND_load_file() no longer counts bytes obtained
8049 by stat(). RAND_load_file(..., -1) is new and uses the complete file
8050 to seed the PRNG (previously an explicit byte count was required).
8051 [Ulf Möller, Bodo Möller]
8053 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
8054 used (char *) instead of (void *) and had casts all over the place.
8057 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
8060 *) Retain source code compatibility for BN_prime_checks macro:
8061 BN_is_prime(..., BN_prime_checks, ...) now uses
8062 BN_prime_checks_for_size to determine the appropriate number of
8063 Rabin-Miller iterations.
8066 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
8067 DH_CHECK_P_NOT_SAFE_PRIME.
8068 (Check if this is true? OpenPGP calls them "strong".)
8071 *) Merge the functionality of "dh" and "gendh" programs into a new program
8072 "dhparam". The old programs are retained for now but will handle DH keys
8073 (instead of parameters) in future.
8076 *) Make the ciphers, s_server and s_client programs check the return values
8077 when a new cipher list is set.
8080 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
8081 ciphers. Before when the 56bit ciphers were enabled the sorting was
8084 The syntax for the cipher sorting has been extended to support sorting by
8085 cipher-strength (using the strength_bits hard coded in the tables).
8086 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
8088 Fix a bug in the cipher-command parser: when supplying a cipher command
8089 string with an "undefined" symbol (neither command nor alphanumeric
8090 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
8091 an error is flagged.
8093 Due to the strength-sorting extension, the code of the
8094 ssl_create_cipher_list() function was completely rearranged. I hope that
8095 the readability was also increased :-)
8096 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8098 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
8099 for the first serial number and places 2 in the serial number file. This
8100 avoids problems when the root CA is created with serial number zero and
8101 the first user certificate has the same issuer name and serial number
8105 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
8106 the new code. Add documentation for this stuff.
8109 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
8110 X509_*() to X509at_*() on the grounds that they don't handle X509
8111 structures and behave in an analagous way to the X509v3 functions:
8112 they shouldn't be called directly but wrapper functions should be used
8115 So we also now have some wrapper functions that call the X509at functions
8116 when passed certificate requests. (TO DO: similar things can be done with
8117 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
8118 things. Some of these need some d2i or i2d and print functionality
8119 because they handle more complex structures.)
8122 *) Add missing #ifndefs that caused missing symbols when building libssl
8123 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
8124 NO_RSA in ssl/s2*.c.
8125 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
8127 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
8128 has a return value which indicates the quality of the random data
8129 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
8130 error queue. New function RAND_pseudo_bytes() generates output that is
8131 guaranteed to be unique but not unpredictable. RAND_add is like
8132 RAND_seed, but takes an extra argument for an entropy estimate
8133 (RAND_seed always assumes full entropy).
8136 *) Do more iterations of Rabin-Miller probable prime test (specifically,
8137 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
8138 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
8139 in crypto/bn/bn_prime.c for the complete table). This guarantees a
8140 false-positive rate of at most 2^-80 for random input.
8143 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
8146 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
8147 in the 0.9.5 release), this returns the chain
8148 from an X509_CTX structure with a dup of the stack and all
8149 the X509 reference counts upped: so the stack will exist
8150 after X509_CTX_cleanup() has been called. Modify pkcs12.c
8153 Also make SSL_SESSION_print() print out the verify return
8157 *) Add manpage for the pkcs12 command. Also change the default
8158 behaviour so MAC iteration counts are used unless the new
8159 -nomaciter option is used. This improves file security and
8160 only older versions of MSIE (4.0 for example) need it.
8163 *) Honor the no-xxx Configure options when creating .DEF files.
8166 *) Add PKCS#10 attributes to field table: challengePassword,
8167 unstructuredName and unstructuredAddress. These are taken from
8168 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
8169 international characters are used.
8171 More changes to X509_ATTRIBUTE code: allow the setting of types
8172 based on strings. Remove the 'loc' parameter when adding
8173 attributes because these will be a SET OF encoding which is sorted
8177 *) Initial changes to the 'req' utility to allow request generation
8178 automation. This will allow an application to just generate a template
8179 file containing all the field values and have req construct the
8182 Initial support for X509_ATTRIBUTE handling. Stacks of these are
8183 used all over the place including certificate requests and PKCS#7
8184 structures. They are currently handled manually where necessary with
8185 some primitive wrappers for PKCS#7. The new functions behave in a
8186 manner analogous to the X509 extension functions: they allow
8187 attributes to be looked up by NID and added.
8189 Later something similar to the X509V3 code would be desirable to
8190 automatically handle the encoding, decoding and printing of the
8191 more complex types. The string types like challengePassword can
8192 be handled by the string table functions.
8194 Also modified the multi byte string table handling. Now there is
8195 a 'global mask' which masks out certain types. The table itself
8196 can use the flag STABLE_NO_MASK to ignore the mask setting: this
8197 is useful when for example there is only one permissible type
8198 (as in countryName) and using the mask might result in no valid
8202 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
8203 SSL_get_peer_finished to allow applications to obtain the latest
8204 Finished messages sent to the peer or expected from the peer,
8205 respectively. (SSL_get_peer_finished is usually the Finished message
8206 actually received from the peer, otherwise the protocol will be aborted.)
8208 As the Finished message are message digests of the complete handshake
8209 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
8210 be used for external authentication procedures when the authentication
8211 provided by SSL/TLS is not desired or is not enough.
8214 *) Enhanced support for Alpha Linux is added. Now ./config checks if
8215 the host supports BWX extension and if Compaq C is present on the
8216 $PATH. Just exploiting of the BWX extension results in 20-30%
8217 performance kick for some algorithms, e.g. DES and RC4 to mention
8218 a couple. Compaq C in turn generates ~20% faster code for MD5 and
8222 *) Add support for MS "fast SGC". This is arguably a violation of the
8223 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
8224 weak crypto and after checking the certificate is SGC a second one
8225 with strong crypto. MS SGC stops the first handshake after receiving
8226 the server certificate message and sends a second client hello. Since
8227 a server will typically do all the time consuming operations before
8228 expecting any further messages from the client (server key exchange
8229 is the most expensive) there is little difference between the two.
8231 To get OpenSSL to support MS SGC we have to permit a second client
8232 hello message after we have sent server done. In addition we have to
8233 reset the MAC if we do get this second client hello.
8236 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
8237 if a DER encoded private key is RSA or DSA traditional format. Changed
8238 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
8239 format DER encoded private key. Newer code should use PKCS#8 format which
8240 has the key type encoded in the ASN1 structure. Added DER private key
8241 support to pkcs8 application.
8244 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
8245 ciphersuites has been selected (as required by the SSL 3/TLS 1
8246 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
8247 is set, we interpret this as a request to violate the specification
8248 (the worst that can happen is a handshake failure, and 'correct'
8249 behaviour would result in a handshake failure anyway).
8252 *) In SSL_CTX_add_session, take into account that there might be multiple
8253 SSL_SESSION structures with the same session ID (e.g. when two threads
8254 concurrently obtain them from an external cache).
8255 The internal cache can handle only one SSL_SESSION with a given ID,
8256 so if there's a conflict, we now throw out the old one to achieve
8260 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
8261 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
8262 some routines that use cipher OIDs: some ciphers do not have OIDs
8263 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
8267 *) Simplify the trust setting structure and code. Now we just have
8268 two sequences of OIDs for trusted and rejected settings. These will
8269 typically have values the same as the extended key usage extension
8270 and any application specific purposes.
8272 The trust checking code now has a default behaviour: it will just
8273 check for an object with the same NID as the passed id. Functions can
8274 be provided to override either the default behaviour or the behaviour
8275 for a given id. SSL client, server and email already have functions
8276 in place for compatibility: they check the NID and also return "trusted"
8277 if the certificate is self signed.
8280 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
8281 traditional format into an EVP_PKEY structure.
8284 *) Add a password callback function PEM_cb() which either prompts for
8285 a password if usr_data is NULL or otherwise assumes it is a null
8286 terminated password. Allow passwords to be passed on command line
8287 environment or config files in a few more utilities.
8290 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
8291 keys. Add some short names for PKCS#8 PBE algorithms and allow them
8292 to be specified on the command line for the pkcs8 and pkcs12 utilities.
8293 Update documentation.
8296 *) Support for ASN1 "NULL" type. This could be handled before by using
8297 ASN1_TYPE but there wasn't any function that would try to read a NULL
8298 and produce an error if it couldn't. For compatibility we also have
8299 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
8300 don't allocate anything because they don't need to.
8303 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
8305 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
8307 *) Rebuild of the memory allocation routines used by OpenSSL code and
8308 possibly others as well. The purpose is to make an interface that
8309 provide hooks so anyone can build a separate set of allocation and
8310 deallocation routines to be used by OpenSSL, for example memory
8311 pool implementations, or something else, which was previously hard
8312 since Malloc(), Realloc() and Free() were defined as macros having
8313 the values malloc, realloc and free, respectively (except for Win32
8314 compilations). The same is provided for memory debugging code.
8315 OpenSSL already comes with functionality to find memory leaks, but
8316 this gives people a chance to debug other memory problems.
8318 With these changes, a new set of functions and macros have appeared:
8320 CRYPTO_set_mem_debug_functions() [F]
8321 CRYPTO_get_mem_debug_functions() [F]
8322 CRYPTO_dbg_set_options() [F]
8323 CRYPTO_dbg_get_options() [F]
8324 CRYPTO_malloc_debug_init() [M]
8326 The memory debug functions are NULL by default, unless the library
8327 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
8328 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
8329 gives the standard debugging functions that come with OpenSSL) or
8330 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
8331 provided by the library user) must be used. When the standard
8332 debugging functions are used, CRYPTO_dbg_set_options can be used to
8333 request additional information:
8334 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
8335 the CRYPTO_MDEBUG_xxx macro when compiling the library.
8337 Also, things like CRYPTO_set_mem_functions will always give the
8338 expected result (the new set of functions is used for allocation
8339 and deallocation) at all times, regardless of platform and compiler
8342 To finish it up, some functions that were never use in any other
8343 way than through macros have a new API and new semantic:
8346 CRYPTO_dbg_realloc()
8349 All macros of value have retained their old syntax.
8350 [Richard Levitte and Bodo Moeller]
8352 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
8353 ordering of SMIMECapabilities wasn't in "strength order" and there
8354 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
8358 *) Some ASN1 types with illegal zero length encoding (INTEGER,
8359 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
8360 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
8362 *) Merge in my S/MIME library for OpenSSL. This provides a simple
8363 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
8364 functionality to handle multipart/signed properly) and a utility
8365 called 'smime' to call all this stuff. This is based on code I
8366 originally wrote for Celo who have kindly allowed it to be
8367 included in OpenSSL.
8370 *) Add variants des_set_key_checked and des_set_key_unchecked of
8371 des_set_key (aka des_key_sched). Global variable des_check_key
8372 decides which of these is called by des_set_key; this way
8373 des_check_key behaves as it always did, but applications and
8374 the library itself, which was buggy for des_check_key == 1,
8375 have a cleaner way to pick the version they need.
8378 *) New function PKCS12_newpass() which changes the password of a
8382 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
8383 dynamic mix. In both cases the ids can be used as an index into the
8384 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
8385 functions so they accept a list of the field values and the
8386 application doesn't need to directly manipulate the X509_TRUST
8390 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
8394 *) Modify the way the V3 extension code looks up extensions. This now
8395 works in a similar way to the object code: we have some "standard"
8396 extensions in a static table which is searched with OBJ_bsearch()
8397 and the application can add dynamic ones if needed. The file
8398 crypto/x509v3/ext_dat.h now has the info: this file needs to be
8399 updated whenever a new extension is added to the core code and kept
8400 in ext_nid order. There is a simple program 'tabtest.c' which checks
8401 this. New extensions are not added too often so this file can readily
8402 be maintained manually.
8404 There are two big advantages in doing things this way. The extensions
8405 can be looked up immediately and no longer need to be "added" using
8406 X509V3_add_standard_extensions(): this function now does nothing.
8407 [Side note: I get *lots* of email saying the extension code doesn't
8408 work because people forget to call this function]
8409 Also no dynamic allocation is done unless new extensions are added:
8410 so if we don't add custom extensions there is no need to call
8411 X509V3_EXT_cleanup().
8414 *) Modify enc utility's salting as follows: make salting the default. Add a
8415 magic header, so unsalted files fail gracefully instead of just decrypting
8416 to garbage. This is because not salting is a big security hole, so people
8417 should be discouraged from doing it.
8420 *) Fixes and enhancements to the 'x509' utility. It allowed a message
8421 digest to be passed on the command line but it only used this
8422 parameter when signing a certificate. Modified so all relevant
8423 operations are affected by the digest parameter including the
8424 -fingerprint and -x509toreq options. Also -x509toreq choked if a
8425 DSA key was used because it didn't fix the digest.
8428 *) Initial certificate chain verify code. Currently tests the untrusted
8429 certificates for consistency with the verify purpose (which is set
8430 when the X509_STORE_CTX structure is set up) and checks the pathlength.
8432 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
8433 this is because it will reject chains with invalid extensions whereas
8434 every previous version of OpenSSL and SSLeay made no checks at all.
8436 Trust code: checks the root CA for the relevant trust settings. Trust
8437 settings have an initial value consistent with the verify purpose: e.g.
8438 if the verify purpose is for SSL client use it expects the CA to be
8439 trusted for SSL client use. However the default value can be changed to
8440 permit custom trust settings: one example of this would be to only trust
8441 certificates from a specific "secure" set of CAs.
8443 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
8444 which should be used for version portability: especially since the
8445 verify structure is likely to change more often now.
8447 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
8448 to set them. If not set then assume SSL clients will verify SSL servers
8451 Two new options to the verify program: -untrusted allows a set of
8452 untrusted certificates to be passed in and -purpose which sets the
8453 intended purpose of the certificate. If a purpose is set then the
8454 new chain verify code is used to check extension consistency.
8457 *) Support for the authority information access extension.
8460 *) Modify RSA and DSA PEM read routines to transparently handle
8461 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
8462 public keys in a format compatible with certificate
8463 SubjectPublicKeyInfo structures. Unfortunately there were already
8464 functions called *_PublicKey_* which used various odd formats so
8465 these are retained for compatibility: however the DSA variants were
8466 never in a public release so they have been deleted. Changed dsa/rsa
8467 utilities to handle the new format: note no releases ever handled public
8468 keys so we should be OK.
8470 The primary motivation for this change is to avoid the same fiasco
8471 that dogs private keys: there are several incompatible private key
8472 formats some of which are standard and some OpenSSL specific and
8473 require various evil hacks to allow partial transparent handling and
8474 even then it doesn't work with DER formats. Given the option anything
8475 other than PKCS#8 should be dumped: but the other formats have to
8476 stay in the name of compatibility.
8478 With public keys and the benefit of hindsight one standard format
8479 is used which works with EVP_PKEY, RSA or DSA structures: though
8480 it clearly returns an error if you try to read the wrong kind of key.
8482 Added a -pubkey option to the 'x509' utility to output the public key.
8483 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
8484 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
8485 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
8486 that do the same as the EVP_PKEY_assign_*() except they up the
8487 reference count of the added key (they don't "swallow" the
8491 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
8492 CRLs would fail if the file contained no certificates or no CRLs:
8493 added a new function to read in both types and return the number
8494 read: this means that if none are read it will be an error. The
8495 DER versions of the certificate and CRL reader would always fail
8496 because it isn't possible to mix certificates and CRLs in DER format
8497 without choking one or the other routine. Changed this to just read
8498 a certificate: this is the best we can do. Also modified the code
8499 in apps/verify.c to take notice of return codes: it was previously
8500 attempting to read in certificates from NULL pointers and ignoring
8501 any errors: this is one reason why the cert and CRL reader seemed
8502 to work. It doesn't check return codes from the default certificate
8503 routines: these may well fail if the certificates aren't installed.
8506 *) Code to support otherName option in GeneralName.
8509 *) First update to verify code. Change the verify utility
8510 so it warns if it is passed a self signed certificate:
8511 for consistency with the normal behaviour. X509_verify
8512 has been modified to it will now verify a self signed
8513 certificate if *exactly* the same certificate appears
8514 in the store: it was previously impossible to trust a
8515 single self signed certificate. This means that:
8516 openssl verify ss.pem
8517 now gives a warning about a self signed certificate but
8518 openssl verify -CAfile ss.pem ss.pem
8522 *) For servers, store verify_result in SSL_SESSION data structure
8523 (and add it to external session representation).
8524 This is needed when client certificate verifications fails,
8525 but an application-provided verification callback (set by
8526 SSL_CTX_set_cert_verify_callback) allows accepting the session
8527 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
8528 but returns 1): When the session is reused, we have to set
8529 ssl->verify_result to the appropriate error code to avoid
8531 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
8533 *) Fix a bug in the new PKCS#7 code: it didn't consider the
8534 case in PKCS7_dataInit() where the signed PKCS7 structure
8535 didn't contain any existing data because it was being created.
8536 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
8538 *) Add a salt to the key derivation routines in enc.c. This
8539 forms the first 8 bytes of the encrypted file. Also add a
8540 -S option to allow a salt to be input on the command line.
8543 *) New function X509_cmp(). Oddly enough there wasn't a function
8544 to compare two certificates. We do this by working out the SHA1
8545 hash and comparing that. X509_cmp() will be needed by the trust
8549 *) SSL_get1_session() is like SSL_get_session(), but increments
8550 the reference count in the SSL_SESSION returned.
8551 [Geoff Thorpe <geoff@eu.c2.net>]
8553 *) Fix for 'req': it was adding a null to request attributes.
8554 Also change the X509_LOOKUP and X509_INFO code to handle
8555 certificate auxiliary information.
8558 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
8562 *) Add the possibility to add extra information to the memory leak
8563 detecting output, to form tracebacks, showing from where each
8564 allocation was originated: CRYPTO_push_info("constant string") adds
8565 the string plus current file name and line number to a per-thread
8566 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
8567 is like calling CYRPTO_pop_info() until the stack is empty.
8568 Also updated memory leak detection code to be multi-thread-safe.
8571 *) Add options -text and -noout to pkcs7 utility and delete the
8572 encryption options which never did anything. Update docs.
8575 *) Add options to some of the utilities to allow the pass phrase
8576 to be included on either the command line (not recommended on
8577 OSes like Unix) or read from the environment. Update the
8578 manpages and fix a few bugs.
8581 *) Add a few manpages for some of the openssl commands.
8584 *) Fix the -revoke option in ca. It was freeing up memory twice,
8585 leaking and not finding already revoked certificates.
8588 *) Extensive changes to support certificate auxiliary information.
8589 This involves the use of X509_CERT_AUX structure and X509_AUX
8590 functions. An X509_AUX function such as PEM_read_X509_AUX()
8591 can still read in a certificate file in the usual way but it
8592 will also read in any additional "auxiliary information". By
8593 doing things this way a fair degree of compatibility can be
8594 retained: existing certificates can have this information added
8595 using the new 'x509' options.
8597 Current auxiliary information includes an "alias" and some trust
8598 settings. The trust settings will ultimately be used in enhanced
8599 certificate chain verification routines: currently a certificate
8600 can only be trusted if it is self signed and then it is trusted
8604 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
8605 The problem was that one of the replacement routines had not been working
8606 since SSLeay releases. For now the offending routine has been replaced
8607 with non-optimised assembler. Even so, this now gives around 95%
8608 performance improvement for 1024 bit RSA signs.
8611 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
8612 handling. Most clients have the effective key size in bits equal to
8613 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
8614 A few however don't do this and instead use the size of the decrypted key
8615 to determine the RC2 key length and the AlgorithmIdentifier to determine
8616 the effective key length. In this case the effective key length can still
8617 be 40 bits but the key length can be 168 bits for example. This is fixed
8618 by manually forcing an RC2 key into the EVP_PKEY structure because the
8619 EVP code can't currently handle unusual RC2 key sizes: it always assumes
8620 the key length and effective key length are equal.
8623 *) Add a bunch of functions that should simplify the creation of
8624 X509_NAME structures. Now you should be able to do:
8625 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
8626 and have it automatically work out the correct field type and fill in
8627 the structures. The more adventurous can try:
8628 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
8629 and it will (hopefully) work out the correct multibyte encoding.
8632 *) Change the 'req' utility to use the new field handling and multibyte
8633 copy routines. Before the DN field creation was handled in an ad hoc
8634 way in req, ca, and x509 which was rather broken and didn't support
8635 BMPStrings or UTF8Strings. Since some software doesn't implement
8636 BMPStrings or UTF8Strings yet, they can be enabled using the config file
8637 using the dirstring_type option. See the new comment in the default
8638 openssl.cnf for more info.
8641 *) Make crypto/rand/md_rand.c more robust:
8642 - Assure unique random numbers after fork().
8643 - Make sure that concurrent threads access the global counter and
8644 md serializably so that we never lose entropy in them
8645 or use exactly the same state in multiple threads.
8646 Access to the large state is not always serializable because
8647 the additional locking could be a performance killer, and
8648 md should be large enough anyway.
8651 *) New file apps/app_rand.c with commonly needed functionality
8652 for handling the random seed file.
8654 Use the random seed file in some applications that previously did not:
8656 dsaparam -genkey (which also ignored its '-rand' option),
8659 x509 (when signing).
8660 Except on systems with /dev/urandom, it is crucial to have a random
8661 seed file at least for key creation, DSA signing, and for DH exchanges;
8662 for RSA signatures we could do without one.
8664 gendh and gendsa (unlike genrsa) used to read only the first byte
8665 of each file listed in the '-rand' option. The function as previously
8666 found in genrsa is now in app_rand.c and is used by all programs
8667 that support '-rand'.
8670 *) In RAND_write_file, use mode 0600 for creating files;
8671 don't just chmod when it may be too late.
8674 *) Report an error from X509_STORE_load_locations
8675 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
8678 *) New function ASN1_mbstring_copy() this copies a string in either
8679 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
8680 into an ASN1_STRING type. A mask of permissible types is passed
8681 and it chooses the "minimal" type to use or an error if not type
8685 *) Add function equivalents to the various macros in asn1.h. The old
8686 macros are retained with an M_ prefix. Code inside the library can
8687 use the M_ macros. External code (including the openssl utility)
8688 should *NOT* in order to be "shared library friendly".
8691 *) Add various functions that can check a certificate's extensions
8692 to see if it usable for various purposes such as SSL client,
8693 server or S/MIME and CAs of these types. This is currently
8694 VERY EXPERIMENTAL but will ultimately be used for certificate chain
8695 verification. Also added a -purpose flag to x509 utility to
8696 print out all the purposes.
8699 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
8703 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
8704 for, obtain and decode and extension and obtain its critical flag.
8705 This allows all the necessary extension code to be handled in a
8706 single function call.
8709 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
8710 platforms. See crypto/rc4/rc4_enc.c for further details.
8713 *) New -noout option to asn1parse. This causes no output to be produced
8714 its main use is when combined with -strparse and -out to extract data
8715 from a file (which may not be in ASN.1 format).
8718 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
8719 when producing the local key id.
8720 [Richard Levitte <levitte@stacken.kth.se>]
8722 *) New option -dhparam in s_server. This allows a DH parameter file to be
8723 stated explicitly. If it is not stated then it tries the first server
8724 certificate file. The previous behaviour hard coded the filename
8728 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
8729 a public key to be input or output. For example:
8730 openssl rsa -in key.pem -pubout -out pubkey.pem
8731 Also added necessary DSA public key functions to handle this.
8734 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
8735 in the message. This was handled by allowing
8736 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8737 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8739 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8740 to the end of the strings whereas this didn't. This would cause problems
8741 if strings read with d2i_ASN1_bytes() were later modified.
8742 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8744 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8745 data and it contains EOF it will end up returning an error. This is
8746 caused by input 46 bytes long. The cause is due to the way base64
8747 BIOs find the start of base64 encoded data. They do this by trying a
8748 trial decode on each line until they find one that works. When they
8749 do a flag is set and it starts again knowing it can pass all the
8750 data directly through the decoder. Unfortunately it doesn't reset
8751 the context it uses. This means that if EOF is reached an attempt
8752 is made to pass two EOFs through the context and this causes the
8753 resulting error. This can also cause other problems as well. As is
8754 usual with these problems it takes *ages* to find and the fix is
8755 trivial: move one line.
8756 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8758 *) Ugly workaround to get s_client and s_server working under Windows. The
8759 old code wouldn't work because it needed to select() on sockets and the
8760 tty (for keypresses and to see if data could be written). Win32 only
8761 supports select() on sockets so we select() with a 1s timeout on the
8762 sockets and then see if any characters are waiting to be read, if none
8763 are present then we retry, we also assume we can always write data to
8764 the tty. This isn't nice because the code then blocks until we've
8765 received a complete line of data and it is effectively polling the
8766 keyboard at 1s intervals: however it's quite a bit better than not
8767 working at all :-) A dedicated Windows application might handle this
8768 with an event loop for example.
8771 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
8772 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
8773 will be called when RSA_sign() and RSA_verify() are used. This is useful
8774 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
8775 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
8776 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
8777 This necessitated the support of an extra signature type NID_md5_sha1
8778 for SSL signatures and modifications to the SSL library to use it instead
8779 of calling RSA_public_decrypt() and RSA_private_encrypt().
8782 *) Add new -verify -CAfile and -CApath options to the crl program, these
8783 will lookup a CRL issuers certificate and verify the signature in a
8784 similar way to the verify program. Tidy up the crl program so it
8785 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
8786 less strict. It will now permit CRL extensions even if it is not
8787 a V2 CRL: this will allow it to tolerate some broken CRLs.
8790 *) Initialize all non-automatic variables each time one of the openssl
8791 sub-programs is started (this is necessary as they may be started
8792 multiple times from the "OpenSSL>" prompt).
8793 [Lennart Bang, Bodo Moeller]
8795 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
8796 removing all other RSA functionality (this is what NO_RSA does). This
8797 is so (for example) those in the US can disable those operations covered
8798 by the RSA patent while allowing storage and parsing of RSA keys and RSA
8802 *) Non-copying interface to BIO pairs.
8803 (still largely untested)
8806 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
8807 ASCII string. This was handled independently in various places before.
8810 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
8811 UTF8 strings a character at a time.
8814 *) Use client_version from client hello to select the protocol
8815 (s23_srvr.c) and for RSA client key exchange verification
8816 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
8819 *) Add various utility functions to handle SPKACs, these were previously
8820 handled by poking round in the structure internals. Added new function
8821 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
8822 print, verify and generate SPKACs. Based on an original idea from
8823 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
8826 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
8829 *) Allow the config file extension section to be overwritten on the
8830 command line. Based on an original idea from Massimiliano Pala
8831 <madwolf@comune.modena.it>. The new option is called -extensions
8832 and can be applied to ca, req and x509. Also -reqexts to override
8833 the request extensions in req and -crlexts to override the crl extensions
8837 *) Add new feature to the SPKAC handling in ca. Now you can include
8838 the same field multiple times by preceding it by "XXXX." for example:
8841 this is the same syntax as used in the req config file.
8844 *) Allow certificate extensions to be added to certificate requests. These
8845 are specified in a 'req_extensions' option of the req section of the
8846 config file. They can be printed out with the -text option to req but
8847 are otherwise ignored at present.
8850 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
8851 data read consists of only the final block it would not decrypted because
8852 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
8853 A misplaced 'break' also meant the decrypted final block might not be
8854 copied until the next read.
8857 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
8858 a few extra parameters to the DH structure: these will be useful if
8859 for example we want the value of 'q' or implement X9.42 DH.
8862 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
8863 provides hooks that allow the default DSA functions or functions on a
8864 "per key" basis to be replaced. This allows hardware acceleration and
8865 hardware key storage to be handled without major modification to the
8866 library. Also added low level modexp hooks and CRYPTO_EX structure and
8867 associated functions.
8870 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
8871 as "read only": it can't be written to and the buffer it points to will
8872 not be freed. Reading from a read only BIO is much more efficient than
8873 a normal memory BIO. This was added because there are several times when
8874 an area of memory needs to be read from a BIO. The previous method was
8875 to create a memory BIO and write the data to it, this results in two
8876 copies of the data and an O(n^2) reading algorithm. There is a new
8877 function BIO_new_mem_buf() which creates a read only memory BIO from
8878 an area of memory. Also modified the PKCS#7 routines to use read only
8882 *) Bugfix: ssl23_get_client_hello did not work properly when called in
8883 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
8884 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
8885 but a retry condition occured while trying to read the rest.
8888 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
8889 NID_pkcs7_encrypted by default: this was wrong since this should almost
8890 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
8891 the encrypted data type: this is a more sensible place to put it and it
8892 allows the PKCS#12 code to be tidied up that duplicated this
8896 *) Changed obj_dat.pl script so it takes its input and output files on
8897 the command line. This should avoid shell escape redirection problems
8901 *) Initial support for certificate extension requests, these are included
8902 in things like Xenroll certificate requests. Included functions to allow
8903 extensions to be obtained and added.
8906 *) -crlf option to s_client and s_server for sending newlines as
8907 CRLF (as required by many protocols).
8910 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
8912 *) Install libRSAglue.a when OpenSSL is built with RSAref.
8913 [Ralf S. Engelschall]
8915 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
8916 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
8918 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
8922 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
8923 DH parameters/keys (q is lost during that conversion, but the resulting
8924 DH parameters contain its length).
8926 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
8927 much faster than DH_generate_parameters (which creates parameters
8928 where p = 2*q + 1), and also the smaller q makes DH computations
8929 much more efficient (160-bit exponentiation instead of 1024-bit
8930 exponentiation); so this provides a convenient way to support DHE
8931 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
8932 utter importance to use
8933 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8935 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8936 when such DH parameters are used, because otherwise small subgroup
8937 attacks may become possible!
8940 *) Avoid memory leak in i2d_DHparams.
8943 *) Allow the -k option to be used more than once in the enc program:
8944 this allows the same encrypted message to be read by multiple recipients.
8947 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
8948 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
8949 it will always use the numerical form of the OID, even if it has a short
8953 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
8954 method only got called if p,q,dmp1,dmq1,iqmp components were present,
8955 otherwise bn_mod_exp was called. In the case of hardware keys for example
8956 no private key components need be present and it might store extra data
8957 in the RSA structure, which cannot be accessed from bn_mod_exp.
8958 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
8959 private key operations.
8962 *) Added support for SPARC Linux.
8965 *) pem_password_cb function type incompatibly changed from
8966 typedef int pem_password_cb(char *buf, int size, int rwflag);
8968 ....(char *buf, int size, int rwflag, void *userdata);
8969 so that applications can pass data to their callbacks:
8970 The PEM[_ASN1]_{read,write}... functions and macros now take an
8971 additional void * argument, which is just handed through whenever
8972 the password callback is called.
8973 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
8975 New function SSL_CTX_set_default_passwd_cb_userdata.
8977 Compatibility note: As many C implementations push function arguments
8978 onto the stack in reverse order, the new library version is likely to
8979 interoperate with programs that have been compiled with the old
8980 pem_password_cb definition (PEM_whatever takes some data that
8981 happens to be on the stack as its last argument, and the callback
8982 just ignores this garbage); but there is no guarantee whatsoever that
8985 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
8986 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
8987 problems not only on Windows, but also on some Unix platforms.
8988 To avoid problematic command lines, these definitions are now in an
8989 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
8990 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
8993 *) MIPS III/IV assembler module is reimplemented.
8996 *) More DES library cleanups: remove references to srand/rand and
8997 delete an unused file.
9000 *) Add support for the the free Netwide assembler (NASM) under Win32,
9001 since not many people have MASM (ml) and it can be hard to obtain.
9002 This is currently experimental but it seems to work OK and pass all
9003 the tests. Check out INSTALL.W32 for info.
9006 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
9007 without temporary keys kept an extra copy of the server key,
9008 and connections with temporary keys did not free everything in case
9012 *) New function RSA_check_key and new openssl rsa option -check
9013 for verifying the consistency of RSA keys.
9014 [Ulf Moeller, Bodo Moeller]
9016 *) Various changes to make Win32 compile work:
9017 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
9018 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
9019 comparison" warnings.
9020 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
9023 *) Add a debugging option to PKCS#5 v2 key generation function: when
9024 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
9025 derived keys are printed to stderr.
9028 *) Copy the flags in ASN1_STRING_dup().
9029 [Roman E. Pavlov <pre@mo.msk.ru>]
9031 *) The x509 application mishandled signing requests containing DSA
9032 keys when the signing key was also DSA and the parameters didn't match.
9034 It was supposed to omit the parameters when they matched the signing key:
9035 the verifying software was then supposed to automatically use the CA's
9036 parameters if they were absent from the end user certificate.
9038 Omitting parameters is no longer recommended. The test was also
9039 the wrong way round! This was probably due to unusual behaviour in
9040 EVP_cmp_parameters() which returns 1 if the parameters match.
9041 This meant that parameters were omitted when they *didn't* match and
9042 the certificate was useless. Certificates signed with 'ca' didn't have
9044 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
9046 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
9047 The interface is as follows:
9048 Applications can use
9049 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
9050 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
9051 "off" is now the default.
9052 The library internally uses
9053 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
9054 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
9055 to disable memory-checking temporarily.
9057 Some inconsistent states that previously were possible (and were
9058 even the default) are now avoided.
9060 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
9061 with each memory chunk allocated; this is occasionally more helpful
9062 than just having a counter.
9064 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
9066 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
9070 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
9071 which largely parallels "options", but is for changing API behaviour,
9072 whereas "options" are about protocol behaviour.
9073 Initial "mode" flags are:
9075 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
9076 a single record has been written.
9077 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
9078 retries use the same buffer location.
9079 (But all of the contents must be
9083 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
9086 *) Fix problems with no-hmac etc.
9087 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
9089 *) New functions RSA_get_default_method(), RSA_set_method() and
9090 RSA_get_method(). These allows replacement of RSA_METHODs without having
9091 to mess around with the internals of an RSA structure.
9094 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
9095 Also really enable memory leak checks in openssl.c and in some
9097 [Chad C. Mulligan, Bodo Moeller]
9099 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
9100 up the length of negative integers. This has now been simplified to just
9101 store the length when it is first determined and use it later, rather
9102 than trying to keep track of where data is copied and updating it to
9104 [Steve Henson, reported by Brien Wheeler
9105 <bwheeler@authentica-security.com>]
9107 *) Add a new function PKCS7_signatureVerify. This allows the verification
9108 of a PKCS#7 signature but with the signing certificate passed to the
9109 function itself. This contrasts with PKCS7_dataVerify which assumes the
9110 certificate is present in the PKCS#7 structure. This isn't always the
9111 case: certificates can be omitted from a PKCS#7 structure and be
9112 distributed by "out of band" means (such as a certificate database).
9115 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
9116 function prototypes in pem.h, also change util/mkdef.pl to add the
9117 necessary function names.
9120 *) mk1mf.pl (used by Windows builds) did not properly read the
9121 options set by Configure in the top level Makefile, and Configure
9122 was not even able to write more than one option correctly.
9123 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
9126 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
9127 file to be loaded from a BIO or FILE pointer. The BIO version will
9128 for example allow memory BIOs to contain config info.
9131 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
9132 Whoever hopes to achieve shared-library compatibility across versions
9133 must use this, not the compile-time macro.
9134 (Exercise 0.9.4: Which is the minimum library version required by
9136 Note: All this applies only to multi-threaded programs, others don't
9140 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
9141 through a BIO pair triggered the default case, i.e.
9142 SSLerr(...,SSL_R_UNKNOWN_STATE).
9145 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
9146 can use the SSL library even if none of the specific BIOs is
9150 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
9151 for the encoded length.
9152 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
9154 *) Add initial documentation of the X509V3 functions.
9157 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
9158 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
9159 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
9160 secure PKCS#8 private key format with a high iteration count.
9163 *) Fix determination of Perl interpreter: A perl or perl5
9164 _directory_ in $PATH was also accepted as the interpreter.
9165 [Ralf S. Engelschall]
9167 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
9168 wrong with it but it was very old and did things like calling
9169 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
9173 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
9174 to use the new extension code.
9177 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
9178 with macros. This should make it easier to change their form, add extra
9179 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
9183 *) Add to configuration table a new entry that can specify an alternative
9184 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
9185 according to Mark Crispin <MRC@Panda.COM>.
9189 *) DES CBC did not update the IV. Weird.
9192 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
9193 Changing the behaviour of the former might break existing programs --
9194 where IV updating is needed, des_ncbc_encrypt can be used.
9197 *) When bntest is run from "make test" it drives bc to check its
9198 calculations, as well as internally checking them. If an internal check
9199 fails, it needs to cause bc to give a non-zero result or make test carries
9200 on without noticing the failure. Fixed.
9203 *) DES library cleanups.
9206 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
9207 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
9208 ciphers. NOTE: although the key derivation function has been verified
9209 against some published test vectors it has not been extensively tested
9210 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
9214 *) Instead of "mkdir -p", which is not fully portable, use new
9215 Perl script "util/mkdir-p.pl".
9218 *) Rewrite the way password based encryption (PBE) is handled. It used to
9219 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
9220 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
9221 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
9222 the 'parameter' field of the AlgorithmIdentifier is passed to the
9223 underlying key generation function so it must do its own ASN1 parsing.
9224 This has also changed the EVP_PBE_CipherInit() function which now has a
9225 'parameter' argument instead of literal salt and iteration count values
9226 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
9229 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
9230 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
9231 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
9232 KEY" because this clashed with PKCS#8 unencrypted string. Since this
9233 value was just used as a "magic string" and not used directly its
9234 value doesn't matter.
9237 *) Introduce some semblance of const correctness to BN. Shame C doesn't
9241 *) "linux-sparc64" configuration (ultrapenguin).
9242 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
9243 "linux-sparc" configuration.
9244 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
9246 *) config now generates no-xxx options for missing ciphers.
9249 *) Support the EBCDIC character set (work in progress).
9250 File ebcdic.c not yet included because it has a different license.
9251 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9253 *) Support BS2000/OSD-POSIX.
9254 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9256 *) Make callbacks for key generation use void * instead of char *.
9259 *) Make S/MIME samples compile (not yet tested).
9262 *) Additional typesafe stacks.
9265 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
9269 Changes between 0.9.3 and 0.9.3a [29 May 1999]
9271 *) New configuration variant "sco5-gcc".
9273 *) Updated some demos.
9274 [Sean O Riordain, Wade Scholine]
9276 *) Add missing BIO_free at exit of pkcs12 application.
9279 *) Fix memory leak in conf.c.
9282 *) Updates for Win32 to assembler version of MD5.
9285 *) Set #! path to perl in apps/der_chop to where we found it
9286 instead of using a fixed path.
9289 *) SHA library changes for irix64-mips4-cc.
9292 *) Improvements for VMS support.
9296 Changes between 0.9.2b and 0.9.3 [24 May 1999]
9298 *) Bignum library bug fix. IRIX 6 passes "make test" now!
9299 This also avoids the problems with SC4.2 and unpatched SC5.
9300 [Andy Polyakov <appro@fy.chalmers.se>]
9302 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
9303 These are required because of the typesafe stack would otherwise break
9304 existing code. If old code used a structure member which used to be STACK
9305 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
9306 sk_num or sk_value it would produce an error because the num, data members
9307 are not present in STACK_OF. Now it just produces a warning. sk_set
9308 replaces the old method of assigning a value to sk_value
9309 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
9310 that does this will no longer work (and should use sk_set instead) but
9311 this could be regarded as a "questionable" behaviour anyway.
9314 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
9315 correctly handle encrypted S/MIME data.
9318 *) Change type of various DES function arguments from des_cblock
9319 (which means, in function argument declarations, pointer to char)
9320 to des_cblock * (meaning pointer to array with 8 char elements),
9321 which allows the compiler to do more typechecking; it was like
9322 that back in SSLeay, but with lots of ugly casts.
9324 Introduce new type const_des_cblock.
9327 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
9328 problems: find RecipientInfo structure that matches recipient certificate
9329 and initialise the ASN1 structures properly based on passed cipher.
9332 *) Belatedly make the BN tests actually check the results.
9335 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
9336 to and from BNs: it was completely broken. New compilation option
9337 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
9338 key elements as negative integers.
9341 *) Reorganize and speed up MD5.
9342 [Andy Polyakov <appro@fy.chalmers.se>]
9345 [Richard Levitte <richard@levitte.org>]
9347 *) New option -out to asn1parse to allow the parsed structure to be
9348 output to a file. This is most useful when combined with the -strparse
9349 option to examine the output of things like OCTET STRINGS.
9352 *) Make SSL library a little more fool-proof by not requiring any longer
9353 that SSL_set_{accept,connect}_state be called before
9354 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
9355 in many applications because usually everything *appeared* to work as
9356 intended anyway -- now it really works as intended).
9359 *) Move openssl.cnf out of lib/.
9362 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
9363 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
9364 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
9365 [Ralf S. Engelschall]
9367 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
9368 handle PKCS#7 enveloped data properly.
9369 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
9371 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
9372 copying pointers. The cert_st handling is changed by this in
9373 various ways (and thus what used to be known as ctx->default_cert
9374 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
9375 any longer when s->cert does not give us what we need).
9376 ssl_cert_instantiate becomes obsolete by this change.
9377 As soon as we've got the new code right (possibly it already is?),
9378 we have solved a couple of bugs of the earlier code where s->cert
9379 was used as if it could not have been shared with other SSL structures.
9381 Note that using the SSL API in certain dirty ways now will result
9382 in different behaviour than observed with earlier library versions:
9383 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
9384 does not influence s as it used to.
9386 In order to clean up things more thoroughly, inside SSL_SESSION
9387 we don't use CERT any longer, but a new structure SESS_CERT
9388 that holds per-session data (if available); currently, this is
9389 the peer's certificate chain and, for clients, the server's certificate
9390 and temporary key. CERT holds only those values that can have
9391 meaningful defaults in an SSL_CTX.
9394 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
9395 from the internal representation. Various PKCS#7 fixes: remove some
9396 evil casts and set the enc_dig_alg field properly based on the signing
9400 *) Allow PKCS#12 password to be set from the command line or the
9401 environment. Let 'ca' get its config file name from the environment
9402 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
9406 *) Allow certificate policies extension to use an IA5STRING for the
9407 organization field. This is contrary to the PKIX definition but
9408 VeriSign uses it and IE5 only recognises this form. Document 'x509'
9412 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
9413 without disallowing inline assembler and the like for non-pedantic builds.
9416 *) Support Borland C++ builder.
9417 [Janez Jere <jj@void.si>, modified by Ulf Möller]
9422 *) SHA-1 cleanups and performance enhancements.
9423 [Andy Polyakov <appro@fy.chalmers.se>]
9425 *) Sparc v8plus assembler for the bignum library.
9426 [Andy Polyakov <appro@fy.chalmers.se>]
9428 *) Accept any -xxx and +xxx compiler options in Configure.
9431 *) Update HPUX configuration.
9434 *) Add missing sk_<type>_unshift() function to safestack.h
9435 [Ralf S. Engelschall]
9437 *) New function SSL_CTX_use_certificate_chain_file that sets the
9438 "extra_cert"s in addition to the certificate. (This makes sense
9439 only for "PEM" format files, as chains as a whole are not
9443 *) Support verify_depth from the SSL API.
9444 x509_vfy.c had what can be considered an off-by-one-error:
9445 Its depth (which was not part of the external interface)
9446 was actually counting the number of certificates in a chain;
9447 now it really counts the depth.
9450 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
9451 instead of X509err, which often resulted in confusing error
9452 messages since the error codes are not globally unique
9453 (e.g. an alleged error in ssl3_accept when a certificate
9454 didn't match the private key).
9456 *) New function SSL_CTX_set_session_id_context that allows to set a default
9457 value (so that you don't need SSL_set_session_id_context for each
9458 connection using the SSL_CTX).
9461 *) OAEP decoding bug fix.
9464 *) Support INSTALL_PREFIX for package builders, as proposed by
9468 *) New Configure options "threads" and "no-threads". For systems
9469 where the proper compiler options are known (currently Solaris
9470 and Linux), "threads" is the default.
9473 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
9476 *) Install various scripts to $(OPENSSLDIR)/misc, not to
9477 $(INSTALLTOP)/bin -- they shouldn't clutter directories
9478 such as /usr/local/bin.
9481 *) "make linux-shared" to build shared libraries.
9482 [Niels Poppe <niels@netbox.org>]
9484 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
9487 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
9488 extension adding in x509 utility.
9491 *) Remove NOPROTO sections and error code comments.
9494 *) Partial rewrite of the DEF file generator to now parse the ANSI
9498 *) New Configure options --prefix=DIR and --openssldir=DIR.
9501 *) Complete rewrite of the error code script(s). It is all now handled
9502 by one script at the top level which handles error code gathering,
9503 header rewriting and C source file generation. It should be much better
9504 than the old method: it now uses a modified version of Ulf's parser to
9505 read the ANSI prototypes in all header files (thus the old K&R definitions
9506 aren't needed for error creation any more) and do a better job of
9507 translating function codes into names. The old 'ASN1 error code imbedded
9508 in a comment' is no longer necessary and it doesn't use .err files which
9509 have now been deleted. Also the error code call doesn't have to appear all
9510 on one line (which resulted in some large lines...).
9513 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
9516 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
9517 0 (which usually indicates a closed connection), but continue reading.
9520 *) Fix some race conditions.
9523 *) Add support for CRL distribution points extension. Add Certificate
9524 Policies and CRL distribution points documentation.
9527 *) Move the autogenerated header file parts to crypto/opensslconf.h.
9530 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
9531 8 of keying material. Merlin has also confirmed interop with this fix
9532 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
9533 [Merlin Hughes <merlin@baltimore.ie>]
9535 *) Fix lots of warnings.
9536 [Richard Levitte <levitte@stacken.kth.se>]
9538 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
9539 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
9540 [Richard Levitte <levitte@stacken.kth.se>]
9542 *) Fix problems with sizeof(long) == 8.
9543 [Andy Polyakov <appro@fy.chalmers.se>]
9545 *) Change functions to ANSI C.
9548 *) Fix typos in error codes.
9549 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
9551 *) Remove defunct assembler files from Configure.
9554 *) SPARC v8 assembler BIGNUM implementation.
9555 [Andy Polyakov <appro@fy.chalmers.se>]
9557 *) Support for Certificate Policies extension: both print and set.
9558 Various additions to support the r2i method this uses.
9561 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
9562 return a const string when you are expecting an allocated buffer.
9565 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
9566 types DirectoryString and DisplayText.
9569 *) Add code to allow r2i extensions to access the configuration database,
9570 add an LHASH database driver and add several ctx helper functions.
9573 *) Fix an evil bug in bn_expand2() which caused various BN functions to
9574 fail when they extended the size of a BIGNUM.
9577 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
9578 support typesafe stack.
9581 *) Fix typo in SSL_[gs]et_options().
9582 [Nils Frostberg <nils@medcom.se>]
9584 *) Delete various functions and files that belonged to the (now obsolete)
9585 old X509V3 handling code.
9588 *) New Configure option "rsaref".
9591 *) Don't auto-generate pem.h.
9594 *) Introduce type-safe ASN.1 SETs.
9597 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
9598 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
9600 *) Introduce type-safe STACKs. This will almost certainly break lots of code
9601 that links with OpenSSL (well at least cause lots of warnings), but fear
9602 not: the conversion is trivial, and it eliminates loads of evil casts. A
9603 few STACKed things have been converted already. Feel free to convert more.
9604 In the fullness of time, I'll do away with the STACK type altogether.
9607 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
9608 specified in <certfile> by updating the entry in the index.txt file.
9609 This way one no longer has to edit the index.txt file manually for
9610 revoking a certificate. The -revoke option does the gory details now.
9611 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
9613 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
9614 `-text' option at all and this way the `-noout -text' combination was
9615 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
9616 [Ralf S. Engelschall]
9618 *) Make sure a corresponding plain text error message exists for the
9619 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
9620 verify callback function determined that a certificate was revoked.
9621 [Ralf S. Engelschall]
9623 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
9624 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
9625 all available cipers including rc5, which was forgotten until now.
9626 In order to let the testing shell script know which algorithms
9627 are available, a new (up to now undocumented) command
9628 "openssl list-cipher-commands" is used.
9631 *) Bugfix: s_client occasionally would sleep in select() when
9632 it should have checked SSL_pending() first.
9635 *) New functions DSA_do_sign and DSA_do_verify to provide access to
9636 the raw DSA values prior to ASN.1 encoding.
9639 *) Tweaks to Configure
9640 [Niels Poppe <niels@netbox.org>]
9642 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
9646 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
9649 *) New config option to avoid instructions that are illegal on the 80386.
9650 The default code is faster, but requires at least a 486.
9653 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
9654 SSL2_SERVER_VERSION (not used at all) macros, which are now the
9655 same as SSL2_VERSION anyway.
9658 *) New "-showcerts" option for s_client.
9661 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
9662 application. Various cleanups and fixes.
9665 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
9666 modify error routines to work internally. Add error codes and PBE init
9667 to library startup routines.
9670 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
9671 packing functions to asn1 and evp. Changed function names and error
9672 codes along the way.
9675 *) PKCS12 integration: and so it begins... First of several patches to
9676 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
9677 objects to objects.h
9680 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
9681 and display support for Thawte strong extranet extension.
9684 *) Add LinuxPPC support.
9685 [Jeff Dubrule <igor@pobox.org>]
9687 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
9688 bn_div_words in alpha.s.
9689 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
9691 *) Make sure the RSA OAEP test is skipped under -DRSAref because
9692 OAEP isn't supported when OpenSSL is built with RSAref.
9693 [Ulf Moeller <ulf@fitug.de>]
9695 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
9696 so they no longer are missing under -DNOPROTO.
9697 [Soren S. Jorvang <soren@t.dk>]
9700 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
9702 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
9703 doesn't work when the session is reused. Coming soon!
9706 *) Fix a security hole, that allows sessions to be reused in the wrong
9707 context thus bypassing client cert protection! All software that uses
9708 client certs and session caches in multiple contexts NEEDS PATCHING to
9709 allow session reuse! A fuller solution is in the works.
9710 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
9712 *) Some more source tree cleanups (removed obsolete files
9713 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
9714 permission on "config" script to be executable) and a fix for the INSTALL
9716 [Ulf Moeller <ulf@fitug.de>]
9718 *) Remove some legacy and erroneous uses of malloc, free instead of
9720 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
9722 *) Make rsa_oaep_test return non-zero on error.
9723 [Ulf Moeller <ulf@fitug.de>]
9725 *) Add support for native Solaris shared libraries. Configure
9726 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
9727 if someone would make that last step automatic.
9728 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
9730 *) ctx_size was not built with the right compiler during "make links". Fixed.
9733 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
9734 except NULL ciphers". This means the default cipher list will no longer
9735 enable NULL ciphers. They need to be specifically enabled e.g. with
9736 the string "DEFAULT:eNULL".
9739 *) Fix to RSA private encryption routines: if p < q then it would
9740 occasionally produce an invalid result. This will only happen with
9741 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9744 *) Be less restrictive and allow also `perl util/perlpath.pl
9745 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9746 because this way one can also use an interpreter named `perl5' (which is
9747 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9748 installed as `perl').
9749 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9751 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9752 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9754 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9755 advapi32.lib to Win32 build and change the pem test comparision
9756 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9757 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9758 and crypto/des/ede_cbcm_enc.c.
9761 *) DES quad checksum was broken on big-endian architectures. Fixed.
9764 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9765 Win32 test batch file so it (might) work again. The Win32 test batch file
9766 is horrible: I feel ill....
9769 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
9770 in e_os.h. Audit of header files to check ANSI and non ANSI
9771 sections: 10 functions were absent from non ANSI section and not exported
9772 from Windows DLLs. Fixed up libeay.num for new functions.
9775 *) Make `openssl version' output lines consistent.
9776 [Ralf S. Engelschall]
9778 *) Fix Win32 symbol export lists for BIO functions: Added
9779 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
9780 to ms/libeay{16,32}.def.
9781 [Ralf S. Engelschall]
9783 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
9784 fine under Unix and passes some trivial tests I've now added. But the
9785 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
9786 added to make sure no one expects that this stuff really works in the
9787 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
9788 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
9790 [Ralf S. Engelschall]
9792 *) Fix the generation of two part addresses in perl.
9793 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
9795 *) Add config entry for Linux on MIPS.
9796 [John Tobey <jtobey@channel1.com>]
9798 *) Make links whenever Configure is run, unless we are on Windoze.
9801 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
9802 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
9806 *) Add a useful kludge to allow package maintainers to specify compiler and
9807 other platforms details on the command line without having to patch the
9808 Configure script everytime: One now can use ``perl Configure
9809 <id>:<details>'', i.e. platform ids are allowed to have details appended
9810 to them (seperated by colons). This is treated as there would be a static
9811 pre-configured entry in Configure's %table under key <id> with value
9812 <details> and ``perl Configure <id>'' is called. So, when you want to
9813 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
9814 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
9815 now, which overrides the FreeBSD-elf entry on-the-fly.
9816 [Ralf S. Engelschall]
9818 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
9821 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
9822 on the `perl Configure ...' command line. This way one can compile
9823 OpenSSL libraries with Position Independent Code (PIC) which is needed
9824 for linking it into DSOs.
9825 [Ralf S. Engelschall]
9827 *) Remarkably, export ciphers were totally broken and no-one had noticed!
9831 *) Cleaned up the LICENSE document: The official contact for any license
9832 questions now is the OpenSSL core team under openssl-core@openssl.org.
9833 And add a paragraph about the dual-license situation to make sure people
9834 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
9835 to the OpenSSL toolkit.
9836 [Ralf S. Engelschall]
9838 *) General source tree makefile cleanups: Made `making xxx in yyy...'
9839 display consistent in the source tree and replaced `/bin/rm' by `rm'.
9840 Additonally cleaned up the `make links' target: Remove unnecessary
9841 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
9842 to speed processing and no longer clutter the display with confusing
9843 stuff. Instead only the actually done links are displayed.
9844 [Ralf S. Engelschall]
9846 *) Permit null encryption ciphersuites, used for authentication only. It used
9847 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
9848 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
9852 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
9853 signed attributes when verifying signatures (this would break them),
9854 the detached data encoding was wrong and public keys obtained using
9855 X509_get_pubkey() weren't freed.
9858 *) Add text documentation for the BUFFER functions. Also added a work around
9859 to a Win95 console bug. This was triggered by the password read stuff: the
9860 last character typed gets carried over to the next fread(). If you were
9861 generating a new cert request using 'req' for example then the last
9862 character of the passphrase would be CR which would then enter the first
9866 *) Added the new `Includes OpenSSL Cryptography Software' button as
9867 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
9868 button and can be used by applications based on OpenSSL to show the
9869 relationship to the OpenSSL project.
9870 [Ralf S. Engelschall]
9872 *) Remove confusing variables in function signatures in files
9873 ssl/ssl_lib.c and ssl/ssl.h.
9874 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9876 *) Don't install bss_file.c under PREFIX/include/
9877 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9879 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
9880 functions that return function pointers and has support for NT specific
9881 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
9882 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
9883 unsigned to signed types: this was killing the Win32 compile.
9886 *) Add new certificate file to stack functions,
9887 SSL_add_dir_cert_subjects_to_stack() and
9888 SSL_add_file_cert_subjects_to_stack(). These largely supplant
9889 SSL_load_client_CA_file(), and can be used to add multiple certs easily
9890 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
9891 This means that Apache-SSL and similar packages don't have to mess around
9892 to add as many CAs as they want to the preferred list.
9895 *) Experiment with doxygen documentation. Currently only partially applied to
9897 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
9898 openssl.doxy as the configuration file.
9901 *) Get rid of remaining C++-style comments which strict C compilers hate.
9902 [Ralf S. Engelschall, pointed out by Carlos Amengual]
9904 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
9905 compiled in by default: it has problems with large keys.
9908 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
9909 DH private keys and/or callback functions which directly correspond to
9910 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
9911 is needed for applications which have to configure certificates on a
9912 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
9914 For the RSA certificate situation is makes no difference, but
9915 for the DSA certificate situation this fixes the "no shared cipher"
9916 problem where the OpenSSL cipher selection procedure failed because the
9917 temporary keys were not overtaken from the context and the API provided
9918 no way to reconfigure them.
9919 The new functions now let applications reconfigure the stuff and they
9920 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
9921 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
9922 non-public-API function ssl_cert_instantiate() is used as a helper
9923 function and also to reduce code redundancy inside ssl_rsa.c.
9924 [Ralf S. Engelschall]
9926 *) Move s_server -dcert and -dkey options out of the undocumented feature
9927 area because they are useful for the DSA situation and should be
9928 recognized by the users.
9929 [Ralf S. Engelschall]
9931 *) Fix the cipher decision scheme for export ciphers: the export bits are
9932 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
9933 SSL_EXP_MASK. So, the original variable has to be used instead of the
9934 already masked variable.
9935 [Richard Levitte <levitte@stacken.kth.se>]
9937 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
9938 [Richard Levitte <levitte@stacken.kth.se>]
9940 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
9941 from `int' to `unsigned int' because it's a length and initialized by
9942 EVP_DigestFinal() which expects an `unsigned int *'.
9943 [Richard Levitte <levitte@stacken.kth.se>]
9945 *) Don't hard-code path to Perl interpreter on shebang line of Configure
9946 script. Instead use the usual Shell->Perl transition trick.
9947 [Ralf S. Engelschall]
9949 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
9950 (in addition to RSA certificates) to match the behaviour of `openssl dsa
9951 -noout -modulus' as it's already the case for `openssl rsa -noout
9952 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
9953 currently the public key is printed (a decision which was already done by
9954 `openssl dsa -modulus' in the past) which serves a similar purpose.
9955 Additionally the NO_RSA no longer completely removes the whole -modulus
9956 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
9958 [Ralf S. Engelschall]
9960 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
9961 BIO. See the source (crypto/evp/bio_ok.c) for more info.
9962 [Arne Ansper <arne@ats.cyber.ee>]
9964 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
9965 to be added. Now both 'req' and 'ca' can use new objects defined in the
9969 *) Add cool BIO that does syslog (or event log on NT).
9970 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
9972 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
9973 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
9974 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
9975 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
9978 *) Add preliminary config info for new extension code.
9981 *) Make RSA_NO_PADDING really use no padding.
9982 [Ulf Moeller <ulf@fitug.de>]
9984 *) Generate errors when private/public key check is done.
9987 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
9988 for some CRL extensions and new objects added.
9991 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
9992 key usage extension and fuller support for authority key id.
9995 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
9996 padding method for RSA, which is recommended for new applications in PKCS
9997 #1 v2.0 (RFC 2437, October 1998).
9998 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
9999 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
10000 against Bleichbacher's attack on RSA.
10001 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
10004 *) Updates to the new SSL compression code
10005 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10007 *) Fix so that the version number in the master secret, when passed
10008 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
10009 (because the server will not accept higher), that the version number
10010 is 0x03,0x01, not 0x03,0x00
10011 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10013 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
10014 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
10015 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
10018 *) Support for RAW extensions where an arbitrary extension can be
10019 created by including its DER encoding. See apps/openssl.cnf for
10023 *) Make sure latest Perl versions don't interpret some generated C array
10024 code as Perl array code in the crypto/err/err_genc.pl script.
10025 [Lars Weber <3weber@informatik.uni-hamburg.de>]
10027 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
10028 not many people have the assembler. Various Win32 compilation fixes and
10029 update to the INSTALL.W32 file with (hopefully) more accurate Win32
10030 build instructions.
10033 *) Modify configure script 'Configure' to automatically create crypto/date.h
10034 file under Win32 and also build pem.h from pem.org. New script
10035 util/mkfiles.pl to create the MINFO file on environments that can't do a
10036 'make files': perl util/mkfiles.pl >MINFO should work.
10039 *) Major rework of DES function declarations, in the pursuit of correctness
10040 and purity. As a result, many evil casts evaporated, and some weirdness,
10041 too. You may find this causes warnings in your code. Zapping your evil
10042 casts will probably fix them. Mostly.
10045 *) Fix for a typo in asn1.h. Bug fix to object creation script
10046 obj_dat.pl. It considered a zero in an object definition to mean
10047 "end of object": none of the objects in objects.h have any zeros
10048 so it wasn't spotted.
10049 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
10051 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
10052 Masking (CBCM). In the absence of test vectors, the best I have been able
10053 to do is check that the decrypt undoes the encrypt, so far. Send me test
10054 vectors if you have them.
10057 *) Correct calculation of key length for export ciphers (too much space was
10058 allocated for null ciphers). This has not been tested!
10061 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
10062 message is now correct (it understands "crypto" and "ssl" on its
10063 command line). There is also now an "update" option. This will update
10064 the util/ssleay.num and util/libeay.num files with any new functions.
10066 perl util/mkdef.pl crypto ssl update
10067 it will update them.
10070 *) Overhauled the Perl interface (perl/*):
10071 - ported BN stuff to OpenSSL's different BN library
10072 - made the perl/ source tree CVS-aware
10073 - renamed the package from SSLeay to OpenSSL (the files still contain
10074 their history because I've copied them in the repository)
10075 - removed obsolete files (the test scripts will be replaced
10076 by better Test::Harness variants in the future)
10077 [Ralf S. Engelschall]
10079 *) First cut for a very conservative source tree cleanup:
10080 1. merge various obsolete readme texts into doc/ssleay.txt
10081 where we collect the old documents and readme texts.
10082 2. remove the first part of files where I'm already sure that we no
10083 longer need them because of three reasons: either they are just temporary
10084 files which were left by Eric or they are preserved original files where
10085 I've verified that the diff is also available in the CVS via "cvs diff
10086 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
10087 the crypto/md/ stuff).
10088 [Ralf S. Engelschall]
10090 *) More extension code. Incomplete support for subject and issuer alt
10091 name, issuer and authority key id. Change the i2v function parameters
10092 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
10093 what that's for :-) Fix to ASN1 macro which messed up
10094 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
10097 *) Preliminary support for ENUMERATED type. This is largely copied from the
10101 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
10102 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10104 *) Make sure `make rehash' target really finds the `openssl' program.
10105 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10107 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
10108 like to hear about it if this slows down other processors.
10111 *) Add CygWin32 platform information to Configure script.
10112 [Alan Batie <batie@aahz.jf.intel.com>]
10114 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
10115 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
10117 *) New program nseq to manipulate netscape certificate sequences
10120 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
10124 *) Fixes to BN code. Previously the default was to define BN_RECURSION
10125 but the BN code had some problems that would cause failures when
10126 doing certificate verification and some other functions.
10127 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10129 *) Add ASN1 and PEM code to support netscape certificate sequences.
10132 *) Add ASN1 and PEM code to support netscape certificate sequences.
10135 *) Add several PKIX and private extended key usage OIDs.
10138 *) Modify the 'ca' program to handle the new extension code. Modify
10139 openssl.cnf for new extension format, add comments.
10142 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
10143 and add a sample to openssl.cnf so req -x509 now adds appropriate
10147 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
10148 error code, add initial support to X509_print() and x509 application.
10151 *) Takes a deep breath and start addding X509 V3 extension support code. Add
10152 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
10153 stuff is currently isolated and isn't even compiled yet.
10156 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
10157 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
10158 Removed the versions check from X509 routines when loading extensions:
10159 this allows certain broken certificates that don't set the version
10160 properly to be processed.
10163 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
10164 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
10165 can still be regenerated with "make depend".
10168 *) Spelling mistake in C version of CAST-128.
10169 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
10171 *) Changes to the error generation code. The perl script err-code.pl
10172 now reads in the old error codes and retains the old numbers, only
10173 adding new ones if necessary. It also only changes the .err files if new
10174 codes are added. The makefiles have been modified to only insert errors
10175 when needed (to avoid needlessly modifying header files). This is done
10176 by only inserting errors if the .err file is newer than the auto generated
10177 C file. To rebuild all the error codes from scratch (the old behaviour)
10178 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
10179 or delete all the .err files.
10182 *) CAST-128 was incorrectly implemented for short keys. The C version has
10183 been fixed, but is untested. The assembler versions are also fixed, but
10184 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
10185 to regenerate it if needed.
10186 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
10187 Hagino <itojun@kame.net>]
10189 *) File was opened incorrectly in randfile.c.
10190 [Ulf Möller <ulf@fitug.de>]
10192 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
10193 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
10194 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
10195 al: it's just almost always a UTCTime. Note this patch adds new error
10196 codes so do a "make errors" if there are problems.
10199 *) Correct Linux 1 recognition in config.
10200 [Ulf Möller <ulf@fitug.de>]
10202 *) Remove pointless MD5 hash when using DSA keys in ca.
10203 [Anonymous <nobody@replay.com>]
10205 *) Generate an error if given an empty string as a cert directory. Also
10206 generate an error if handed NULL (previously returned 0 to indicate an
10207 error, but didn't set one).
10208 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
10210 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
10213 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
10214 parameters. This was causing a warning which killed off the Win32 compile.
10217 *) Remove C++ style comments from crypto/bn/bn_local.h.
10218 [Neil Costigan <neil.costigan@celocom.com>]
10220 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
10221 based on a text string, looking up short and long names and finally
10222 "dot" format. The "dot" format stuff didn't work. Added new function
10223 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
10224 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
10225 OID is not part of the table.
10228 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
10229 X509_LOOKUP_by_alias().
10232 *) Sort openssl functions by name.
10235 *) Get the gendsa program working (hopefully) and add it to app list. Remove
10236 encryption from sample DSA keys (in case anyone is interested the password
10240 *) Make _all_ *_free functions accept a NULL pointer.
10241 [Frans Heymans <fheymans@isaserver.be>]
10243 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
10245 [Anonymous <nobody@replay.com>]
10247 *) s_server should send the CAfile as acceptable CAs, not its own cert.
10248 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10250 *) Don't blow it for numeric -newkey arguments to apps/req.
10251 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10253 *) Temp key "for export" tests were wrong in s3_srvr.c.
10254 [Anonymous <nobody@replay.com>]
10256 *) Add prototype for temp key callback functions
10257 SSL_CTX_set_tmp_{rsa,dh}_callback().
10260 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
10261 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
10264 *) X509_name_add_entry() freed the wrong thing after an error.
10265 [Arne Ansper <arne@ats.cyber.ee>]
10267 *) rsa_eay.c would attempt to free a NULL context.
10268 [Arne Ansper <arne@ats.cyber.ee>]
10270 *) BIO_s_socket() had a broken should_retry() on Windoze.
10271 [Arne Ansper <arne@ats.cyber.ee>]
10273 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
10274 [Arne Ansper <arne@ats.cyber.ee>]
10276 *) Make sure the already existing X509_STORE->depth variable is initialized
10277 in X509_STORE_new(), but document the fact that this variable is still
10278 unused in the certificate verification process.
10279 [Ralf S. Engelschall]
10281 *) Fix the various library and apps files to free up pkeys obtained from
10282 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
10285 *) Fix reference counting in X509_PUBKEY_get(). This makes
10286 demos/maurice/example2.c work, amongst others, probably.
10287 [Steve Henson and Ben Laurie]
10289 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
10290 `openssl' and second, the shortcut symlinks for the `openssl <command>'
10291 are no longer created. This way we have a single and consistent command
10292 line interface `openssl <command>', similar to `cvs <command>'.
10293 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
10295 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
10296 BIT STRING wrapper always have zero unused bits.
10299 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
10302 *) Make the top-level INSTALL documentation easier to understand.
10305 *) Makefiles updated to exit if an error occurs in a sub-directory
10306 make (including if user presses ^C) [Paul Sutton]
10308 *) Make Montgomery context stuff explicit in RSA data structure.
10311 *) Fix build order of pem and err to allow for generated pem.h.
10314 *) Fix renumbering bug in X509_NAME_delete_entry().
10317 *) Enhanced the err-ins.pl script so it makes the error library number
10318 global and can add a library name. This is needed for external ASN1 and
10319 other error libraries.
10322 *) Fixed sk_insert which never worked properly.
10325 *) Fix ASN1 macros so they can handle indefinite length construted
10326 EXPLICIT tags. Some non standard certificates use these: they can now
10330 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
10331 into a single doc/ssleay.txt bundle. This way the information is still
10332 preserved but no longer messes up this directory. Now it's new room for
10333 the new set of documenation files.
10334 [Ralf S. Engelschall]
10336 *) SETs were incorrectly DER encoded. This was a major pain, because they
10337 shared code with SEQUENCEs, which aren't coded the same. This means that
10338 almost everything to do with SETs or SEQUENCEs has either changed name or
10339 number of arguments.
10340 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
10342 *) Fix test data to work with the above.
10345 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
10346 was already fixed by Eric for 0.9.1 it seems.
10347 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
10349 *) Autodetect FreeBSD3.
10352 *) Fix various bugs in Configure. This affects the following platforms:
10356 unixware-2.0-pentium
10360 *) Eliminate generated files from CVS. Reorder tests to regenerate files
10361 before they are needed.
10364 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
10368 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
10370 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
10371 changed SSLeay to OpenSSL in version strings.
10372 [Ralf S. Engelschall]
10374 *) Some fixups to the top-level documents.
10377 *) Fixed the nasty bug where rsaref.h was not found under compile-time
10378 because the symlink to include/ was missing.
10379 [Ralf S. Engelschall]
10381 *) Incorporated the popular no-RSA/DSA-only patches
10382 which allow to compile a RSA-free SSLeay.
10383 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
10385 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
10386 when "ssleay" is still not found.
10387 [Ralf S. Engelschall]
10389 *) Added more platforms to Configure: Cray T3E, HPUX 11,
10390 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
10392 *) Updated the README file.
10393 [Ralf S. Engelschall]
10395 *) Added various .cvsignore files in the CVS repository subdirs
10396 to make a "cvs update" really silent.
10397 [Ralf S. Engelschall]
10399 *) Recompiled the error-definition header files and added
10400 missing symbols to the Win32 linker tables.
10401 [Ralf S. Engelschall]
10403 *) Cleaned up the top-level documents;
10404 o new files: CHANGES and LICENSE
10405 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
10406 o merged COPYRIGHT into LICENSE
10407 o removed obsolete TODO file
10408 o renamed MICROSOFT to INSTALL.W32
10409 [Ralf S. Engelschall]
10411 *) Removed dummy files from the 0.9.1b source tree:
10412 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
10413 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
10414 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
10415 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
10416 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
10417 [Ralf S. Engelschall]
10419 *) Added various platform portability fixes.
10422 *) The Genesis of the OpenSSL rpject:
10423 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
10424 Young and Tim J. Hudson created while they were working for C2Net until
10426 [The OpenSSL Project]
10429 Changes between 0.9.0b and 0.9.1b [not released]
10431 *) Updated a few CA certificates under certs/
10434 *) Changed some BIGNUM api stuff.
10437 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
10438 DGUX x86, Linux Alpha, etc.
10441 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
10442 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
10446 *) Add -strparse option to asn1pars program which parses nested
10448 [Dr Stephen Henson <shenson@bigfoot.com>]
10450 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
10453 *) DSA fix for "ca" program.
10456 *) Added "-genkey" option to "dsaparam" program.
10459 *) Added RIPE MD160 (rmd160) message digest.
10462 *) Added -a (all) option to "ssleay version" command.
10465 *) Added PLATFORM define which is the id given to Configure.
10468 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
10471 *) Extended the ASN.1 parser routines.
10474 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
10477 *) Added a BN_CTX to the BN library.
10480 *) Fixed the weak key values in DES library
10483 *) Changed API in EVP library for cipher aliases.
10486 *) Added support for RC2/64bit cipher.
10489 *) Converted the lhash library to the crypto/mem.c functions.
10492 *) Added more recognized ASN.1 object ids.
10495 *) Added more RSA padding checks for SSL/TLS.
10498 *) Added BIO proxy/filter functionality.
10501 *) Added extra_certs to SSL_CTX which can be used
10502 send extra CA certificates to the client in the CA cert chain sending
10503 process. It can be configured with SSL_CTX_add_extra_chain_cert().
10506 *) Now Fortezza is denied in the authentication phase because
10507 this is key exchange mechanism is not supported by SSLeay at all.
10510 *) Additional PKCS1 checks.
10513 *) Support the string "TLSv1" for all TLS v1 ciphers.
10516 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
10517 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
10520 *) Fixed a few memory leaks.
10523 *) Fixed various code and comment typos.
10526 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
10527 bytes sent in the client random.
10528 [Edward Bishop <ebishop@spyglass.com>]