5 Changes between 1.0.1r and 1.0.1s [xx XXX xxxx]
9 Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
11 *) Protection for DH small subgroup attacks
13 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
14 switched on by default and cannot be disabled. This could have some
18 *) SSLv2 doesn't block disabled ciphers
20 A malicious client can negotiate SSLv2 ciphers that have been disabled on
21 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
22 been disabled, provided that the SSLv2 protocol was not also disabled via
25 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
26 and Sebastian Schinzel.
30 *) Reject DH handshakes with parameters shorter than 1024 bits.
33 Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
35 *) Certificate verify crash with missing PSS parameter
37 The signature verification routines will crash with a NULL pointer
38 dereference if presented with an ASN.1 signature using the RSA PSS
39 algorithm and absent mask generation function parameter. Since these
40 routines are used to verify certificate signature algorithms this can be
41 used to crash any certificate verification operation and exploited in a
42 DoS attack. Any application which performs certificate verification is
43 vulnerable including OpenSSL clients and servers which enable client
46 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
50 *) X509_ATTRIBUTE memory leak
52 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
53 memory. This structure is used by the PKCS#7 and CMS routines so any
54 application which reads PKCS#7 or CMS data from untrusted sources is
55 affected. SSL/TLS is not affected.
57 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
62 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
63 This changes the decoding behaviour for some invalid messages,
64 though the change is mostly in the more lenient direction, and
65 legacy behaviour is preserved as much as possible.
68 *) In DSA_generate_parameters_ex, if the provided seed is too short,
69 use a random seed, as already documented.
70 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
72 Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
74 *) Alternate chains certificate forgery
76 During certificate verfification, OpenSSL will attempt to find an
77 alternative certificate chain if the first attempt to build such a chain
78 fails. An error in the implementation of this logic can mean that an
79 attacker could cause certain checks on untrusted certificates to be
80 bypassed, such as the CA flag, enabling them to use a valid leaf
81 certificate to act as a CA and "issue" an invalid certificate.
83 This issue was reported to OpenSSL by Adam Langley/David Benjamin
88 *) Race condition handling PSK identify hint
90 If PSK identity hints are received by a multi-threaded client then
91 the values are wrongly updated in the parent SSL_CTX structure. This can
92 result in a race condition potentially leading to a double free of the
97 Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
98 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
99 incompatibility in the handling of HMAC. The previous ABI has now been
102 Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
104 *) Malformed ECParameters causes infinite loop
106 When processing an ECParameters structure OpenSSL enters an infinite loop
107 if the curve specified is over a specially malformed binary polynomial
110 This can be used to perform denial of service against any
111 system which processes public keys, certificate requests or
112 certificates. This includes TLS clients and TLS servers with
113 client authentication enabled.
115 This issue was reported to OpenSSL by Joseph Barr-Pixton.
119 *) Exploitable out-of-bounds read in X509_cmp_time
121 X509_cmp_time does not properly check the length of the ASN1_TIME
122 string and can read a few bytes out of bounds. In addition,
123 X509_cmp_time accepts an arbitrary number of fractional seconds in the
126 An attacker can use this to craft malformed certificates and CRLs of
127 various sizes and potentially cause a segmentation fault, resulting in
128 a DoS on applications that verify certificates or CRLs. TLS clients
129 that verify CRLs are affected. TLS clients and servers with client
130 authentication enabled may be affected if they use custom verification
133 This issue was reported to OpenSSL by Robert Swiecki (Google), and
134 independently by Hanno Böck.
138 *) PKCS7 crash with missing EnvelopedContent
140 The PKCS#7 parsing code does not handle missing inner EncryptedContent
141 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
142 with missing content and trigger a NULL pointer dereference on parsing.
144 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
145 structures from untrusted sources are affected. OpenSSL clients and
146 servers are not affected.
148 This issue was reported to OpenSSL by Michal Zalewski (Google).
152 *) CMS verify infinite loop with unknown hash function
154 When verifying a signedData message the CMS code can enter an infinite loop
155 if presented with an unknown hash function OID. This can be used to perform
156 denial of service against any system which verifies signedData messages using
158 This issue was reported to OpenSSL by Johannes Bauer.
162 *) Race condition handling NewSessionTicket
164 If a NewSessionTicket is received by a multi-threaded client when attempting to
165 reuse a previous ticket then a race condition can occur potentially leading to
166 a double free of the ticket data.
170 *) Reject DH handshakes with parameters shorter than 768 bits.
171 [Kurt Roeckx and Emilia Kasper]
173 *) dhparam: generate 2048-bit parameters by default.
174 [Kurt Roeckx and Emilia Kasper]
176 Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
178 *) Segmentation fault in ASN1_TYPE_cmp fix
180 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
181 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
182 certificate signature algorithm consistency this can be used to crash any
183 certificate verification operation and exploited in a DoS attack. Any
184 application which performs certificate verification is vulnerable including
185 OpenSSL clients and servers which enable client authentication.
189 *) ASN.1 structure reuse memory corruption fix
191 Reusing a structure in ASN.1 parsing may allow an attacker to cause
192 memory corruption via an invalid write. Such reuse is and has been
193 strongly discouraged and is believed to be rare.
195 Applications that parse structures containing CHOICE or ANY DEFINED BY
196 components may be affected. Certificate parsing (d2i_X509 and related
197 functions) are however not affected. OpenSSL clients and servers are
202 *) PKCS7 NULL pointer dereferences fix
204 The PKCS#7 parsing code does not handle missing outer ContentInfo
205 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
206 missing content and trigger a NULL pointer dereference on parsing.
208 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
209 otherwise parse PKCS#7 structures from untrusted sources are
210 affected. OpenSSL clients and servers are not affected.
212 This issue was reported to OpenSSL by Michal Zalewski (Google).
216 *) DoS via reachable assert in SSLv2 servers fix
218 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
219 servers that both support SSLv2 and enable export cipher suites by sending
220 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
222 This issue was discovered by Sean Burford (Google) and Emilia Käsper
223 (OpenSSL development team).
227 *) Use After Free following d2i_ECPrivatekey error fix
229 A malformed EC private key file consumed via the d2i_ECPrivateKey function
230 could cause a use after free condition. This, in turn, could cause a double
231 free in several private key parsing functions (such as d2i_PrivateKey
232 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
233 for applications that receive EC private keys from untrusted
234 sources. This scenario is considered rare.
236 This issue was discovered by the BoringSSL project and fixed in their
241 *) X509_to_X509_REQ NULL pointer deref fix
243 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
244 the certificate key is invalid. This function is rarely used in practice.
246 This issue was discovered by Brian Carpenter.
250 *) Removed the export ciphers from the DEFAULT ciphers
253 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
255 *) Build fixes for the Windows and OpenVMS platforms
256 [Matt Caswell and Richard Levitte]
258 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
260 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
261 message can cause a segmentation fault in OpenSSL due to a NULL pointer
262 dereference. This could lead to a Denial Of Service attack. Thanks to
263 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
267 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
268 dtls1_buffer_record function under certain conditions. In particular this
269 could occur if an attacker sent repeated DTLS records with the same
270 sequence number but for the next epoch. The memory leak could be exploited
271 by an attacker in a Denial of Service attack through memory exhaustion.
272 Thanks to Chris Mueller for reporting this issue.
276 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
277 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
278 method would be set to NULL which could later result in a NULL pointer
279 dereference. Thanks to Frank Schmirler for reporting this issue.
283 *) Abort handshake if server key exchange message is omitted for ephemeral
286 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
287 reporting this issue.
291 *) Remove non-export ephemeral RSA code on client and server. This code
292 violated the TLS standard by allowing the use of temporary RSA keys in
293 non-export ciphersuites and could be used by a server to effectively
294 downgrade the RSA key length used to a value smaller than the server
295 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
296 INRIA or reporting this issue.
300 *) Fixed issue where DH client certificates are accepted without verification.
301 An OpenSSL server will accept a DH certificate for client authentication
302 without the certificate verify message. This effectively allows a client to
303 authenticate without the use of a private key. This only affects servers
304 which trust a client certificate authority which issues certificates
305 containing DH keys: these are extremely rare and hardly ever encountered.
306 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
311 *) Ensure that the session ID context of an SSL is updated when its
312 SSL_CTX is updated via SSL_set_SSL_CTX.
314 The session ID context is typically set from the parent SSL_CTX,
315 and can vary with the CTX.
318 *) Fix various certificate fingerprint issues.
320 By using non-DER or invalid encodings outside the signed portion of a
321 certificate the fingerprint can be changed without breaking the signature.
322 Although no details of the signed portion of the certificate can be changed
323 this can cause problems with some applications: e.g. those using the
324 certificate fingerprint for blacklists.
326 1. Reject signatures with non zero unused bits.
328 If the BIT STRING containing the signature has non zero unused bits reject
329 the signature. All current signature algorithms require zero unused bits.
331 2. Check certificate algorithm consistency.
333 Check the AlgorithmIdentifier inside TBS matches the one in the
334 certificate signature. NB: this will result in signature failure
335 errors for some broken certificates.
337 Thanks to Konrad Kraszewski from Google for reporting this issue.
339 3. Check DSA/ECDSA signatures use DER.
341 Reencode DSA/ECDSA signatures and compare with the original received
342 signature. Return an error if there is a mismatch.
344 This will reject various cases including garbage after signature
345 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
346 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
347 (negative or with leading zeroes).
349 Further analysis was conducted and fixes were developed by Stephen Henson
350 of the OpenSSL core team.
355 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
356 results on some platforms, including x86_64. This bug occurs at random
357 with a very low probability, and is not known to be exploitable in any
358 way, though its exact impact is difficult to determine. Thanks to Pieter
359 Wuille (Blockstream) who reported this issue and also suggested an initial
360 fix. Further analysis was conducted by the OpenSSL development team and
361 Adam Langley of Google. The final fix was developed by Andy Polyakov of
362 the OpenSSL core team.
366 *) Do not resume sessions on the server if the negotiated protocol
367 version does not match the session's version. Resuming with a different
368 version, while not strictly forbidden by the RFC, is of questionable
369 sanity and breaks all known clients.
370 [David Benjamin, Emilia Käsper]
372 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
373 early CCS messages during renegotiation. (Note that because
374 renegotiation is encrypted, this early CCS was not exploitable.)
377 *) Tighten client-side session ticket handling during renegotiation:
378 ensure that the client only accepts a session ticket if the server sends
379 the extension anew in the ServerHello. Previously, a TLS client would
380 reuse the old extension state and thus accept a session ticket if one was
381 announced in the initial ServerHello.
383 Similarly, ensure that the client requires a session ticket if one
384 was advertised in the ServerHello. Previously, a TLS client would
385 ignore a missing NewSessionTicket message.
388 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
392 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
393 sends a carefully crafted handshake message, to cause OpenSSL to fail
394 to free up to 64k of memory causing a memory leak. This could be
395 exploited in a Denial Of Service attack. This issue affects OpenSSL
396 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
397 whether SRTP is used or configured. Implementations of OpenSSL that
398 have been compiled with OPENSSL_NO_SRTP defined are not affected.
400 The fix was developed by the OpenSSL team.
404 *) Session Ticket Memory Leak.
406 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
407 integrity of that ticket is first verified. In the event of a session
408 ticket integrity check failing, OpenSSL will fail to free memory
409 causing a memory leak. By sending a large number of invalid session
410 tickets an attacker could exploit this issue in a Denial Of Service
415 *) Build option no-ssl3 is incomplete.
417 When OpenSSL is configured with "no-ssl3" as a build option, servers
418 could accept and complete a SSL 3.0 handshake, and clients could be
419 configured to send them.
421 [Akamai and the OpenSSL team]
423 *) Add support for TLS_FALLBACK_SCSV.
424 Client applications doing fallback retries should call
425 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
427 [Adam Langley, Bodo Moeller]
429 *) Add additional DigestInfo checks.
431 Reencode DigestInto in DER and check against the original when
432 verifying RSA signature: this will reject any improperly encoded
433 DigestInfo structures.
435 Note: this is a precautionary measure and no attacks are currently known.
439 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
441 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
442 SRP code can be overrun an internal buffer. Add sanity check that
443 g, A, B < N to SRP code.
445 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
446 Group for discovering this issue.
450 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
451 TLS 1.0 instead of higher protocol versions when the ClientHello message
452 is badly fragmented. This allows a man-in-the-middle attacker to force a
453 downgrade to TLS 1.0 even if both the server and the client support a
454 higher protocol version, by modifying the client's TLS records.
456 Thanks to David Benjamin and Adam Langley (Google) for discovering and
457 researching this issue.
461 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
462 to a denial of service attack. A malicious server can crash the client
463 with a null pointer dereference (read) by specifying an anonymous (EC)DH
464 ciphersuite and sending carefully crafted handshake messages.
466 Thanks to Felix Gröbert (Google) for discovering and researching this
471 *) By sending carefully crafted DTLS packets an attacker could cause openssl
472 to leak memory. This can be exploited through a Denial of Service attack.
473 Thanks to Adam Langley for discovering and researching this issue.
477 *) An attacker can force openssl to consume large amounts of memory whilst
478 processing DTLS handshake messages. This can be exploited through a
479 Denial of Service attack.
480 Thanks to Adam Langley for discovering and researching this issue.
484 *) An attacker can force an error condition which causes openssl to crash
485 whilst processing DTLS packets due to memory being freed twice. This
486 can be exploited through a Denial of Service attack.
487 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
492 *) If a multithreaded client connects to a malicious server using a resumed
493 session and the server sends an ec point format extension it could write
494 up to 255 bytes to freed memory.
496 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
501 *) A malicious server can crash an OpenSSL client with a null pointer
502 dereference (read) by specifying an SRP ciphersuite even though it was not
503 properly negotiated with the client. This can be exploited through a
504 Denial of Service attack.
506 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
507 discovering and researching this issue.
511 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
512 X509_name_oneline, X509_name_print_ex et al. to leak some information
513 from the stack. Applications may be affected if they echo pretty printing
514 output to the attacker.
516 Thanks to Ivan Fratric (Google) for discovering this issue.
518 [Emilia Käsper, and Steve Henson]
520 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
521 for corner cases. (Certain input points at infinity could lead to
522 bogus results, with non-infinity inputs mapped to infinity too.)
525 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
527 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
528 handshake can force the use of weak keying material in OpenSSL
529 SSL/TLS clients and servers.
531 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
532 researching this issue. (CVE-2014-0224)
533 [KIKUCHI Masashi, Steve Henson]
535 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
536 OpenSSL DTLS client the code can be made to recurse eventually crashing
539 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
541 [Imre Rad, Steve Henson]
543 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
544 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
545 client or server. This is potentially exploitable to run arbitrary
546 code on a vulnerable client or server.
548 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
549 [Jüri Aedla, Steve Henson]
551 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
552 are subject to a denial of service attack.
554 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
555 this issue. (CVE-2014-3470)
556 [Felix Gröbert, Ivan Fratric, Steve Henson]
558 *) Harmonize version and its documentation. -f flag is used to display
560 [mancha <mancha1@zoho.com>]
562 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
564 [mancha <mancha1@zoho.com>]
566 *) Fix some double frees. These are not thought to be exploitable.
567 [mancha <mancha1@zoho.com>]
569 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
571 *) A missing bounds check in the handling of the TLS heartbeat extension
572 can be used to reveal up to 64k of memory to a connected client or
575 Thanks for Neel Mehta of Google Security for discovering this bug and to
576 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
577 preparing the fix (CVE-2014-0160)
578 [Adam Langley, Bodo Moeller]
580 *) Fix for the attack described in the paper "Recovering OpenSSL
581 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
582 by Yuval Yarom and Naomi Benger. Details can be obtained from:
583 http://eprint.iacr.org/2014/140
585 Thanks to Yuval Yarom and Naomi Benger for discovering this
586 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
587 [Yuval Yarom and Naomi Benger]
589 *) TLS pad extension: draft-agl-tls-padding-03
591 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
592 TLS client Hello record length value would otherwise be > 255 and
593 less that 512 pad with a dummy extension containing zeroes so it
594 is at least 512 bytes long.
596 [Adam Langley, Steve Henson]
598 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
600 *) Fix for TLS record tampering bug. A carefully crafted invalid
601 handshake could crash OpenSSL with a NULL pointer exception.
602 Thanks to Anton Johansson for reporting this issues.
605 *) Keep original DTLS digest and encryption contexts in retransmission
606 structures so we can use the previous session parameters if they need
607 to be resent. (CVE-2013-6450)
610 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
611 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
612 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
613 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
614 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
615 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
616 [Rob Stradling, Adam Langley]
618 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
620 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
621 supporting platforms or when small records were transferred.
622 [Andy Polyakov, Steve Henson]
624 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
626 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
628 This addresses the flaw in CBC record processing discovered by
629 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
630 at: http://www.isg.rhul.ac.uk/tls/
632 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
633 Security Group at Royal Holloway, University of London
634 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
635 Emilia Käsper for the initial patch.
637 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
639 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
640 ciphersuites which can be exploited in a denial of service attack.
641 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
642 and detecting this bug and to Wolfgang Ettlinger
643 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
647 *) Return an error when checking OCSP signatures when key is NULL.
648 This fixes a DoS attack. (CVE-2013-0166)
651 *) Make openssl verify return errors.
652 [Chris Palmer <palmer@google.com> and Ben Laurie]
654 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
655 the right response is stapled. Also change SSL_get_certificate()
656 so it returns the certificate actually sent.
657 See http://rt.openssl.org/Ticket/Display.html?id=2836.
658 [Rob Stradling <rob.stradling@comodo.com>]
660 *) Fix possible deadlock when decoding public keys.
663 *) Don't use TLS 1.0 record version number in initial client hello
667 Changes between 1.0.1b and 1.0.1c [10 May 2012]
669 *) Sanity check record length before skipping explicit IV in TLS
670 1.2, 1.1 and DTLS to fix DoS attack.
672 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
673 fuzzing as a service testing platform.
677 *) Initialise tkeylen properly when encrypting CMS messages.
678 Thanks to Solar Designer of Openwall for reporting this issue.
681 *) In FIPS mode don't try to use composite ciphers as they are not
685 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
687 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
688 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
689 mean any application compiled against OpenSSL 1.0.0 headers setting
690 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
691 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
692 0x10000000L Any application which was previously compiled against
693 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
694 will need to be recompiled as a result. Letting be results in
695 inability to disable specifically TLS 1.1 and in client context,
696 in unlike event, limit maximum offered version to TLS 1.0 [see below].
699 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
700 disable just protocol X, but all protocols above X *if* there are
701 protocols *below* X still enabled. In more practical terms it means
702 that if application wants to disable TLS1.0 in favor of TLS1.1 and
703 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
704 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
708 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
710 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
711 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
712 in CRYPTO_realloc_clean.
714 Thanks to Tavis Ormandy, Google Security Team, for discovering this
715 issue and to Adam Langley <agl@chromium.org> for fixing it.
717 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
719 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
722 *) Workarounds for some broken servers that "hang" if a client hello
723 record length exceeds 255 bytes.
725 1. Do not use record version number > TLS 1.0 in initial client
726 hello: some (but not all) hanging servers will now work.
727 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
728 the number of ciphers sent in the client hello. This should be
729 set to an even number, such as 50, for example by passing:
730 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
731 Most broken servers should now work.
732 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
733 TLS 1.2 client support entirely.
736 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
739 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
741 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
742 STRING form instead of a DigestInfo.
745 *) The format used for MDC2 RSA signatures is inconsistent between EVP
746 and the RSA_sign/RSA_verify functions. This was made more apparent when
747 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
748 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
749 the correct format in RSA_verify so both forms transparently work.
752 *) Some servers which support TLS 1.0 can choke if we initially indicate
753 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
754 encrypted premaster secret. As a workaround use the maximum pemitted
755 client version in client hello, this should keep such servers happy
756 and still work with previous versions of OpenSSL.
759 *) Add support for TLS/DTLS heartbeats.
760 [Robin Seggelmann <seggelmann@fh-muenster.de>]
762 *) Add support for SCTP.
763 [Robin Seggelmann <seggelmann@fh-muenster.de>]
765 *) Improved PRNG seeding for VOS.
766 [Paul Green <Paul.Green@stratus.com>]
768 *) Extensive assembler packs updates, most notably:
770 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
771 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
772 - x86_64: bit-sliced AES implementation;
773 - ARM: NEON support, contemporary platforms optimizations;
774 - s390x: z196 support;
775 - *: GHASH and GF(2^m) multiplication implementations;
779 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
780 (removal of unnecessary code)
781 [Peter Sylvester <peter.sylvester@edelweb.fr>]
783 *) Add TLS key material exporter from RFC 5705.
786 *) Add DTLS-SRTP negotiation from RFC 5764.
789 *) Add Next Protocol Negotiation,
790 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
791 disabled with a no-npn flag to config or Configure. Code donated
793 [Adam Langley <agl@google.com> and Ben Laurie]
795 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
796 NIST-P256, NIST-P521, with constant-time single point multiplication on
797 typical inputs. Compiler support for the nonstandard type __uint128_t is
798 required to use this (present in gcc 4.4 and later, for 64-bit builds).
799 Code made available under Apache License version 2.0.
801 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
802 line to include this in your build of OpenSSL, and run "make depend" (or
803 "make update"). This enables the following EC_METHODs:
805 EC_GFp_nistp224_method()
806 EC_GFp_nistp256_method()
807 EC_GFp_nistp521_method()
809 EC_GROUP_new_by_curve_name() will automatically use these (while
810 EC_GROUP_new_curve_GFp() currently prefers the more flexible
812 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
814 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
815 all platforms. Move ssize_t definition from e_os.h to the public
816 header file e_os2.h as it now appears in public header file cms.h
819 *) New -sigopt option to the ca, req and x509 utilities. Additional
820 signature parameters can be passed using this option and in
824 *) Add RSA PSS signing function. This will generate and set the
825 appropriate AlgorithmIdentifiers for PSS based on those in the
826 corresponding EVP_MD_CTX structure. No application support yet.
829 *) Support for companion algorithm specific ASN1 signing routines.
830 New function ASN1_item_sign_ctx() signs a pre-initialised
831 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
832 the appropriate parameters.
835 *) Add new algorithm specific ASN1 verification initialisation function
836 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
837 handling will be the same no matter what EVP_PKEY_METHOD is used.
838 Add a PSS handler to support verification of PSS signatures: checked
839 against a number of sample certificates.
842 *) Add signature printing for PSS. Add PSS OIDs.
843 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
845 *) Add algorithm specific signature printing. An individual ASN1 method
846 can now print out signatures instead of the standard hex dump.
848 More complex signatures (e.g. PSS) can print out more meaningful
849 information. Include DSA version that prints out the signature
853 *) Password based recipient info support for CMS library: implementing
857 *) Split password based encryption into PBES2 and PBKDF2 functions. This
858 neatly separates the code into cipher and PBE sections and is required
859 for some algorithms that split PBES2 into separate pieces (such as
863 *) Session-handling fixes:
864 - Fix handling of connections that are resuming with a session ID,
865 but also support Session Tickets.
866 - Fix a bug that suppressed issuing of a new ticket if the client
867 presented a ticket with an expired session.
868 - Try to set the ticket lifetime hint to something reasonable.
869 - Make tickets shorter by excluding irrelevant information.
870 - On the client side, don't ignore renewed tickets.
871 [Adam Langley, Bodo Moeller (Google)]
873 *) Fix PSK session representation.
876 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
878 This work was sponsored by Intel.
881 *) Add GCM support to TLS library. Some custom code is needed to split
882 the IV between the fixed (from PRF) and explicit (from TLS record)
883 portions. This adds all GCM ciphersuites supported by RFC5288 and
884 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
885 add a special AESGCM string for GCM only.
888 *) Expand range of ctrls for AES GCM. Permit setting invocation
889 field on decrypt and retrieval of invocation field only on encrypt.
892 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
893 As required by RFC5289 these ciphersuites cannot be used if for
894 versions of TLS earlier than 1.2.
897 *) For FIPS capable OpenSSL interpret a NULL default public key method
898 as unset and return the appopriate default but do *not* set the default.
899 This means we can return the appopriate method in applications that
900 swicth between FIPS and non-FIPS modes.
903 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
904 ENGINE is used then we cannot handle that in the FIPS module so we
905 keep original code iff non-FIPS operations are allowed.
908 *) Add -attime option to openssl utilities.
909 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
911 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
914 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
915 FIPS EC methods unconditionally for now.
918 *) New build option no-ec2m to disable characteristic 2 code.
921 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
922 all cases can be covered as some introduce binary incompatibilities.
925 *) Redirect RSA operations to FIPS module including keygen,
926 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
929 *) Add similar low level API blocking to ciphers.
932 *) Low level digest APIs are not approved in FIPS mode: any attempt
933 to use these will cause a fatal error. Applications that *really* want
934 to use them can use the private_* version instead.
937 *) Redirect cipher operations to FIPS module for FIPS builds.
940 *) Redirect digest operations to FIPS module for FIPS builds.
943 *) Update build system to add "fips" flag which will link in fipscanister.o
944 for static and shared library builds embedding a signature if needed.
947 *) Output TLS supported curves in preference order instead of numerical
948 order. This is currently hardcoded for the highest order curves first.
949 This should be configurable so applications can judge speed vs strength.
952 *) Add TLS v1.2 server support for client authentication.
955 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
959 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
960 FIPS modules versions.
963 *) Add TLS v1.2 client side support for client authentication. Keep cache
964 of handshake records longer as we don't know the hash algorithm to use
965 until after the certificate request message is received.
968 *) Initial TLS v1.2 client support. Add a default signature algorithms
969 extension including all the algorithms we support. Parse new signature
970 format in client key exchange. Relax some ECC signing restrictions for
971 TLS v1.2 as indicated in RFC5246.
974 *) Add server support for TLS v1.2 signature algorithms extension. Switch
975 to new signature format when needed using client digest preference.
976 All server ciphersuites should now work correctly in TLS v1.2. No client
977 support yet and no support for client certificates.
980 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
981 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
982 ciphersuites. At present only RSA key exchange ciphersuites work with
983 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
984 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
985 and version checking.
988 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
989 with this defined it will not be affected by any changes to ssl internal
990 structures. Add several utility functions to allow openssl application
991 to work with OPENSSL_NO_SSL_INTERN defined.
995 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
997 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
1000 *) Permit abbreviated handshakes when renegotiating using the function
1001 SSL_renegotiate_abbreviated().
1002 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1004 *) Add call to ENGINE_register_all_complete() to
1005 ENGINE_load_builtin_engines(), so some implementations get used
1006 automatically instead of needing explicit application support.
1009 *) Add support for TLS key exporter as described in RFC5705.
1010 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
1012 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
1013 a few changes are required:
1015 Add SSL_OP_NO_TLSv1_1 flag.
1016 Add TLSv1_1 methods.
1017 Update version checking logic to handle version 1.1.
1018 Add explicit IV handling (ported from DTLS code).
1019 Add command line options to s_client/s_server.
1022 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
1024 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1025 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1026 content decryption and always return the same error. Note: this attack
1027 needs on average 2^20 messages so it only affects automated senders. The
1028 old behaviour can be reenabled in the CMS code by setting the
1029 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1030 an MMA defence is not necessary.
1031 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1032 this issue. (CVE-2012-0884)
1035 *) Fix CVE-2011-4619: make sure we really are receiving a
1036 client hello before rejecting multiple SGC restarts. Thanks to
1037 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1040 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
1042 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1043 Thanks to Antonio Martin, Enterprise Secure Access Research and
1044 Development, Cisco Systems, Inc. for discovering this bug and
1045 preparing a fix. (CVE-2012-0050)
1048 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
1050 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1051 of the Vaudenay padding oracle attack on CBC mode encryption
1052 which enables an efficient plaintext recovery attack against
1053 the OpenSSL implementation of DTLS. Their attack exploits timing
1054 differences arising during decryption processing. A research
1055 paper describing this attack can be found at:
1056 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1057 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1058 Security Group at Royal Holloway, University of London
1059 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1060 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1061 for preparing the fix. (CVE-2011-4108)
1062 [Robin Seggelmann, Michael Tuexen]
1064 *) Clear bytes used for block padding of SSL 3.0 records.
1066 [Adam Langley (Google)]
1068 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1069 Kadianakis <desnacked@gmail.com> for discovering this issue and
1070 Adam Langley for preparing the fix. (CVE-2011-4619)
1071 [Adam Langley (Google)]
1073 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
1074 [Andrey Kulikov <amdeich@gmail.com>]
1076 *) Prevent malformed RFC3779 data triggering an assertion failure.
1077 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1078 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1079 [Rob Austein <sra@hactrn.net>]
1081 *) Improved PRNG seeding for VOS.
1082 [Paul Green <Paul.Green@stratus.com>]
1084 *) Fix ssl_ciph.c set-up race.
1085 [Adam Langley (Google)]
1087 *) Fix spurious failures in ecdsatest.c.
1088 [Emilia Käsper (Google)]
1090 *) Fix the BIO_f_buffer() implementation (which was mixing different
1091 interpretations of the '..._len' fields).
1092 [Adam Langley (Google)]
1094 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1095 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1096 threads won't reuse the same blinding coefficients.
1098 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1099 lock to call BN_BLINDING_invert_ex, and avoids one use of
1100 BN_BLINDING_update for each BN_BLINDING structure (previously,
1101 the last update always remained unused).
1102 [Emilia Käsper (Google)]
1104 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
1105 [Bob Buckholz (Google)]
1107 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
1109 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
1110 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
1111 [Kaspar Brand <ossl@velox.ch>]
1113 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1114 for multi-threaded use of ECDH. (CVE-2011-3210)
1115 [Adam Langley (Google)]
1117 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1120 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
1121 signature public key algorithm by using OID xref utilities instead.
1122 Before this you could only use some ECC ciphersuites with SHA1 only.
1125 *) Add protection against ECDSA timing attacks as mentioned in the paper
1126 by Billy Bob Brumley and Nicola Tuveri, see:
1128 http://eprint.iacr.org/2011/232.pdf
1130 [Billy Bob Brumley and Nicola Tuveri]
1132 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
1134 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1135 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1137 *) Fix bug in string printing code: if *any* escaping is enabled we must
1138 escape the escape character (backslash) or the resulting string is
1142 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
1144 *) Disable code workaround for ancient and obsolete Netscape browsers
1145 and servers: an attacker can use it in a ciphersuite downgrade attack.
1146 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1149 *) Fixed J-PAKE implementation error, originally discovered by
1150 Sebastien Martini, further info and confirmation from Stefan
1151 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1154 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
1156 *) Fix extension code to avoid race conditions which can result in a buffer
1157 overrun vulnerability: resumed sessions must not be modified as they can
1158 be shared by multiple threads. CVE-2010-3864
1161 *) Fix WIN32 build system to correctly link an ENGINE directory into
1165 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
1167 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
1169 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
1171 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
1173 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
1174 context. The operation can be customised via the ctrl mechanism in
1175 case ENGINEs want to include additional functionality.
1178 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
1181 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
1182 output hashes compatible with older versions of OpenSSL.
1183 [Willy Weisz <weisz@vcpc.univie.ac.at>]
1185 *) Fix compression algorithm handling: if resuming a session use the
1186 compression algorithm of the resumed session instead of determining
1187 it from client hello again. Don't allow server to change algorithm.
1190 *) Add load_crls() function to apps tidying load_certs() too. Add option
1191 to verify utility to allow additional CRLs to be included.
1194 *) Update OCSP request code to permit adding custom headers to the request:
1195 some responders need this.
1198 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
1200 [Julia Lawall <julia@diku.dk>]
1202 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
1203 needlessly dereferenced structures, used obsolete functions and
1204 didn't handle all updated verify codes correctly.
1207 *) Disable MD2 in the default configuration.
1210 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
1211 indicate the initial BIO being pushed or popped. This makes it possible
1212 to determine whether the BIO is the one explicitly called or as a result
1213 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
1214 it handles reference counts correctly and doesn't zero out the I/O bio
1215 when it is not being explicitly popped. WARNING: applications which
1216 included workarounds for the old buggy behaviour will need to be modified
1217 or they could free up already freed BIOs.
1220 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
1221 renaming to all platforms (within the 0.9.8 branch, this was
1222 done conditionally on Netware platforms to avoid a name clash).
1223 [Guenter <lists@gknw.net>]
1225 *) Add ECDHE and PSK support to DTLS.
1226 [Michael Tuexen <tuexen@fh-muenster.de>]
1228 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
1232 *) Add "missing" function EVP_MD_flags() (without this the only way to
1233 retrieve a digest flags is by accessing the structure directly. Update
1234 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
1235 or cipher is registered as in the "from" argument. Print out all
1236 registered digests in the dgst usage message instead of manually
1237 attempting to work them out.
1240 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
1241 this allows the use of compression and extensions. Change default cipher
1242 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
1243 by default unless an application cipher string requests it.
1246 *) Alter match criteria in PKCS12_parse(). It used to try to use local
1247 key ids to find matching certificates and keys but some PKCS#12 files
1248 don't follow the (somewhat unwritten) rules and this strategy fails.
1249 Now just gather all certificates together and the first private key
1250 then look for the first certificate that matches the key.
1253 *) Support use of registered digest and cipher names for dgst and cipher
1254 commands instead of having to add each one as a special case. So now
1261 openssl dgst -sha256 foo
1263 and this works for ENGINE based algorithms too.
1267 *) Update Gost ENGINE to support parameter files.
1268 [Victor B. Wagner <vitus@cryptocom.ru>]
1270 *) Support GeneralizedTime in ca utility.
1271 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
1273 *) Enhance the hash format used for certificate directory links. The new
1274 form uses the canonical encoding (meaning equivalent names will work
1275 even if they aren't identical) and uses SHA1 instead of MD5. This form
1276 is incompatible with the older format and as a result c_rehash should
1277 be used to rebuild symbolic links.
1280 *) Make PKCS#8 the default write format for private keys, replacing the
1281 traditional format. This form is standardised, more secure and doesn't
1282 include an implicit MD5 dependency.
1285 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
1286 committed to OpenSSL should pass this lot as a minimum.
1289 *) Add session ticket override functionality for use by EAP-FAST.
1290 [Jouni Malinen <j@w1.fi>]
1292 *) Modify HMAC functions to return a value. Since these can be implemented
1293 in an ENGINE errors can occur.
1296 *) Type-checked OBJ_bsearch_ex.
1299 *) Type-checked OBJ_bsearch. Also some constification necessitated
1300 by type-checking. Still to come: TXT_DB, bsearch(?),
1301 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
1305 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
1306 seconds to a tm structure directly, instead of going through OS
1307 specific date routines. This avoids any issues with OS routines such
1308 as the year 2038 bug. New *_adj() functions for ASN1 time structures
1309 and X509_time_adj_ex() to cover the extended range. The existing
1310 X509_time_adj() is still usable and will no longer have any date issues.
1313 *) Delta CRL support. New use deltas option which will attempt to locate
1314 and search any appropriate delta CRLs available.
1316 This work was sponsored by Google.
1319 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
1320 code and add additional score elements. Validate alternate CRL paths
1321 as part of the CRL checking and indicate a new error "CRL path validation
1322 error" in this case. Applications wanting additional details can use
1323 the verify callback and check the new "parent" field. If this is not
1324 NULL CRL path validation is taking place. Existing applications wont
1325 see this because it requires extended CRL support which is off by
1328 This work was sponsored by Google.
1331 *) Support for freshest CRL extension.
1333 This work was sponsored by Google.
1336 *) Initial indirect CRL support. Currently only supported in the CRLs
1337 passed directly and not via lookup. Process certificate issuer
1338 CRL entry extension and lookup CRL entries by bother issuer name
1339 and serial number. Check and process CRL issuer entry in IDP extension.
1341 This work was sponsored by Google.
1344 *) Add support for distinct certificate and CRL paths. The CRL issuer
1345 certificate is validated separately in this case. Only enabled if
1346 an extended CRL support flag is set: this flag will enable additional
1347 CRL functionality in future.
1349 This work was sponsored by Google.
1352 *) Add support for policy mappings extension.
1354 This work was sponsored by Google.
1357 *) Fixes to pathlength constraint, self issued certificate handling,
1358 policy processing to align with RFC3280 and PKITS tests.
1360 This work was sponsored by Google.
1363 *) Support for name constraints certificate extension. DN, email, DNS
1364 and URI types are currently supported.
1366 This work was sponsored by Google.
1369 *) To cater for systems that provide a pointer-based thread ID rather
1370 than numeric, deprecate the current numeric thread ID mechanism and
1371 replace it with a structure and associated callback type. This
1372 mechanism allows a numeric "hash" to be extracted from a thread ID in
1373 either case, and on platforms where pointers are larger than 'long',
1374 mixing is done to help ensure the numeric 'hash' is usable even if it
1375 can't be guaranteed unique. The default mechanism is to use "&errno"
1376 as a pointer-based thread ID to distinguish between threads.
1378 Applications that want to provide their own thread IDs should now use
1379 CRYPTO_THREADID_set_callback() to register a callback that will call
1380 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
1382 Note that ERR_remove_state() is now deprecated, because it is tied
1383 to the assumption that thread IDs are numeric. ERR_remove_state(0)
1384 to free the current thread's error state should be replaced by
1385 ERR_remove_thread_state(NULL).
1387 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
1388 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
1389 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
1390 application was previously providing a numeric thread callback that
1391 was inappropriate for distinguishing threads, then uniqueness might
1392 have been obtained with &errno that happened immediately in the
1393 intermediate development versions of OpenSSL; this is no longer the
1394 case, the numeric thread callback will now override the automatic use
1396 [Geoff Thorpe, with help from Bodo Moeller]
1398 *) Initial support for different CRL issuing certificates. This covers a
1399 simple case where the self issued certificates in the chain exist and
1400 the real CRL issuer is higher in the existing chain.
1402 This work was sponsored by Google.
1405 *) Removed effectively defunct crypto/store from the build.
1408 *) Revamp of STACK to provide stronger type-checking. Still to come:
1409 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
1410 ASN1_STRING, CONF_VALUE.
1413 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
1414 RAM on SSL connections. This option can save about 34k per idle SSL.
1417 *) Revamp of LHASH to provide stronger type-checking. Still to come:
1418 STACK, TXT_DB, bsearch, qsort.
1421 *) Initial support for Cryptographic Message Syntax (aka CMS) based
1422 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
1423 support for data, signedData, compressedData, digestedData and
1424 encryptedData, envelopedData types included. Scripts to check against
1425 RFC4134 examples draft and interop and consistency checks of many
1426 content types and variants.
1429 *) Add options to enc utility to support use of zlib compression BIO.
1432 *) Extend mk1mf to support importing of options and assembly language
1433 files from Configure script, currently only included in VC-WIN32.
1434 The assembly language rules can now optionally generate the source
1435 files from the associated perl scripts.
1438 *) Implement remaining functionality needed to support GOST ciphersuites.
1439 Interop testing has been performed using CryptoPro implementations.
1440 [Victor B. Wagner <vitus@cryptocom.ru>]
1442 *) s390x assembler pack.
1445 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
1449 *) Implement Opaque PRF Input TLS extension as specified in
1450 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
1451 official specification yet and no extension type assignment by
1452 IANA exists, this extension (for now) will have to be explicitly
1453 enabled when building OpenSSL by providing the extension number
1454 to use. For example, specify an option
1456 -DTLSEXT_TYPE_opaque_prf_input=0x9527
1458 to the "config" or "Configure" script to enable the extension,
1459 assuming extension number 0x9527 (which is a completely arbitrary
1460 and unofficial assignment based on the MD5 hash of the Internet
1461 Draft). Note that by doing so, you potentially lose
1462 interoperability with other TLS implementations since these might
1463 be using the same extension number for other purposes.
1465 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
1466 opaque PRF input value to use in the handshake. This will create
1467 an interal copy of the length-'len' string at 'src', and will
1468 return non-zero for success.
1470 To get more control and flexibility, provide a callback function
1473 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
1474 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
1478 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
1481 Callback function 'cb' will be called in handshakes, and is
1482 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
1483 Argument 'arg' is for application purposes (the value as given to
1484 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
1485 be provided to the callback function). The callback function
1486 has to return non-zero to report success: usually 1 to use opaque
1487 PRF input just if possible, or 2 to enforce use of the opaque PRF
1488 input. In the latter case, the library will abort the handshake
1489 if opaque PRF input is not successfully negotiated.
1491 Arguments 'peerinput' and 'len' given to the callback function
1492 will always be NULL and 0 in the case of a client. A server will
1493 see the client's opaque PRF input through these variables if
1494 available (NULL and 0 otherwise). Note that if the server
1495 provides an opaque PRF input, the length must be the same as the
1496 length of the client's opaque PRF input.
1498 Note that the callback function will only be called when creating
1499 a new session (session resumption can resume whatever was
1500 previously negotiated), and will not be called in SSL 2.0
1501 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
1502 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
1503 for applications that need to enforce opaque PRF input.
1507 *) Update ssl code to support digests other than SHA1+MD5 for handshake
1510 [Victor B. Wagner <vitus@cryptocom.ru>]
1512 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1513 RFC4507bis. The encrypted ticket format is an encrypted encoded
1514 SSL_SESSION structure, that way new session features are automatically
1517 If a client application caches session in an SSL_SESSION structure
1518 support is transparent because tickets are now stored in the encoded
1521 The SSL_CTX structure automatically generates keys for ticket
1522 protection in servers so again support should be possible
1523 with no application modification.
1525 If a client or server wishes to disable RFC4507 support then the option
1526 SSL_OP_NO_TICKET can be set.
1528 Add a TLS extension debugging callback to allow the contents of any client
1529 or server extensions to be examined.
1531 This work was sponsored by Google.
1534 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
1535 OpenSSL should now compile cleanly on gcc 4.2
1536 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
1538 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
1539 support including streaming MAC support: this is required for GOST
1540 ciphersuite support.
1541 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
1543 *) Add option -stream to use PKCS#7 streaming in smime utility. New
1544 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
1545 to output in BER and PEM format.
1548 *) Experimental support for use of HMAC via EVP_PKEY interface. This
1549 allows HMAC to be handled via the EVP_DigestSign*() interface. The
1550 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
1551 ENGINE support for HMAC keys which are unextractable. New -mac and
1552 -macopt options to dgst utility.
1555 *) New option -sigopt to dgst utility. Update dgst to use
1556 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
1557 alternative signing paramaters such as X9.31 or PSS in the dgst
1561 *) Change ssl_cipher_apply_rule(), the internal function that does
1562 the work each time a ciphersuite string requests enabling
1563 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
1564 removing ("!foo+bar") a class of ciphersuites: Now it maintains
1565 the order of disabled ciphersuites such that those ciphersuites
1566 that most recently went from enabled to disabled not only stay
1567 in order with respect to each other, but also have higher priority
1568 than other disabled ciphersuites the next time ciphersuites are
1571 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
1572 the same ciphersuites as with "HIGH" alone, but in a specific
1573 order where the PSK ciphersuites come first (since they are the
1574 most recently disabled ciphersuites when "HIGH" is parsed).
1576 Also, change ssl_create_cipher_list() (using this new
1577 funcionality) such that between otherwise identical
1578 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
1582 *) Change ssl_create_cipher_list() so that it automatically
1583 arranges the ciphersuites in reasonable order before starting
1584 to process the rule string. Thus, the definition for "DEFAULT"
1585 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
1586 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
1587 This makes it much easier to arrive at a reasonable default order
1588 in applications for which anonymous ciphers are OK (meaning
1589 that you can't actually use DEFAULT).
1590 [Bodo Moeller; suggested by Victor Duchovni]
1592 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
1593 processing) into multiple integers instead of setting
1594 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
1595 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
1596 (These masks as well as the individual bit definitions are hidden
1597 away into the non-exported interface ssl/ssl_locl.h, so this
1598 change to the definition of the SSL_CIPHER structure shouldn't
1599 affect applications.) This give us more bits for each of these
1600 categories, so there is no longer a need to coagulate AES128 and
1601 AES256 into a single algorithm bit, and to coagulate Camellia128
1602 and Camellia256 into a single algorithm bit, which has led to all
1605 Thus, among other things, the kludge introduced in 0.9.7m and
1606 0.9.8e for masking out AES256 independently of AES128 or masking
1607 out Camellia256 independently of AES256 is not needed here in 0.9.9.
1609 With the change, we also introduce new ciphersuite aliases that
1610 so far were missing: "AES128", "AES256", "CAMELLIA128", and
1614 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
1615 Use the leftmost N bytes of the signature input if the input is
1616 larger than the prime q (with N being the size in bytes of q).
1619 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
1620 it yet and it is largely untested.
1623 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
1626 *) Initial incomplete changes to avoid need for function casts in OpenSSL
1627 some compilers (gcc 4.2 and later) reject their use. Safestack is
1628 reimplemented. Update ASN1 to avoid use of legacy functions.
1631 *) Win32/64 targets are linked with Winsock2.
1634 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
1635 to external functions. This can be used to increase CRL handling
1636 efficiency especially when CRLs are very large by (for example) storing
1637 the CRL revoked certificates in a database.
1640 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
1641 new CRLs added to a directory can be used. New command line option
1642 -verify_return_error to s_client and s_server. This causes real errors
1643 to be returned by the verify callback instead of carrying on no matter
1644 what. This reflects the way a "real world" verify callback would behave.
1647 *) GOST engine, supporting several GOST algorithms and public key formats.
1648 Kindly donated by Cryptocom.
1651 *) Partial support for Issuing Distribution Point CRL extension. CRLs
1652 partitioned by DP are handled but no indirect CRL or reason partitioning
1653 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
1654 selected via a scoring technique which handles IDP and AKID in CRLs.
1657 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
1658 will ultimately be used for all verify operations: this will remove the
1659 X509_STORE dependency on certificate verification and allow alternative
1660 lookup methods. X509_STORE based implementations of these two callbacks.
1663 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
1664 Modify get_crl() to find a valid (unexpired) CRL if possible.
1667 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
1668 this would be called X509_CRL_cmp() but that name is already used by
1669 a function that just compares CRL issuer names. Cache several CRL
1670 extensions in X509_CRL structure and cache CRLDP in X509.
1673 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
1674 this maps equivalent X509_NAME structures into a consistent structure.
1675 Name comparison can then be performed rapidly using memcmp().
1678 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
1682 *) Allow digests to supply their own micalg string for S/MIME type using
1683 the ctrl EVP_MD_CTRL_MICALG.
1686 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
1687 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
1688 ctrl. It can then customise the structure before and/or after signing
1692 *) New function OBJ_add_sigid() to allow application defined signature OIDs
1693 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
1694 to free up any added signature OIDs.
1697 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
1698 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
1699 digest and cipher tables. New options added to openssl utility:
1700 list-message-digest-algorithms and list-cipher-algorithms.
1703 *) Change the array representation of binary polynomials: the list
1704 of degrees of non-zero coefficients is now terminated with -1.
1705 Previously it was terminated with 0, which was also part of the
1706 value; thus, the array representation was not applicable to
1707 polynomials where t^0 has coefficient zero. This change makes
1708 the array representation useful in a more general context.
1711 *) Various modifications and fixes to SSL/TLS cipher string
1712 handling. For ECC, the code now distinguishes between fixed ECDH
1713 with RSA certificates on the one hand and with ECDSA certificates
1714 on the other hand, since these are separate ciphersuites. The
1715 unused code for Fortezza ciphersuites has been removed.
1717 For consistency with EDH, ephemeral ECDH is now called "EECDH"
1718 (not "ECDHE"). For consistency with the code for DH
1719 certificates, use of ECDH certificates is now considered ECDH
1720 authentication, not RSA or ECDSA authentication (the latter is
1721 merely the CA's signing algorithm and not actively used in the
1724 The temporary ciphersuite alias "ECCdraft" is no longer
1725 available, and ECC ciphersuites are no longer excluded from "ALL"
1726 and "DEFAULT". The following aliases now exist for RFC 4492
1727 ciphersuites, most of these by analogy with the DH case:
1729 kECDHr - ECDH cert, signed with RSA
1730 kECDHe - ECDH cert, signed with ECDSA
1731 kECDH - ECDH cert (signed with either RSA or ECDSA)
1732 kEECDH - ephemeral ECDH
1733 ECDH - ECDH cert or ephemeral ECDH
1739 AECDH - anonymous ECDH
1740 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
1744 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
1745 Use correct micalg parameters depending on digest(s) in signed message.
1748 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
1749 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
1752 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
1753 an engine to register a method. Add ENGINE lookups for methods and
1754 functional reference processing.
1757 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
1758 EVP_{Sign,Verify}* which allow an application to customise the signature
1762 *) New -resign option to smime utility. This adds one or more signers
1763 to an existing PKCS#7 signedData structure. Also -md option to use an
1764 alternative message digest algorithm for signing.
1767 *) Tidy up PKCS#7 routines and add new functions to make it easier to
1768 create PKCS7 structures containing multiple signers. Update smime
1769 application to support multiple signers.
1772 *) New -macalg option to pkcs12 utility to allow setting of an alternative
1776 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
1777 Reorganize PBE internals to lookup from a static table using NIDs,
1778 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
1779 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
1780 PRF which will be automatically used with PBES2.
1783 *) Replace the algorithm specific calls to generate keys in "req" with the
1787 *) Update PKCS#7 enveloped data routines to use new API. This is now
1788 supported by any public key method supporting the encrypt operation. A
1789 ctrl is added to allow the public key algorithm to examine or modify
1790 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
1794 *) Add a ctrl to asn1 method to allow a public key algorithm to express
1795 a default digest type to use. In most cases this will be SHA1 but some
1796 algorithms (such as GOST) need to specify an alternative digest. The
1797 return value indicates how strong the prefernce is 1 means optional and
1798 2 is mandatory (that is it is the only supported type). Modify
1799 ASN1_item_sign() to accept a NULL digest argument to indicate it should
1800 use the default md. Update openssl utilities to use the default digest
1801 type for signing if it is not explicitly indicated.
1804 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
1805 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
1806 signing method from the key type. This effectively removes the link
1807 between digests and public key types.
1810 *) Add an OID cross reference table and utility functions. Its purpose is to
1811 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
1812 rsaEncryption. This will allow some of the algorithm specific hackery
1813 needed to use the correct OID to be removed.
1816 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
1817 structures for PKCS7_sign(). They are now set up by the relevant public
1821 *) Add provisional EC pkey method with support for ECDSA and ECDH.
1824 *) Add support for key derivation (agreement) in the API, DH method and
1828 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
1829 public and private key formats. As a side effect these add additional
1830 command line functionality not previously available: DSA signatures can be
1831 generated and verified using pkeyutl and DH key support and generation in
1836 [Oliver Tappe <zooey@hirschkaefer.de>]
1838 *) New make target "install_html_docs" installs HTML renditions of the
1840 [Oliver Tappe <zooey@hirschkaefer.de>]
1842 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
1843 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
1844 support key and parameter generation and add initial key generation
1845 functionality for RSA.
1848 *) Add functions for main EVP_PKEY_method operations. The undocumented
1849 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
1850 EVP_PKEY_{encrypt,decrypt}_old.
1853 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
1854 key API, doesn't do much yet.
1857 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
1858 public key algorithms. New option to openssl utility:
1859 "list-public-key-algorithms" to print out info.
1862 *) Implement the Supported Elliptic Curves Extension for
1863 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1866 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
1867 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
1870 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
1871 utilities such as rsa, dsa, dsaparam etc except they process any key
1875 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1876 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1877 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
1881 *) Initial support for pluggable public key ASN1.
1882 De-spaghettify the public key ASN1 handling. Move public and private
1883 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
1884 algorithm specific handling to a single module within the relevant
1885 algorithm directory. Add functions to allow (near) opaque processing
1886 of public and private key structures.
1889 *) Implement the Supported Point Formats Extension for
1890 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1893 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
1894 for the psk identity [hint] and the psk callback functions to the
1895 SSL_SESSION, SSL and SSL_CTX structure.
1898 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
1902 SSL_CTX_use_psk_identity_hint
1903 SSL_get_psk_identity_hint
1904 SSL_get_psk_identity
1905 SSL_use_psk_identity_hint
1907 [Mika Kousa and Pasi Eronen of Nokia Corporation]
1909 *) Add RFC 3161 compliant time stamp request creation, response generation
1910 and response verification functionality.
1911 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
1913 *) Add initial support for TLS extensions, specifically for the server_name
1914 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1915 have new members for a host name. The SSL data structure has an
1916 additional member SSL_CTX *initial_ctx so that new sessions can be
1917 stored in that context to allow for session resumption, even after the
1918 SSL has been switched to a new SSL_CTX in reaction to a client's
1919 server_name extension.
1921 New functions (subject to change):
1923 SSL_get_servername()
1924 SSL_get_servername_type()
1927 New CTRL codes and macros (subject to change):
1929 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1930 - SSL_CTX_set_tlsext_servername_callback()
1931 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1932 - SSL_CTX_set_tlsext_servername_arg()
1933 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1935 openssl s_client has a new '-servername ...' option.
1937 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1938 '-key2 ...', '-servername_fatal' (subject to change). This allows
1939 testing the HostName extension for a specific single host name ('-cert'
1940 and '-key' remain fallbacks for handshakes without HostName
1941 negotiation). If the unrecogninzed_name alert has to be sent, this by
1942 default is a warning; it becomes fatal with the '-servername_fatal'
1945 [Peter Sylvester, Remy Allais, Christophe Renou]
1947 *) Whirlpool hash implementation is added.
1950 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1951 bn(64,32). Because of instruction set limitations it doesn't have
1952 any negative impact on performance. This was done mostly in order
1953 to make it possible to share assembler modules, such as bn_mul_mont
1954 implementations, between 32- and 64-bit builds without hassle.
1957 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1958 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1962 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1963 dedicated Montgomery multiplication procedure, is introduced.
1964 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1965 "64-bit" performance on certain 32-bit targets.
1968 *) New option SSL_OP_NO_COMP to disable use of compression selectively
1969 in SSL structures. New SSL ctrl to set maximum send fragment size.
1970 Save memory by seeting the I/O buffer sizes dynamically instead of
1971 using the maximum available value.
1974 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
1975 in addition to the text details.
1978 *) Very, very preliminary EXPERIMENTAL support for printing of general
1979 ASN1 structures. This currently produces rather ugly output and doesn't
1980 handle several customised structures at all.
1983 *) Integrated support for PVK file format and some related formats such
1984 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
1985 these in the 'rsa' and 'dsa' utilities.
1988 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
1991 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
1992 place for the (very old) "NETSCAPE" format certificates which are now
1993 handled using new ASN1 code equivalents.
1996 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
1997 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
1998 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
2001 *) Modify CRL distribution points extension code to print out previously
2002 unsupported fields. Enhance extension setting code to allow setting of
2006 *) Add print and set support for Issuing Distribution Point CRL extension.
2009 *) Change 'Configure' script to enable Camellia by default.
2012 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
2014 *) When rejecting SSL/TLS records due to an incorrect version number, never
2015 update s->server with a new major version number. As of
2016 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
2017 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
2018 the previous behavior could result in a read attempt at NULL when
2019 receiving specific incorrect SSL/TLS records once record payload
2020 protection is active. (CVE-2010-0740)
2021 [Bodo Moeller, Adam Langley <agl@chromium.org>]
2023 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
2024 could be crashed if the relevant tables were not present (e.g. chrooted).
2025 [Tomas Hoger <thoger@redhat.com>]
2027 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
2029 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
2030 [Martin Olsson, Neel Mehta]
2032 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
2033 accommodate for stack sorting, always a write lock!).
2036 *) On some versions of WIN32 Heap32Next is very slow. This can cause
2037 excessive delays in the RAND_poll(): over a minute. As a workaround
2038 include a time check in the inner Heap32Next loop too.
2041 *) The code that handled flushing of data in SSL/TLS originally used the
2042 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
2043 the problem outlined in PR#1949. The fix suggested there however can
2044 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
2045 of Apache). So instead simplify the code to flush unconditionally.
2046 This should be fine since flushing with no data to flush is a no op.
2049 *) Handle TLS versions 2.0 and later properly and correctly use the
2050 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
2051 off ancient servers have a habit of sticking around for a while...
2054 *) Modify compression code so it frees up structures without using the
2055 ex_data callbacks. This works around a problem where some applications
2056 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
2057 restarting) then use compression (e.g. SSL with compression) later.
2058 This results in significant per-connection memory leaks and
2059 has caused some security issues including CVE-2008-1678 and
2063 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
2064 change when encrypting or decrypting.
2067 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
2068 connect and renegotiate with servers which do not support RI.
2069 Until RI is more widely deployed this option is enabled by default.
2072 *) Add "missing" ssl ctrls to clear options and mode.
2075 *) If client attempts to renegotiate and doesn't support RI respond with
2076 a no_renegotiation alert as required by RFC5746. Some renegotiating
2077 TLS clients will continue a connection gracefully when they receive
2078 the alert. Unfortunately OpenSSL mishandled this alert and would hang
2079 waiting for a server hello which it will never receive. Now we treat a
2080 received no_renegotiation alert as a fatal error. This is because
2081 applications requesting a renegotiation might well expect it to succeed
2082 and would have no code in place to handle the server denying it so the
2083 only safe thing to do is to terminate the connection.
2086 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
2087 peer supports secure renegotiation and 0 otherwise. Print out peer
2088 renegotiation support in s_client/s_server.
2091 *) Replace the highly broken and deprecated SPKAC certification method with
2092 the updated NID creation version. This should correctly handle UTF8.
2095 *) Implement RFC5746. Re-enable renegotiation but require the extension
2096 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
2097 turns out to be a bad idea. It has been replaced by
2098 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
2099 SSL_CTX_set_options(). This is really not recommended unless you
2100 know what you are doing.
2101 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
2103 *) Fixes to stateless session resumption handling. Use initial_ctx when
2104 issuing and attempting to decrypt tickets in case it has changed during
2105 servername handling. Use a non-zero length session ID when attempting
2106 stateless session resumption: this makes it possible to determine if
2107 a resumption has occurred immediately after receiving server hello
2108 (several places in OpenSSL subtly assume this) instead of later in
2112 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
2113 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
2114 fixes for a few places where the return code is not checked
2116 [Julia Lawall <julia@diku.dk>]
2118 *) Add --strict-warnings option to Configure script to include devteam
2119 warnings in other configurations.
2122 *) Add support for --libdir option and LIBDIR variable in makefiles. This
2123 makes it possible to install openssl libraries in locations which
2124 have names other than "lib", for example "/usr/lib64" which some
2126 [Steve Henson, based on patch from Jeremy Utley]
2128 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
2129 X690 8.9.12 and can produce some misleading textual output of OIDs.
2130 [Steve Henson, reported by Dan Kaminsky]
2132 *) Delete MD2 from algorithm tables. This follows the recommendation in
2133 several standards that it is not used in new applications due to
2134 several cryptographic weaknesses. For binary compatibility reasons
2135 the MD2 API is still compiled in by default.
2138 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
2142 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
2143 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
2145 [Guenter <lists@gknw.net>]
2147 *) Fix the server certificate chain building code to use X509_verify_cert(),
2148 it used to have an ad-hoc builder which was unable to cope with anything
2149 other than a simple chain.
2150 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
2152 *) Don't check self signed certificate signatures in X509_verify_cert()
2153 by default (a flag can override this): it just wastes time without
2154 adding any security. As a useful side effect self signed root CAs
2155 with non-FIPS digests are now usable in FIPS mode.
2158 *) In dtls1_process_out_of_seq_message() the check if the current message
2159 is already buffered was missing. For every new message was memory
2160 allocated, allowing an attacker to perform an denial of service attack
2161 with sending out of seq handshake messages until there is no memory
2162 left. Additionally every future messege was buffered, even if the
2163 sequence number made no sense and would be part of another handshake.
2164 So only messages with sequence numbers less than 10 in advance will be
2165 buffered. (CVE-2009-1378)
2166 [Robin Seggelmann, discovered by Daniel Mentz]
2168 *) Records are buffered if they arrive with a future epoch to be
2169 processed after finishing the corresponding handshake. There is
2170 currently no limitation to this buffer allowing an attacker to perform
2171 a DOS attack with sending records with future epochs until there is no
2172 memory left. This patch adds the pqueue_size() function to detemine
2173 the size of a buffer and limits the record buffer to 100 entries.
2175 [Robin Seggelmann, discovered by Daniel Mentz]
2177 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
2178 parent structure is freed. (CVE-2009-1379)
2181 *) Handle non-blocking I/O properly in SSL_shutdown() call.
2182 [Darryl Miles <darryl-mailinglists@netbauds.net>]
2185 [Ilya O. <vrghost@gmail.com>]
2187 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
2189 *) Disable renegotiation completely - this fixes a severe security
2190 problem (CVE-2009-3555) at the cost of breaking all
2191 renegotiation. Renegotiation can be re-enabled by setting
2192 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
2193 run-time. This is really not recommended unless you know what
2197 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
2199 *) Don't set val to NULL when freeing up structures, it is freed up by
2200 underlying code. If sizeof(void *) > sizeof(long) this can result in
2201 zeroing past the valid field. (CVE-2009-0789)
2202 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
2204 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
2205 checked correctly. This would allow some invalid signed attributes to
2206 appear to verify correctly. (CVE-2009-0591)
2207 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2209 *) Reject UniversalString and BMPString types with invalid lengths. This
2210 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
2211 a legal length. (CVE-2009-0590)
2214 *) Set S/MIME signing as the default purpose rather than setting it
2215 unconditionally. This allows applications to override it at the store
2219 *) Permit restricted recursion of ASN1 strings. This is needed in practice
2220 to handle some structures.
2223 *) Improve efficiency of mem_gets: don't search whole buffer each time
2225 [Jeremy Shapiro <jnshapir@us.ibm.com>]
2227 *) New -hex option for openssl rand.
2230 *) Print out UTF8String and NumericString when parsing ASN1.
2233 *) Support NumericString type for name components.
2236 *) Allow CC in the environment to override the automatically chosen
2237 compiler. Note that nothing is done to ensure flags work with the
2241 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
2243 *) Properly check EVP_VerifyFinal() and similar return values
2245 [Ben Laurie, Bodo Moeller, Google Security Team]
2247 *) Enable TLS extensions by default.
2250 *) Allow the CHIL engine to be loaded, whether the application is
2251 multithreaded or not. (This does not release the developer from the
2252 obligation to set up the dynamic locking callbacks.)
2253 [Sander Temme <sander@temme.net>]
2255 *) Use correct exit code if there is an error in dgst command.
2256 [Steve Henson; problem pointed out by Roland Dirlewanger]
2258 *) Tweak Configure so that you need to say "experimental-jpake" to enable
2259 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
2262 *) Add experimental JPAKE support, including demo authentication in
2263 s_client and s_server.
2266 *) Set the comparison function in v3_addr_canonize().
2267 [Rob Austein <sra@hactrn.net>]
2269 *) Add support for XMPP STARTTLS in s_client.
2270 [Philip Paeps <philip@freebsd.org>]
2272 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
2273 to ensure that even with this option, only ciphersuites in the
2274 server's preference list will be accepted. (Note that the option
2275 applies only when resuming a session, so the earlier behavior was
2276 just about the algorithm choice for symmetric cryptography.)
2279 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
2281 *) Fix NULL pointer dereference if a DTLS server received
2282 ChangeCipherSpec as first record (CVE-2009-1386).
2285 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
2286 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
2289 *) The fix in 0.9.8c that supposedly got rid of unsafe
2290 double-checked locking was incomplete for RSA blinding,
2291 addressing just one layer of what turns out to have been
2292 doubly unsafe triple-checked locking.
2294 So now fix this for real by retiring the MONT_HELPER macro
2295 in crypto/rsa/rsa_eay.c.
2297 [Bodo Moeller; problem pointed out by Marius Schilder]
2299 *) Various precautionary measures:
2301 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
2303 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
2304 (NB: This would require knowledge of the secret session ticket key
2305 to exploit, in which case you'd be SOL either way.)
2307 - Change bn_nist.c so that it will properly handle input BIGNUMs
2308 outside the expected range.
2310 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
2313 [Neel Mehta, Bodo Moeller]
2315 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
2316 the load fails. Useful for distros.
2317 [Ben Laurie and the FreeBSD team]
2319 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
2322 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
2325 *) Expand ENGINE to support engine supplied SSL client certificate functions.
2327 This work was sponsored by Logica.
2330 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
2331 keystores. Support for SSL/TLS client authentication too.
2332 Not compiled unless enable-capieng specified to Configure.
2334 This work was sponsored by Logica.
2337 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
2338 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
2339 attribute creation routines such as certifcate requests and PKCS#12
2343 Changes between 0.9.8g and 0.9.8h [28 May 2008]
2345 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
2346 handshake which could lead to a cilent crash as found using the
2347 Codenomicon TLS test suite (CVE-2008-1672)
2348 [Steve Henson, Mark Cox]
2350 *) Fix double free in TLS server name extensions which could lead to
2351 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
2354 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
2356 Clear the error queue to ensure that error entries left from
2357 older function calls do not interfere with the correct operation.
2358 [Lutz Jaenicke, Erik de Castro Lopo]
2360 *) Remove root CA certificates of commercial CAs:
2362 The OpenSSL project does not recommend any specific CA and does not
2363 have any policy with respect to including or excluding any CA.
2364 Therefore it does not make any sense to ship an arbitrary selection
2365 of root CA certificates with the OpenSSL software.
2368 *) RSA OAEP patches to fix two separate invalid memory reads.
2369 The first one involves inputs when 'lzero' is greater than
2370 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
2371 before the beginning of from). The second one involves inputs where
2372 the 'db' section contains nothing but zeroes (there is a one-byte
2373 invalid read after the end of 'db').
2374 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2376 *) Partial backport from 0.9.9-dev:
2378 Introduce bn_mul_mont (dedicated Montgomery multiplication
2379 procedure) as a candidate for BIGNUM assembler implementation.
2380 While 0.9.9-dev uses assembler for various architectures, only
2381 x86_64 is available by default here in the 0.9.8 branch, and
2382 32-bit x86 is available through a compile-time setting.
2384 To try the 32-bit x86 assembler implementation, use Configure
2385 option "enable-montasm" (which exists only for this backport).
2387 As "enable-montasm" for 32-bit x86 disclaims code stability
2388 anyway, in this constellation we activate additional code
2389 backported from 0.9.9-dev for further performance improvements,
2390 namely BN_from_montgomery_word. (To enable this otherwise,
2391 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
2393 [Andy Polyakov (backport partially by Bodo Moeller)]
2395 *) Add TLS session ticket callback. This allows an application to set
2396 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
2397 values. This is useful for key rollover for example where several key
2398 sets may exist with different names.
2401 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
2402 This was broken until now in 0.9.8 releases, such that the only way
2403 a registered ENGINE could be used (assuming it initialises
2404 successfully on the host) was to explicitly set it as the default
2405 for the relevant algorithms. This is in contradiction with 0.9.7
2406 behaviour and the documentation. With this fix, when an ENGINE is
2407 registered into a given algorithm's table of implementations, the
2408 'uptodate' flag is reset so that auto-discovery will be used next
2409 time a new context for that algorithm attempts to select an
2411 [Ian Lister (tweaked by Geoff Thorpe)]
2413 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
2414 implemention in the following ways:
2416 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
2419 Lack of BER streaming support means one pass streaming processing is
2420 only supported if data is detached: setting the streaming flag is
2421 ignored for embedded content.
2423 CMS support is disabled by default and must be explicitly enabled
2424 with the enable-cms configuration option.
2427 *) Update the GMP engine glue to do direct copies between BIGNUM and
2428 mpz_t when openssl and GMP use the same limb size. Otherwise the
2429 existing "conversion via a text string export" trick is still used.
2430 [Paul Sheer <paulsheer@gmail.com>]
2432 *) Zlib compression BIO. This is a filter BIO which compressed and
2433 uncompresses any data passed through it.
2436 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
2437 RFC3394 compatible AES key wrapping.
2440 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
2441 sets string data without copying. X509_ALGOR_set0() and
2442 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
2443 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
2444 from an X509_ATTRIBUTE structure optionally checking it occurs only
2445 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
2449 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
2450 to get the expected BN_FLG_CONSTTIME behavior.
2451 [Bodo Moeller (Google)]
2455 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
2456 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
2457 - added some more tests to do_tests.pl
2458 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
2459 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
2460 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
2461 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
2462 - various changes to netware.pl to enable gcc-cross builds on Win32
2464 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
2465 - various changes to fix missing prototype warnings
2466 - fixed x86nasm.pl to create correct asm files for NASM COFF output
2467 - added AES, WHIRLPOOL and CPUID assembler code to build files
2468 - added missing AES assembler make rules to mk1mf.pl
2469 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
2470 [Guenter Knauf <eflash@gmx.net>]
2472 *) Implement certificate status request TLS extension defined in RFC3546.
2473 A client can set the appropriate parameters and receive the encoded
2474 OCSP response via a callback. A server can query the supplied parameters
2475 and set the encoded OCSP response in the callback. Add simplified examples
2476 to s_client and s_server.
2479 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
2481 *) Fix various bugs:
2482 + Binary incompatibility of ssl_ctx_st structure
2483 + DTLS interoperation with non-compliant servers
2484 + Don't call get_session_cb() without proposed session
2485 + Fix ia64 assembler code
2486 [Andy Polyakov, Steve Henson]
2488 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
2490 *) DTLS Handshake overhaul. There were longstanding issues with
2491 OpenSSL DTLS implementation, which were making it impossible for
2492 RFC 4347 compliant client to communicate with OpenSSL server.
2493 Unfortunately just fixing these incompatibilities would "cut off"
2494 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
2495 server keeps tolerating non RFC compliant syntax. The opposite is
2496 not true, 0.9.8f client can not communicate with earlier server.
2497 This update even addresses CVE-2007-4995.
2500 *) Changes to avoid need for function casts in OpenSSL: some compilers
2501 (gcc 4.2 and later) reject their use.
2502 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
2505 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2506 RFC4507bis. The encrypted ticket format is an encrypted encoded
2507 SSL_SESSION structure, that way new session features are automatically
2510 If a client application caches session in an SSL_SESSION structure
2511 support is transparent because tickets are now stored in the encoded
2514 The SSL_CTX structure automatically generates keys for ticket
2515 protection in servers so again support should be possible
2516 with no application modification.
2518 If a client or server wishes to disable RFC4507 support then the option
2519 SSL_OP_NO_TICKET can be set.
2521 Add a TLS extension debugging callback to allow the contents of any client
2522 or server extensions to be examined.
2524 This work was sponsored by Google.
2527 *) Add initial support for TLS extensions, specifically for the server_name
2528 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2529 have new members for a host name. The SSL data structure has an
2530 additional member SSL_CTX *initial_ctx so that new sessions can be
2531 stored in that context to allow for session resumption, even after the
2532 SSL has been switched to a new SSL_CTX in reaction to a client's
2533 server_name extension.
2535 New functions (subject to change):
2537 SSL_get_servername()
2538 SSL_get_servername_type()
2541 New CTRL codes and macros (subject to change):
2543 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2544 - SSL_CTX_set_tlsext_servername_callback()
2545 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2546 - SSL_CTX_set_tlsext_servername_arg()
2547 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2549 openssl s_client has a new '-servername ...' option.
2551 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2552 '-key2 ...', '-servername_fatal' (subject to change). This allows
2553 testing the HostName extension for a specific single host name ('-cert'
2554 and '-key' remain fallbacks for handshakes without HostName
2555 negotiation). If the unrecogninzed_name alert has to be sent, this by
2556 default is a warning; it becomes fatal with the '-servername_fatal'
2559 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
2561 *) Add AES and SSE2 assembly language support to VC++ build.
2564 *) Mitigate attack on final subtraction in Montgomery reduction.
2567 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
2568 (which previously caused an internal error).
2571 *) Squeeze another 10% out of IGE mode when in != out.
2574 *) AES IGE mode speedup.
2575 [Dean Gaudet (Google)]
2577 *) Add the Korean symmetric 128-bit cipher SEED (see
2578 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
2579 add SEED ciphersuites from RFC 4162:
2581 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
2582 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
2583 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
2584 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
2586 To minimize changes between patchlevels in the OpenSSL 0.9.8
2587 series, SEED remains excluded from compilation unless OpenSSL
2588 is configured with 'enable-seed'.
2589 [KISA, Bodo Moeller]
2591 *) Mitigate branch prediction attacks, which can be practical if a
2592 single processor is shared, allowing a spy process to extract
2593 information. For detailed background information, see
2594 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
2595 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
2596 and Necessary Software Countermeasures"). The core of the change
2597 are new versions BN_div_no_branch() and
2598 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
2599 respectively, which are slower, but avoid the security-relevant
2600 conditional branches. These are automatically called by BN_div()
2601 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
2602 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
2603 remove a conditional branch.
2605 BN_FLG_CONSTTIME is the new name for the previous
2606 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
2607 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
2608 in the exponent causes BN_mod_exp_mont() to use the alternative
2609 implementation in BN_mod_exp_mont_consttime().) The old name
2610 remains as a deprecated alias.
2612 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
2613 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
2614 constant-time implementations for more than just exponentiation.
2615 Here too the old name is kept as a deprecated alias.
2617 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2618 the BN_BLINDING structure gets an independent copy of the
2619 modulus. This means that the previous "BIGNUM *m" argument to
2620 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2621 essentially becomes "const BIGNUM *m", although we can't actually
2622 change this in the header file before 0.9.9. It allows
2623 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2624 enable BN_FLG_CONSTTIME.
2626 [Matthew D Wood (Intel Corp)]
2628 *) In the SSL/TLS server implementation, be strict about session ID
2629 context matching (which matters if an application uses a single
2630 external cache for different purposes). Previously,
2631 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2632 set. This did ensure strict client verification, but meant that,
2633 with applications using a single external cache for quite
2634 different requirements, clients could circumvent ciphersuite
2635 restrictions for a given session ID context by starting a session
2636 in a different context.
2639 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2640 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2641 authentication-only ciphersuites.
2644 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
2645 not complete and could lead to a possible single byte overflow
2646 (CVE-2007-5135) [Ben Laurie]
2648 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
2650 *) Since AES128 and AES256 (and similarly Camellia128 and
2651 Camellia256) share a single mask bit in the logic of
2652 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2653 kludge to work properly if AES128 is available and AES256 isn't
2654 (or if Camellia128 is available and Camellia256 isn't).
2657 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
2658 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
2659 When a point or a seed is encoded in a BIT STRING, we need to
2660 prevent the removal of trailing zero bits to get the proper DER
2661 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
2662 of a NamedBitList, for which trailing 0 bits need to be removed.)
2665 *) Have SSL/TLS server implementation tolerate "mismatched" record
2666 protocol version while receiving ClientHello even if the
2667 ClientHello is fragmented. (The server can't insist on the
2668 particular protocol version it has chosen before the ServerHello
2669 message has informed the client about his choice.)
2672 *) Add RFC 3779 support.
2673 [Rob Austein for ARIN, Ben Laurie]
2675 *) Load error codes if they are not already present instead of using a
2676 static variable. This allows them to be cleanly unloaded and reloaded.
2677 Improve header file function name parsing.
2680 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
2681 or CAPABILITY handshake as required by RFCs.
2684 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
2686 *) Introduce limits to prevent malicious keys being able to
2687 cause a denial of service. (CVE-2006-2940)
2688 [Steve Henson, Bodo Moeller]
2690 *) Fix ASN.1 parsing of certain invalid structures that can result
2691 in a denial of service. (CVE-2006-2937) [Steve Henson]
2693 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2694 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2696 *) Fix SSL client code which could crash if connecting to a
2697 malicious SSLv2 server. (CVE-2006-4343)
2698 [Tavis Ormandy and Will Drewry, Google Security Team]
2700 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
2701 match only those. Before that, "AES256-SHA" would be interpreted
2702 as a pattern and match "AES128-SHA" too (since AES128-SHA got
2703 the same strength classification in 0.9.7h) as we currently only
2704 have a single AES bit in the ciphersuite description bitmap.
2705 That change, however, also applied to ciphersuite strings such as
2706 "RC4-MD5" that intentionally matched multiple ciphersuites --
2707 namely, SSL 2.0 ciphersuites in addition to the more common ones
2708 from SSL 3.0/TLS 1.0.
2710 So we change the selection algorithm again: Naming an explicit
2711 ciphersuite selects this one ciphersuite, and any other similar
2712 ciphersuite (same bitmap) from *other* protocol versions.
2713 Thus, "RC4-MD5" again will properly select both the SSL 2.0
2714 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
2716 Since SSL 2.0 does not have any ciphersuites for which the
2717 128/256 bit distinction would be relevant, this works for now.
2718 The proper fix will be to use different bits for AES128 and
2719 AES256, which would have avoided the problems from the beginning;
2720 however, bits are scarce, so we can only do this in a new release
2721 (not just a patchlevel) when we can change the SSL_CIPHER
2722 definition to split the single 'unsigned long mask' bitmap into
2723 multiple values to extend the available space.
2727 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2729 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2730 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2732 *) Add AES IGE and biIGE modes.
2735 *) Change the Unix randomness entropy gathering to use poll() when
2736 possible instead of select(), since the latter has some
2737 undesirable limitations.
2738 [Darryl Miles via Richard Levitte and Bodo Moeller]
2740 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2741 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2742 cannot be implicitly activated as part of, e.g., the "AES" alias.
2743 However, please upgrade to OpenSSL 0.9.9[-dev] for
2744 non-experimental use of the ECC ciphersuites to get TLS extension
2745 support, which is required for curve and point format negotiation
2746 to avoid potential handshake problems.
2749 *) Disable rogue ciphersuites:
2751 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2752 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2753 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2755 The latter two were purportedly from
2756 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2759 Also deactivate the remaining ciphersuites from
2760 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2761 unofficial, and the ID has long expired.
2764 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2765 dual-core machines) and other potential thread-safety issues.
2768 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
2769 versions), which is now available for royalty-free use
2770 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
2771 Also, add Camellia TLS ciphersuites from RFC 4132.
2773 To minimize changes between patchlevels in the OpenSSL 0.9.8
2774 series, Camellia remains excluded from compilation unless OpenSSL
2775 is configured with 'enable-camellia'.
2778 *) Disable the padding bug check when compression is in use. The padding
2779 bug check assumes the first packet is of even length, this is not
2780 necessarily true if compresssion is enabled and can result in false
2781 positives causing handshake failure. The actual bug test is ancient
2782 code so it is hoped that implementations will either have fixed it by
2783 now or any which still have the bug do not support compression.
2786 Changes between 0.9.8a and 0.9.8b [04 May 2006]
2788 *) When applying a cipher rule check to see if string match is an explicit
2789 cipher suite and only match that one cipher suite if it is.
2792 *) Link in manifests for VC++ if needed.
2793 [Austin Ziegler <halostatue@gmail.com>]
2795 *) Update support for ECC-based TLS ciphersuites according to
2796 draft-ietf-tls-ecc-12.txt with proposed changes (but without
2797 TLS extensions, which are supported starting with the 0.9.9
2798 branch, not in the OpenSSL 0.9.8 branch).
2801 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
2802 opaque EVP_CIPHER_CTX handling.
2805 *) Fixes and enhancements to zlib compression code. We now only use
2806 "zlib1.dll" and use the default __cdecl calling convention on Win32
2807 to conform with the standards mentioned here:
2808 http://www.zlib.net/DLL_FAQ.txt
2809 Static zlib linking now works on Windows and the new --with-zlib-include
2810 --with-zlib-lib options to Configure can be used to supply the location
2811 of the headers and library. Gracefully handle case where zlib library
2815 *) Several fixes and enhancements to the OID generation code. The old code
2816 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
2817 handle numbers larger than ULONG_MAX, truncated printing and had a
2818 non standard OBJ_obj2txt() behaviour.
2821 *) Add support for building of engines under engine/ as shared libraries
2822 under VC++ build system.
2825 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
2826 Hopefully, we will not see any false combination of paths any more.
2829 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
2831 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2832 (part of SSL_OP_ALL). This option used to disable the
2833 countermeasure against man-in-the-middle protocol-version
2834 rollback in the SSL 2.0 server implementation, which is a bad
2835 idea. (CVE-2005-2969)
2837 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2838 for Information Security, National Institute of Advanced Industrial
2839 Science and Technology [AIST], Japan)]
2841 *) Add two function to clear and return the verify parameter flags.
2844 *) Keep cipherlists sorted in the source instead of sorting them at
2845 runtime, thus removing the need for a lock.
2848 *) Avoid some small subgroup attacks in Diffie-Hellman.
2849 [Nick Mathewson and Ben Laurie]
2851 *) Add functions for well-known primes.
2854 *) Extended Windows CE support.
2855 [Satoshi Nakamura and Andy Polyakov]
2857 *) Initialize SSL_METHOD structures at compile time instead of during
2858 runtime, thus removing the need for a lock.
2861 *) Make PKCS7_decrypt() work even if no certificate is supplied by
2862 attempting to decrypt each encrypted key in turn. Add support to
2866 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
2868 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
2871 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
2874 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
2875 key into the same file any more.
2878 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
2881 *) Add -utf8 command line and config file option to 'ca'.
2882 [Stefan <stf@udoma.org]
2884 *) Removed the macro des_crypt(), as it seems to conflict with some
2885 libraries. Use DES_crypt().
2888 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
2889 involves renaming the source and generated shared-libs for
2890 both. The engines will accept the corrected or legacy ids
2891 ('ncipher' and '4758_cca' respectively) when binding. NB,
2892 this only applies when building 'shared'.
2893 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
2895 *) Add attribute functions to EVP_PKEY structure. Modify
2896 PKCS12_create() to recognize a CSP name attribute and
2897 use it. Make -CSP option work again in pkcs12 utility.
2900 *) Add new functionality to the bn blinding code:
2901 - automatic re-creation of the BN_BLINDING parameters after
2902 a fixed number of uses (currently 32)
2903 - add new function for parameter creation
2904 - introduce flags to control the update behaviour of the
2905 BN_BLINDING parameters
2906 - hide BN_BLINDING structure
2907 Add a second BN_BLINDING slot to the RSA structure to improve
2908 performance when a single RSA object is shared among several
2912 *) Add support for DTLS.
2913 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
2915 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
2916 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
2919 *) Remove buggy and incompletet DH cert support from
2920 ssl/ssl_rsa.c and ssl/s3_both.c
2923 *) Use SHA-1 instead of MD5 as the default digest algorithm for
2924 the apps/openssl applications.
2927 *) Compile clean with "-Wall -Wmissing-prototypes
2928 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
2929 DEBUG_SAFESTACK must also be set.
2932 *) Change ./Configure so that certain algorithms can be disabled by default.
2933 The new counterpiece to "no-xxx" is "enable-xxx".
2935 The patented RC5 and MDC2 algorithms will now be disabled unless
2936 "enable-rc5" and "enable-mdc2", respectively, are specified.
2938 (IDEA remains enabled despite being patented. This is because IDEA
2939 is frequently required for interoperability, and there is no license
2940 fee for non-commercial use. As before, "no-idea" can be used to
2941 avoid this algorithm.)
2945 *) Add processing of proxy certificates (see RFC 3820). This work was
2946 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
2947 EGEE (Enabling Grids for E-science in Europe).
2950 *) RC4 performance overhaul on modern architectures/implementations, such
2951 as Intel P4, IA-64 and AMD64.
2954 *) New utility extract-section.pl. This can be used specify an alternative
2955 section number in a pod file instead of having to treat each file as
2956 a separate case in Makefile. This can be done by adding two lines to the
2959 =for comment openssl_section:XXX
2961 The blank line is mandatory.
2965 *) New arguments -certform, -keyform and -pass for s_client and s_server
2966 to allow alternative format key and certificate files and passphrase
2970 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
2971 update associated structures and add various utility functions.
2973 Add new policy related verify parameters, include policy checking in
2974 standard verify code. Enhance 'smime' application with extra parameters
2975 to support policy checking and print out.
2978 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
2979 Nehemiah processors. These extensions support AES encryption in hardware
2980 as well as RNG (though RNG support is currently disabled).
2981 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
2983 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
2986 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
2987 [Andy Polyakov and a number of other people]
2989 *) Improved PowerPC platform support. Most notably BIGNUM assembler
2990 implementation contributed by IBM.
2991 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
2993 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
2994 exponent rather than 'unsigned long'. There is a corresponding change to
2995 the new 'rsa_keygen' element of the RSA_METHOD structure.
2996 [Jelte Jansen, Geoff Thorpe]
2998 *) Functionality for creating the initial serial number file is now
2999 moved from CA.pl to the 'ca' utility with a new option -create_serial.
3001 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
3002 number file to 1, which is bound to cause problems. To avoid
3003 the problems while respecting compatibility between different 0.9.7
3004 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
3005 CA.pl for serial number initialization. With the new release 0.9.8,
3006 we can fix the problem directly in the 'ca' utility.)
3009 *) Reduced header interdepencies by declaring more opaque objects in
3010 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
3011 give fewer recursive includes, which could break lazy source code - so
3012 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
3013 developers should define this symbol when building and using openssl to
3014 ensure they track the recommended behaviour, interfaces, [etc], but
3015 backwards-compatible behaviour prevails when this isn't defined.
3018 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
3021 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
3022 This will generate a random key of the appropriate length based on the
3023 cipher context. The EVP_CIPHER can provide its own random key generation
3024 routine to support keys of a specific form. This is used in the des and
3025 3des routines to generate a key of the correct parity. Update S/MIME
3026 code to use new functions and hence generate correct parity DES keys.
3027 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
3028 valid (weak or incorrect parity).
3031 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
3032 as looking them up. This is useful when the verified structure may contain
3033 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
3034 present unless the new PKCS7_NO_CRL flag is asserted.
3037 *) Extend ASN1 oid configuration module. It now additionally accepts the
3040 shortName = some long name, 1.2.3.4
3043 *) Reimplemented the BN_CTX implementation. There is now no more static
3044 limitation on the number of variables it can handle nor the depth of the
3045 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
3046 information can now expand as required, and rather than having a single
3047 static array of bignums, BN_CTX now uses a linked-list of such arrays
3048 allowing it to expand on demand whilst maintaining the usefulness of
3049 BN_CTX's "bundling".
3052 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
3053 to allow all RSA operations to function using a single BN_CTX.
3056 *) Preliminary support for certificate policy evaluation and checking. This
3057 is initially intended to pass the tests outlined in "Conformance Testing
3058 of Relying Party Client Certificate Path Processing Logic" v1.07.
3061 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
3062 remained unused and not that useful. A variety of other little bignum
3063 tweaks and fixes have also been made continuing on from the audit (see
3067 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
3068 associated ASN1, EVP and SSL functions and old ASN1 macros.
3071 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
3072 and this should never fail. So the return value from the use of
3073 BN_set_word() (which can fail due to needless expansion) is now deprecated;
3074 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
3077 *) BN_CTX_get() should return zero-valued bignums, providing the same
3078 initialised value as BN_new().
3079 [Geoff Thorpe, suggested by Ulf Möller]
3081 *) Support for inhibitAnyPolicy certificate extension.
3084 *) An audit of the BIGNUM code is underway, for which debugging code is
3085 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
3086 is considered valid when processing BIGNUMs, and causes execution to
3087 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
3088 further steps are taken to deliberately pollute unused data in BIGNUM
3089 structures to try and expose faulty code further on. For now, openssl will
3090 (in its default mode of operation) continue to tolerate the inconsistent
3091 forms that it has tolerated in the past, but authors and packagers should
3092 consider trying openssl and their own applications when compiled with
3093 these debugging symbols defined. It will help highlight potential bugs in
3094 their own code, and will improve the test coverage for OpenSSL itself. At
3095 some point, these tighter rules will become openssl's default to improve
3096 maintainability, though the assert()s and other overheads will remain only
3097 in debugging configurations. See bn.h for more details.
3098 [Geoff Thorpe, Nils Larsch, Ulf Möller]
3100 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
3101 that can only be obtained through BN_CTX_new() (which implicitly
3102 initialises it). The presence of this function only made it possible
3103 to overwrite an existing structure (and cause memory leaks).
3106 *) Because of the callback-based approach for implementing LHASH as a
3107 template type, lh_insert() adds opaque objects to hash-tables and
3108 lh_doall() or lh_doall_arg() are typically used with a destructor callback
3109 to clean up those corresponding objects before destroying the hash table
3110 (and losing the object pointers). So some over-zealous constifications in
3111 LHASH have been relaxed so that lh_insert() does not take (nor store) the
3112 objects as "const" and the lh_doall[_arg] callback wrappers are not
3113 prototyped to have "const" restrictions on the object pointers they are
3114 given (and so aren't required to cast them away any more).
3117 *) The tmdiff.h API was so ugly and minimal that our own timing utility
3118 (speed) prefers to use its own implementation. The two implementations
3119 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
3120 its object type properly exposed (MS_TM) instead of casting to/from "char
3121 *". This may still change yet if someone realises MS_TM and "ms_time_***"
3122 aren't necessarily the greatest nomenclatures - but this is what was used
3123 internally to the implementation so I've used that for now.
3126 *) Ensure that deprecated functions do not get compiled when
3127 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
3128 the self-tests were still using deprecated key-generation functions so
3129 these have been updated also.
3132 *) Reorganise PKCS#7 code to separate the digest location functionality
3133 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
3134 New function PKCS7_set_digest() to set the digest type for PKCS#7
3135 digestedData type. Add additional code to correctly generate the
3136 digestedData type and add support for this type in PKCS7 initialization
3140 *) New function PKCS7_set0_type_other() this initializes a PKCS7
3141 structure of type "other".
3144 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
3145 sure the loop does correctly stop and breaking ("division by zero")
3146 modulus operations are not performed. The (pre-generated) prime
3147 table crypto/bn/bn_prime.h was already correct, but it could not be
3148 re-generated on some platforms because of the "division by zero"
3149 situation in the script.
3150 [Ralf S. Engelschall]
3152 *) Update support for ECC-based TLS ciphersuites according to
3153 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
3154 SHA-1 now is only used for "small" curves (where the
3155 representation of a field element takes up to 24 bytes); for
3156 larger curves, the field element resulting from ECDH is directly
3157 used as premaster secret.
3158 [Douglas Stebila (Sun Microsystems Laboratories)]
3160 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
3161 curve secp160r1 to the tests.
3162 [Douglas Stebila (Sun Microsystems Laboratories)]
3164 *) Add the possibility to load symbols globally with DSO.
3165 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
3167 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
3168 control of the error stack.
3171 *) Add support for STORE in ENGINE.
3174 *) Add the STORE type. The intention is to provide a common interface
3175 to certificate and key stores, be they simple file-based stores, or
3176 HSM-type store, or LDAP stores, or...
3177 NOTE: The code is currently UNTESTED and isn't really used anywhere.
3180 *) Add a generic structure called OPENSSL_ITEM. This can be used to
3181 pass a list of arguments to any function as well as provide a way
3182 for a function to pass data back to the caller.
3185 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
3186 works like BUF_strdup() but can be used to duplicate a portion of
3187 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
3191 *) Add the function sk_find_ex() which works like sk_find(), but will
3192 return an index to an element even if an exact match couldn't be
3193 found. The index is guaranteed to point at the element where the
3194 searched-for key would be inserted to preserve sorting order.
3197 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
3198 takes an extra flags argument for optional functionality. Currently,
3199 the following flags are defined:
3201 OBJ_BSEARCH_VALUE_ON_NOMATCH
3202 This one gets OBJ_bsearch_ex() to return a pointer to the first
3203 element where the comparing function returns a negative or zero
3206 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
3207 This one gets OBJ_bsearch_ex() to return a pointer to the first
3208 element where the comparing function returns zero. This is useful
3209 if there are more than one element where the comparing function
3213 *) Make it possible to create self-signed certificates with 'openssl ca'
3214 in such a way that the self-signed certificate becomes part of the
3215 CA database and uses the same mechanisms for serial number generation
3216 as all other certificate signing. The new flag '-selfsign' enables
3217 this functionality. Adapt CA.sh and CA.pl.in.
3220 *) Add functionality to check the public key of a certificate request
3221 against a given private. This is useful to check that a certificate
3222 request can be signed by that key (self-signing).
3225 *) Make it possible to have multiple active certificates with the same
3226 subject in the CA index file. This is done only if the keyword
3227 'unique_subject' is set to 'no' in the main CA section (default
3228 if 'CA_default') of the configuration file. The value is saved
3229 with the database itself in a separate index attribute file,
3230 named like the index file with '.attr' appended to the name.
3233 *) Generate muti valued AVAs using '+' notation in config files for
3237 *) Support for nameConstraints certificate extension.
3240 *) Support for policyConstraints certificate extension.
3243 *) Support for policyMappings certificate extension.
3246 *) Make sure the default DSA_METHOD implementation only uses its
3247 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
3248 and change its own handlers to be NULL so as to remove unnecessary
3249 indirection. This lets alternative implementations fallback to the
3250 default implementation more easily.
3253 *) Support for directoryName in GeneralName related extensions
3257 *) Make it possible to link applications using Makefile.shared.
3258 Make that possible even when linking against static libraries!
3261 *) Support for single pass processing for S/MIME signing. This now
3262 means that S/MIME signing can be done from a pipe, in addition
3263 cleartext signing (multipart/signed type) is effectively streaming
3264 and the signed data does not need to be all held in memory.
3266 This is done with a new flag PKCS7_STREAM. When this flag is set
3267 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
3268 is done after the data is output (and digests calculated) in
3269 SMIME_write_PKCS7().
3272 *) Add full support for -rpath/-R, both in shared libraries and
3273 applications, at least on the platforms where it's known how
3277 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
3278 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
3279 will now compute a table of multiples of the generator that
3280 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
3281 faster (notably in the case of a single point multiplication,
3282 scalar * generator).
3283 [Nils Larsch, Bodo Moeller]
3285 *) IPv6 support for certificate extensions. The various extensions
3286 which use the IP:a.b.c.d can now take IPv6 addresses using the
3287 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
3291 *) Added an ENGINE that implements RSA by performing private key
3292 exponentiations with the GMP library. The conversions to and from
3293 GMP's mpz_t format aren't optimised nor are any montgomery forms
3294 cached, and on x86 it appears OpenSSL's own performance has caught up.
3295 However there are likely to be other architectures where GMP could
3296 provide a boost. This ENGINE is not built in by default, but it can be
3297 specified at Configure time and should be accompanied by the necessary
3298 linker additions, eg;
3299 ./config -DOPENSSL_USE_GMP -lgmp
3302 *) "openssl engine" will not display ENGINE/DSO load failure errors when
3303 testing availability of engines with "-t" - the old behaviour is
3304 produced by increasing the feature's verbosity with "-tt".
3307 *) ECDSA routines: under certain error conditions uninitialized BN objects
3308 could be freed. Solution: make sure initialization is performed early
3309 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
3313 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
3314 and DH_METHOD (eg. by ENGINE implementations) to override the normal
3315 software implementations. For DSA and DH, parameter generation can
3316 also be overriden by providing the appropriate method callbacks.
3319 *) Change the "progress" mechanism used in key-generation and
3320 primality testing to functions that take a new BN_GENCB pointer in
3321 place of callback/argument pairs. The new API functions have "_ex"
3322 postfixes and the older functions are reimplemented as wrappers for
3323 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
3324 declarations of the old functions to help (graceful) attempts to
3325 migrate to the new functions. Also, the new key-generation API
3326 functions operate on a caller-supplied key-structure and return
3327 success/failure rather than returning a key or NULL - this is to
3328 help make "keygen" another member function of RSA_METHOD etc.
3330 Example for using the new callback interface:
3332 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
3336 BN_GENCB_set(&my_cb, my_callback, my_arg);
3338 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
3339 /* For the meaning of a, b in calls to my_callback(), see the
3340 * documentation of the function that calls the callback.
3341 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
3342 * my_callback should return 1 if it wants BN_is_prime_ex()
3343 * to continue, or 0 to stop.
3348 *) Change the ZLIB compression method to be stateful, and make it
3349 available to TLS with the number defined in
3350 draft-ietf-tls-compression-04.txt.
3353 *) Add the ASN.1 structures and functions for CertificatePair, which
3354 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
3356 CertificatePair ::= SEQUENCE {
3357 forward [0] Certificate OPTIONAL,
3358 reverse [1] Certificate OPTIONAL,
3359 -- at least one of the pair shall be present -- }
3361 Also implement the PEM functions to read and write certificate
3362 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
3364 This needed to be defined, mostly for the sake of the LDAP
3365 attribute crossCertificatePair, but may prove useful elsewhere as
3369 *) Make it possible to inhibit symlinking of shared libraries in
3370 Makefile.shared, for Cygwin's sake.
3373 *) Extend the BIGNUM API by creating a function
3374 void BN_set_negative(BIGNUM *a, int neg);
3375 and a macro that behave like
3376 int BN_is_negative(const BIGNUM *a);
3378 to avoid the need to access 'a->neg' directly in applications.
3381 *) Implement fast modular reduction for pseudo-Mersenne primes
3382 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
3383 EC_GROUP_new_curve_GFp() will now automatically use this
3385 [Nils Larsch <nla@trustcenter.de>]
3387 *) Add new lock type (CRYPTO_LOCK_BN).
3390 *) Change the ENGINE framework to automatically load engines
3391 dynamically from specific directories unless they could be
3392 found to already be built in or loaded. Move all the
3393 current engines except for the cryptodev one to a new
3395 The engines in engines/ are built as shared libraries if
3396 the "shared" options was given to ./Configure or ./config.
3397 Otherwise, they are inserted in libcrypto.a.
3398 /usr/local/ssl/engines is the default directory for dynamic
3399 engines, but that can be overriden at configure time through
3400 the usual use of --prefix and/or --openssldir, and at run
3401 time with the environment variable OPENSSL_ENGINES.
3402 [Geoff Thorpe and Richard Levitte]
3404 *) Add Makefile.shared, a helper makefile to build shared
3405 libraries. Addapt Makefile.org.
3408 *) Add version info to Win32 DLLs.
3409 [Peter 'Luna' Runestig" <peter@runestig.com>]
3411 *) Add new 'medium level' PKCS#12 API. Certificates and keys
3412 can be added using this API to created arbitrary PKCS#12
3413 files while avoiding the low level API.
3415 New options to PKCS12_create(), key or cert can be NULL and
3416 will then be omitted from the output file. The encryption
3417 algorithm NIDs can be set to -1 for no encryption, the mac
3418 iteration count can be set to 0 to omit the mac.
3420 Enhance pkcs12 utility by making the -nokeys and -nocerts
3421 options work when creating a PKCS#12 file. New option -nomac
3422 to omit the mac, NONE can be set for an encryption algorithm.
3423 New code is modified to use the enhanced PKCS12_create()
3424 instead of the low level API.
3427 *) Extend ASN1 encoder to support indefinite length constructed
3428 encoding. This can output sequences tags and octet strings in
3429 this form. Modify pk7_asn1.c to support indefinite length
3430 encoding. This is experimental and needs additional code to
3431 be useful, such as an ASN1 bio and some enhanced streaming
3434 Extend template encode functionality so that tagging is passed
3435 down to the template encoder.
3438 *) Let 'openssl req' fail if an argument to '-newkey' is not
3439 recognized instead of using RSA as a default.
3442 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
3443 As these are not official, they are not included in "ALL";
3444 the "ECCdraft" ciphersuite group alias can be used to select them.
3445 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
3447 *) Add ECDH engine support.
3448 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
3450 *) Add ECDH in new directory crypto/ecdh/.
3451 [Douglas Stebila (Sun Microsystems Laboratories)]
3453 *) Let BN_rand_range() abort with an error after 100 iterations
3454 without success (which indicates a broken PRNG).
3457 *) Change BN_mod_sqrt() so that it verifies that the input value
3458 is really the square of the return value. (Previously,
3459 BN_mod_sqrt would show GIGO behaviour.)
3462 *) Add named elliptic curves over binary fields from X9.62, SECG,
3463 and WAP/WTLS; add OIDs that were still missing.
3465 [Sheueling Chang Shantz and Douglas Stebila
3466 (Sun Microsystems Laboratories)]
3468 *) Extend the EC library for elliptic curves over binary fields
3469 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
3472 EC_GF2m_simple_method
3476 EC_GROUP_new_curve_GF2m
3477 EC_GROUP_set_curve_GF2m
3478 EC_GROUP_get_curve_GF2m
3479 EC_POINT_set_affine_coordinates_GF2m
3480 EC_POINT_get_affine_coordinates_GF2m
3481 EC_POINT_set_compressed_coordinates_GF2m
3483 Point compression for binary fields is disabled by default for
3484 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
3487 As binary polynomials are represented as BIGNUMs, various members
3488 of the EC_GROUP and EC_POINT data structures can be shared
3489 between the implementations for prime fields and binary fields;
3490 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
3491 are essentially identical to their ..._GFp counterparts.
3492 (For simplicity, the '..._GFp' prefix has been dropped from
3493 various internal method names.)
3495 An internal 'field_div' method (similar to 'field_mul' and
3496 'field_sqr') has been added; this is used only for binary fields.
3498 [Sheueling Chang Shantz and Douglas Stebila
3499 (Sun Microsystems Laboratories)]
3501 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
3502 through methods ('mul', 'precompute_mult').
3504 The generic implementations (now internally called 'ec_wNAF_mul'
3505 and 'ec_wNAF_precomputed_mult') remain the default if these
3506 methods are undefined.
3508 [Sheueling Chang Shantz and Douglas Stebila
3509 (Sun Microsystems Laboratories)]
3511 *) New function EC_GROUP_get_degree, which is defined through
3512 EC_METHOD. For curves over prime fields, this returns the bit
3513 length of the modulus.
3515 [Sheueling Chang Shantz and Douglas Stebila
3516 (Sun Microsystems Laboratories)]
3518 *) New functions EC_GROUP_dup, EC_POINT_dup.
3519 (These simply call ..._new and ..._copy).
3521 [Sheueling Chang Shantz and Douglas Stebila
3522 (Sun Microsystems Laboratories)]
3524 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
3525 Polynomials are represented as BIGNUMs (where the sign bit is not
3526 used) in the following functions [macros]:
3529 BN_GF2m_sub [= BN_GF2m_add]
3530 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
3531 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
3532 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
3534 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
3535 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
3536 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
3537 BN_GF2m_cmp [= BN_ucmp]
3539 (Note that only the 'mod' functions are actually for fields GF(2^m).
3540 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
3542 For some functions, an the irreducible polynomial defining a
3543 field can be given as an 'unsigned int[]' with strictly
3544 decreasing elements giving the indices of those bits that are set;
3545 i.e., p[] represents the polynomial
3546 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
3548 p[0] > p[1] > ... > p[k] = 0.
3549 This applies to the following functions:
3554 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
3555 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
3557 BN_GF2m_mod_sqrt_arr
3558 BN_GF2m_mod_solve_quad_arr
3562 Conversion can be performed by the following functions:
3567 bntest.c has additional tests for binary polynomial arithmetic.
3569 Two implementations for BN_GF2m_mod_div() are available.
3570 The default algorithm simply uses BN_GF2m_mod_inv() and
3571 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
3572 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
3573 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
3575 [Sheueling Chang Shantz and Douglas Stebila
3576 (Sun Microsystems Laboratories)]
3578 *) Add new error code 'ERR_R_DISABLED' that can be used when some
3579 functionality is disabled at compile-time.
3580 [Douglas Stebila <douglas.stebila@sun.com>]
3582 *) Change default behaviour of 'openssl asn1parse' so that more
3583 information is visible when viewing, e.g., a certificate:
3585 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
3586 mode the content of non-printable OCTET STRINGs is output in a
3587 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
3588 avoid the appearance of a printable string.
3589 [Nils Larsch <nla@trustcenter.de>]
3591 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
3593 EC_GROUP_set_asn1_flag()
3594 EC_GROUP_get_asn1_flag()
3595 EC_GROUP_set_point_conversion_form()
3596 EC_GROUP_get_point_conversion_form()
3597 These control ASN1 encoding details:
3598 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
3599 has been set to OPENSSL_EC_NAMED_CURVE.
3600 - Points are encoded in uncompressed form by default; options for
3601 asn1_for are as for point2oct, namely
3602 POINT_CONVERSION_COMPRESSED
3603 POINT_CONVERSION_UNCOMPRESSED
3604 POINT_CONVERSION_HYBRID
3606 Also add 'seed' and 'seed_len' members to EC_GROUP with access
3609 EC_GROUP_get0_seed()
3610 EC_GROUP_get_seed_len()
3611 This is used only for ASN1 purposes (so far).
3612 [Nils Larsch <nla@trustcenter.de>]
3614 *) Add 'field_type' member to EC_METHOD, which holds the NID
3615 of the appropriate field type OID. The new function
3616 EC_METHOD_get_field_type() returns this value.
3617 [Nils Larsch <nla@trustcenter.de>]
3622 EC_POINT_point2hex()
3623 EC_POINT_hex2point()
3624 providing useful interfaces to EC_POINT_point2oct() and
3625 EC_POINT_oct2point().
3626 [Nils Larsch <nla@trustcenter.de>]
3628 *) Change internals of the EC library so that the functions
3629 EC_GROUP_set_generator()
3630 EC_GROUP_get_generator()
3631 EC_GROUP_get_order()
3632 EC_GROUP_get_cofactor()
3633 are implemented directly in crypto/ec/ec_lib.c and not dispatched
3634 to methods, which would lead to unnecessary code duplication when
3635 adding different types of curves.
3636 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
3638 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
3639 arithmetic, and such that modified wNAFs are generated
3640 (which avoid length expansion in many cases).
3643 *) Add a function EC_GROUP_check_discriminant() (defined via
3644 EC_METHOD) that verifies that the curve discriminant is non-zero.
3646 Add a function EC_GROUP_check() that makes some sanity tests
3647 on a EC_GROUP, its generator and order. This includes
3648 EC_GROUP_check_discriminant().
3649 [Nils Larsch <nla@trustcenter.de>]
3651 *) Add ECDSA in new directory crypto/ecdsa/.
3653 Add applications 'openssl ecparam' and 'openssl ecdsa'
3654 (these are based on 'openssl dsaparam' and 'openssl dsa').
3656 ECDSA support is also included in various other files across the
3657 library. Most notably,
3658 - 'openssl req' now has a '-newkey ecdsa:file' option;
3659 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
3660 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
3661 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
3662 them suitable for ECDSA where domain parameters must be
3663 extracted before the specific public key;
3664 - ECDSA engine support has been added.
3665 [Nils Larsch <nla@trustcenter.de>]
3667 *) Include some named elliptic curves, and add OIDs from X9.62,
3668 SECG, and WAP/WTLS. Each curve can be obtained from the new
3670 EC_GROUP_new_by_curve_name(),
3671 and the list of available named curves can be obtained with
3672 EC_get_builtin_curves().
3673 Also add a 'curve_name' member to EC_GROUP objects, which can be
3675 EC_GROUP_set_curve_name()
3676 EC_GROUP_get_curve_name()
3677 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
3679 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3680 was actually never needed) and in BN_mul(). The removal in BN_mul()
3681 required a small change in bn_mul_part_recursive() and the addition
3682 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3683 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3684 bn_sub_words() and bn_add_words() except they take arrays with
3688 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
3690 *) Cleanse PEM buffers before freeing them since they may contain
3692 [Benjamin Bennett <ben@psc.edu>]
3694 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3695 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3696 authentication-only ciphersuites.
3699 *) Since AES128 and AES256 share a single mask bit in the logic of
3700 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3701 kludge to work properly if AES128 is available and AES256 isn't.
3704 *) Expand security boundary to match 1.1.1 module.
3707 *) Remove redundant features: hash file source, editing of test vectors
3708 modify fipsld to use external fips_premain.c signature.
3711 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
3712 run algorithm test programs.
3715 *) Make algorithm test programs more tolerant of whitespace.
3718 *) Have SSL/TLS server implementation tolerate "mismatched" record
3719 protocol version while receiving ClientHello even if the
3720 ClientHello is fragmented. (The server can't insist on the
3721 particular protocol version it has chosen before the ServerHello
3722 message has informed the client about his choice.)
3725 *) Load error codes if they are not already present instead of using a
3726 static variable. This allows them to be cleanly unloaded and reloaded.
3729 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3731 *) Introduce limits to prevent malicious keys being able to
3732 cause a denial of service. (CVE-2006-2940)
3733 [Steve Henson, Bodo Moeller]
3735 *) Fix ASN.1 parsing of certain invalid structures that can result
3736 in a denial of service. (CVE-2006-2937) [Steve Henson]
3738 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3739 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3741 *) Fix SSL client code which could crash if connecting to a
3742 malicious SSLv2 server. (CVE-2006-4343)
3743 [Tavis Ormandy and Will Drewry, Google Security Team]
3745 *) Change ciphersuite string processing so that an explicit
3746 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3747 will no longer include "AES128-SHA"), and any other similar
3748 ciphersuite (same bitmap) from *other* protocol versions (so that
3749 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3750 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3751 changes from 0.9.8b and 0.9.8d.
3754 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
3756 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3757 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3759 *) Change the Unix randomness entropy gathering to use poll() when
3760 possible instead of select(), since the latter has some
3761 undesirable limitations.
3762 [Darryl Miles via Richard Levitte and Bodo Moeller]
3764 *) Disable rogue ciphersuites:
3766 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3767 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3768 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3770 The latter two were purportedly from
3771 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3774 Also deactive the remaining ciphersuites from
3775 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3776 unofficial, and the ID has long expired.
3779 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3780 dual-core machines) and other potential thread-safety issues.
3783 Changes between 0.9.7i and 0.9.7j [04 May 2006]
3785 *) Adapt fipsld and the build system to link against the validated FIPS
3786 module in FIPS mode.
3789 *) Fixes for VC++ 2005 build under Windows.
3792 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
3793 from a Windows bash shell such as MSYS. It is autodetected from the
3794 "config" script when run from a VC++ environment. Modify standard VC++
3795 build to use fipscanister.o from the GNU make build.
3798 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
3800 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
3801 The value now differs depending on if you build for FIPS or not.
3802 BEWARE! A program linked with a shared FIPSed libcrypto can't be
3803 safely run with a non-FIPSed libcrypto, as it may crash because of
3804 the difference induced by this change.
3807 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
3809 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3810 (part of SSL_OP_ALL). This option used to disable the
3811 countermeasure against man-in-the-middle protocol-version
3812 rollback in the SSL 2.0 server implementation, which is a bad
3813 idea. (CVE-2005-2969)
3815 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3816 for Information Security, National Institute of Advanced Industrial
3817 Science and Technology [AIST], Japan)]
3819 *) Minimal support for X9.31 signatures and PSS padding modes. This is
3820 mainly for FIPS compliance and not fully integrated at this stage.
3823 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
3824 the exponentiation using a fixed-length exponent. (Otherwise,
3825 the information leaked through timing could expose the secret key
3826 after many signatures; cf. Bleichenbacher's attack on DSA with
3830 *) Make a new fixed-window mod_exp implementation the default for
3831 RSA, DSA, and DH private-key operations so that the sequence of
3832 squares and multiplies and the memory access pattern are
3833 independent of the particular secret key. This will mitigate
3834 cache-timing and potential related attacks.
3836 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
3837 and this is automatically used by BN_mod_exp_mont() if the new flag
3838 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
3839 will use this BN flag for private exponents unless the flag
3840 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
3841 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
3843 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
3845 *) Change the client implementation for SSLv23_method() and
3846 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
3847 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
3848 (Previously, the SSL 2.0 backwards compatible Client Hello
3849 message format would be used even with SSL_OP_NO_SSLv2.)
3852 *) Add support for smime-type MIME parameter in S/MIME messages which some
3856 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
3857 a threadsafe manner. Modify rsa code to use new function and add calls
3858 to dsa and dh code (which had race conditions before).
3861 *) Include the fixed error library code in the C error file definitions
3862 instead of fixing them up at runtime. This keeps the error code
3863 structures constant.
3866 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
3868 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
3871 *) Fixes for newer kerberos headers. NB: the casts are needed because
3872 the 'length' field is signed on one version and unsigned on another
3873 with no (?) obvious way to tell the difference, without these VC++
3874 complains. Also the "definition" of FAR (blank) is no longer included
3875 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
3876 some needed definitions.
3879 *) Undo Cygwin change.
3882 *) Added support for proxy certificates according to RFC 3820.
3883 Because they may be a security thread to unaware applications,
3884 they must be explicitely allowed in run-time. See
3885 docs/HOWTO/proxy_certificates.txt for further information.
3888 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
3890 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
3891 server and client random values. Previously
3892 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
3893 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
3895 This change has negligible security impact because:
3897 1. Server and client random values still have 24 bytes of pseudo random
3900 2. Server and client random values are sent in the clear in the initial
3903 3. The master secret is derived using the premaster secret (48 bytes in
3904 size for static RSA ciphersuites) as well as client server and random
3907 The OpenSSL team would like to thank the UK NISCC for bringing this issue
3910 [Stephen Henson, reported by UK NISCC]
3912 *) Use Windows randomness collection on Cygwin.
3915 *) Fix hang in EGD/PRNGD query when communication socket is closed
3916 prematurely by EGD/PRNGD.
3917 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
3919 *) Prompt for pass phrases when appropriate for PKCS12 input format.
3922 *) Back-port of selected performance improvements from development
3923 branch, as well as improved support for PowerPC platforms.
3926 *) Add lots of checks for memory allocation failure, error codes to indicate
3927 failure and freeing up memory if a failure occurs.
3928 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
3930 *) Add new -passin argument to dgst.
3933 *) Perform some character comparisons of different types in X509_NAME_cmp:
3934 this is needed for some certificates that reencode DNs into UTF8Strings
3935 (in violation of RFC3280) and can't or wont issue name rollover
3939 *) Make an explicit check during certificate validation to see that
3940 the CA setting in each certificate on the chain is correct. As a
3941 side effect always do the following basic checks on extensions,
3942 not just when there's an associated purpose to the check:
3944 - if there is an unhandled critical extension (unless the user
3945 has chosen to ignore this fault)
3946 - if the path length has been exceeded (if one is set at all)
3947 - that certain extensions fit the associated purpose (if one has
3951 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
3953 *) Avoid a race condition when CRLs are checked in a multi threaded
3954 environment. This would happen due to the reordering of the revoked
3955 entries during signature checking and serial number lookup. Now the
3956 encoding is cached and the serial number sort performed under a lock.
3957 Add new STACK function sk_is_sorted().
3960 *) Add Delta CRL to the extension code.
3963 *) Various fixes to s3_pkt.c so alerts are sent properly.
3964 [David Holmes <d.holmes@f5.com>]
3966 *) Reduce the chances of duplicate issuer name and serial numbers (in
3967 violation of RFC3280) using the OpenSSL certificate creation utilities.
3968 This is done by creating a random 64 bit value for the initial serial
3969 number when a serial number file is created or when a self signed
3970 certificate is created using 'openssl req -x509'. The initial serial
3971 number file is created using 'openssl x509 -next_serial' in CA.pl
3972 rather than being initialized to 1.
3975 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
3977 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3978 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3979 [Joe Orton, Steve Henson]
3981 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
3983 [Joe Orton, Steve Henson]
3985 *) Make it possible to have multiple active certificates with the same
3986 subject in the CA index file. This is done only if the keyword
3987 'unique_subject' is set to 'no' in the main CA section (default
3988 if 'CA_default') of the configuration file. The value is saved
3989 with the database itself in a separate index attribute file,
3990 named like the index file with '.attr' appended to the name.
3993 *) X509 verify fixes. Disable broken certificate workarounds when
3994 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
3995 keyUsage extension present. Don't accept CRLs with unhandled critical
3996 extensions: since verify currently doesn't process CRL extensions this
3997 rejects a CRL with *any* critical extensions. Add new verify error codes
4001 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
4002 A clarification of RFC2560 will require the use of OCTET STRINGs and
4003 some implementations cannot handle the current raw format. Since OpenSSL
4004 copies and compares OCSP nonces as opaque blobs without any attempt at
4005 parsing them this should not create any compatibility issues.
4008 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
4009 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
4010 this HMAC (and other) operations are several times slower than OpenSSL
4014 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
4015 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4017 *) Use the correct content when signing type "other".
4020 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
4022 *) Fix various bugs revealed by running the NISCC test suite:
4024 Stop out of bounds reads in the ASN1 code when presented with
4025 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4027 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
4029 If verify callback ignores invalid public key errors don't try to check
4030 certificate signature with the NULL public key.
4034 *) New -ignore_err option in ocsp application to stop the server
4035 exiting on the first error in a request.
4038 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4039 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4043 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4044 extra data after the compression methods not only for TLS 1.0
4045 but also for SSL 3.0 (as required by the specification).
4046 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4048 *) Change X509_certificate_type() to mark the key as exported/exportable
4049 when it's 512 *bits* long, not 512 bytes.
4052 *) Change AES_cbc_encrypt() so it outputs exact multiple of
4053 blocks during encryption.
4056 *) Various fixes to base64 BIO and non blocking I/O. On write
4057 flushes were not handled properly if the BIO retried. On read
4058 data was not being buffered properly and had various logic bugs.
4059 This also affects blocking I/O when the data being decoded is a
4063 *) Various S/MIME bugfixes and compatibility changes:
4064 output correct application/pkcs7 MIME type if
4065 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
4066 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
4067 of files as .eml work). Correctly handle very long lines in MIME
4071 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
4073 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4074 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4075 a protocol version number mismatch like a decryption error
4076 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4079 *) Turn on RSA blinding by default in the default implementation
4080 to avoid a timing attack. Applications that don't want it can call
4081 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4082 They would be ill-advised to do so in most cases.
4083 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4085 *) Change RSA blinding code so that it works when the PRNG is not
4086 seeded (in this case, the secret RSA exponent is abused as
4087 an unpredictable seed -- if it is not unpredictable, there
4088 is no point in blinding anyway). Make RSA blinding thread-safe
4089 by remembering the creator's thread ID in rsa->blinding and
4090 having all other threads use local one-time blinding factors
4091 (this requires more computation than sharing rsa->blinding, but
4092 avoids excessive locking; and if an RSA object is not shared
4093 between threads, blinding will still be very fast).
4096 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
4097 ENGINE as defaults for all supported algorithms irrespective of
4098 the 'flags' parameter. 'flags' is now honoured, so applications
4099 should make sure they are passing it correctly.
4102 *) Target "mingw" now allows native Windows code to be generated in
4103 the Cygwin environment as well as with the MinGW compiler.
4106 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
4108 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4109 via timing by performing a MAC computation even if incorrrect
4110 block cipher padding has been found. This is a countermeasure
4111 against active attacks where the attacker has to distinguish
4112 between bad padding and a MAC verification error. (CVE-2003-0078)
4114 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4115 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4116 Martin Vuagnoux (EPFL, Ilion)]
4118 *) Make the no-err option work as intended. The intention with no-err
4119 is not to have the whole error stack handling routines removed from
4120 libcrypto, it's only intended to remove all the function name and
4121 reason texts, thereby removing some of the footprint that may not
4122 be interesting if those errors aren't displayed anyway.
4124 NOTE: it's still possible for any application or module to have it's
4125 own set of error texts inserted. The routines are there, just not
4126 used by default when no-err is given.
4129 *) Add support for FreeBSD on IA64.
4130 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
4132 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
4133 Kerberos function mit_des_cbc_cksum(). Before this change,
4134 the value returned by DES_cbc_cksum() was like the one from
4135 mit_des_cbc_cksum(), except the bytes were swapped.
4136 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
4138 *) Allow an application to disable the automatic SSL chain building.
4139 Before this a rather primitive chain build was always performed in
4140 ssl3_output_cert_chain(): an application had no way to send the
4141 correct chain if the automatic operation produced an incorrect result.
4143 Now the chain builder is disabled if either:
4145 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
4147 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
4149 The reasoning behind this is that an application would not want the
4150 auto chain building to take place if extra chain certificates are
4151 present and it might also want a means of sending no additional
4152 certificates (for example the chain has two certificates and the
4156 *) Add the possibility to build without the ENGINE framework.
4157 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4159 *) Under Win32 gmtime() can return NULL: check return value in
4160 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
4163 *) DSA routines: under certain error conditions uninitialized BN objects
4164 could be freed. Solution: make sure initialization is performed early
4165 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
4166 Nils Larsch <nla@trustcenter.de> via PR#459)
4169 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
4170 checked on reconnect on the client side, therefore session resumption
4171 could still fail with a "ssl session id is different" error. This
4172 behaviour is masked when SSL_OP_ALL is used due to
4173 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
4174 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4175 followup to PR #377.
4178 *) IA-32 assembler support enhancements: unified ELF targets, support
4179 for SCO/Caldera platforms, fix for Cygwin shared build.
4182 *) Add support for FreeBSD on sparc64. As a consequence, support for
4183 FreeBSD on non-x86 processors is separate from x86 processors on
4184 the config script, much like the NetBSD support.
4185 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
4187 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
4189 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
4192 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
4193 code (06) was taken as the first octet of the session ID and the last
4194 octet was ignored consequently. As a result SSLv2 client side session
4195 caching could not have worked due to the session ID mismatch between
4197 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4201 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
4202 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
4206 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
4207 seems that in spite of existing for more than a year, many application
4208 author have done nothing to provide the necessary callbacks, which
4209 means that this particular engine will not work properly anywhere.
4210 This is a very unfortunate situation which forces us, in the name
4211 of usability, to give the hw_ncipher.c a static lock, which is part
4213 NOTE: This is for the 0.9.7 series ONLY. This hack will never
4214 appear in 0.9.8 or later. We EXPECT application authors to have
4215 dealt properly with this when 0.9.8 is released (unless we actually
4216 make such changes in the libcrypto locking code that changes will
4217 have to be made anyway).
4220 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
4221 octets have been read, EOF or an error occurs. Without this change
4222 some truncated ASN1 structures will not produce an error.
4225 *) Disable Heimdal support, since it hasn't been fully implemented.
4226 Still give the possibility to force the use of Heimdal, but with
4227 warnings and a request that patches get sent to openssl-dev.
4230 *) Add the VC-CE target, introduce the WINCE sysname, and add
4231 INSTALL.WCE and appropriate conditionals to make it build.
4232 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4234 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
4235 cygssl-x.y.z.dll, where x, y and z are the major, minor and
4236 edit numbers of the version.
4237 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4239 *) Introduce safe string copy and catenation functions
4240 (BUF_strlcpy() and BUF_strlcat()).
4241 [Ben Laurie (CHATS) and Richard Levitte]
4243 *) Avoid using fixed-size buffers for one-line DNs.
4244 [Ben Laurie (CHATS)]
4246 *) Add BUF_MEM_grow_clean() to avoid information leakage when
4247 resizing buffers containing secrets, and use where appropriate.
4248 [Ben Laurie (CHATS)]
4250 *) Avoid using fixed size buffers for configuration file location.
4251 [Ben Laurie (CHATS)]
4253 *) Avoid filename truncation for various CA files.
4254 [Ben Laurie (CHATS)]
4256 *) Use sizeof in preference to magic numbers.
4257 [Ben Laurie (CHATS)]
4259 *) Avoid filename truncation in cert requests.
4260 [Ben Laurie (CHATS)]
4262 *) Add assertions to check for (supposedly impossible) buffer
4264 [Ben Laurie (CHATS)]
4266 *) Don't cache truncated DNS entries in the local cache (this could
4267 potentially lead to a spoofing attack).
4268 [Ben Laurie (CHATS)]
4270 *) Fix various buffers to be large enough for hex/decimal
4271 representations in a platform independent manner.
4272 [Ben Laurie (CHATS)]
4274 *) Add CRYPTO_realloc_clean() to avoid information leakage when
4275 resizing buffers containing secrets, and use where appropriate.
4276 [Ben Laurie (CHATS)]
4278 *) Add BIO_indent() to avoid much slightly worrying code to do
4280 [Ben Laurie (CHATS)]
4282 *) Convert sprintf()/BIO_puts() to BIO_printf().
4283 [Ben Laurie (CHATS)]
4285 *) buffer_gets() could terminate with the buffer only half
4287 [Ben Laurie (CHATS)]
4289 *) Add assertions to prevent user-supplied crypto functions from
4290 overflowing internal buffers by having large block sizes, etc.
4291 [Ben Laurie (CHATS)]
4293 *) New OPENSSL_assert() macro (similar to assert(), but enabled
4295 [Ben Laurie (CHATS)]
4297 *) Eliminate unused copy of key in RC4.
4298 [Ben Laurie (CHATS)]
4300 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
4301 [Ben Laurie (CHATS)]
4303 *) Fix off-by-one error in EGD path.
4304 [Ben Laurie (CHATS)]
4306 *) If RANDFILE path is too long, ignore instead of truncating.
4307 [Ben Laurie (CHATS)]
4309 *) Eliminate unused and incorrectly sized X.509 structure
4311 [Ben Laurie (CHATS)]
4313 *) Eliminate unused and dangerous function knumber().
4314 [Ben Laurie (CHATS)]
4316 *) Eliminate unused and dangerous structure, KSSL_ERR.
4317 [Ben Laurie (CHATS)]
4319 *) Protect against overlong session ID context length in an encoded
4320 session object. Since these are local, this does not appear to be
4322 [Ben Laurie (CHATS)]
4324 *) Change from security patch (see 0.9.6e below) that did not affect
4325 the 0.9.6 release series:
4327 Remote buffer overflow in SSL3 protocol - an attacker could
4328 supply an oversized master key in Kerberos-enabled versions.
4330 [Ben Laurie (CHATS)]
4332 *) Change the SSL kerb5 codes to match RFC 2712.
4335 *) Make -nameopt work fully for req and add -reqopt switch.
4336 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
4338 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
4339 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
4341 *) Make sure tests can be performed even if the corresponding algorithms
4342 have been removed entirely. This was also the last step to make
4343 OpenSSL compilable with DJGPP under all reasonable conditions.
4344 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
4346 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
4347 to allow version independent disabling of normally unselected ciphers,
4348 which may be activated as a side-effect of selecting a single cipher.
4350 (E.g., cipher list string "RSA" enables ciphersuites that are left
4351 out of "ALL" because they do not provide symmetric encryption.
4352 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
4353 [Lutz Jaenicke, Bodo Moeller]
4355 *) Add appropriate support for separate platform-dependent build
4356 directories. The recommended way to make a platform-dependent
4357 build directory is the following (tested on Linux), maybe with
4360 # Place yourself outside of the OpenSSL source tree. In
4361 # this example, the environment variable OPENSSL_SOURCE
4362 # is assumed to contain the absolute OpenSSL source directory.
4363 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
4364 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
4365 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
4366 mkdir -p `dirname $F`
4367 ln -s $OPENSSL_SOURCE/$F $F
4370 To be absolutely sure not to disturb the source tree, a "make clean"
4371 is a good thing. If it isn't successfull, don't worry about it,
4372 it probably means the source directory is very clean.
4375 *) Make sure any ENGINE control commands make local copies of string
4376 pointers passed to them whenever necessary. Otherwise it is possible
4377 the caller may have overwritten (or deallocated) the original string
4378 data when a later ENGINE operation tries to use the stored values.
4379 [Götz Babin-Ebell <babinebell@trustcenter.de>]
4381 *) Improve diagnostics in file reading and command-line digests.
4382 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
4384 *) Add AES modes CFB and OFB to the object database. Correct an
4385 error in AES-CFB decryption.
4388 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
4389 allows existing EVP_CIPHER_CTX structures to be reused after
4390 calling EVP_*Final(). This behaviour is used by encryption
4391 BIOs and some applications. This has the side effect that
4392 applications must explicitly clean up cipher contexts with
4393 EVP_CIPHER_CTX_cleanup() or they will leak memory.
4396 *) Check the values of dna and dnb in bn_mul_recursive before calling
4397 bn_mul_comba (a non zero value means the a or b arrays do not contain
4398 n2 elements) and fallback to bn_mul_normal if either is not zero.
4401 *) Fix escaping of non-ASCII characters when using the -subj option
4402 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
4405 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
4406 form for "surname", serialNumber has no short form.
4407 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
4408 therefore remove "mail" short name for "internet 7".
4409 The OID for unique identifiers in X509 certificates is
4410 x500UniqueIdentifier, not uniqueIdentifier.
4411 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
4414 *) Add an "init" command to the ENGINE config module and auto initialize
4415 ENGINEs. Without any "init" command the ENGINE will be initialized
4416 after all ctrl commands have been executed on it. If init=1 the
4417 ENGINE is initailized at that point (ctrls before that point are run
4418 on the uninitialized ENGINE and after on the initialized one). If
4419 init=0 then the ENGINE will not be iniatialized at all.
4422 *) Fix the 'app_verify_callback' interface so that the user-defined
4423 argument is actually passed to the callback: In the
4424 SSL_CTX_set_cert_verify_callback() prototype, the callback
4425 declaration has been changed from
4428 int (*cb)(X509_STORE_CTX *,void *);
4429 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
4430 i=s->ctx->app_verify_callback(&ctx)
4431 has been changed into
4432 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
4434 To update applications using SSL_CTX_set_cert_verify_callback(),
4435 a dummy argument can be added to their callback functions.
4436 [D. K. Smetters <smetters@parc.xerox.com>]
4438 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
4439 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
4441 *) Add and OPENSSL_LOAD_CONF define which will cause
4442 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
4443 This allows older applications to transparently support certain
4444 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
4445 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
4446 load the config file and OPENSSL_add_all_algorithms_conf() which will
4447 always load it have also been added.
4450 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
4451 Adjust NIDs and EVP layer.
4452 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4454 *) Config modules support in openssl utility.
4456 Most commands now load modules from the config file,
4457 though in a few (such as version) this isn't done
4458 because it couldn't be used for anything.
4460 In the case of ca and req the config file used is
4461 the same as the utility itself: that is the -config
4462 command line option can be used to specify an
4466 *) Move default behaviour from OPENSSL_config(). If appname is NULL
4467 use "openssl_conf" if filename is NULL use default openssl config file.
4470 *) Add an argument to OPENSSL_config() to allow the use of an alternative
4471 config section name. Add a new flag to tolerate a missing config file
4472 and move code to CONF_modules_load_file().
4475 *) Support for crypto accelerator cards from Accelerated Encryption
4476 Processing, www.aep.ie. (Use engine 'aep')
4477 The support was copied from 0.9.6c [engine] and adapted/corrected
4478 to work with the new engine framework.
4479 [AEP Inc. and Richard Levitte]
4481 *) Support for SureWare crypto accelerator cards from Baltimore
4482 Technologies. (Use engine 'sureware')
4483 The support was copied from 0.9.6c [engine] and adapted
4484 to work with the new engine framework.
4487 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
4488 make the newer ENGINE framework commands for the CHIL engine work.
4489 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
4491 *) Make it possible to produce shared libraries on ReliantUNIX.
4492 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
4494 *) Add the configuration target debug-linux-ppro.
4495 Make 'openssl rsa' use the general key loading routines
4496 implemented in apps.c, and make those routines able to
4497 handle the key format FORMAT_NETSCAPE and the variant
4499 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4501 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4502 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4504 *) Add -keyform to rsautl, and document -engine.
4505 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
4507 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
4508 BIO_R_NO_SUCH_FILE error code rather than the generic
4509 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
4512 *) Add new functions
4514 ERR_peek_last_error_line
4515 ERR_peek_last_error_line_data.
4516 These are similar to
4519 ERR_peek_error_line_data,
4520 but report on the latest error recorded rather than the first one
4521 still in the error queue.
4522 [Ben Laurie, Bodo Moeller]
4524 *) default_algorithms option in ENGINE config module. This allows things
4526 default_algorithms = ALL
4527 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
4530 *) Prelminary ENGINE config module.
4533 *) New experimental application configuration code.
4536 *) Change the AES code to follow the same name structure as all other
4537 symmetric ciphers, and behave the same way. Move everything to
4538 the directory crypto/aes, thereby obsoleting crypto/rijndael.
4539 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4541 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
4542 [Ben Laurie and Theo de Raadt]
4544 *) Add option to output public keys in req command.
4545 [Massimiliano Pala madwolf@openca.org]
4547 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
4548 (up to about 10% better than before for P-192 and P-224).
4551 *) New functions/macros
4553 SSL_CTX_set_msg_callback(ctx, cb)
4554 SSL_CTX_set_msg_callback_arg(ctx, arg)
4555 SSL_set_msg_callback(ssl, cb)
4556 SSL_set_msg_callback_arg(ssl, arg)
4558 to request calling a callback function
4560 void cb(int write_p, int version, int content_type,
4561 const void *buf, size_t len, SSL *ssl, void *arg)
4563 whenever a protocol message has been completely received
4564 (write_p == 0) or sent (write_p == 1). Here 'version' is the
4565 protocol version according to which the SSL library interprets
4566 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
4567 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
4568 the content type as defined in the SSL 3.0/TLS 1.0 protocol
4569 specification (change_cipher_spec(20), alert(21), handshake(22)).
4570 'buf' and 'len' point to the actual message, 'ssl' to the
4571 SSL object, and 'arg' is the application-defined value set by
4572 SSL[_CTX]_set_msg_callback_arg().
4574 'openssl s_client' and 'openssl s_server' have new '-msg' options
4575 to enable a callback that displays all protocol messages.
4578 *) Change the shared library support so shared libraries are built as
4579 soon as the corresponding static library is finished, and thereby get
4580 openssl and the test programs linked against the shared library.
4581 This still only happens when the keyword "shard" has been given to
4582 the configuration scripts.
4584 NOTE: shared library support is still an experimental thing, and
4585 backward binary compatibility is still not guaranteed.
4586 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
4588 *) Add support for Subject Information Access extension.
4589 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4591 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
4592 additional bytes when new memory had to be allocated, not just
4593 when reusing an existing buffer.
4596 *) New command line and configuration option 'utf8' for the req command.
4597 This allows field values to be specified as UTF8 strings.
4600 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
4601 runs for the former and machine-readable output for the latter.
4604 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
4605 of the e-mail address in the DN (i.e., it will go into a certificate
4606 extension only). The new configuration file option 'email_in_dn = no'
4607 has the same effect.
4608 [Massimiliano Pala madwolf@openca.org]
4610 *) Change all functions with names starting with des_ to be starting
4611 with DES_ instead. Add wrappers that are compatible with libdes,
4612 but are named _ossl_old_des_*. Finally, add macros that map the
4613 des_* symbols to the corresponding _ossl_old_des_* if libdes
4614 compatibility is desired. If OpenSSL 0.9.6c compatibility is
4615 desired, the des_* symbols will be mapped to DES_*, with one
4618 Since we provide two compatibility mappings, the user needs to
4619 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
4620 compatibility is desired. The default (i.e., when that macro
4621 isn't defined) is OpenSSL 0.9.6c compatibility.
4623 There are also macros that enable and disable the support of old
4624 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
4625 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
4626 are defined, the default will apply: to support the old des routines.
4628 In either case, one must include openssl/des.h to get the correct
4629 definitions. Do not try to just include openssl/des_old.h, that
4632 NOTE: This is a major break of an old API into a new one. Software
4633 authors are encouraged to switch to the DES_ style functions. Some
4634 time in the future, des_old.h and the libdes compatibility functions
4635 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
4636 default), and then completely removed.
4639 *) Test for certificates which contain unsupported critical extensions.
4640 If such a certificate is found during a verify operation it is
4641 rejected by default: this behaviour can be overridden by either
4642 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
4643 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
4644 X509_supported_extension() has also been added which returns 1 if a
4645 particular extension is supported.
4648 *) Modify the behaviour of EVP cipher functions in similar way to digests
4649 to retain compatibility with existing code.
4652 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
4653 compatibility with existing code. In particular the 'ctx' parameter does
4654 not have to be to be initialized before the call to EVP_DigestInit() and
4655 it is tidied up after a call to EVP_DigestFinal(). New function
4656 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
4657 EVP_MD_CTX_copy() changed to not require the destination to be
4658 initialized valid and new function EVP_MD_CTX_copy_ex() added which
4659 requires the destination to be valid.
4661 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
4662 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
4665 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
4666 so that complete 'Handshake' protocol structures are kept in memory
4667 instead of overwriting 'msg_type' and 'length' with 'body' data.
4670 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
4671 [Massimo Santin via Richard Levitte]
4673 *) Major restructuring to the underlying ENGINE code. This includes
4674 reduction of linker bloat, separation of pure "ENGINE" manipulation
4675 (initialisation, etc) from functionality dealing with implementations
4676 of specific crypto iterfaces. This change also introduces integrated
4677 support for symmetric ciphers and digest implementations - so ENGINEs
4678 can now accelerate these by providing EVP_CIPHER and EVP_MD
4679 implementations of their own. This is detailed in crypto/engine/README
4680 as it couldn't be adequately described here. However, there are a few
4681 API changes worth noting - some RSA, DSA, DH, and RAND functions that
4682 were changed in the original introduction of ENGINE code have now
4683 reverted back - the hooking from this code to ENGINE is now a good
4684 deal more passive and at run-time, operations deal directly with
4685 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
4686 dereferencing through an ENGINE pointer any more. Also, the ENGINE
4687 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
4688 they were not being used by the framework as there is no concept of a
4689 BIGNUM_METHOD and they could not be generalised to the new
4690 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
4691 ENGINE_cpy() has been removed as it cannot be consistently defined in
4695 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
4698 *) Change mkdef.pl to sort symbols that get the same entry number,
4699 and make sure the automatically generated functions ERR_load_*
4700 become part of libeay.num as well.
4703 *) New function SSL_renegotiate_pending(). This returns true once
4704 renegotiation has been requested (either SSL_renegotiate() call
4705 or HelloRequest/ClientHello receveived from the peer) and becomes
4706 false once a handshake has been completed.
4707 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
4708 sends a HelloRequest, but does not ensure that a handshake takes
4709 place. SSL_renegotiate_pending() is useful for checking if the
4710 client has followed the request.)
4713 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
4714 By default, clients may request session resumption even during
4715 renegotiation (if session ID contexts permit); with this option,
4716 session resumption is possible only in the first handshake.
4718 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4719 more bits available for options that should not be part of
4720 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4723 *) Add some demos for certificate and certificate request creation.
4726 *) Make maximum certificate chain size accepted from the peer application
4727 settable (SSL*_get/set_max_cert_list()), as proposed by
4728 "Douglas E. Engert" <deengert@anl.gov>.
4731 *) Add support for shared libraries for Unixware-7
4732 (Boyd Lynn Gerber <gerberb@zenez.com>).
4735 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4736 be done prior to destruction. Use this to unload error strings from
4737 ENGINEs that load their own error strings. NB: This adds two new API
4738 functions to "get" and "set" this destroy handler in an ENGINE.
4741 *) Alter all existing ENGINE implementations (except "openssl" and
4742 "openbsd") to dynamically instantiate their own error strings. This
4743 makes them more flexible to be built both as statically-linked ENGINEs
4744 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4745 Also, add stub code to each that makes building them as self-contained
4746 shared-libraries easier (see README.ENGINE).
4749 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4750 implementations into applications that are completely implemented in
4751 self-contained shared-libraries. The "dynamic" ENGINE exposes control
4752 commands that can be used to configure what shared-library to load and
4753 to control aspects of the way it is handled. Also, made an update to
4754 the README.ENGINE file that brings its information up-to-date and
4755 provides some information and instructions on the "dynamic" ENGINE
4756 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
4759 *) Make it possible to unload ranges of ERR strings with a new
4760 "ERR_unload_strings" function.
4763 *) Add a copy() function to EVP_MD.
4766 *) Make EVP_MD routines take a context pointer instead of just the
4767 md_data void pointer.
4770 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
4771 that the digest can only process a single chunk of data
4772 (typically because it is provided by a piece of
4773 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
4774 is only going to provide a single chunk of data, and hence the
4775 framework needn't accumulate the data for oneshot drivers.
4778 *) As with "ERR", make it possible to replace the underlying "ex_data"
4779 functions. This change also alters the storage and management of global
4780 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
4781 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
4782 index counters. The API functions that use this state have been changed
4783 to take a "class_index" rather than pointers to the class's local STACK
4784 and counter, and there is now an API function to dynamically create new
4785 classes. This centralisation allows us to (a) plug a lot of the
4786 thread-safety problems that existed, and (b) makes it possible to clean
4787 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
4788 such data would previously have always leaked in application code and
4789 workarounds were in place to make the memory debugging turn a blind eye
4790 to it. Application code that doesn't use this new function will still
4791 leak as before, but their memory debugging output will announce it now
4792 rather than letting it slide.
4794 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
4795 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
4796 has a return value to indicate success or failure.
4799 *) Make it possible to replace the underlying "ERR" functions such that the
4800 global state (2 LHASH tables and 2 locks) is only used by the "default"
4801 implementation. This change also adds two functions to "get" and "set"
4802 the implementation prior to it being automatically set the first time
4803 any other ERR function takes place. Ie. an application can call "get",
4804 pass the return value to a module it has just loaded, and that module
4805 can call its own "set" function using that value. This means the
4806 module's "ERR" operations will use (and modify) the error state in the
4807 application and not in its own statically linked copy of OpenSSL code.
4810 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
4811 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
4812 the operation, and provides a more encapsulated way for external code
4813 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
4814 to use these functions rather than manually incrementing the counts.
4816 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
4819 *) Add EVP test program.
4822 *) Add symmetric cipher support to ENGINE. Expect the API to change!
4825 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
4826 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
4827 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
4828 These allow a CRL to be built without having to access X509_CRL fields
4829 directly. Modify 'ca' application to use new functions.
4832 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
4833 bug workarounds. Rollback attack detection is a security feature.
4834 The problem will only arise on OpenSSL servers when TLSv1 is not
4835 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
4836 Software authors not wanting to support TLSv1 will have special reasons
4837 for their choice and can explicitly enable this option.
4838 [Bodo Moeller, Lutz Jaenicke]
4840 *) Rationalise EVP so it can be extended: don't include a union of
4841 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
4842 (similar to those existing for EVP_CIPHER_CTX).
4847 EVP_MD_CTX_init(&md); /* new function call */
4848 EVP_DigestInit(&md, EVP_sha1());
4849 EVP_DigestUpdate(&md, in, len);
4850 EVP_DigestFinal(&md, out, NULL);
4851 EVP_MD_CTX_cleanup(&md); /* new function call */
4855 *) Make DES key schedule conform to the usual scheme, as well as
4856 correcting its structure. This means that calls to DES functions
4857 now have to pass a pointer to a des_key_schedule instead of a
4858 plain des_key_schedule (which was actually always a pointer
4861 des_key_schedule ks;
4863 des_set_key_checked(..., &ks);
4864 des_ncbc_encrypt(..., &ks, ...);
4866 (Note that a later change renames 'des_...' into 'DES_...'.)
4869 *) Initial reduction of linker bloat: the use of some functions, such as
4870 PEM causes large amounts of unused functions to be linked in due to
4871 poor organisation. For example pem_all.c contains every PEM function
4872 which has a knock on effect of linking in large amounts of (unused)
4873 ASN1 code. Grouping together similar functions and splitting unrelated
4874 functions prevents this.
4877 *) Cleanup of EVP macros.
4880 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
4881 correct _ecb suffix.
4884 *) Add initial OCSP responder support to ocsp application. The
4885 revocation information is handled using the text based index
4886 use by the ca application. The responder can either handle
4887 requests generated internally, supplied in files (for example
4888 via a CGI script) or using an internal minimal server.
4891 *) Add configuration choices to get zlib compression for TLS.
4894 *) Changes to Kerberos SSL for RFC 2712 compliance:
4895 1. Implemented real KerberosWrapper, instead of just using
4896 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
4897 2. Implemented optional authenticator field of KerberosWrapper.
4899 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
4900 and authenticator structs; see crypto/krb5/.
4902 Generalized Kerberos calls to support multiple Kerberos libraries.
4903 [Vern Staats <staatsvr@asc.hpc.mil>,
4904 Jeffrey Altman <jaltman@columbia.edu>
4905 via Richard Levitte]
4907 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
4908 already does with RSA. testdsa.h now has 'priv_key/pub_key'
4909 values for each of the key sizes rather than having just
4910 parameters (and 'speed' generating keys each time).
4913 *) Speed up EVP routines.
4916 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
4917 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
4918 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
4919 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
4921 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
4922 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
4923 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
4926 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
4928 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
4931 *) Added the OS2-EMX target.
4932 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
4934 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
4935 to support NCONF routines in extension code. New function CONF_set_nconf()
4936 to allow functions which take an NCONF to also handle the old LHASH
4937 structure: this means that the old CONF compatible routines can be
4938 retained (in particular wrt extensions) without having to duplicate the
4939 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
4942 *) Enhance the general user interface with mechanisms for inner control
4943 and with possibilities to have yes/no kind of prompts.
4946 *) Change all calls to low level digest routines in the library and
4947 applications to use EVP. Add missing calls to HMAC_cleanup() and
4948 don't assume HMAC_CTX can be copied using memcpy().
4949 [Verdon Walker <VWalker@novell.com>, Steve Henson]
4951 *) Add the possibility to control engines through control names but with
4952 arbitrary arguments instead of just a string.
4953 Change the key loaders to take a UI_METHOD instead of a callback
4954 function pointer. NOTE: this breaks binary compatibility with earlier
4955 versions of OpenSSL [engine].
4956 Adapt the nCipher code for these new conditions and add a card insertion
4960 *) Enhance the general user interface with mechanisms to better support
4961 dialog box interfaces, application-defined prompts, the possibility
4962 to use defaults (for example default passwords from somewhere else)
4963 and interrupts/cancellations.
4966 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
4967 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
4970 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
4971 tidy up some unnecessarily weird code in 'sk_new()').
4972 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
4974 *) Change the key loading routines for ENGINEs to use the same kind
4975 callback (pem_password_cb) as all other routines that need this
4979 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
4980 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
4981 than this minimum value is recommended.
4984 *) New random seeder for OpenVMS, using the system process statistics
4985 that are easily reachable.
4988 *) Windows apparently can't transparently handle global
4989 variables defined in DLLs. Initialisations such as:
4991 const ASN1_ITEM *it = &ASN1_INTEGER_it;
4993 wont compile. This is used by the any applications that need to
4994 declare their own ASN1 modules. This was fixed by adding the option
4995 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
4996 needed for static libraries under Win32.
4999 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
5000 setting of purpose and trust fields. New X509_STORE trust and
5001 purpose functions and tidy up setting in other SSL functions.
5004 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
5005 structure. These are inherited by X509_STORE_CTX when it is
5006 initialised. This allows various defaults to be set in the
5007 X509_STORE structure (such as flags for CRL checking and custom
5008 purpose or trust settings) for functions which only use X509_STORE_CTX
5009 internally such as S/MIME.
5011 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
5012 trust settings if they are not set in X509_STORE. This allows X509_STORE
5013 purposes and trust (in S/MIME for example) to override any set by default.
5015 Add command line options for CRL checking to smime, s_client and s_server
5019 *) Initial CRL based revocation checking. If the CRL checking flag(s)
5020 are set then the CRL is looked up in the X509_STORE structure and
5021 its validity and signature checked, then if the certificate is found
5022 in the CRL the verify fails with a revoked error.
5024 Various new CRL related callbacks added to X509_STORE_CTX structure.
5026 Command line options added to 'verify' application to support this.
5028 This needs some additional work, such as being able to handle multiple
5029 CRLs with different times, extension based lookup (rather than just
5030 by subject name) and ultimately more complete V2 CRL extension
5034 *) Add a general user interface API (crypto/ui/). This is designed
5035 to replace things like des_read_password and friends (backward
5036 compatibility functions using this new API are provided).
5037 The purpose is to remove prompting functions from the DES code
5038 section as well as provide for prompting through dialog boxes in
5039 a window system and the like.
5042 *) Add "ex_data" support to ENGINE so implementations can add state at a
5043 per-structure level rather than having to store it globally.
5046 *) Make it possible for ENGINE structures to be copied when retrieved by
5047 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
5048 This causes the "original" ENGINE structure to act like a template,
5049 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
5050 operational state can be localised to each ENGINE structure, despite the
5051 fact they all share the same "methods". New ENGINE structures returned in
5052 this case have no functional references and the return value is the single
5053 structural reference. This matches the single structural reference returned
5054 by ENGINE_by_id() normally, when it is incremented on the pre-existing
5058 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
5059 needs to match any other type at all we need to manually clear the
5063 *) Changes to the "openssl engine" utility to include;
5064 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
5065 about an ENGINE's available control commands.
5066 - executing control commands from command line arguments using the
5067 '-pre' and '-post' switches. '-post' is only used if '-t' is
5068 specified and the ENGINE is successfully initialised. The syntax for
5069 the individual commands are colon-separated, for example;
5070 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
5073 *) New dynamic control command support for ENGINEs. ENGINEs can now
5074 declare their own commands (numbers), names (strings), descriptions,
5075 and input types for run-time discovery by calling applications. A
5076 subset of these commands are implicitly classed as "executable"
5077 depending on their input type, and only these can be invoked through
5078 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
5079 can be based on user input, config files, etc). The distinction is
5080 that "executable" commands cannot return anything other than a boolean
5081 result and can only support numeric or string input, whereas some
5082 discoverable commands may only be for direct use through
5083 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
5084 pointers, or other custom uses. The "executable" commands are to
5085 support parameterisations of ENGINE behaviour that can be
5086 unambiguously defined by ENGINEs and used consistently across any
5087 OpenSSL-based application. Commands have been added to all the
5088 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
5089 control over shared-library paths without source code alterations.
5092 *) Changed all ENGINE implementations to dynamically allocate their
5093 ENGINEs rather than declaring them statically. Apart from this being
5094 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
5095 this also allows the implementations to compile without using the
5096 internal engine_int.h header.
5099 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
5100 'const' value. Any code that should be able to modify a RAND_METHOD
5101 should already have non-const pointers to it (ie. they should only
5102 modify their own ones).
5105 *) Made a variety of little tweaks to the ENGINE code.
5106 - "atalla" and "ubsec" string definitions were moved from header files
5107 to C code. "nuron" string definitions were placed in variables
5108 rather than hard-coded - allowing parameterisation of these values
5109 later on via ctrl() commands.
5110 - Removed unused "#if 0"'d code.
5111 - Fixed engine list iteration code so it uses ENGINE_free() to release
5112 structural references.
5113 - Constified the RAND_METHOD element of ENGINE structures.
5114 - Constified various get/set functions as appropriate and added
5115 missing functions (including a catch-all ENGINE_cpy that duplicates
5116 all ENGINE values onto a new ENGINE except reference counts/state).
5117 - Removed NULL parameter checks in get/set functions. Setting a method
5118 or function to NULL is a way of cancelling out a previously set
5119 value. Passing a NULL ENGINE parameter is just plain stupid anyway
5120 and doesn't justify the extra error symbols and code.
5121 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
5122 flags from engine_int.h to engine.h.
5123 - Changed prototypes for ENGINE handler functions (init(), finish(),
5124 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
5127 *) Implement binary inversion algorithm for BN_mod_inverse in addition
5128 to the algorithm using long division. The binary algorithm can be
5129 used only if the modulus is odd. On 32-bit systems, it is faster
5130 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
5131 roughly 5-15% for 256-bit moduli), so we use it only for moduli
5132 up to 450 bits. In 64-bit environments, the binary algorithm
5133 appears to be advantageous for much longer moduli; here we use it
5134 for moduli up to 2048 bits.
5137 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
5138 could not support the combine flag in choice fields.
5141 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
5142 extensions from a certificate request to the certificate.
5145 *) Allow multiple 'certopt' and 'nameopt' options to be separated
5146 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
5147 file: this allows the display of the certificate about to be
5148 signed to be customised, to allow certain fields to be included
5149 or excluded and extension details. The old system didn't display
5150 multicharacter strings properly, omitted fields not in the policy
5151 and couldn't display additional details such as extensions.
5154 *) Function EC_POINTs_mul for multiple scalar multiplication
5155 of an arbitrary number of elliptic curve points
5156 \sum scalars[i]*points[i],
5157 optionally including the generator defined for the EC_GROUP:
5158 scalar*generator + \sum scalars[i]*points[i].
5160 EC_POINT_mul is a simple wrapper function for the typical case
5161 that the point list has just one item (besides the optional
5165 *) First EC_METHODs for curves over GF(p):
5167 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
5168 operations and provides various method functions that can also
5169 operate with faster implementations of modular arithmetic.
5171 EC_GFp_mont_method() reuses most functions that are part of
5172 EC_GFp_simple_method, but uses Montgomery arithmetic.
5174 [Bodo Moeller; point addition and point doubling
5175 implementation directly derived from source code provided by
5176 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
5178 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
5179 crypto/ec/ec_lib.c):
5181 Curves are EC_GROUP objects (with an optional group generator)
5182 based on EC_METHODs that are built into the library.
5184 Points are EC_POINT objects based on EC_GROUP objects.
5186 Most of the framework would be able to handle curves over arbitrary
5187 finite fields, but as there are no obvious types for fields other
5188 than GF(p), some functions are limited to that for now.
5191 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
5192 that the file contains a complete HTTP response.
5195 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
5196 change the def and num file printf format specifier from "%-40sXXX"
5197 to "%-39s XXX". The latter will always guarantee a space after the
5198 field while the former will cause them to run together if the field
5199 is 40 of more characters long.
5202 *) Constify the cipher and digest 'method' functions and structures
5203 and modify related functions to take constant EVP_MD and EVP_CIPHER
5207 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
5208 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
5211 *) Modify EVP_Digest*() routines so they now return values. Although the
5212 internal software routines can never fail additional hardware versions
5216 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
5218 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
5219 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
5222 ERR_R_NESTED_ASN1_ERROR
5224 ERR_R_MISSING_ASN1_EOS
5225 were 4 .. 9, conflicting with
5226 ERR_LIB_RSA (= ERR_R_RSA_LIB)
5228 ERR_LIB_PEM (= ERR_R_PEM_LIB).
5229 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
5231 Add new error code 'ERR_R_INTERNAL_ERROR'.
5234 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
5238 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
5239 sets the subject name for a new request or supersedes the
5240 subject name in a given request. Formats that can be parsed are
5241 'CN=Some Name, OU=myOU, C=IT'
5243 'CN=Some Name/OU=myOU/C=IT'.
5245 Add options '-batch' and '-verbose' to 'openssl req'.
5246 [Massimiliano Pala <madwolf@hackmasters.net>]
5248 *) Introduce the possibility to access global variables through
5249 functions on platform were that's the best way to handle exporting
5250 global variables in shared libraries. To enable this functionality,
5251 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
5252 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
5253 is normally done by Configure or something similar).
5255 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
5256 in the source file (foo.c) like this:
5258 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
5259 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
5261 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
5262 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
5264 OPENSSL_DECLARE_GLOBAL(int,foo);
5265 #define foo OPENSSL_GLOBAL_REF(foo)
5266 OPENSSL_DECLARE_GLOBAL(double,bar);
5267 #define bar OPENSSL_GLOBAL_REF(bar)
5269 The #defines are very important, and therefore so is including the
5270 header file everywhere where the defined globals are used.
5272 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
5273 of ASN.1 items, but that structure is a bit different.
5275 The largest change is in util/mkdef.pl which has been enhanced with
5276 better and easier to understand logic to choose which symbols should
5277 go into the Windows .def files as well as a number of fixes and code
5278 cleanup (among others, algorithm keywords are now sorted
5279 lexicographically to avoid constant rewrites).
5282 *) In BN_div() keep a copy of the sign of 'num' before writing the
5283 result to 'rm' because if rm==num the value will be overwritten
5284 and produce the wrong result if 'num' is negative: this caused
5285 problems with BN_mod() and BN_nnmod().
5288 *) Function OCSP_request_verify(). This checks the signature on an
5289 OCSP request and verifies the signer certificate. The signer
5290 certificate is just checked for a generic purpose and OCSP request
5294 *) Add OCSP_check_validity() function to check the validity of OCSP
5295 responses. OCSP responses are prepared in real time and may only
5296 be a few seconds old. Simply checking that the current time lies
5297 between thisUpdate and nextUpdate max reject otherwise valid responses
5298 caused by either OCSP responder or client clock inaccuracy. Instead
5299 we allow thisUpdate and nextUpdate to fall within a certain period of
5300 the current time. The age of the response can also optionally be
5301 checked. Two new options -validity_period and -status_age added to
5305 *) If signature or public key algorithm is unrecognized print out its
5306 OID rather that just UNKNOWN.
5309 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
5310 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
5311 ID to be generated from the issuer certificate alone which can then be
5312 passed to OCSP_id_issuer_cmp().
5315 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
5316 ASN1 modules to export functions returning ASN1_ITEM pointers
5317 instead of the ASN1_ITEM structures themselves. This adds several
5318 new macros which allow the underlying ASN1 function/structure to
5319 be accessed transparently. As a result code should not use ASN1_ITEM
5320 references directly (such as &X509_it) but instead use the relevant
5321 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
5322 use of the new ASN1 code on platforms where exporting structures
5323 is problematical (for example in shared libraries) but exporting
5324 functions returning pointers to structures is not.
5327 *) Add support for overriding the generation of SSL/TLS session IDs.
5328 These callbacks can be registered either in an SSL_CTX or per SSL.
5329 The purpose of this is to allow applications to control, if they wish,
5330 the arbitrary values chosen for use as session IDs, particularly as it
5331 can be useful for session caching in multiple-server environments. A
5332 command-line switch for testing this (and any client code that wishes
5333 to use such a feature) has been added to "s_server".
5334 [Geoff Thorpe, Lutz Jaenicke]
5336 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
5337 of the form '#if defined(...) || defined(...) || ...' and
5338 '#if !defined(...) && !defined(...) && ...'. This also avoids
5339 the growing number of special cases it was previously handling.
5342 *) Make all configuration macros available for application by making
5343 sure they are available in opensslconf.h, by giving them names starting
5344 with "OPENSSL_" to avoid conflicts with other packages and by making
5345 sure e_os2.h will cover all platform-specific cases together with
5347 Additionally, it is now possible to define configuration/platform-
5348 specific names (called "system identities"). In the C code, these
5349 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
5350 macro with the name beginning with "OPENSSL_SYS_", which is determined
5351 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
5355 *) New option -set_serial to 'req' and 'x509' this allows the serial
5356 number to use to be specified on the command line. Previously self
5357 signed certificates were hard coded with serial number 0 and the
5358 CA options of 'x509' had to use a serial number in a file which was
5362 *) New options to 'ca' utility to support V2 CRL entry extensions.
5363 Currently CRL reason, invalidity date and hold instruction are
5364 supported. Add new CRL extensions to V3 code and some new objects.
5367 *) New function EVP_CIPHER_CTX_set_padding() this is used to
5368 disable standard block padding (aka PKCS#5 padding) in the EVP
5369 API, which was previously mandatory. This means that the data is
5370 not padded in any way and so the total length much be a multiple
5371 of the block size, otherwise an error occurs.
5374 *) Initial (incomplete) OCSP SSL support.
5377 *) New function OCSP_parse_url(). This splits up a URL into its host,
5378 port and path components: primarily to parse OCSP URLs. New -url
5379 option to ocsp utility.
5382 *) New nonce behavior. The return value of OCSP_check_nonce() now
5383 reflects the various checks performed. Applications can decide
5384 whether to tolerate certain situations such as an absent nonce
5385 in a response when one was present in a request: the ocsp application
5386 just prints out a warning. New function OCSP_add1_basic_nonce()
5387 this is to allow responders to include a nonce in a response even if
5388 the request is nonce-less.
5391 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
5392 skipped when using openssl x509 multiple times on a single input file,
5393 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
5396 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
5397 set string type: to handle setting ASN1_TIME structures. Fix ca
5398 utility to correctly initialize revocation date of CRLs.
5401 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
5402 the clients preferred ciphersuites and rather use its own preferences.
5403 Should help to work around M$ SGC (Server Gated Cryptography) bug in
5404 Internet Explorer by ensuring unchanged hash method during stepup.
5405 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
5408 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
5409 to aes and add a new 'exist' option to print out symbols that don't
5413 *) Additional options to ocsp utility to allow flags to be set and
5414 additional certificates supplied.
5417 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
5418 OCSP client a number of certificate to only verify the response
5422 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
5423 handle the new API. Currently only ECB, CBC modes supported. Add new
5426 Add TLS AES ciphersuites as described in RFC3268, "Advanced
5427 Encryption Standard (AES) Ciphersuites for Transport Layer
5428 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
5429 not enabled by default and were not part of the "ALL" ciphersuite
5430 alias because they were not yet official; they could be
5431 explicitly requested by specifying the "AESdraft" ciphersuite
5432 group alias. In the final release of OpenSSL 0.9.7, the group
5433 alias is called "AES" and is part of "ALL".)
5434 [Ben Laurie, Steve Henson, Bodo Moeller]
5436 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
5437 request to response.
5440 *) Functions for OCSP responders. OCSP_request_onereq_count(),
5441 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
5442 extract information from a certificate request. OCSP_response_create()
5443 creates a response and optionally adds a basic response structure.
5444 OCSP_basic_add1_status() adds a complete single response to a basic
5445 response and returns the OCSP_SINGLERESP structure just added (to allow
5446 extensions to be included for example). OCSP_basic_add1_cert() adds a
5447 certificate to a basic response and OCSP_basic_sign() signs a basic
5448 response with various flags. New helper functions ASN1_TIME_check()
5449 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
5450 (converts ASN1_TIME to GeneralizedTime).
5453 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
5454 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
5455 structure from a certificate. X509_pubkey_digest() digests the public_key
5456 contents: this is used in various key identifiers.
5459 *) Make sk_sort() tolerate a NULL argument.
5460 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
5462 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
5463 passed by the function are trusted implicitly. If any of them signed the
5464 response then it is assumed to be valid and is not verified.
5467 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
5468 to data. This was previously part of the PKCS7 ASN1 code. This
5469 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
5470 [Steve Henson, reported by Kenneth R. Robinette
5471 <support@securenetterm.com>]
5473 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
5474 routines: without these tracing memory leaks is very painful.
5475 Fix leaks in PKCS12 and PKCS7 routines.
5478 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
5479 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
5480 effectively meant GeneralizedTime would never be used. Now it
5481 is initialised to -1 but X509_time_adj() now has to check the value
5482 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
5483 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
5484 [Steve Henson, reported by Kenneth R. Robinette
5485 <support@securenetterm.com>]
5487 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
5488 result in a zero length in the ASN1_INTEGER structure which was
5489 not consistent with the structure when d2i_ASN1_INTEGER() was used
5490 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
5491 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
5492 where it did not print out a minus for negative ASN1_INTEGER.
5495 *) Add summary printout to ocsp utility. The various functions which
5496 convert status values to strings have been renamed to:
5497 OCSP_response_status_str(), OCSP_cert_status_str() and
5498 OCSP_crl_reason_str() and are no longer static. New options
5499 to verify nonce values and to disable verification. OCSP response
5500 printout format cleaned up.
5503 *) Add additional OCSP certificate checks. These are those specified
5504 in RFC2560. This consists of two separate checks: the CA of the
5505 certificate being checked must either be the OCSP signer certificate
5506 or the issuer of the OCSP signer certificate. In the latter case the
5507 OCSP signer certificate must contain the OCSP signing extended key
5508 usage. This check is performed by attempting to match the OCSP
5509 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
5510 in the OCSP_CERTID structures of the response.
5513 *) Initial OCSP certificate verification added to OCSP_basic_verify()
5514 and related routines. This uses the standard OpenSSL certificate
5515 verify routines to perform initial checks (just CA validity) and
5516 to obtain the certificate chain. Then additional checks will be
5517 performed on the chain. Currently the root CA is checked to see
5518 if it is explicitly trusted for OCSP signing. This is used to set
5519 a root CA as a global signing root: that is any certificate that
5520 chains to that CA is an acceptable OCSP signing certificate.
5523 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
5524 extensions from a separate configuration file.
5525 As when reading extensions from the main configuration file,
5526 the '-extensions ...' option may be used for specifying the
5528 [Massimiliano Pala <madwolf@comune.modena.it>]
5530 *) New OCSP utility. Allows OCSP requests to be generated or
5531 read. The request can be sent to a responder and the output
5532 parsed, outputed or printed in text form. Not complete yet:
5533 still needs to check the OCSP response validity.
5536 *) New subcommands for 'openssl ca':
5537 'openssl ca -status <serial>' prints the status of the cert with
5538 the given serial number (according to the index file).
5539 'openssl ca -updatedb' updates the expiry status of certificates
5541 [Massimiliano Pala <madwolf@comune.modena.it>]
5543 *) New '-newreq-nodes' command option to CA.pl. This is like
5544 '-newreq', but calls 'openssl req' with the '-nodes' option
5545 so that the resulting key is not encrypted.
5546 [Damien Miller <djm@mindrot.org>]
5548 *) New configuration for the GNU Hurd.
5549 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
5551 *) Initial code to implement OCSP basic response verify. This
5552 is currently incomplete. Currently just finds the signer's
5553 certificate and verifies the signature on the response.
5556 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
5557 value of OPENSSLDIR. This is available via the new '-d' option
5558 to 'openssl version', and is also included in 'openssl version -a'.
5561 *) Allowing defining memory allocation callbacks that will be given
5562 file name and line number information in additional arguments
5563 (a const char* and an int). The basic functionality remains, as
5564 well as the original possibility to just replace malloc(),
5565 realloc() and free() by functions that do not know about these
5566 additional arguments. To register and find out the current
5567 settings for extended allocation functions, the following
5568 functions are provided:
5570 CRYPTO_set_mem_ex_functions
5571 CRYPTO_set_locked_mem_ex_functions
5572 CRYPTO_get_mem_ex_functions
5573 CRYPTO_get_locked_mem_ex_functions
5575 These work the same way as CRYPTO_set_mem_functions and friends.
5576 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
5577 extended allocation function is enabled.
5578 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
5579 a conventional allocation function is enabled.
5580 [Richard Levitte, Bodo Moeller]
5582 *) Finish off removing the remaining LHASH function pointer casts.
5583 There should no longer be any prototype-casting required when using
5584 the LHASH abstraction, and any casts that remain are "bugs". See
5585 the callback types and macros at the head of lhash.h for details
5586 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
5589 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
5590 If /dev/[u]random devices are not available or do not return enough
5591 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
5593 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
5594 /etc/entropy will be queried once each in this sequence, quering stops
5595 when enough entropy was collected without querying more sockets.
5598 *) Change the Unix RAND_poll() variant to be able to poll several
5599 random devices, as specified by DEVRANDOM, until a sufficient amount
5600 of data has been collected. We spend at most 10 ms on each file
5601 (select timeout) and read in non-blocking mode. DEVRANDOM now
5602 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
5603 (previously it was just the string "/dev/urandom"), so on typical
5604 platforms the 10 ms delay will never occur.
5605 Also separate out the Unix variant to its own file, rand_unix.c.
5606 For VMS, there's a currently-empty rand_vms.c.
5609 *) Move OCSP client related routines to ocsp_cl.c. These
5610 provide utility functions which an application needing
5611 to issue a request to an OCSP responder and analyse the
5612 response will typically need: as opposed to those which an
5613 OCSP responder itself would need which will be added later.
5615 OCSP_request_sign() signs an OCSP request with an API similar
5616 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
5617 response. OCSP_response_get1_basic() extracts basic response
5618 from response. OCSP_resp_find_status(): finds and extracts status
5619 information from an OCSP_CERTID structure (which will be created
5620 when the request structure is built). These are built from lower
5621 level functions which work on OCSP_SINGLERESP structures but
5622 wont normally be used unless the application wishes to examine
5623 extensions in the OCSP response for example.
5625 Replace nonce routines with a pair of functions.
5626 OCSP_request_add1_nonce() adds a nonce value and optionally
5627 generates a random value. OCSP_check_nonce() checks the
5628 validity of the nonce in an OCSP response.
5631 *) Change function OCSP_request_add() to OCSP_request_add0_id().
5632 This doesn't copy the supplied OCSP_CERTID and avoids the
5633 need to free up the newly created id. Change return type
5634 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
5635 This can then be used to add extensions to the request.
5636 Deleted OCSP_request_new(), since most of its functionality
5637 is now in OCSP_REQUEST_new() (and the case insensitive name
5638 clash) apart from the ability to set the request name which
5639 will be added elsewhere.
5642 *) Update OCSP API. Remove obsolete extensions argument from
5643 various functions. Extensions are now handled using the new
5644 OCSP extension code. New simple OCSP HTTP function which
5645 can be used to send requests and parse the response.
5648 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
5649 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
5650 uses the special reorder version of SET OF to sort the attributes
5651 and reorder them to match the encoded order. This resolves a long
5652 standing problem: a verify on a PKCS7 structure just after signing
5653 it used to fail because the attribute order did not match the
5654 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
5655 it uses the received order. This is necessary to tolerate some broken
5656 software that does not order SET OF. This is handled by encoding
5657 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
5658 to produce the required SET OF.
5661 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
5662 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
5663 files to get correct declarations of the ASN.1 item variables.
5666 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
5667 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
5668 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
5669 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
5670 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
5671 ASN1_ITEM and no wrapper functions.
5674 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
5675 replace the old function pointer based I/O routines. Change most of
5676 the *_d2i_bio() and *_d2i_fp() functions to use these.
5679 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
5680 lines, recognice more "algorithms" that can be deselected, and make
5681 it complain about algorithm deselection that isn't recognised.
5684 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
5685 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
5686 to use new functions. Add NO_ASN1_OLD which can be set to remove
5687 some old style ASN1 functions: this can be used to determine if old
5688 code will still work when these eventually go away.
5691 *) New extension functions for OCSP structures, these follow the
5692 same conventions as certificates and CRLs.
5695 *) New function X509V3_add1_i2d(). This automatically encodes and
5696 adds an extension. Its behaviour can be customised with various
5697 flags to append, replace or delete. Various wrappers added for
5698 certifcates and CRLs.
5701 *) Fix to avoid calling the underlying ASN1 print routine when
5702 an extension cannot be parsed. Correct a typo in the
5703 OCSP_SERVICELOC extension. Tidy up print OCSP format.
5706 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
5707 entries for variables.
5710 *) Add functionality to apps/openssl.c for detecting locking
5711 problems: As the program is single-threaded, all we have
5712 to do is register a locking callback using an array for
5713 storing which locks are currently held by the program.
5716 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
5717 SSL_get_ex_data_X509_STORE_idx(), which is used in
5718 ssl_verify_cert_chain() and thus can be called at any time
5719 during TLS/SSL handshakes so that thread-safety is essential.
5720 Unfortunately, the ex_data design is not at all suited
5721 for multi-threaded use, so it probably should be abolished.
5724 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5725 [Broadcom, tweaked and integrated by Geoff Thorpe]
5727 *) Move common extension printing code to new function
5728 X509V3_print_extensions(). Reorganise OCSP print routines and
5729 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5732 *) New function X509_signature_print() to remove duplication in some
5736 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5737 set (this was treated exactly the same as SET OF previously). This
5738 is used to reorder the STACK representing the structure to match the
5739 encoding. This will be used to get round a problem where a PKCS7
5740 structure which was signed could not be verified because the STACK
5741 order did not reflect the encoded order.
5744 *) Reimplement the OCSP ASN1 module using the new code.
5747 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5748 for its ASN1 operations. The old style function pointers still exist
5749 for now but they will eventually go away.
5752 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
5753 completely replaces the old ASN1 functionality with a table driven
5754 encoder and decoder which interprets an ASN1_ITEM structure describing
5755 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
5756 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
5757 has also been converted to the new form.
5760 *) Change BN_mod_exp_recp so that negative moduli are tolerated
5761 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
5762 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
5763 for negative moduli.
5766 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
5767 of not touching the result's sign bit.
5770 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
5774 *) Changed the LHASH code to use prototypes for callbacks, and created
5775 macros to declare and implement thin (optionally static) functions
5776 that provide type-safety and avoid function pointer casting for the
5777 type-specific callbacks.
5780 *) Added Kerberos Cipher Suites to be used with TLS, as written in
5782 [Veers Staats <staatsvr@asc.hpc.mil>,
5783 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
5785 *) Reformat the FAQ so the different questions and answers can be divided
5786 in sections depending on the subject.
5789 *) Have the zlib compression code load ZLIB.DLL dynamically under
5793 *) New function BN_mod_sqrt for computing square roots modulo a prime
5794 (using the probabilistic Tonelli-Shanks algorithm unless
5795 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
5796 be handled deterministically).
5797 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5799 *) Make BN_mod_inverse faster by explicitly handling small quotients
5800 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
5801 512 bits], about 30% for larger ones [1024 or 2048 bits].)
5804 *) New function BN_kronecker.
5807 *) Fix BN_gcd so that it works on negative inputs; the result is
5808 positive unless both parameters are zero.
5809 Previously something reasonably close to an infinite loop was
5810 possible because numbers could be growing instead of shrinking
5811 in the implementation of Euclid's algorithm.
5814 *) Fix BN_is_word() and BN_is_one() macros to take into account the
5815 sign of the number in question.
5817 Fix BN_is_word(a,w) to work correctly for w == 0.
5819 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
5820 because its test if the absolute value of 'a' equals 'w'.
5821 Note that BN_abs_is_word does *not* handle w == 0 reliably;
5822 it exists mostly for use in the implementations of BN_is_zero(),
5823 BN_is_one(), and BN_is_word().
5826 *) New function BN_swap.
5829 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
5830 the exponentiation functions are more likely to produce reasonable
5831 results on negative inputs.
5834 *) Change BN_mod_mul so that the result is always non-negative.
5835 Previously, it could be negative if one of the factors was negative;
5836 I don't think anyone really wanted that behaviour.
5839 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
5840 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
5841 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
5842 and add new functions:
5851 BN_mod_lshift1_quick
5855 These functions always generate non-negative results.
5857 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
5858 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
5860 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
5861 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
5862 be reduced modulo m.
5863 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5866 The following entry accidentily appeared in the CHANGES file
5867 distributed with OpenSSL 0.9.7. The modifications described in
5868 it do *not* apply to OpenSSL 0.9.7.
5870 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5871 was actually never needed) and in BN_mul(). The removal in BN_mul()
5872 required a small change in bn_mul_part_recursive() and the addition
5873 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5874 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5875 bn_sub_words() and bn_add_words() except they take arrays with
5880 *) In 'openssl passwd', verify passwords read from the terminal
5881 unless the '-salt' option is used (which usually means that
5882 verification would just waste user's time since the resulting
5883 hash is going to be compared with some given password hash)
5884 or the new '-noverify' option is used.
5886 This is an incompatible change, but it does not affect
5887 non-interactive use of 'openssl passwd' (passwords on the command
5888 line, '-stdin' option, '-in ...' option) and thus should not
5892 *) Remove all references to RSAref, since there's no more need for it.
5895 *) Make DSO load along a path given through an environment variable
5896 (SHLIB_PATH) with shl_load().
5899 *) Constify the ENGINE code as a result of BIGNUM constification.
5900 Also constify the RSA code and most things related to it. In a
5901 few places, most notable in the depth of the ASN.1 code, ugly
5902 casts back to non-const were required (to be solved at a later
5906 *) Make it so the openssl application has all engines loaded by default.
5909 *) Constify the BIGNUM routines a little more.
5912 *) Add the following functions:
5914 ENGINE_load_cswift()
5916 ENGINE_load_atalla()
5918 ENGINE_load_builtin_engines()
5920 That way, an application can itself choose if external engines that
5921 are built-in in OpenSSL shall ever be used or not. The benefit is
5922 that applications won't have to be linked with libdl or other dso
5923 libraries unless it's really needed.
5925 Changed 'openssl engine' to load all engines on demand.
5926 Changed the engine header files to avoid the duplication of some
5927 declarations (they differed!).
5930 *) 'openssl engine' can now list capabilities.
5933 *) Better error reporting in 'openssl engine'.
5936 *) Never call load_dh_param(NULL) in s_server.
5939 *) Add engine application. It can currently list engines by name and
5940 identity, and test if they are actually available.
5943 *) Improve RPM specification file by forcing symbolic linking and making
5944 sure the installed documentation is also owned by root.root.
5945 [Damien Miller <djm@mindrot.org>]
5947 *) Give the OpenSSL applications more possibilities to make use of
5948 keys (public as well as private) handled by engines.
5951 *) Add OCSP code that comes from CertCo.
5954 *) Add VMS support for the Rijndael code.
5957 *) Added untested support for Nuron crypto accelerator.
5960 *) Add support for external cryptographic devices. This code was
5961 previously distributed separately as the "engine" branch.
5962 [Geoff Thorpe, Richard Levitte]
5964 *) Rework the filename-translation in the DSO code. It is now possible to
5965 have far greater control over how a "name" is turned into a filename
5966 depending on the operating environment and any oddities about the
5967 different shared library filenames on each system.
5970 *) Support threads on FreeBSD-elf in Configure.
5973 *) Fix for SHA1 assembly problem with MASM: it produces
5974 warnings about corrupt line number information when assembling
5975 with debugging information. This is caused by the overlapping
5977 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
5980 NCONF_get_number() has no error checking at all. As a replacement,
5981 NCONF_get_number_e() is defined (_e for "error checking") and is
5982 promoted strongly. The old NCONF_get_number is kept around for
5983 binary backward compatibility.
5984 Make it possible for methods to load from something other than a BIO,
5985 by providing a function pointer that is given a name instead of a BIO.
5986 For example, this could be used to load configuration data from an
5990 *) Fix for non blocking accept BIOs. Added new I/O special reason
5991 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
5992 with non blocking I/O was not possible because no retry code was
5993 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
5997 *) Added the beginnings of Rijndael support.
6000 *) Fix for bug in DirectoryString mask setting. Add support for
6001 X509_NAME_print_ex() in 'req' and X509_print_ex() function
6002 to allow certificate printing to more controllable, additional
6003 'certopt' option to 'x509' to allow new printing options to be
6007 *) Clean old EAY MD5 hack from e_os.h.
6010 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
6012 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6013 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6014 [Joe Orton, Steve Henson]
6016 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
6018 *) Fix additional bug revealed by the NISCC test suite:
6020 Stop bug triggering large recursion when presented with
6021 certain ASN.1 tags (CVE-2003-0851)
6024 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
6026 *) Fix various bugs revealed by running the NISCC test suite:
6028 Stop out of bounds reads in the ASN1 code when presented with
6029 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6031 If verify callback ignores invalid public key errors don't try to check
6032 certificate signature with the NULL public key.
6036 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6037 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6041 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6042 extra data after the compression methods not only for TLS 1.0
6043 but also for SSL 3.0 (as required by the specification).
6044 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6046 *) Change X509_certificate_type() to mark the key as exported/exportable
6047 when it's 512 *bits* long, not 512 bytes.
6050 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
6052 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6053 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6054 a protocol version number mismatch like a decryption error
6055 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6058 *) Turn on RSA blinding by default in the default implementation
6059 to avoid a timing attack. Applications that don't want it can call
6060 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6061 They would be ill-advised to do so in most cases.
6062 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6064 *) Change RSA blinding code so that it works when the PRNG is not
6065 seeded (in this case, the secret RSA exponent is abused as
6066 an unpredictable seed -- if it is not unpredictable, there
6067 is no point in blinding anyway). Make RSA blinding thread-safe
6068 by remembering the creator's thread ID in rsa->blinding and
6069 having all other threads use local one-time blinding factors
6070 (this requires more computation than sharing rsa->blinding, but
6071 avoids excessive locking; and if an RSA object is not shared
6072 between threads, blinding will still be very fast).
6075 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
6077 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6078 via timing by performing a MAC computation even if incorrrect
6079 block cipher padding has been found. This is a countermeasure
6080 against active attacks where the attacker has to distinguish
6081 between bad padding and a MAC verification error. (CVE-2003-0078)
6083 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6084 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6085 Martin Vuagnoux (EPFL, Ilion)]
6087 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
6089 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
6090 memory from it's contents. This is done with a counter that will
6091 place alternating values in each byte. This can be used to solve
6092 two issues: 1) the removal of calls to memset() by highly optimizing
6093 compilers, and 2) cleansing with other values than 0, since those can
6094 be read through on certain media, for example a swap space on disk.
6097 *) Bugfix: client side session caching did not work with external caching,
6098 because the session->cipher setting was not restored when reloading
6099 from the external cache. This problem was masked, when
6100 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
6101 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
6104 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
6105 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
6106 [Zeev Lieber <zeev-l@yahoo.com>]
6108 *) Undo an undocumented change introduced in 0.9.6e which caused
6109 repeated calls to OpenSSL_add_all_ciphers() and
6110 OpenSSL_add_all_digests() to be ignored, even after calling
6114 *) Change the default configuration reader to deal with last line not
6115 being properly terminated.
6118 *) Change X509_NAME_cmp() so it applies the special rules on handling
6119 DN values that are of type PrintableString, as well as RDNs of type
6120 emailAddress where the value has the type ia5String.
6121 [stefank@valicert.com via Richard Levitte]
6123 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
6124 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
6125 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
6126 the bitwise-OR of the two for use by the majority of applications
6127 wanting this behaviour, and update the docs. The documented
6128 behaviour and actual behaviour were inconsistent and had been
6129 changing anyway, so this is more a bug-fix than a behavioural
6131 [Geoff Thorpe, diagnosed by Nadav Har'El]
6133 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
6134 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
6137 *) Fix initialization code race conditions in
6138 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
6139 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
6140 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
6141 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
6142 ssl2_get_cipher_by_char(),
6143 ssl3_get_cipher_by_char().
6144 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
6146 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
6147 the cached sessions are flushed, as the remove_cb() might use ex_data
6148 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
6149 (see [openssl.org #212]).
6150 [Geoff Thorpe, Lutz Jaenicke]
6152 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
6153 length, instead of the encoding length to d2i_ASN1_OBJECT.
6156 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
6158 *) [In 0.9.6g-engine release:]
6159 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
6160 [Lynn Gazis <lgazis@rainbow.com>]
6162 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
6164 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
6165 and get fix the header length calculation.
6166 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
6167 Alon Kantor <alonk@checkpoint.com> (and others),
6170 *) Use proper error handling instead of 'assertions' in buffer
6171 overflow checks added in 0.9.6e. This prevents DoS (the
6172 assertions could call abort()).
6173 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
6175 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
6177 *) Add various sanity checks to asn1_get_length() to reject
6178 the ASN1 length bytes if they exceed sizeof(long), will appear
6179 negative or the content length exceeds the length of the
6181 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6183 *) Fix cipher selection routines: ciphers without encryption had no flags
6184 for the cipher strength set and where therefore not handled correctly
6185 by the selection routines (PR #130).
6188 *) Fix EVP_dsa_sha macro.
6192 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
6193 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
6194 that was added in OpenSSL 0.9.6d.
6196 As the countermeasure turned out to be incompatible with some
6197 broken SSL implementations, the new option is part of SSL_OP_ALL.
6198 SSL_OP_ALL is usually employed when compatibility with weird SSL
6199 implementations is desired (e.g. '-bugs' option to 's_client' and
6200 's_server'), so the new option is automatically set in many
6204 *) Changes in security patch:
6206 Changes marked "(CHATS)" were sponsored by the Defense Advanced
6207 Research Projects Agency (DARPA) and Air Force Research Laboratory,
6208 Air Force Materiel Command, USAF, under agreement number
6211 *) Add various sanity checks to asn1_get_length() to reject
6212 the ASN1 length bytes if they exceed sizeof(long), will appear
6213 negative or the content length exceeds the length of the
6214 supplied buffer. (CVE-2002-0659)
6215 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6217 *) Assertions for various potential buffer overflows, not known to
6219 [Ben Laurie (CHATS)]
6221 *) Various temporary buffers to hold ASCII versions of integers were
6222 too small for 64 bit platforms. (CVE-2002-0655)
6223 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
6225 *) Remote buffer overflow in SSL3 protocol - an attacker could
6226 supply an oversized session ID to a client. (CVE-2002-0656)
6227 [Ben Laurie (CHATS)]
6229 *) Remote buffer overflow in SSL2 protocol - an attacker could
6230 supply an oversized client master key. (CVE-2002-0656)
6231 [Ben Laurie (CHATS)]
6233 Changes between 0.9.6c and 0.9.6d [9 May 2002]
6235 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
6236 encoded as NULL) with id-dsa-with-sha1.
6237 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
6239 *) Check various X509_...() return values in apps/req.c.
6240 [Nils Larsch <nla@trustcenter.de>]
6242 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
6243 an end-of-file condition would erronously be flagged, when the CRLF
6244 was just at the end of a processed block. The bug was discovered when
6245 processing data through a buffering memory BIO handing the data to a
6246 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
6247 <ptsekov@syntrex.com> and Nedelcho Stanev.
6250 *) Implement a countermeasure against a vulnerability recently found
6251 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
6252 before application data chunks to avoid the use of known IVs
6253 with data potentially chosen by the attacker.
6256 *) Fix length checks in ssl3_get_client_hello().
6259 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
6260 to prevent ssl3_read_internal() from incorrectly assuming that
6261 ssl3_read_bytes() found application data while handshake
6262 processing was enabled when in fact s->s3->in_read_app_data was
6263 merely automatically cleared during the initial handshake.
6264 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
6266 *) Fix object definitions for Private and Enterprise: they were not
6267 recognized in their shortname (=lowercase) representation. Extend
6268 obj_dat.pl to issue an error when using undefined keywords instead
6269 of silently ignoring the problem (Svenning Sorensen
6270 <sss@sss.dnsalias.net>).
6273 *) Fix DH_generate_parameters() so that it works for 'non-standard'
6274 generators, i.e. generators other than 2 and 5. (Previously, the
6275 code did not properly initialise the 'add' and 'rem' values to
6276 BN_generate_prime().)
6278 In the new general case, we do not insist that 'generator' is
6279 actually a primitive root: This requirement is rather pointless;
6280 a generator of the order-q subgroup is just as good, if not
6284 *) Map new X509 verification errors to alerts. Discovered and submitted by
6285 Tom Wu <tom@arcot.com>.
6288 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
6289 returning non-zero before the data has been completely received
6290 when using non-blocking I/O.
6291 [Bodo Moeller; problem pointed out by John Hughes]
6293 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
6294 [Ben Laurie, Lutz Jaenicke]
6296 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
6297 Yoram Zahavi <YoramZ@gilian.com>).
6300 *) Add information about CygWin 1.3 and on, and preserve proper
6301 configuration for the versions before that.
6302 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6304 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
6305 check whether we deal with a copy of a session and do not delete from
6306 the cache in this case. Problem reported by "Izhar Shoshani Levi"
6307 <izhar@checkpoint.com>.
6310 *) Do not store session data into the internal session cache, if it
6311 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
6312 flag is set). Proposed by Aslam <aslam@funk.com>.
6315 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
6319 *) [In 0.9.6d-engine release:]
6320 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6321 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6323 *) Add the configuration target linux-s390x.
6324 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
6326 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
6327 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
6328 variable as an indication that a ClientHello message has been
6329 received. As the flag value will be lost between multiple
6330 invocations of ssl3_accept when using non-blocking I/O, the
6331 function may not be aware that a handshake has actually taken
6332 place, thus preventing a new session from being added to the
6335 To avoid this problem, we now set s->new_session to 2 instead of
6336 using a local variable.
6337 [Lutz Jaenicke, Bodo Moeller]
6339 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
6340 if the SSL_R_LENGTH_MISMATCH error is detected.
6341 [Geoff Thorpe, Bodo Moeller]
6343 *) New 'shared_ldflag' column in Configure platform table.
6346 *) Fix EVP_CIPHER_mode macro.
6347 ["Dan S. Camper" <dan@bti.net>]
6349 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
6350 type, we must throw them away by setting rr->length to 0.
6351 [D P Chang <dpc@qualys.com>]
6353 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
6355 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
6356 <Dominikus.Scherkl@biodata.com>. (The previous implementation
6357 worked incorrectly for those cases where range = 10..._2 and
6358 3*range is two bits longer than range.)
6361 *) Only add signing time to PKCS7 structures if it is not already
6365 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
6366 OBJ_ld_ce should be OBJ_id_ce.
6367 Also some ip-pda OIDs in crypto/objects/objects.txt were
6368 incorrect (cf. RFC 3039).
6369 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
6371 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
6372 returns early because it has nothing to do.
6373 [Andy Schneider <andy.schneider@bjss.co.uk>]
6375 *) [In 0.9.6c-engine release:]
6376 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
6377 [Andy Schneider <andy.schneider@bjss.co.uk>]
6379 *) [In 0.9.6c-engine release:]
6380 Add support for Cryptographic Appliance's keyserver technology.
6381 (Use engine 'keyclient')
6382 [Cryptographic Appliances and Geoff Thorpe]
6384 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
6385 is called via tools/c89.sh because arguments have to be
6386 rearranged (all '-L' options must appear before the first object
6388 [Richard Shapiro <rshapiro@abinitio.com>]
6390 *) [In 0.9.6c-engine release:]
6391 Add support for Broadcom crypto accelerator cards, backported
6393 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
6395 *) [In 0.9.6c-engine release:]
6396 Add support for SureWare crypto accelerator cards from
6397 Baltimore Technologies. (Use engine 'sureware')
6398 [Baltimore Technologies and Mark Cox]
6400 *) [In 0.9.6c-engine release:]
6401 Add support for crypto accelerator cards from Accelerated
6402 Encryption Processing, www.aep.ie. (Use engine 'aep')
6403 [AEP Inc. and Mark Cox]
6405 *) Add a configuration entry for gcc on UnixWare.
6406 [Gary Benson <gbenson@redhat.com>]
6408 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
6409 messages are stored in a single piece (fixed-length part and
6410 variable-length part combined) and fix various bugs found on the way.
6413 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
6414 instead. BIO_gethostbyname() does not know what timeouts are
6415 appropriate, so entries would stay in cache even when they have
6417 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
6419 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
6420 faced with a pathologically small ClientHello fragment that does
6421 not contain client_version: Instead of aborting with an error,
6422 simply choose the highest available protocol version (i.e.,
6423 TLS 1.0 unless it is disabled). In practice, ClientHello
6424 messages are never sent like this, but this change gives us
6425 strictly correct behaviour at least for TLS.
6428 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
6429 never resets s->method to s->ctx->method when called from within
6430 one of the SSL handshake functions.
6431 [Bodo Moeller; problem pointed out by Niko Baric]
6433 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
6434 (sent using the client's version number) if client_version is
6435 smaller than the protocol version in use. Also change
6436 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
6437 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
6438 the client will at least see that alert.
6441 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
6445 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
6446 client receives HelloRequest while in a handshake.
6447 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
6449 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
6450 should end in 'break', not 'goto end' which circuments various
6451 cleanups done in state SSL_ST_OK. But session related stuff
6452 must be disabled for SSL_ST_OK in the case that we just sent a
6455 Also avoid some overhead by not calling ssl_init_wbio_buffer()
6456 before just sending a HelloRequest.
6457 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
6459 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
6460 reveal whether illegal block cipher padding was found or a MAC
6461 verification error occured. (Neither SSLerr() codes nor alerts
6462 are directly visible to potential attackers, but the information
6463 may leak via logfiles.)
6465 Similar changes are not required for the SSL 2.0 implementation
6466 because the number of padding bytes is sent in clear for SSL 2.0,
6467 and the extra bytes are just ignored. However ssl/s2_pkt.c
6468 failed to verify that the purported number of padding bytes is in
6472 *) Add OpenUNIX-8 support including shared libraries
6473 (Boyd Lynn Gerber <gerberb@zenez.com>).
6476 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
6477 'wristwatch attack' using huge encoding parameters (cf.
6478 James H. Manger's CRYPTO 2001 paper). Note that the
6479 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
6480 encoding parameters and hence was not vulnerable.
6483 *) BN_sqr() bug fix.
6484 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
6486 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
6487 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
6488 followed by modular reduction.
6489 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
6491 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
6492 equivalent based on BN_pseudo_rand() instead of BN_rand().
6495 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
6496 This function was broken, as the check for a new client hello message
6497 to handle SGC did not allow these large messages.
6498 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
6501 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
6504 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
6505 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
6508 *) Rework the configuration and shared library support for Tru64 Unix.
6509 The configuration part makes use of modern compiler features and
6510 still retains old compiler behavior for those that run older versions
6511 of the OS. The shared library support part includes a variant that
6512 uses the RPATH feature, and is available through the special
6513 configuration target "alpha-cc-rpath", which will never be selected
6515 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
6517 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
6518 with the same message size as in ssl3_get_certificate_request().
6519 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
6520 messages might inadvertently be reject as too long.
6521 [Petr Lampa <lampa@fee.vutbr.cz>]
6523 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
6526 *) Modified SSL library such that the verify_callback that has been set
6527 specificly for an SSL object with SSL_set_verify() is actually being
6528 used. Before the change, a verify_callback set with this function was
6529 ignored and the verify_callback() set in the SSL_CTX at the time of
6530 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
6531 to allow the necessary settings.
6534 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
6535 explicitly to NULL, as at least on Solaris 8 this seems not always to be
6536 done automatically (in contradiction to the requirements of the C
6537 standard). This made problems when used from OpenSSH.
6540 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
6541 dh->length and always used
6543 BN_rand_range(priv_key, dh->p).
6545 BN_rand_range() is not necessary for Diffie-Hellman, and this
6546 specific range makes Diffie-Hellman unnecessarily inefficient if
6547 dh->length (recommended exponent length) is much smaller than the
6548 length of dh->p. We could use BN_rand_range() if the order of
6549 the subgroup was stored in the DH structure, but we only have
6554 BN_rand(priv_key, l, ...)
6556 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
6562 RSA_eay_public_encrypt
6563 RSA_eay_private_decrypt
6564 RSA_eay_private_encrypt (signing)
6565 RSA_eay_public_decrypt (signature verification)
6567 (default implementations for RSA_public_encrypt,
6568 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
6569 always reject numbers >= n.
6572 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
6573 to synchronize access to 'locking_thread'. This is necessary on
6574 systems where access to 'locking_thread' (an 'unsigned long'
6575 variable) is not atomic.
6578 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
6579 *before* setting the 'crypto_lock_rand' flag. The previous code had
6580 a race condition if 0 is a valid thread ID.
6581 [Travis Vitek <vitek@roguewave.com>]
6583 *) Add support for shared libraries under Irix.
6584 [Albert Chin-A-Young <china@thewrittenword.com>]
6586 *) Add configuration option to build on Linux on both big-endian and
6588 [Ralf Baechle <ralf@uni-koblenz.de>]
6590 *) Add the possibility to create shared libraries on HP-UX.
6593 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
6595 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
6596 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
6597 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
6598 PRNG state recovery was possible based on the output of
6599 one PRNG request appropriately sized to gain knowledge on
6600 'md' followed by enough consecutive 1-byte PRNG requests
6601 to traverse all of 'state'.
6603 1. When updating 'md_local' (the current thread's copy of 'md')
6604 during PRNG output generation, hash all of the previous
6605 'md_local' value, not just the half used for PRNG output.
6607 2. Make the number of bytes from 'state' included into the hash
6608 independent from the number of PRNG bytes requested.
6610 The first measure alone would be sufficient to avoid
6611 Markku-Juhani's attack. (Actually it had never occurred
6612 to me that the half of 'md_local' used for chaining was the
6613 half from which PRNG output bytes were taken -- I had always
6614 assumed that the secret half would be used.) The second
6615 measure makes sure that additional data from 'state' is never
6616 mixed into 'md_local' in small portions; this heuristically
6617 further strengthens the PRNG.
6620 *) Fix crypto/bn/asm/mips3.s.
6623 *) When only the key is given to "enc", the IV is undefined. Print out
6624 an error message in this case.
6627 *) Handle special case when X509_NAME is empty in X509 printing routines.
6630 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
6631 positive and less than q.
6634 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
6635 used: it isn't thread safe and the add_lock_callback should handle
6637 [Paul Rose <Paul.Rose@bridge.com>]
6639 *) Verify that incoming data obeys the block size in
6640 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
6644 [Ulf Möller, Bodo Möller]
6646 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
6647 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
6648 when fixing the server behaviour for backwards-compatible 'client
6649 hello' messages. (Note that the attack is impractical against
6650 SSL 3.0 and TLS 1.0 anyway because length and version checking
6651 means that the probability of guessing a valid ciphertext is
6652 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
6655 Before 0.9.5, the countermeasure (hide the error by generating a
6656 random 'decryption result') did not work properly because
6657 ERR_clear_error() was missing, meaning that SSL_get_error() would
6658 detect the supposedly ignored error.
6660 Both problems are now fixed.
6663 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
6664 (previously it was 1024).
6667 *) Fix for compatibility mode trust settings: ignore trust settings
6668 unless some valid trust or reject settings are present.
6671 *) Fix for blowfish EVP: its a variable length cipher.
6674 *) Fix various bugs related to DSA S/MIME verification. Handle missing
6675 parameters in DSA public key structures and return an error in the
6676 DSA routines if parameters are absent.
6679 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
6680 in the current directory if neither $RANDFILE nor $HOME was set.
6681 RAND_file_name() in 0.9.6a returned NULL in this case. This has
6682 caused some confusion to Windows users who haven't defined $HOME.
6683 Thus RAND_file_name() is changed again: e_os.h can define a
6684 DEFAULT_HOME, which will be used if $HOME is not set.
6685 For Windows, we use "C:"; on other platforms, we still require
6686 environment variables.
6688 *) Move 'if (!initialized) RAND_poll()' into regions protected by
6689 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
6690 having multiple threads call RAND_poll() concurrently.
6693 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
6694 combination of a flag and a thread ID variable.
6695 Otherwise while one thread is in ssleay_rand_bytes (which sets the
6696 flag), *other* threads can enter ssleay_add_bytes without obeying
6697 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
6698 that they do not hold after the first thread unsets add_do_not_lock).
6701 *) Change bctest again: '-x' expressions are not available in all
6705 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
6707 *) Fix a couple of memory leaks in PKCS7_dataDecode()
6708 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
6710 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
6711 the default extension for executables, if any. Also, make the perl
6712 scripts that use symlink() to test if it really exists and use "cp"
6713 if it doesn't. All this made OpenSSL compilable and installable in
6717 *) Fix for asn1_GetSequence() for indefinite length constructed data.
6718 If SEQUENCE is length is indefinite just set c->slen to the total
6719 amount of data available.
6720 [Steve Henson, reported by shige@FreeBSD.org]
6721 [This change does not apply to 0.9.7.]
6723 *) Change bctest to avoid here-documents inside command substitution
6724 (workaround for FreeBSD /bin/sh bug).
6725 For compatibility with Ultrix, avoid shell functions (introduced
6726 in the bctest version that searches along $PATH).
6729 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6730 with des_encrypt() defined on some operating systems, like Solaris
6734 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6735 On the Importance of Eliminating Errors in Cryptographic
6736 Computations, J. Cryptology 14 (2001) 2, 101-119,
6737 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6740 *) MIPS assembler BIGNUM division bug fix.
6743 *) Disabled incorrect Alpha assembler code.
6746 *) Fix PKCS#7 decode routines so they correctly update the length
6747 after reading an EOC for the EXPLICIT tag.
6749 [This change does not apply to 0.9.7.]
6751 *) Fix bug in PKCS#12 key generation routines. This was triggered
6752 if a 3DES key was generated with a 0 initial byte. Include
6753 PKCS12_BROKEN_KEYGEN compilation option to retain the old
6754 (but broken) behaviour.
6757 *) Enhance bctest to search for a working bc along $PATH and print
6759 [Tim Rice <tim@multitalents.net> via Richard Levitte]
6761 *) Fix memory leaks in err.c: free err_data string if necessary;
6762 don't write to the wrong index in ERR_set_error_data.
6765 *) Implement ssl23_peek (analogous to ssl23_read), which previously
6769 *) Replace rdtsc with _emit statements for VC++ version 5.
6770 [Jeremy Cooper <jeremy@baymoo.org>]
6772 *) Make it possible to reuse SSLv2 sessions.
6775 *) In copy_email() check for >= 0 as a return value for
6776 X509_NAME_get_index_by_NID() since 0 is a valid index.
6777 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
6779 *) Avoid coredump with unsupported or invalid public keys by checking if
6780 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
6781 PKCS7_verify() fails with non detached data.
6784 *) Don't use getenv in library functions when run as setuid/setgid.
6785 New function OPENSSL_issetugid().
6788 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
6789 due to incorrect handling of multi-threading:
6791 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
6793 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
6795 3. Count how many times MemCheck_off() has been called so that
6796 nested use can be treated correctly. This also avoids
6797 inband-signalling in the previous code (which relied on the
6798 assumption that thread ID 0 is impossible).
6801 *) Add "-rand" option also to s_client and s_server.
6804 *) Fix CPU detection on Irix 6.x.
6805 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
6806 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6808 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
6811 [This change does not apply to 0.9.7.]
6813 *) Use the cached encoding of an X509_NAME structure rather than
6814 copying it. This is apparently the reason for the libsafe "errors"
6815 but the code is actually correct.
6818 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
6819 Bleichenbacher's DSA attack.
6820 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
6821 to be set and top=0 forces the highest bit to be set; top=-1 is new
6822 and leaves the highest bit random.
6823 [Ulf Moeller, Bodo Moeller]
6825 *) In the NCONF_...-based implementations for CONF_... queries
6826 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
6827 a temporary CONF structure with the data component set to NULL
6828 (which gives segmentation faults in lh_retrieve).
6829 Instead, use NULL for the CONF pointer in CONF_get_string and
6830 CONF_get_number (which may use environment variables) and directly
6831 return NULL from CONF_get_section.
6834 *) Fix potential buffer overrun for EBCDIC.
6837 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
6838 keyUsage if basicConstraints absent for a CA.
6841 *) Make SMIME_write_PKCS7() write mail header values with a format that
6842 is more generally accepted (no spaces before the semicolon), since
6843 some programs can't parse those values properly otherwise. Also make
6844 sure BIO's that break lines after each write do not create invalid
6848 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
6849 macros previously used would not encode an empty SEQUENCE OF
6850 and break the signature.
6852 [This change does not apply to 0.9.7.]
6854 *) Zero the premaster secret after deriving the master secret in
6858 *) Add some EVP_add_digest_alias registrations (as found in
6859 OpenSSL_add_all_digests()) to SSL_library_init()
6860 aka OpenSSL_add_ssl_algorithms(). This provides improved
6861 compatibility with peers using X.509 certificates
6862 with unconventional AlgorithmIdentifier OIDs.
6865 *) Fix for Irix with NO_ASM.
6866 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6868 *) ./config script fixes.
6869 [Ulf Moeller, Richard Levitte]
6871 *) Fix 'openssl passwd -1'.
6874 *) Change PKCS12_key_gen_asc() so it can cope with non null
6875 terminated strings whose length is passed in the passlen
6876 parameter, for example from PEM callbacks. This was done
6877 by adding an extra length parameter to asc2uni().
6878 [Steve Henson, reported by <oddissey@samsung.co.kr>]
6880 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
6881 call failed, free the DSA structure.
6884 *) Fix to uni2asc() to cope with zero length Unicode strings.
6885 These are present in some PKCS#12 files.
6888 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
6889 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
6890 when writing a 32767 byte record.
6891 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
6893 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
6894 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
6896 (RSA objects have a reference count access to which is protected
6897 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
6898 so they are meant to be shared between threads.)
6899 [Bodo Moeller, Geoff Thorpe; original patch submitted by
6900 "Reddie, Steven" <Steven.Reddie@ca.com>]
6902 *) Fix a deadlock in CRYPTO_mem_leaks().
6905 *) Use better test patterns in bntest.
6908 *) rand_win.c fix for Borland C.
6911 *) BN_rshift bugfix for n == 0.
6914 *) Add a 'bctest' script that checks for some known 'bc' bugs
6915 so that 'make test' does not abort just because 'bc' is broken.
6918 *) Store verify_result within SSL_SESSION also for client side to
6919 avoid potential security hole. (Re-used sessions on the client side
6920 always resulted in verify_result==X509_V_OK, not using the original
6921 result of the server certificate verification.)
6924 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
6925 SSL3_RT_APPLICATION_DATA, return 0.
6926 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
6930 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
6931 releases, have been re-implemented by renaming the previous
6932 implementations of ssl2_read and ssl3_read to ssl2_read_internal
6933 and ssl3_read_internal, respectively, and adding 'peek' parameters
6934 to them. The new ssl[23]_{read,peek} functions are calls to
6935 ssl[23]_read_internal with the 'peek' flag set appropriately.
6936 A 'peek' parameter has also been added to ssl3_read_bytes, which
6937 does the actual work for ssl3_read_internal.
6940 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
6941 the method-specific "init()" handler. Also clean up ex_data after
6942 calling the method-specific "finish()" handler. Previously, this was
6943 happening the other way round.
6946 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
6947 The previous value, 12, was not always sufficient for BN_mod_exp().
6950 *) Make sure that shared libraries get the internal name engine with
6951 the full version number and not just 0. This should mark the
6952 shared libraries as not backward compatible. Of course, this should
6953 be changed again when we can guarantee backward binary compatibility.
6956 *) Fix typo in get_cert_by_subject() in by_dir.c
6957 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
6959 *) Rework the system to generate shared libraries:
6961 - Make note of the expected extension for the shared libraries and
6962 if there is a need for symbolic links from for example libcrypto.so.0
6963 to libcrypto.so.0.9.7. There is extended info in Configure for
6966 - Make as few rebuilds of the shared libraries as possible.
6968 - Still avoid linking the OpenSSL programs with the shared libraries.
6970 - When installing, install the shared libraries separately from the
6974 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
6976 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
6977 and not in SSL_clear because the latter is also used by the
6978 accept/connect functions; previously, the settings made by
6979 SSL_set_read_ahead would be lost during the handshake.
6980 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
6982 *) Correct util/mkdef.pl to be selective about disabled algorithms.
6983 Previously, it would create entries for disableed algorithms no
6987 *) Added several new manual pages for SSL_* function.
6990 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
6992 *) In ssl23_get_client_hello, generate an error message when faced
6993 with an initial SSL 3.0/TLS record that is too small to contain the
6994 first two bytes of the ClientHello message, i.e. client_version.
6995 (Note that this is a pathologic case that probably has never happened
6996 in real life.) The previous approach was to use the version number
6997 from the record header as a substitute; but our protocol choice
6998 should not depend on that one because it is not authenticated
6999 by the Finished messages.
7002 *) More robust randomness gathering functions for Windows.
7003 [Jeffrey Altman <jaltman@columbia.edu>]
7005 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
7006 not set then we don't setup the error code for issuer check errors
7007 to avoid possibly overwriting other errors which the callback does
7008 handle. If an application does set the flag then we assume it knows
7009 what it is doing and can handle the new informational codes
7013 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
7014 a general "ANY" type, as such it should be able to decode anything
7015 including tagged types. However it didn't check the class so it would
7016 wrongly interpret tagged types in the same way as their universal
7017 counterpart and unknown types were just rejected. Changed so that the
7018 tagged and unknown types are handled in the same way as a SEQUENCE:
7019 that is the encoding is stored intact. There is also a new type
7020 "V_ASN1_OTHER" which is used when the class is not universal, in this
7021 case we have no idea what the actual type is so we just lump them all
7025 *) On VMS, stdout may very well lead to a file that is written to
7026 in a record-oriented fashion. That means that every write() will
7027 write a separate record, which will be read separately by the
7028 programs trying to read from it. This can be very confusing.
7030 The solution is to put a BIO filter in the way that will buffer
7031 text until a linefeed is reached, and then write everything a
7032 line at a time, so every record written will be an actual line,
7033 not chunks of lines and not (usually doesn't happen, but I've
7034 seen it once) several lines in one record. BIO_f_linebuffer() is
7037 Currently, it's a VMS-only method, because that's where it has
7038 been tested well enough.
7041 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
7042 it can return incorrect results.
7043 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
7044 but it was in 0.9.6-beta[12].)
7047 *) Disable the check for content being present when verifying detached
7048 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
7049 include zero length content when signing messages.
7052 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
7053 BIO_ctrl (for BIO pairs).
7056 *) Add DSO method for VMS.
7059 *) Bug fix: Montgomery multiplication could produce results with the
7063 *) Add RPM specification openssl.spec and modify it to build three
7064 packages. The default package contains applications, application
7065 documentation and run-time libraries. The devel package contains
7066 include files, static libraries and function documentation. The
7067 doc package contains the contents of the doc directory. The original
7068 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
7071 *) Add a large number of documentation files for many SSL routines.
7072 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7074 *) Add a configuration entry for Sony News 4.
7075 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
7077 *) Don't set the two most significant bits to one when generating a
7078 random number < q in the DSA library.
7081 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
7082 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
7083 the underlying transport is blocking) if a handshake took place.
7084 (The default behaviour is needed by applications such as s_client
7085 and s_server that use select() to determine when to use SSL_read;
7086 but for applications that know in advance when to expect data, it
7087 just makes things more complicated.)
7090 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
7094 *) Add a few more EBCDIC conditionals that make `req' and `x509'
7095 work better on such systems.
7096 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7098 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
7099 Update PKCS12_parse() so it copies the friendlyName and the
7100 keyid to the certificates aux info.
7103 *) Fix bug in PKCS7_verify() which caused an infinite loop
7104 if there was more than one signature.
7105 [Sven Uszpelkat <su@celocom.de>]
7107 *) Major change in util/mkdef.pl to include extra information
7108 about each symbol, as well as presentig variables as well
7109 as functions. This change means that there's n more need
7110 to rebuild the .num files when some algorithms are excluded.
7113 *) Allow the verify time to be set by an application,
7114 rather than always using the current time.
7117 *) Phase 2 verify code reorganisation. The certificate
7118 verify code now looks up an issuer certificate by a
7119 number of criteria: subject name, authority key id
7120 and key usage. It also verifies self signed certificates
7121 by the same criteria. The main comparison function is
7122 X509_check_issued() which performs these checks.
7124 Lot of changes were necessary in order to support this
7125 without completely rewriting the lookup code.
7127 Authority and subject key identifier are now cached.
7129 The LHASH 'certs' is X509_STORE has now been replaced
7130 by a STACK_OF(X509_OBJECT). This is mainly because an
7131 LHASH can't store or retrieve multiple objects with
7132 the same hash value.
7134 As a result various functions (which were all internal
7135 use only) have changed to handle the new X509_STORE
7136 structure. This will break anything that messed round
7137 with X509_STORE internally.
7139 The functions X509_STORE_add_cert() now checks for an
7140 exact match, rather than just subject name.
7142 The X509_STORE API doesn't directly support the retrieval
7143 of multiple certificates matching a given criteria, however
7144 this can be worked round by performing a lookup first
7145 (which will fill the cache with candidate certificates)
7146 and then examining the cache for matches. This is probably
7147 the best we can do without throwing out X509_LOOKUP
7148 entirely (maybe later...).
7150 The X509_VERIFY_CTX structure has been enhanced considerably.
7152 All certificate lookup operations now go via a get_issuer()
7153 callback. Although this currently uses an X509_STORE it
7154 can be replaced by custom lookups. This is a simple way
7155 to bypass the X509_STORE hackery necessary to make this
7156 work and makes it possible to use more efficient techniques
7157 in future. A very simple version which uses a simple
7158 STACK for its trusted certificate store is also provided
7159 using X509_STORE_CTX_trusted_stack().
7161 The verify_cb() and verify() callbacks now have equivalents
7162 in the X509_STORE_CTX structure.
7164 X509_STORE_CTX also has a 'flags' field which can be used
7165 to customise the verify behaviour.
7168 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
7169 excludes S/MIME capabilities.
7172 *) When a certificate request is read in keep a copy of the
7173 original encoding of the signed data and use it when outputing
7174 again. Signatures then use the original encoding rather than
7175 a decoded, encoded version which may cause problems if the
7176 request is improperly encoded.
7179 *) For consistency with other BIO_puts implementations, call
7180 buffer_write(b, ...) directly in buffer_puts instead of calling
7183 In BIO_puts, increment b->num_write as in BIO_write.
7184 [Peter.Sylvester@EdelWeb.fr]
7186 *) Fix BN_mul_word for the case where the word is 0. (We have to use
7187 BN_zero, we may not return a BIGNUM with an array consisting of
7191 *) Avoid calling abort() from within the library when problems are
7192 detected, except if preprocessor symbols have been defined
7193 (such as REF_CHECK, BN_DEBUG etc.).
7196 *) New openssl application 'rsautl'. This utility can be
7197 used for low level RSA operations. DER public key
7198 BIO/fp routines also added.
7201 *) New Configure entry and patches for compiling on QNX 4.
7202 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
7204 *) A demo state-machine implementation was sponsored by
7205 Nuron (http://www.nuron.com/) and is now available in
7206 demos/state_machine.
7209 *) New options added to the 'dgst' utility for signature
7210 generation and verification.
7213 *) Unrecognized PKCS#7 content types are now handled via a
7214 catch all ASN1_TYPE structure. This allows unsupported
7215 types to be stored as a "blob" and an application can
7216 encode and decode it manually.
7219 *) Fix various signed/unsigned issues to make a_strex.c
7221 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
7223 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
7224 length if passed a buffer. ASN1_INTEGER_to_BN failed
7225 if passed a NULL BN and its argument was negative.
7226 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
7228 *) Modification to PKCS#7 encoding routines to output definite
7229 length encoding. Since currently the whole structures are in
7230 memory there's not real point in using indefinite length
7231 constructed encoding. However if OpenSSL is compiled with
7232 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
7235 *) Added BIO_vprintf() and BIO_vsnprintf().
7238 *) Added more prefixes to parse for in the the strings written
7239 through a logging bio, to cover all the levels that are available
7240 through syslog. The prefixes are now:
7242 PANIC, EMERG, EMR => LOG_EMERG
7243 ALERT, ALR => LOG_ALERT
7244 CRIT, CRI => LOG_CRIT
7245 ERROR, ERR => LOG_ERR
7246 WARNING, WARN, WAR => LOG_WARNING
7247 NOTICE, NOTE, NOT => LOG_NOTICE
7248 INFO, INF => LOG_INFO
7249 DEBUG, DBG => LOG_DEBUG
7251 and as before, if none of those prefixes are present at the
7252 beginning of the string, LOG_ERR is chosen.
7254 On Win32, the LOG_* levels are mapped according to this:
7256 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
7257 LOG_WARNING => EVENTLOG_WARNING_TYPE
7258 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
7262 *) Made it possible to reconfigure with just the configuration
7263 argument "reconf" or "reconfigure". The command line arguments
7264 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
7265 and are retrieved from there when reconfiguring.
7269 [Assar Westerlund <assar@sics.se>, Richard Levitte]
7271 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
7274 *) The obj_dat.pl script was messing up the sorting of object
7275 names. The reason was that it compared the quoted version
7276 of strings as a result "OCSP" > "OCSP Signing" because
7277 " > SPACE. Changed script to store unquoted versions of
7278 names and add quotes on output. It was also omitting some
7279 names from the lookup table if they were given a default
7280 value (that is if SN is missing it is given the same
7281 value as LN and vice versa), these are now added on the
7282 grounds that if an object has a name we should be able to
7283 look it up. Finally added warning output when duplicate
7284 short or long names are found.
7287 *) Changes needed for Tandem NSK.
7288 [Scott Uroff <scott@xypro.com>]
7290 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
7291 RSA_padding_check_SSLv23(), special padding was never detected
7292 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
7293 version rollback attacks was not effective.
7295 In s23_clnt.c, don't use special rollback-attack detection padding
7296 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
7297 client; similarly, in s23_srvr.c, don't do the rollback check if
7298 SSL 2.0 is the only protocol enabled in the server.
7301 *) Make it possible to get hexdumps of unprintable data with 'openssl
7302 asn1parse'. By implication, the functions ASN1_parse_dump() and
7303 BIO_dump_indent() are added.
7306 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
7307 these print out strings and name structures based on various
7308 flags including RFC2253 support and proper handling of
7309 multibyte characters. Added options to the 'x509' utility
7310 to allow the various flags to be set.
7313 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
7314 Also change the functions X509_cmp_current_time() and
7315 X509_gmtime_adj() work with an ASN1_TIME structure,
7316 this will enable certificates using GeneralizedTime in validity
7317 dates to be checked.
7320 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
7321 negative public key encodings) on by default,
7322 NO_NEG_PUBKEY_BUG can be set to disable it.
7325 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
7326 content octets. An i2c_ASN1_OBJECT is unnecessary because
7327 the encoding can be trivially obtained from the structure.
7330 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
7331 not read locks (CRYPTO_r_[un]lock).
7334 *) A first attempt at creating official support for shared
7335 libraries through configuration. I've kept it so the
7336 default is static libraries only, and the OpenSSL programs
7337 are always statically linked for now, but there are
7338 preparations for dynamic linking in place.
7339 This has been tested on Linux and Tru64.
7342 *) Randomness polling function for Win9x, as described in:
7343 Peter Gutmann, Software Generation of Practically Strong
7347 *) Fix so PRNG is seeded in req if using an already existing
7351 *) New options to smime application. -inform and -outform
7352 allow alternative formats for the S/MIME message including
7353 PEM and DER. The -content option allows the content to be
7354 specified separately. This should allow things like Netscape
7355 form signing output easier to verify.
7358 *) Fix the ASN1 encoding of tags using the 'long form'.
7361 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
7362 STRING types. These convert content octets to and from the
7363 underlying type. The actual tag and length octets are
7364 already assumed to have been read in and checked. These
7365 are needed because all other string types have virtually
7366 identical handling apart from the tag. By having versions
7367 of the ASN1 functions that just operate on content octets
7368 IMPLICIT tagging can be handled properly. It also allows
7369 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
7370 and ASN1_INTEGER are identical apart from the tag.
7373 *) Change the handling of OID objects as follows:
7375 - New object identifiers are inserted in objects.txt, following
7376 the syntax given in objects.README.
7377 - objects.pl is used to process obj_mac.num and create a new
7379 - obj_dat.pl is used to create a new obj_dat.h, using the data in
7382 This is currently kind of a hack, and the perl code in objects.pl
7383 isn't very elegant, but it works as I intended. The simplest way
7384 to check that it worked correctly is to look in obj_dat.h and
7385 check the array nid_objs and make sure the objects haven't moved
7386 around (this is important!). Additions are OK, as well as
7387 consistent name changes.
7390 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
7393 *) Addition of the command line parameter '-rand file' to 'openssl req'.
7394 The given file adds to whatever has already been seeded into the
7395 random pool through the RANDFILE configuration file option or
7396 environment variable, or the default random state file.
7399 *) mkstack.pl now sorts each macro group into lexical order.
7400 Previously the output order depended on the order the files
7401 appeared in the directory, resulting in needless rewriting
7405 *) Patches to make OpenSSL compile under Win32 again. Mostly
7406 work arounds for the VC++ problem that it treats func() as
7407 func(void). Also stripped out the parts of mkdef.pl that
7408 added extra typesafe functions: these no longer exist.
7411 *) Reorganisation of the stack code. The macros are now all
7412 collected in safestack.h . Each macro is defined in terms of
7413 a "stack macro" of the form SKM_<name>(type, a, b). The
7414 DEBUG_SAFESTACK is now handled in terms of function casts,
7415 this has the advantage of retaining type safety without the
7416 use of additional functions. If DEBUG_SAFESTACK is not defined
7417 then the non typesafe macros are used instead. Also modified the
7418 mkstack.pl script to handle the new form. Needs testing to see
7419 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
7420 the default if no major problems. Similar behaviour for ASN1_SET_OF
7421 and PKCS12_STACK_OF.
7424 *) When some versions of IIS use the 'NET' form of private key the
7425 key derivation algorithm is different. Normally MD5(password) is
7426 used as a 128 bit RC4 key. In the modified case
7427 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
7428 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
7429 as the old Netscape_RSA functions except they have an additional
7430 'sgckey' parameter which uses the modified algorithm. Also added
7431 an -sgckey command line option to the rsa utility. Thanks to
7432 Adrian Peck <bertie@ncipher.com> for posting details of the modified
7433 algorithm to openssl-dev.
7436 *) The evp_local.h macros were using 'c.##kname' which resulted in
7437 invalid expansion on some systems (SCO 5.0.5 for example).
7438 Corrected to 'c.kname'.
7439 [Phillip Porch <root@theporch.com>]
7441 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
7442 a STACK of email addresses from a certificate or request, these look
7443 in the subject name and the subject alternative name extensions and
7444 omit any duplicate addresses.
7447 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
7448 This makes DSA verification about 2 % faster.
7451 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
7452 (meaning that now 2^5 values will be precomputed, which is only 4 KB
7453 plus overhead for 1024 bit moduli).
7454 This makes exponentiations about 0.5 % faster for 1024 bit
7455 exponents (as measured by "openssl speed rsa2048").
7458 *) Rename memory handling macros to avoid conflicts with other
7460 Malloc => OPENSSL_malloc
7461 Malloc_locked => OPENSSL_malloc_locked
7462 Realloc => OPENSSL_realloc
7463 Free => OPENSSL_free
7466 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
7467 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
7470 *) CygWin32 support.
7471 [John Jarvie <jjarvie@newsguy.com>]
7473 *) The type-safe stack code has been rejigged. It is now only compiled
7474 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
7475 by default all type-specific stack functions are "#define"d back to
7476 standard stack functions. This results in more streamlined output
7477 but retains the type-safety checking possibilities of the original
7481 *) The STACK code has been cleaned up, and certain type declarations
7482 that didn't make a lot of sense have been brought in line. This has
7483 also involved a cleanup of sorts in safestack.h to more correctly
7484 map type-safe stack functions onto their plain stack counterparts.
7485 This work has also resulted in a variety of "const"ifications of
7486 lots of the code, especially "_cmp" operations which should normally
7487 be prototyped with "const" parameters anyway.
7490 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
7491 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
7492 (The PRNG state consists of two parts, the large pool 'state' and 'md',
7493 where all of 'md' is used each time the PRNG is used, but 'state'
7494 is used only indexed by a cyclic counter. As entropy may not be
7495 well distributed from the beginning, 'md' is important as a
7496 chaining variable. However, the output function chains only half
7497 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
7498 all of 'md', and seeding with STATE_SIZE dummy bytes will result
7499 in all of 'state' being rewritten, with the new values depending
7500 on virtually all of 'md'. This overcomes the 80 bit limitation.)
7503 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
7504 the handshake is continued after ssl_verify_cert_chain();
7505 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
7506 can lead to 'unexplainable' connection aborts later.
7507 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
7509 *) Major EVP API cipher revision.
7510 Add hooks for extra EVP features. This allows various cipher
7511 parameters to be set in the EVP interface. Support added for variable
7512 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
7513 setting of RC2 and RC5 parameters.
7515 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
7518 Remove lots of duplicated code from the EVP library. For example *every*
7519 cipher init() function handles the 'iv' in the same way according to the
7520 cipher mode. They also all do nothing if the 'key' parameter is NULL and
7521 for CFB and OFB modes they zero ctx->num.
7523 New functionality allows removal of S/MIME code RC2 hack.
7525 Most of the routines have the same form and so can be declared in terms
7528 By shifting this to the top level EVP_CipherInit() it can be removed from
7529 all individual ciphers. If the cipher wants to handle IVs or keys
7530 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
7533 Change lots of functions like EVP_EncryptUpdate() to now return a
7534 value: although software versions of the algorithms cannot fail
7535 any installed hardware versions can.
7538 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
7539 this option is set, tolerate broken clients that send the negotiated
7540 protocol version number instead of the requested protocol version
7544 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
7545 i.e. non-zero for export ciphersuites, zero otherwise.
7546 Previous versions had this flag inverted, inconsistent with
7547 rsa_tmp_cb (..._TMP_RSA_CB).
7548 [Bodo Moeller; problem reported by Amit Chopra]
7550 *) Add missing DSA library text string. Work around for some IIS
7551 key files with invalid SEQUENCE encoding.
7554 *) Add a document (doc/standards.txt) that list all kinds of standards
7555 and so on that are implemented in OpenSSL.
7558 *) Enhance c_rehash script. Old version would mishandle certificates
7559 with the same subject name hash and wouldn't handle CRLs at all.
7560 Added -fingerprint option to crl utility, to support new c_rehash
7564 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
7567 *) Fix for SSL server purpose checking. Server checking was
7568 rejecting certificates which had extended key usage present
7569 but no ssl client purpose.
7570 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
7572 *) Make PKCS#12 code work with no password. The PKCS#12 spec
7573 is a little unclear about how a blank password is handled.
7574 Since the password in encoded as a BMPString with terminating
7575 double NULL a zero length password would end up as just the
7576 double NULL. However no password at all is different and is
7577 handled differently in the PKCS#12 key generation code. NS
7578 treats a blank password as zero length. MSIE treats it as no
7579 password on export: but it will try both on import. We now do
7580 the same: PKCS12_parse() tries zero length and no password if
7581 the password is set to "" or NULL (NULL is now a valid password:
7582 it wasn't before) as does the pkcs12 application.
7585 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
7586 perror when PEM_read_bio_X509_REQ fails, the error message must
7587 be obtained from the error queue.
7590 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
7591 it in ERR_remove_state if appropriate, and change ERR_get_state
7592 accordingly to avoid race conditions (this is necessary because
7593 thread_hash is no longer constant once set).
7596 *) Bugfix for linux-elf makefile.one.
7599 *) RSA_get_default_method() will now cause a default
7600 RSA_METHOD to be chosen if one doesn't exist already.
7601 Previously this was only set during a call to RSA_new()
7602 or RSA_new_method(NULL) meaning it was possible for
7603 RSA_get_default_method() to return NULL.
7606 *) Added native name translation to the existing DSO code
7607 that will convert (if the flag to do so is set) filenames
7608 that are sufficiently small and have no path information
7609 into a canonical native form. Eg. "blah" converted to
7610 "libblah.so" or "blah.dll" etc.
7613 *) New function ERR_error_string_n(e, buf, len) which is like
7614 ERR_error_string(e, buf), but writes at most 'len' bytes
7615 including the 0 terminator. For ERR_error_string_n, 'buf'
7617 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
7619 *) CONF library reworked to become more general. A new CONF
7620 configuration file reader "class" is implemented as well as a
7621 new functions (NCONF_*, for "New CONF") to handle it. The now
7622 old CONF_* functions are still there, but are reimplemented to
7623 work in terms of the new functions. Also, a set of functions
7624 to handle the internal storage of the configuration data is
7625 provided to make it easier to write new configuration file
7626 reader "classes" (I can definitely see something reading a
7627 configuration file in XML format, for example), called _CONF_*,
7628 or "the configuration storage API"...
7630 The new configuration file reading functions are:
7632 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
7633 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
7635 NCONF_default, NCONF_WIN32
7637 NCONF_dump_fp, NCONF_dump_bio
7639 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
7640 NCONF_new creates a new CONF object. This works in the same way
7641 as other interfaces in OpenSSL, like the BIO interface.
7642 NCONF_dump_* dump the internal storage of the configuration file,
7643 which is useful for debugging. All other functions take the same
7644 arguments as the old CONF_* functions wth the exception of the
7645 first that must be a `CONF *' instead of a `LHASH *'.
7647 To make it easer to use the new classes with the old CONF_* functions,
7648 the function CONF_set_default_method is provided.
7651 *) Add '-tls1' option to 'openssl ciphers', which was already
7652 mentioned in the documentation but had not been implemented.
7653 (This option is not yet really useful because even the additional
7654 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
7657 *) Initial DSO code added into libcrypto for letting OpenSSL (and
7658 OpenSSL-based applications) load shared libraries and bind to
7659 them in a portable way.
7660 [Geoff Thorpe, with contributions from Richard Levitte]
7662 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
7664 *) Make sure _lrotl and _lrotr are only used with MSVC.
7666 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
7667 (the default implementation of RAND_status).
7669 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
7670 to '-clrext' (= clear extensions), as intended and documented.
7671 [Bodo Moeller; inconsistency pointed out by Michael Attili
7674 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
7675 was larger than the MD block size.
7676 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
7678 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
7679 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
7680 using the passed key: if the passed key was a private key the result
7681 of X509_print(), for example, would be to print out all the private key
7685 *) des_quad_cksum() byte order bug fix.
7686 [Ulf Möller, using the problem description in krb4-0.9.7, where
7687 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
7689 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
7691 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
7693 *) For easily testing in shell scripts whether some command
7694 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
7695 returns with exit code 0 iff no command of the given name is available.
7696 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
7697 the output goes to stdout and nothing is printed to stderr.
7698 Additional arguments are always ignored.
7700 Since for each cipher there is a command of the same name,
7701 the 'no-cipher' compilation switches can be tested this way.
7703 ('openssl no-XXX' is not able to detect pseudo-commands such
7704 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
7707 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
7710 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
7711 is set; it will be thrown away anyway because each handshake creates
7713 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
7714 to parameters -- in previous versions (since OpenSSL 0.9.3) the
7715 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
7716 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7719 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7720 'Q' and 'R' lose their special meanings (quit/renegotiate).
7721 This is part of what -quiet does; unlike -quiet, -ign_eof
7722 does not suppress any output.
7725 *) Add compatibility options to the purpose and trust code. The
7726 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7727 accepts a certificate or CA, this was the previous behaviour,
7728 with all the associated security issues.
7730 X509_TRUST_COMPAT is the old trust behaviour: only and
7731 automatically trust self signed roots in certificate store. A
7732 new trust setting X509_TRUST_DEFAULT is used to specify that
7733 a purpose has no associated trust setting and it should instead
7734 use the value in the default purpose.
7737 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7738 and fix a memory leak.
7741 *) In util/mkerr.pl (which implements 'make errors'), preserve
7742 reason strings from the previous version of the .c file, as
7743 the default to have only downcase letters (and digits) in
7744 automatically generated reasons codes is not always appropriate.
7747 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7748 using strerror. Previously, ERR_reason_error_string() returned
7749 library names as reason strings for SYSerr; but SYSerr is a special
7750 case where small numbers are errno values, not library numbers.
7753 *) Add '-dsaparam' option to 'openssl dhparam' application. This
7754 converts DSA parameters into DH parameters. (When creating parameters,
7755 DSA_generate_parameters is used.)
7758 *) Include 'length' (recommended exponent length) in C code generated
7759 by 'openssl dhparam -C'.
7762 *) The second argument to set_label in perlasm was already being used
7763 so couldn't be used as a "file scope" flag. Moved to third argument
7767 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
7768 instead of RAND_bytes for encryption IVs and salts.
7771 *) Include RAND_status() into RAND_METHOD instead of implementing
7772 it only for md_rand.c Otherwise replacing the PRNG by calling
7773 RAND_set_rand_method would be impossible.
7776 *) Don't let DSA_generate_key() enter an infinite loop if the random
7777 number generation fails.
7780 *) New 'rand' application for creating pseudo-random output.
7783 *) Added configuration support for Linux/IA64
7784 [Rolf Haberrecker <rolf@suse.de>]
7786 *) Assembler module support for Mingw32.
7789 *) Shared library support for HPUX (in shlib/).
7790 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
7792 *) Shared library support for Solaris gcc.
7793 [Lutz Behnke <behnke@trustcenter.de>]
7795 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
7797 *) PKCS7_encrypt() was adding text MIME headers twice because they
7798 were added manually and by SMIME_crlf_copy().
7801 *) In bntest.c don't call BN_rand with zero bits argument.
7802 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
7804 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
7805 case was implemented. This caused BN_div_recp() to fail occasionally.
7808 *) Add an optional second argument to the set_label() in the perl
7809 assembly language builder. If this argument exists and is set
7810 to 1 it signals that the assembler should use a symbol whose
7811 scope is the entire file, not just the current function. This
7812 is needed with MASM which uses the format label:: for this scope.
7813 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
7815 *) Change the ASN1 types so they are typedefs by default. Before
7816 almost all types were #define'd to ASN1_STRING which was causing
7817 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
7821 *) Change names of new functions to the new get1/get0 naming
7822 convention: After 'get1', the caller owns a reference count
7823 and has to call ..._free; 'get0' returns a pointer to some
7824 data structure without incrementing reference counters.
7825 (Some of the existing 'get' functions increment a reference
7826 counter, some don't.)
7827 Similarly, 'set1' and 'add1' functions increase reference
7828 counters or duplicate objects.
7831 *) Allow for the possibility of temp RSA key generation failure:
7832 the code used to assume it always worked and crashed on failure.
7835 *) Fix potential buffer overrun problem in BIO_printf().
7836 [Ulf Möller, using public domain code by Patrick Powell; problem
7837 pointed out by David Sacerdote <das33@cornell.edu>]
7839 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
7840 RAND_egd() and RAND_status(). In the command line application,
7841 the EGD socket can be specified like a seed file using RANDFILE
7845 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
7846 Some CAs (e.g. Verisign) distribute certificates in this form.
7849 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
7850 list to exclude them. This means that no special compilation option
7851 is needed to use anonymous DH: it just needs to be included in the
7855 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
7856 EVP_MD_type. The old functionality is available in a new macro called
7857 EVP_MD_md(). Change code that uses it and update docs.
7860 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
7861 where the 'void *' argument is replaced by a function pointer argument.
7862 Previously 'void *' was abused to point to functions, which works on
7863 many platforms, but is not correct. As these functions are usually
7864 called by macros defined in OpenSSL header files, most source code
7865 should work without changes.
7868 *) <openssl/opensslconf.h> (which is created by Configure) now contains
7869 sections with information on -D... compiler switches used for
7870 compiling the library so that applications can see them. To enable
7871 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
7872 must be defined. E.g.,
7873 #define OPENSSL_ALGORITHM_DEFINES
7874 #include <openssl/opensslconf.h>
7875 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
7876 [Richard Levitte, Ulf and Bodo Möller]
7878 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
7882 *) Change the 'other' type in certificate aux info to a STACK_OF
7883 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
7884 the required ASN1 format: arbitrary types determined by an OID.
7887 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
7888 argument to 'req'. This is not because the function is newer or
7889 better than others it just uses the work 'NEW' in the certificate
7890 request header lines. Some software needs this.
7893 *) Reorganise password command line arguments: now passwords can be
7894 obtained from various sources. Delete the PEM_cb function and make
7895 it the default behaviour: i.e. if the callback is NULL and the
7896 usrdata argument is not NULL interpret it as a null terminated pass
7897 phrase. If usrdata and the callback are NULL then the pass phrase
7898 is prompted for as usual.
7901 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
7902 the support is automatically enabled. The resulting binaries will
7903 autodetect the card and use it if present.
7904 [Ben Laurie and Compaq Inc.]
7906 *) Work around for Netscape hang bug. This sends certificate request
7907 and server done in one record. Since this is perfectly legal in the
7908 SSL/TLS protocol it isn't a "bug" option and is on by default. See
7909 the bugs/SSLv3 entry for more info.
7912 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
7915 *) Add -rand argument to smime and pkcs12 applications and read/write
7919 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
7922 *) Add command line password options to the remaining applications.
7925 *) Bug fix for BN_div_recp() for numerators with an even number of
7929 *) More tests in bntest.c, and changed test_bn output.
7932 *) ./config recognizes MacOS X now.
7935 *) Bug fix for BN_div() when the first words of num and divsor are
7936 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
7939 *) Add support for various broken PKCS#8 formats, and command line
7940 options to produce them.
7943 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
7944 get temporary BIGNUMs from a BN_CTX.
7947 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
7951 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
7952 include a #define from the old name to the new. The original intent
7953 was that statically linked binaries could for example just call
7954 SSLeay_add_all_ciphers() to just add ciphers to the table and not
7955 link with digests. This never worked becayse SSLeay_add_all_digests()
7956 and SSLeay_add_all_ciphers() were in the same source file so calling
7957 one would link with the other. They are now in separate source files.
7960 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
7963 *) Use a less unusual form of the Miller-Rabin primality test (it used
7964 a binary algorithm for exponentiation integrated into the Miller-Rabin
7965 loop, our standard modexp algorithms are faster).
7968 *) Support for the EBCDIC character set completed.
7969 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
7971 *) Source code cleanups: use const where appropriate, eliminate casts,
7972 use void * instead of char * in lhash.
7975 *) Bugfix: ssl3_send_server_key_exchange was not restartable
7976 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
7977 this the server could overwrite ephemeral keys that the client
7981 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
7982 using 50 iterations of the Rabin-Miller test.
7984 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
7985 iterations of the Rabin-Miller test as required by the appendix
7986 to FIPS PUB 186[-1]) instead of DSA_is_prime.
7987 As BN_is_prime_fasttest includes trial division, DSA parameter
7988 generation becomes much faster.
7990 This implies a change for the callback functions in DSA_is_prime
7991 and DSA_generate_parameters: The callback function is called once
7992 for each positive witness in the Rabin-Miller test, not just
7993 occasionally in the inner loop; and the parameters to the
7994 callback function now provide an iteration count for the outer
7995 loop rather than for the current invocation of the inner loop.
7996 DSA_generate_parameters additionally can call the callback
7997 function with an 'iteration count' of -1, meaning that a
7998 candidate has passed the trial division test (when q is generated
7999 from an application-provided seed, trial division is skipped).
8002 *) New function BN_is_prime_fasttest that optionally does trial
8003 division before starting the Rabin-Miller test and has
8004 an additional BN_CTX * argument (whereas BN_is_prime always
8005 has to allocate at least one BN_CTX).
8006 'callback(1, -1, cb_arg)' is called when a number has passed the
8007 trial division stage.
8010 *) Fix for bug in CRL encoding. The validity dates weren't being handled
8014 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
8017 *) New function BN_pseudo_rand().
8020 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
8021 bignum version of BN_from_montgomery() with the working code from
8022 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
8026 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
8027 made it impossible to use the same SSL_SESSION data structure in
8028 SSL2 clients in multiple threads.
8031 *) The return value of RAND_load_file() no longer counts bytes obtained
8032 by stat(). RAND_load_file(..., -1) is new and uses the complete file
8033 to seed the PRNG (previously an explicit byte count was required).
8034 [Ulf Möller, Bodo Möller]
8036 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
8037 used (char *) instead of (void *) and had casts all over the place.
8040 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
8043 *) Retain source code compatibility for BN_prime_checks macro:
8044 BN_is_prime(..., BN_prime_checks, ...) now uses
8045 BN_prime_checks_for_size to determine the appropriate number of
8046 Rabin-Miller iterations.
8049 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
8050 DH_CHECK_P_NOT_SAFE_PRIME.
8051 (Check if this is true? OpenPGP calls them "strong".)
8054 *) Merge the functionality of "dh" and "gendh" programs into a new program
8055 "dhparam". The old programs are retained for now but will handle DH keys
8056 (instead of parameters) in future.
8059 *) Make the ciphers, s_server and s_client programs check the return values
8060 when a new cipher list is set.
8063 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
8064 ciphers. Before when the 56bit ciphers were enabled the sorting was
8067 The syntax for the cipher sorting has been extended to support sorting by
8068 cipher-strength (using the strength_bits hard coded in the tables).
8069 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
8071 Fix a bug in the cipher-command parser: when supplying a cipher command
8072 string with an "undefined" symbol (neither command nor alphanumeric
8073 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
8074 an error is flagged.
8076 Due to the strength-sorting extension, the code of the
8077 ssl_create_cipher_list() function was completely rearranged. I hope that
8078 the readability was also increased :-)
8079 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8081 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
8082 for the first serial number and places 2 in the serial number file. This
8083 avoids problems when the root CA is created with serial number zero and
8084 the first user certificate has the same issuer name and serial number
8088 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
8089 the new code. Add documentation for this stuff.
8092 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
8093 X509_*() to X509at_*() on the grounds that they don't handle X509
8094 structures and behave in an analagous way to the X509v3 functions:
8095 they shouldn't be called directly but wrapper functions should be used
8098 So we also now have some wrapper functions that call the X509at functions
8099 when passed certificate requests. (TO DO: similar things can be done with
8100 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
8101 things. Some of these need some d2i or i2d and print functionality
8102 because they handle more complex structures.)
8105 *) Add missing #ifndefs that caused missing symbols when building libssl
8106 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
8107 NO_RSA in ssl/s2*.c.
8108 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
8110 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
8111 has a return value which indicates the quality of the random data
8112 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
8113 error queue. New function RAND_pseudo_bytes() generates output that is
8114 guaranteed to be unique but not unpredictable. RAND_add is like
8115 RAND_seed, but takes an extra argument for an entropy estimate
8116 (RAND_seed always assumes full entropy).
8119 *) Do more iterations of Rabin-Miller probable prime test (specifically,
8120 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
8121 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
8122 in crypto/bn/bn_prime.c for the complete table). This guarantees a
8123 false-positive rate of at most 2^-80 for random input.
8126 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
8129 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
8130 in the 0.9.5 release), this returns the chain
8131 from an X509_CTX structure with a dup of the stack and all
8132 the X509 reference counts upped: so the stack will exist
8133 after X509_CTX_cleanup() has been called. Modify pkcs12.c
8136 Also make SSL_SESSION_print() print out the verify return
8140 *) Add manpage for the pkcs12 command. Also change the default
8141 behaviour so MAC iteration counts are used unless the new
8142 -nomaciter option is used. This improves file security and
8143 only older versions of MSIE (4.0 for example) need it.
8146 *) Honor the no-xxx Configure options when creating .DEF files.
8149 *) Add PKCS#10 attributes to field table: challengePassword,
8150 unstructuredName and unstructuredAddress. These are taken from
8151 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
8152 international characters are used.
8154 More changes to X509_ATTRIBUTE code: allow the setting of types
8155 based on strings. Remove the 'loc' parameter when adding
8156 attributes because these will be a SET OF encoding which is sorted
8160 *) Initial changes to the 'req' utility to allow request generation
8161 automation. This will allow an application to just generate a template
8162 file containing all the field values and have req construct the
8165 Initial support for X509_ATTRIBUTE handling. Stacks of these are
8166 used all over the place including certificate requests and PKCS#7
8167 structures. They are currently handled manually where necessary with
8168 some primitive wrappers for PKCS#7. The new functions behave in a
8169 manner analogous to the X509 extension functions: they allow
8170 attributes to be looked up by NID and added.
8172 Later something similar to the X509V3 code would be desirable to
8173 automatically handle the encoding, decoding and printing of the
8174 more complex types. The string types like challengePassword can
8175 be handled by the string table functions.
8177 Also modified the multi byte string table handling. Now there is
8178 a 'global mask' which masks out certain types. The table itself
8179 can use the flag STABLE_NO_MASK to ignore the mask setting: this
8180 is useful when for example there is only one permissible type
8181 (as in countryName) and using the mask might result in no valid
8185 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
8186 SSL_get_peer_finished to allow applications to obtain the latest
8187 Finished messages sent to the peer or expected from the peer,
8188 respectively. (SSL_get_peer_finished is usually the Finished message
8189 actually received from the peer, otherwise the protocol will be aborted.)
8191 As the Finished message are message digests of the complete handshake
8192 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
8193 be used for external authentication procedures when the authentication
8194 provided by SSL/TLS is not desired or is not enough.
8197 *) Enhanced support for Alpha Linux is added. Now ./config checks if
8198 the host supports BWX extension and if Compaq C is present on the
8199 $PATH. Just exploiting of the BWX extension results in 20-30%
8200 performance kick for some algorithms, e.g. DES and RC4 to mention
8201 a couple. Compaq C in turn generates ~20% faster code for MD5 and
8205 *) Add support for MS "fast SGC". This is arguably a violation of the
8206 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
8207 weak crypto and after checking the certificate is SGC a second one
8208 with strong crypto. MS SGC stops the first handshake after receiving
8209 the server certificate message and sends a second client hello. Since
8210 a server will typically do all the time consuming operations before
8211 expecting any further messages from the client (server key exchange
8212 is the most expensive) there is little difference between the two.
8214 To get OpenSSL to support MS SGC we have to permit a second client
8215 hello message after we have sent server done. In addition we have to
8216 reset the MAC if we do get this second client hello.
8219 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
8220 if a DER encoded private key is RSA or DSA traditional format. Changed
8221 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
8222 format DER encoded private key. Newer code should use PKCS#8 format which
8223 has the key type encoded in the ASN1 structure. Added DER private key
8224 support to pkcs8 application.
8227 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
8228 ciphersuites has been selected (as required by the SSL 3/TLS 1
8229 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
8230 is set, we interpret this as a request to violate the specification
8231 (the worst that can happen is a handshake failure, and 'correct'
8232 behaviour would result in a handshake failure anyway).
8235 *) In SSL_CTX_add_session, take into account that there might be multiple
8236 SSL_SESSION structures with the same session ID (e.g. when two threads
8237 concurrently obtain them from an external cache).
8238 The internal cache can handle only one SSL_SESSION with a given ID,
8239 so if there's a conflict, we now throw out the old one to achieve
8243 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
8244 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
8245 some routines that use cipher OIDs: some ciphers do not have OIDs
8246 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
8250 *) Simplify the trust setting structure and code. Now we just have
8251 two sequences of OIDs for trusted and rejected settings. These will
8252 typically have values the same as the extended key usage extension
8253 and any application specific purposes.
8255 The trust checking code now has a default behaviour: it will just
8256 check for an object with the same NID as the passed id. Functions can
8257 be provided to override either the default behaviour or the behaviour
8258 for a given id. SSL client, server and email already have functions
8259 in place for compatibility: they check the NID and also return "trusted"
8260 if the certificate is self signed.
8263 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
8264 traditional format into an EVP_PKEY structure.
8267 *) Add a password callback function PEM_cb() which either prompts for
8268 a password if usr_data is NULL or otherwise assumes it is a null
8269 terminated password. Allow passwords to be passed on command line
8270 environment or config files in a few more utilities.
8273 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
8274 keys. Add some short names for PKCS#8 PBE algorithms and allow them
8275 to be specified on the command line for the pkcs8 and pkcs12 utilities.
8276 Update documentation.
8279 *) Support for ASN1 "NULL" type. This could be handled before by using
8280 ASN1_TYPE but there wasn't any function that would try to read a NULL
8281 and produce an error if it couldn't. For compatibility we also have
8282 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
8283 don't allocate anything because they don't need to.
8286 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
8288 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
8290 *) Rebuild of the memory allocation routines used by OpenSSL code and
8291 possibly others as well. The purpose is to make an interface that
8292 provide hooks so anyone can build a separate set of allocation and
8293 deallocation routines to be used by OpenSSL, for example memory
8294 pool implementations, or something else, which was previously hard
8295 since Malloc(), Realloc() and Free() were defined as macros having
8296 the values malloc, realloc and free, respectively (except for Win32
8297 compilations). The same is provided for memory debugging code.
8298 OpenSSL already comes with functionality to find memory leaks, but
8299 this gives people a chance to debug other memory problems.
8301 With these changes, a new set of functions and macros have appeared:
8303 CRYPTO_set_mem_debug_functions() [F]
8304 CRYPTO_get_mem_debug_functions() [F]
8305 CRYPTO_dbg_set_options() [F]
8306 CRYPTO_dbg_get_options() [F]
8307 CRYPTO_malloc_debug_init() [M]
8309 The memory debug functions are NULL by default, unless the library
8310 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
8311 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
8312 gives the standard debugging functions that come with OpenSSL) or
8313 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
8314 provided by the library user) must be used. When the standard
8315 debugging functions are used, CRYPTO_dbg_set_options can be used to
8316 request additional information:
8317 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
8318 the CRYPTO_MDEBUG_xxx macro when compiling the library.
8320 Also, things like CRYPTO_set_mem_functions will always give the
8321 expected result (the new set of functions is used for allocation
8322 and deallocation) at all times, regardless of platform and compiler
8325 To finish it up, some functions that were never use in any other
8326 way than through macros have a new API and new semantic:
8329 CRYPTO_dbg_realloc()
8332 All macros of value have retained their old syntax.
8333 [Richard Levitte and Bodo Moeller]
8335 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
8336 ordering of SMIMECapabilities wasn't in "strength order" and there
8337 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
8341 *) Some ASN1 types with illegal zero length encoding (INTEGER,
8342 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
8343 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
8345 *) Merge in my S/MIME library for OpenSSL. This provides a simple
8346 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
8347 functionality to handle multipart/signed properly) and a utility
8348 called 'smime' to call all this stuff. This is based on code I
8349 originally wrote for Celo who have kindly allowed it to be
8350 included in OpenSSL.
8353 *) Add variants des_set_key_checked and des_set_key_unchecked of
8354 des_set_key (aka des_key_sched). Global variable des_check_key
8355 decides which of these is called by des_set_key; this way
8356 des_check_key behaves as it always did, but applications and
8357 the library itself, which was buggy for des_check_key == 1,
8358 have a cleaner way to pick the version they need.
8361 *) New function PKCS12_newpass() which changes the password of a
8365 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
8366 dynamic mix. In both cases the ids can be used as an index into the
8367 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
8368 functions so they accept a list of the field values and the
8369 application doesn't need to directly manipulate the X509_TRUST
8373 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
8377 *) Modify the way the V3 extension code looks up extensions. This now
8378 works in a similar way to the object code: we have some "standard"
8379 extensions in a static table which is searched with OBJ_bsearch()
8380 and the application can add dynamic ones if needed. The file
8381 crypto/x509v3/ext_dat.h now has the info: this file needs to be
8382 updated whenever a new extension is added to the core code and kept
8383 in ext_nid order. There is a simple program 'tabtest.c' which checks
8384 this. New extensions are not added too often so this file can readily
8385 be maintained manually.
8387 There are two big advantages in doing things this way. The extensions
8388 can be looked up immediately and no longer need to be "added" using
8389 X509V3_add_standard_extensions(): this function now does nothing.
8390 [Side note: I get *lots* of email saying the extension code doesn't
8391 work because people forget to call this function]
8392 Also no dynamic allocation is done unless new extensions are added:
8393 so if we don't add custom extensions there is no need to call
8394 X509V3_EXT_cleanup().
8397 *) Modify enc utility's salting as follows: make salting the default. Add a
8398 magic header, so unsalted files fail gracefully instead of just decrypting
8399 to garbage. This is because not salting is a big security hole, so people
8400 should be discouraged from doing it.
8403 *) Fixes and enhancements to the 'x509' utility. It allowed a message
8404 digest to be passed on the command line but it only used this
8405 parameter when signing a certificate. Modified so all relevant
8406 operations are affected by the digest parameter including the
8407 -fingerprint and -x509toreq options. Also -x509toreq choked if a
8408 DSA key was used because it didn't fix the digest.
8411 *) Initial certificate chain verify code. Currently tests the untrusted
8412 certificates for consistency with the verify purpose (which is set
8413 when the X509_STORE_CTX structure is set up) and checks the pathlength.
8415 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
8416 this is because it will reject chains with invalid extensions whereas
8417 every previous version of OpenSSL and SSLeay made no checks at all.
8419 Trust code: checks the root CA for the relevant trust settings. Trust
8420 settings have an initial value consistent with the verify purpose: e.g.
8421 if the verify purpose is for SSL client use it expects the CA to be
8422 trusted for SSL client use. However the default value can be changed to
8423 permit custom trust settings: one example of this would be to only trust
8424 certificates from a specific "secure" set of CAs.
8426 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
8427 which should be used for version portability: especially since the
8428 verify structure is likely to change more often now.
8430 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
8431 to set them. If not set then assume SSL clients will verify SSL servers
8434 Two new options to the verify program: -untrusted allows a set of
8435 untrusted certificates to be passed in and -purpose which sets the
8436 intended purpose of the certificate. If a purpose is set then the
8437 new chain verify code is used to check extension consistency.
8440 *) Support for the authority information access extension.
8443 *) Modify RSA and DSA PEM read routines to transparently handle
8444 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
8445 public keys in a format compatible with certificate
8446 SubjectPublicKeyInfo structures. Unfortunately there were already
8447 functions called *_PublicKey_* which used various odd formats so
8448 these are retained for compatibility: however the DSA variants were
8449 never in a public release so they have been deleted. Changed dsa/rsa
8450 utilities to handle the new format: note no releases ever handled public
8451 keys so we should be OK.
8453 The primary motivation for this change is to avoid the same fiasco
8454 that dogs private keys: there are several incompatible private key
8455 formats some of which are standard and some OpenSSL specific and
8456 require various evil hacks to allow partial transparent handling and
8457 even then it doesn't work with DER formats. Given the option anything
8458 other than PKCS#8 should be dumped: but the other formats have to
8459 stay in the name of compatibility.
8461 With public keys and the benefit of hindsight one standard format
8462 is used which works with EVP_PKEY, RSA or DSA structures: though
8463 it clearly returns an error if you try to read the wrong kind of key.
8465 Added a -pubkey option to the 'x509' utility to output the public key.
8466 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
8467 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
8468 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
8469 that do the same as the EVP_PKEY_assign_*() except they up the
8470 reference count of the added key (they don't "swallow" the
8474 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
8475 CRLs would fail if the file contained no certificates or no CRLs:
8476 added a new function to read in both types and return the number
8477 read: this means that if none are read it will be an error. The
8478 DER versions of the certificate and CRL reader would always fail
8479 because it isn't possible to mix certificates and CRLs in DER format
8480 without choking one or the other routine. Changed this to just read
8481 a certificate: this is the best we can do. Also modified the code
8482 in apps/verify.c to take notice of return codes: it was previously
8483 attempting to read in certificates from NULL pointers and ignoring
8484 any errors: this is one reason why the cert and CRL reader seemed
8485 to work. It doesn't check return codes from the default certificate
8486 routines: these may well fail if the certificates aren't installed.
8489 *) Code to support otherName option in GeneralName.
8492 *) First update to verify code. Change the verify utility
8493 so it warns if it is passed a self signed certificate:
8494 for consistency with the normal behaviour. X509_verify
8495 has been modified to it will now verify a self signed
8496 certificate if *exactly* the same certificate appears
8497 in the store: it was previously impossible to trust a
8498 single self signed certificate. This means that:
8499 openssl verify ss.pem
8500 now gives a warning about a self signed certificate but
8501 openssl verify -CAfile ss.pem ss.pem
8505 *) For servers, store verify_result in SSL_SESSION data structure
8506 (and add it to external session representation).
8507 This is needed when client certificate verifications fails,
8508 but an application-provided verification callback (set by
8509 SSL_CTX_set_cert_verify_callback) allows accepting the session
8510 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
8511 but returns 1): When the session is reused, we have to set
8512 ssl->verify_result to the appropriate error code to avoid
8514 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
8516 *) Fix a bug in the new PKCS#7 code: it didn't consider the
8517 case in PKCS7_dataInit() where the signed PKCS7 structure
8518 didn't contain any existing data because it was being created.
8519 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
8521 *) Add a salt to the key derivation routines in enc.c. This
8522 forms the first 8 bytes of the encrypted file. Also add a
8523 -S option to allow a salt to be input on the command line.
8526 *) New function X509_cmp(). Oddly enough there wasn't a function
8527 to compare two certificates. We do this by working out the SHA1
8528 hash and comparing that. X509_cmp() will be needed by the trust
8532 *) SSL_get1_session() is like SSL_get_session(), but increments
8533 the reference count in the SSL_SESSION returned.
8534 [Geoff Thorpe <geoff@eu.c2.net>]
8536 *) Fix for 'req': it was adding a null to request attributes.
8537 Also change the X509_LOOKUP and X509_INFO code to handle
8538 certificate auxiliary information.
8541 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
8545 *) Add the possibility to add extra information to the memory leak
8546 detecting output, to form tracebacks, showing from where each
8547 allocation was originated: CRYPTO_push_info("constant string") adds
8548 the string plus current file name and line number to a per-thread
8549 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
8550 is like calling CYRPTO_pop_info() until the stack is empty.
8551 Also updated memory leak detection code to be multi-thread-safe.
8554 *) Add options -text and -noout to pkcs7 utility and delete the
8555 encryption options which never did anything. Update docs.
8558 *) Add options to some of the utilities to allow the pass phrase
8559 to be included on either the command line (not recommended on
8560 OSes like Unix) or read from the environment. Update the
8561 manpages and fix a few bugs.
8564 *) Add a few manpages for some of the openssl commands.
8567 *) Fix the -revoke option in ca. It was freeing up memory twice,
8568 leaking and not finding already revoked certificates.
8571 *) Extensive changes to support certificate auxiliary information.
8572 This involves the use of X509_CERT_AUX structure and X509_AUX
8573 functions. An X509_AUX function such as PEM_read_X509_AUX()
8574 can still read in a certificate file in the usual way but it
8575 will also read in any additional "auxiliary information". By
8576 doing things this way a fair degree of compatibility can be
8577 retained: existing certificates can have this information added
8578 using the new 'x509' options.
8580 Current auxiliary information includes an "alias" and some trust
8581 settings. The trust settings will ultimately be used in enhanced
8582 certificate chain verification routines: currently a certificate
8583 can only be trusted if it is self signed and then it is trusted
8587 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
8588 The problem was that one of the replacement routines had not been working
8589 since SSLeay releases. For now the offending routine has been replaced
8590 with non-optimised assembler. Even so, this now gives around 95%
8591 performance improvement for 1024 bit RSA signs.
8594 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
8595 handling. Most clients have the effective key size in bits equal to
8596 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
8597 A few however don't do this and instead use the size of the decrypted key
8598 to determine the RC2 key length and the AlgorithmIdentifier to determine
8599 the effective key length. In this case the effective key length can still
8600 be 40 bits but the key length can be 168 bits for example. This is fixed
8601 by manually forcing an RC2 key into the EVP_PKEY structure because the
8602 EVP code can't currently handle unusual RC2 key sizes: it always assumes
8603 the key length and effective key length are equal.
8606 *) Add a bunch of functions that should simplify the creation of
8607 X509_NAME structures. Now you should be able to do:
8608 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
8609 and have it automatically work out the correct field type and fill in
8610 the structures. The more adventurous can try:
8611 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
8612 and it will (hopefully) work out the correct multibyte encoding.
8615 *) Change the 'req' utility to use the new field handling and multibyte
8616 copy routines. Before the DN field creation was handled in an ad hoc
8617 way in req, ca, and x509 which was rather broken and didn't support
8618 BMPStrings or UTF8Strings. Since some software doesn't implement
8619 BMPStrings or UTF8Strings yet, they can be enabled using the config file
8620 using the dirstring_type option. See the new comment in the default
8621 openssl.cnf for more info.
8624 *) Make crypto/rand/md_rand.c more robust:
8625 - Assure unique random numbers after fork().
8626 - Make sure that concurrent threads access the global counter and
8627 md serializably so that we never lose entropy in them
8628 or use exactly the same state in multiple threads.
8629 Access to the large state is not always serializable because
8630 the additional locking could be a performance killer, and
8631 md should be large enough anyway.
8634 *) New file apps/app_rand.c with commonly needed functionality
8635 for handling the random seed file.
8637 Use the random seed file in some applications that previously did not:
8639 dsaparam -genkey (which also ignored its '-rand' option),
8642 x509 (when signing).
8643 Except on systems with /dev/urandom, it is crucial to have a random
8644 seed file at least for key creation, DSA signing, and for DH exchanges;
8645 for RSA signatures we could do without one.
8647 gendh and gendsa (unlike genrsa) used to read only the first byte
8648 of each file listed in the '-rand' option. The function as previously
8649 found in genrsa is now in app_rand.c and is used by all programs
8650 that support '-rand'.
8653 *) In RAND_write_file, use mode 0600 for creating files;
8654 don't just chmod when it may be too late.
8657 *) Report an error from X509_STORE_load_locations
8658 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
8661 *) New function ASN1_mbstring_copy() this copies a string in either
8662 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
8663 into an ASN1_STRING type. A mask of permissible types is passed
8664 and it chooses the "minimal" type to use or an error if not type
8668 *) Add function equivalents to the various macros in asn1.h. The old
8669 macros are retained with an M_ prefix. Code inside the library can
8670 use the M_ macros. External code (including the openssl utility)
8671 should *NOT* in order to be "shared library friendly".
8674 *) Add various functions that can check a certificate's extensions
8675 to see if it usable for various purposes such as SSL client,
8676 server or S/MIME and CAs of these types. This is currently
8677 VERY EXPERIMENTAL but will ultimately be used for certificate chain
8678 verification. Also added a -purpose flag to x509 utility to
8679 print out all the purposes.
8682 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
8686 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
8687 for, obtain and decode and extension and obtain its critical flag.
8688 This allows all the necessary extension code to be handled in a
8689 single function call.
8692 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
8693 platforms. See crypto/rc4/rc4_enc.c for further details.
8696 *) New -noout option to asn1parse. This causes no output to be produced
8697 its main use is when combined with -strparse and -out to extract data
8698 from a file (which may not be in ASN.1 format).
8701 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
8702 when producing the local key id.
8703 [Richard Levitte <levitte@stacken.kth.se>]
8705 *) New option -dhparam in s_server. This allows a DH parameter file to be
8706 stated explicitly. If it is not stated then it tries the first server
8707 certificate file. The previous behaviour hard coded the filename
8711 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
8712 a public key to be input or output. For example:
8713 openssl rsa -in key.pem -pubout -out pubkey.pem
8714 Also added necessary DSA public key functions to handle this.
8717 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
8718 in the message. This was handled by allowing
8719 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8720 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8722 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8723 to the end of the strings whereas this didn't. This would cause problems
8724 if strings read with d2i_ASN1_bytes() were later modified.
8725 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8727 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8728 data and it contains EOF it will end up returning an error. This is
8729 caused by input 46 bytes long. The cause is due to the way base64
8730 BIOs find the start of base64 encoded data. They do this by trying a
8731 trial decode on each line until they find one that works. When they
8732 do a flag is set and it starts again knowing it can pass all the
8733 data directly through the decoder. Unfortunately it doesn't reset
8734 the context it uses. This means that if EOF is reached an attempt
8735 is made to pass two EOFs through the context and this causes the
8736 resulting error. This can also cause other problems as well. As is
8737 usual with these problems it takes *ages* to find and the fix is
8738 trivial: move one line.
8739 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8741 *) Ugly workaround to get s_client and s_server working under Windows. The
8742 old code wouldn't work because it needed to select() on sockets and the
8743 tty (for keypresses and to see if data could be written). Win32 only
8744 supports select() on sockets so we select() with a 1s timeout on the
8745 sockets and then see if any characters are waiting to be read, if none
8746 are present then we retry, we also assume we can always write data to
8747 the tty. This isn't nice because the code then blocks until we've
8748 received a complete line of data and it is effectively polling the
8749 keyboard at 1s intervals: however it's quite a bit better than not
8750 working at all :-) A dedicated Windows application might handle this
8751 with an event loop for example.
8754 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
8755 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
8756 will be called when RSA_sign() and RSA_verify() are used. This is useful
8757 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
8758 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
8759 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
8760 This necessitated the support of an extra signature type NID_md5_sha1
8761 for SSL signatures and modifications to the SSL library to use it instead
8762 of calling RSA_public_decrypt() and RSA_private_encrypt().
8765 *) Add new -verify -CAfile and -CApath options to the crl program, these
8766 will lookup a CRL issuers certificate and verify the signature in a
8767 similar way to the verify program. Tidy up the crl program so it
8768 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
8769 less strict. It will now permit CRL extensions even if it is not
8770 a V2 CRL: this will allow it to tolerate some broken CRLs.
8773 *) Initialize all non-automatic variables each time one of the openssl
8774 sub-programs is started (this is necessary as they may be started
8775 multiple times from the "OpenSSL>" prompt).
8776 [Lennart Bang, Bodo Moeller]
8778 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
8779 removing all other RSA functionality (this is what NO_RSA does). This
8780 is so (for example) those in the US can disable those operations covered
8781 by the RSA patent while allowing storage and parsing of RSA keys and RSA
8785 *) Non-copying interface to BIO pairs.
8786 (still largely untested)
8789 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
8790 ASCII string. This was handled independently in various places before.
8793 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
8794 UTF8 strings a character at a time.
8797 *) Use client_version from client hello to select the protocol
8798 (s23_srvr.c) and for RSA client key exchange verification
8799 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
8802 *) Add various utility functions to handle SPKACs, these were previously
8803 handled by poking round in the structure internals. Added new function
8804 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
8805 print, verify and generate SPKACs. Based on an original idea from
8806 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
8809 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
8812 *) Allow the config file extension section to be overwritten on the
8813 command line. Based on an original idea from Massimiliano Pala
8814 <madwolf@comune.modena.it>. The new option is called -extensions
8815 and can be applied to ca, req and x509. Also -reqexts to override
8816 the request extensions in req and -crlexts to override the crl extensions
8820 *) Add new feature to the SPKAC handling in ca. Now you can include
8821 the same field multiple times by preceding it by "XXXX." for example:
8824 this is the same syntax as used in the req config file.
8827 *) Allow certificate extensions to be added to certificate requests. These
8828 are specified in a 'req_extensions' option of the req section of the
8829 config file. They can be printed out with the -text option to req but
8830 are otherwise ignored at present.
8833 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
8834 data read consists of only the final block it would not decrypted because
8835 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
8836 A misplaced 'break' also meant the decrypted final block might not be
8837 copied until the next read.
8840 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
8841 a few extra parameters to the DH structure: these will be useful if
8842 for example we want the value of 'q' or implement X9.42 DH.
8845 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
8846 provides hooks that allow the default DSA functions or functions on a
8847 "per key" basis to be replaced. This allows hardware acceleration and
8848 hardware key storage to be handled without major modification to the
8849 library. Also added low level modexp hooks and CRYPTO_EX structure and
8850 associated functions.
8853 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
8854 as "read only": it can't be written to and the buffer it points to will
8855 not be freed. Reading from a read only BIO is much more efficient than
8856 a normal memory BIO. This was added because there are several times when
8857 an area of memory needs to be read from a BIO. The previous method was
8858 to create a memory BIO and write the data to it, this results in two
8859 copies of the data and an O(n^2) reading algorithm. There is a new
8860 function BIO_new_mem_buf() which creates a read only memory BIO from
8861 an area of memory. Also modified the PKCS#7 routines to use read only
8865 *) Bugfix: ssl23_get_client_hello did not work properly when called in
8866 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
8867 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
8868 but a retry condition occured while trying to read the rest.
8871 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
8872 NID_pkcs7_encrypted by default: this was wrong since this should almost
8873 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
8874 the encrypted data type: this is a more sensible place to put it and it
8875 allows the PKCS#12 code to be tidied up that duplicated this
8879 *) Changed obj_dat.pl script so it takes its input and output files on
8880 the command line. This should avoid shell escape redirection problems
8884 *) Initial support for certificate extension requests, these are included
8885 in things like Xenroll certificate requests. Included functions to allow
8886 extensions to be obtained and added.
8889 *) -crlf option to s_client and s_server for sending newlines as
8890 CRLF (as required by many protocols).
8893 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
8895 *) Install libRSAglue.a when OpenSSL is built with RSAref.
8896 [Ralf S. Engelschall]
8898 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
8899 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
8901 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
8905 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
8906 DH parameters/keys (q is lost during that conversion, but the resulting
8907 DH parameters contain its length).
8909 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
8910 much faster than DH_generate_parameters (which creates parameters
8911 where p = 2*q + 1), and also the smaller q makes DH computations
8912 much more efficient (160-bit exponentiation instead of 1024-bit
8913 exponentiation); so this provides a convenient way to support DHE
8914 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
8915 utter importance to use
8916 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8918 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8919 when such DH parameters are used, because otherwise small subgroup
8920 attacks may become possible!
8923 *) Avoid memory leak in i2d_DHparams.
8926 *) Allow the -k option to be used more than once in the enc program:
8927 this allows the same encrypted message to be read by multiple recipients.
8930 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
8931 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
8932 it will always use the numerical form of the OID, even if it has a short
8936 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
8937 method only got called if p,q,dmp1,dmq1,iqmp components were present,
8938 otherwise bn_mod_exp was called. In the case of hardware keys for example
8939 no private key components need be present and it might store extra data
8940 in the RSA structure, which cannot be accessed from bn_mod_exp.
8941 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
8942 private key operations.
8945 *) Added support for SPARC Linux.
8948 *) pem_password_cb function type incompatibly changed from
8949 typedef int pem_password_cb(char *buf, int size, int rwflag);
8951 ....(char *buf, int size, int rwflag, void *userdata);
8952 so that applications can pass data to their callbacks:
8953 The PEM[_ASN1]_{read,write}... functions and macros now take an
8954 additional void * argument, which is just handed through whenever
8955 the password callback is called.
8956 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
8958 New function SSL_CTX_set_default_passwd_cb_userdata.
8960 Compatibility note: As many C implementations push function arguments
8961 onto the stack in reverse order, the new library version is likely to
8962 interoperate with programs that have been compiled with the old
8963 pem_password_cb definition (PEM_whatever takes some data that
8964 happens to be on the stack as its last argument, and the callback
8965 just ignores this garbage); but there is no guarantee whatsoever that
8968 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
8969 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
8970 problems not only on Windows, but also on some Unix platforms.
8971 To avoid problematic command lines, these definitions are now in an
8972 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
8973 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
8976 *) MIPS III/IV assembler module is reimplemented.
8979 *) More DES library cleanups: remove references to srand/rand and
8980 delete an unused file.
8983 *) Add support for the the free Netwide assembler (NASM) under Win32,
8984 since not many people have MASM (ml) and it can be hard to obtain.
8985 This is currently experimental but it seems to work OK and pass all
8986 the tests. Check out INSTALL.W32 for info.
8989 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
8990 without temporary keys kept an extra copy of the server key,
8991 and connections with temporary keys did not free everything in case
8995 *) New function RSA_check_key and new openssl rsa option -check
8996 for verifying the consistency of RSA keys.
8997 [Ulf Moeller, Bodo Moeller]
8999 *) Various changes to make Win32 compile work:
9000 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
9001 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
9002 comparison" warnings.
9003 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
9006 *) Add a debugging option to PKCS#5 v2 key generation function: when
9007 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
9008 derived keys are printed to stderr.
9011 *) Copy the flags in ASN1_STRING_dup().
9012 [Roman E. Pavlov <pre@mo.msk.ru>]
9014 *) The x509 application mishandled signing requests containing DSA
9015 keys when the signing key was also DSA and the parameters didn't match.
9017 It was supposed to omit the parameters when they matched the signing key:
9018 the verifying software was then supposed to automatically use the CA's
9019 parameters if they were absent from the end user certificate.
9021 Omitting parameters is no longer recommended. The test was also
9022 the wrong way round! This was probably due to unusual behaviour in
9023 EVP_cmp_parameters() which returns 1 if the parameters match.
9024 This meant that parameters were omitted when they *didn't* match and
9025 the certificate was useless. Certificates signed with 'ca' didn't have
9027 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
9029 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
9030 The interface is as follows:
9031 Applications can use
9032 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
9033 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
9034 "off" is now the default.
9035 The library internally uses
9036 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
9037 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
9038 to disable memory-checking temporarily.
9040 Some inconsistent states that previously were possible (and were
9041 even the default) are now avoided.
9043 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
9044 with each memory chunk allocated; this is occasionally more helpful
9045 than just having a counter.
9047 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
9049 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
9053 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
9054 which largely parallels "options", but is for changing API behaviour,
9055 whereas "options" are about protocol behaviour.
9056 Initial "mode" flags are:
9058 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
9059 a single record has been written.
9060 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
9061 retries use the same buffer location.
9062 (But all of the contents must be
9066 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
9069 *) Fix problems with no-hmac etc.
9070 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
9072 *) New functions RSA_get_default_method(), RSA_set_method() and
9073 RSA_get_method(). These allows replacement of RSA_METHODs without having
9074 to mess around with the internals of an RSA structure.
9077 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
9078 Also really enable memory leak checks in openssl.c and in some
9080 [Chad C. Mulligan, Bodo Moeller]
9082 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
9083 up the length of negative integers. This has now been simplified to just
9084 store the length when it is first determined and use it later, rather
9085 than trying to keep track of where data is copied and updating it to
9087 [Steve Henson, reported by Brien Wheeler
9088 <bwheeler@authentica-security.com>]
9090 *) Add a new function PKCS7_signatureVerify. This allows the verification
9091 of a PKCS#7 signature but with the signing certificate passed to the
9092 function itself. This contrasts with PKCS7_dataVerify which assumes the
9093 certificate is present in the PKCS#7 structure. This isn't always the
9094 case: certificates can be omitted from a PKCS#7 structure and be
9095 distributed by "out of band" means (such as a certificate database).
9098 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
9099 function prototypes in pem.h, also change util/mkdef.pl to add the
9100 necessary function names.
9103 *) mk1mf.pl (used by Windows builds) did not properly read the
9104 options set by Configure in the top level Makefile, and Configure
9105 was not even able to write more than one option correctly.
9106 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
9109 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
9110 file to be loaded from a BIO or FILE pointer. The BIO version will
9111 for example allow memory BIOs to contain config info.
9114 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
9115 Whoever hopes to achieve shared-library compatibility across versions
9116 must use this, not the compile-time macro.
9117 (Exercise 0.9.4: Which is the minimum library version required by
9119 Note: All this applies only to multi-threaded programs, others don't
9123 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
9124 through a BIO pair triggered the default case, i.e.
9125 SSLerr(...,SSL_R_UNKNOWN_STATE).
9128 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
9129 can use the SSL library even if none of the specific BIOs is
9133 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
9134 for the encoded length.
9135 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
9137 *) Add initial documentation of the X509V3 functions.
9140 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
9141 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
9142 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
9143 secure PKCS#8 private key format with a high iteration count.
9146 *) Fix determination of Perl interpreter: A perl or perl5
9147 _directory_ in $PATH was also accepted as the interpreter.
9148 [Ralf S. Engelschall]
9150 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
9151 wrong with it but it was very old and did things like calling
9152 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
9156 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
9157 to use the new extension code.
9160 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
9161 with macros. This should make it easier to change their form, add extra
9162 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
9166 *) Add to configuration table a new entry that can specify an alternative
9167 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
9168 according to Mark Crispin <MRC@Panda.COM>.
9172 *) DES CBC did not update the IV. Weird.
9175 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
9176 Changing the behaviour of the former might break existing programs --
9177 where IV updating is needed, des_ncbc_encrypt can be used.
9180 *) When bntest is run from "make test" it drives bc to check its
9181 calculations, as well as internally checking them. If an internal check
9182 fails, it needs to cause bc to give a non-zero result or make test carries
9183 on without noticing the failure. Fixed.
9186 *) DES library cleanups.
9189 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
9190 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
9191 ciphers. NOTE: although the key derivation function has been verified
9192 against some published test vectors it has not been extensively tested
9193 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
9197 *) Instead of "mkdir -p", which is not fully portable, use new
9198 Perl script "util/mkdir-p.pl".
9201 *) Rewrite the way password based encryption (PBE) is handled. It used to
9202 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
9203 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
9204 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
9205 the 'parameter' field of the AlgorithmIdentifier is passed to the
9206 underlying key generation function so it must do its own ASN1 parsing.
9207 This has also changed the EVP_PBE_CipherInit() function which now has a
9208 'parameter' argument instead of literal salt and iteration count values
9209 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
9212 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
9213 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
9214 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
9215 KEY" because this clashed with PKCS#8 unencrypted string. Since this
9216 value was just used as a "magic string" and not used directly its
9217 value doesn't matter.
9220 *) Introduce some semblance of const correctness to BN. Shame C doesn't
9224 *) "linux-sparc64" configuration (ultrapenguin).
9225 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
9226 "linux-sparc" configuration.
9227 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
9229 *) config now generates no-xxx options for missing ciphers.
9232 *) Support the EBCDIC character set (work in progress).
9233 File ebcdic.c not yet included because it has a different license.
9234 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9236 *) Support BS2000/OSD-POSIX.
9237 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9239 *) Make callbacks for key generation use void * instead of char *.
9242 *) Make S/MIME samples compile (not yet tested).
9245 *) Additional typesafe stacks.
9248 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
9252 Changes between 0.9.3 and 0.9.3a [29 May 1999]
9254 *) New configuration variant "sco5-gcc".
9256 *) Updated some demos.
9257 [Sean O Riordain, Wade Scholine]
9259 *) Add missing BIO_free at exit of pkcs12 application.
9262 *) Fix memory leak in conf.c.
9265 *) Updates for Win32 to assembler version of MD5.
9268 *) Set #! path to perl in apps/der_chop to where we found it
9269 instead of using a fixed path.
9272 *) SHA library changes for irix64-mips4-cc.
9275 *) Improvements for VMS support.
9279 Changes between 0.9.2b and 0.9.3 [24 May 1999]
9281 *) Bignum library bug fix. IRIX 6 passes "make test" now!
9282 This also avoids the problems with SC4.2 and unpatched SC5.
9283 [Andy Polyakov <appro@fy.chalmers.se>]
9285 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
9286 These are required because of the typesafe stack would otherwise break
9287 existing code. If old code used a structure member which used to be STACK
9288 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
9289 sk_num or sk_value it would produce an error because the num, data members
9290 are not present in STACK_OF. Now it just produces a warning. sk_set
9291 replaces the old method of assigning a value to sk_value
9292 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
9293 that does this will no longer work (and should use sk_set instead) but
9294 this could be regarded as a "questionable" behaviour anyway.
9297 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
9298 correctly handle encrypted S/MIME data.
9301 *) Change type of various DES function arguments from des_cblock
9302 (which means, in function argument declarations, pointer to char)
9303 to des_cblock * (meaning pointer to array with 8 char elements),
9304 which allows the compiler to do more typechecking; it was like
9305 that back in SSLeay, but with lots of ugly casts.
9307 Introduce new type const_des_cblock.
9310 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
9311 problems: find RecipientInfo structure that matches recipient certificate
9312 and initialise the ASN1 structures properly based on passed cipher.
9315 *) Belatedly make the BN tests actually check the results.
9318 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
9319 to and from BNs: it was completely broken. New compilation option
9320 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
9321 key elements as negative integers.
9324 *) Reorganize and speed up MD5.
9325 [Andy Polyakov <appro@fy.chalmers.se>]
9328 [Richard Levitte <richard@levitte.org>]
9330 *) New option -out to asn1parse to allow the parsed structure to be
9331 output to a file. This is most useful when combined with the -strparse
9332 option to examine the output of things like OCTET STRINGS.
9335 *) Make SSL library a little more fool-proof by not requiring any longer
9336 that SSL_set_{accept,connect}_state be called before
9337 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
9338 in many applications because usually everything *appeared* to work as
9339 intended anyway -- now it really works as intended).
9342 *) Move openssl.cnf out of lib/.
9345 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
9346 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
9347 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
9348 [Ralf S. Engelschall]
9350 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
9351 handle PKCS#7 enveloped data properly.
9352 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
9354 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
9355 copying pointers. The cert_st handling is changed by this in
9356 various ways (and thus what used to be known as ctx->default_cert
9357 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
9358 any longer when s->cert does not give us what we need).
9359 ssl_cert_instantiate becomes obsolete by this change.
9360 As soon as we've got the new code right (possibly it already is?),
9361 we have solved a couple of bugs of the earlier code where s->cert
9362 was used as if it could not have been shared with other SSL structures.
9364 Note that using the SSL API in certain dirty ways now will result
9365 in different behaviour than observed with earlier library versions:
9366 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
9367 does not influence s as it used to.
9369 In order to clean up things more thoroughly, inside SSL_SESSION
9370 we don't use CERT any longer, but a new structure SESS_CERT
9371 that holds per-session data (if available); currently, this is
9372 the peer's certificate chain and, for clients, the server's certificate
9373 and temporary key. CERT holds only those values that can have
9374 meaningful defaults in an SSL_CTX.
9377 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
9378 from the internal representation. Various PKCS#7 fixes: remove some
9379 evil casts and set the enc_dig_alg field properly based on the signing
9383 *) Allow PKCS#12 password to be set from the command line or the
9384 environment. Let 'ca' get its config file name from the environment
9385 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
9389 *) Allow certificate policies extension to use an IA5STRING for the
9390 organization field. This is contrary to the PKIX definition but
9391 VeriSign uses it and IE5 only recognises this form. Document 'x509'
9395 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
9396 without disallowing inline assembler and the like for non-pedantic builds.
9399 *) Support Borland C++ builder.
9400 [Janez Jere <jj@void.si>, modified by Ulf Möller]
9405 *) SHA-1 cleanups and performance enhancements.
9406 [Andy Polyakov <appro@fy.chalmers.se>]
9408 *) Sparc v8plus assembler for the bignum library.
9409 [Andy Polyakov <appro@fy.chalmers.se>]
9411 *) Accept any -xxx and +xxx compiler options in Configure.
9414 *) Update HPUX configuration.
9417 *) Add missing sk_<type>_unshift() function to safestack.h
9418 [Ralf S. Engelschall]
9420 *) New function SSL_CTX_use_certificate_chain_file that sets the
9421 "extra_cert"s in addition to the certificate. (This makes sense
9422 only for "PEM" format files, as chains as a whole are not
9426 *) Support verify_depth from the SSL API.
9427 x509_vfy.c had what can be considered an off-by-one-error:
9428 Its depth (which was not part of the external interface)
9429 was actually counting the number of certificates in a chain;
9430 now it really counts the depth.
9433 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
9434 instead of X509err, which often resulted in confusing error
9435 messages since the error codes are not globally unique
9436 (e.g. an alleged error in ssl3_accept when a certificate
9437 didn't match the private key).
9439 *) New function SSL_CTX_set_session_id_context that allows to set a default
9440 value (so that you don't need SSL_set_session_id_context for each
9441 connection using the SSL_CTX).
9444 *) OAEP decoding bug fix.
9447 *) Support INSTALL_PREFIX for package builders, as proposed by
9451 *) New Configure options "threads" and "no-threads". For systems
9452 where the proper compiler options are known (currently Solaris
9453 and Linux), "threads" is the default.
9456 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
9459 *) Install various scripts to $(OPENSSLDIR)/misc, not to
9460 $(INSTALLTOP)/bin -- they shouldn't clutter directories
9461 such as /usr/local/bin.
9464 *) "make linux-shared" to build shared libraries.
9465 [Niels Poppe <niels@netbox.org>]
9467 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
9470 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
9471 extension adding in x509 utility.
9474 *) Remove NOPROTO sections and error code comments.
9477 *) Partial rewrite of the DEF file generator to now parse the ANSI
9481 *) New Configure options --prefix=DIR and --openssldir=DIR.
9484 *) Complete rewrite of the error code script(s). It is all now handled
9485 by one script at the top level which handles error code gathering,
9486 header rewriting and C source file generation. It should be much better
9487 than the old method: it now uses a modified version of Ulf's parser to
9488 read the ANSI prototypes in all header files (thus the old K&R definitions
9489 aren't needed for error creation any more) and do a better job of
9490 translating function codes into names. The old 'ASN1 error code imbedded
9491 in a comment' is no longer necessary and it doesn't use .err files which
9492 have now been deleted. Also the error code call doesn't have to appear all
9493 on one line (which resulted in some large lines...).
9496 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
9499 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
9500 0 (which usually indicates a closed connection), but continue reading.
9503 *) Fix some race conditions.
9506 *) Add support for CRL distribution points extension. Add Certificate
9507 Policies and CRL distribution points documentation.
9510 *) Move the autogenerated header file parts to crypto/opensslconf.h.
9513 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
9514 8 of keying material. Merlin has also confirmed interop with this fix
9515 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
9516 [Merlin Hughes <merlin@baltimore.ie>]
9518 *) Fix lots of warnings.
9519 [Richard Levitte <levitte@stacken.kth.se>]
9521 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
9522 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
9523 [Richard Levitte <levitte@stacken.kth.se>]
9525 *) Fix problems with sizeof(long) == 8.
9526 [Andy Polyakov <appro@fy.chalmers.se>]
9528 *) Change functions to ANSI C.
9531 *) Fix typos in error codes.
9532 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
9534 *) Remove defunct assembler files from Configure.
9537 *) SPARC v8 assembler BIGNUM implementation.
9538 [Andy Polyakov <appro@fy.chalmers.se>]
9540 *) Support for Certificate Policies extension: both print and set.
9541 Various additions to support the r2i method this uses.
9544 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
9545 return a const string when you are expecting an allocated buffer.
9548 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
9549 types DirectoryString and DisplayText.
9552 *) Add code to allow r2i extensions to access the configuration database,
9553 add an LHASH database driver and add several ctx helper functions.
9556 *) Fix an evil bug in bn_expand2() which caused various BN functions to
9557 fail when they extended the size of a BIGNUM.
9560 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
9561 support typesafe stack.
9564 *) Fix typo in SSL_[gs]et_options().
9565 [Nils Frostberg <nils@medcom.se>]
9567 *) Delete various functions and files that belonged to the (now obsolete)
9568 old X509V3 handling code.
9571 *) New Configure option "rsaref".
9574 *) Don't auto-generate pem.h.
9577 *) Introduce type-safe ASN.1 SETs.
9580 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
9581 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
9583 *) Introduce type-safe STACKs. This will almost certainly break lots of code
9584 that links with OpenSSL (well at least cause lots of warnings), but fear
9585 not: the conversion is trivial, and it eliminates loads of evil casts. A
9586 few STACKed things have been converted already. Feel free to convert more.
9587 In the fullness of time, I'll do away with the STACK type altogether.
9590 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
9591 specified in <certfile> by updating the entry in the index.txt file.
9592 This way one no longer has to edit the index.txt file manually for
9593 revoking a certificate. The -revoke option does the gory details now.
9594 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
9596 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
9597 `-text' option at all and this way the `-noout -text' combination was
9598 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
9599 [Ralf S. Engelschall]
9601 *) Make sure a corresponding plain text error message exists for the
9602 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
9603 verify callback function determined that a certificate was revoked.
9604 [Ralf S. Engelschall]
9606 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
9607 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
9608 all available cipers including rc5, which was forgotten until now.
9609 In order to let the testing shell script know which algorithms
9610 are available, a new (up to now undocumented) command
9611 "openssl list-cipher-commands" is used.
9614 *) Bugfix: s_client occasionally would sleep in select() when
9615 it should have checked SSL_pending() first.
9618 *) New functions DSA_do_sign and DSA_do_verify to provide access to
9619 the raw DSA values prior to ASN.1 encoding.
9622 *) Tweaks to Configure
9623 [Niels Poppe <niels@netbox.org>]
9625 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
9629 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
9632 *) New config option to avoid instructions that are illegal on the 80386.
9633 The default code is faster, but requires at least a 486.
9636 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
9637 SSL2_SERVER_VERSION (not used at all) macros, which are now the
9638 same as SSL2_VERSION anyway.
9641 *) New "-showcerts" option for s_client.
9644 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
9645 application. Various cleanups and fixes.
9648 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
9649 modify error routines to work internally. Add error codes and PBE init
9650 to library startup routines.
9653 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
9654 packing functions to asn1 and evp. Changed function names and error
9655 codes along the way.
9658 *) PKCS12 integration: and so it begins... First of several patches to
9659 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
9660 objects to objects.h
9663 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
9664 and display support for Thawte strong extranet extension.
9667 *) Add LinuxPPC support.
9668 [Jeff Dubrule <igor@pobox.org>]
9670 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
9671 bn_div_words in alpha.s.
9672 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
9674 *) Make sure the RSA OAEP test is skipped under -DRSAref because
9675 OAEP isn't supported when OpenSSL is built with RSAref.
9676 [Ulf Moeller <ulf@fitug.de>]
9678 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
9679 so they no longer are missing under -DNOPROTO.
9680 [Soren S. Jorvang <soren@t.dk>]
9683 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
9685 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
9686 doesn't work when the session is reused. Coming soon!
9689 *) Fix a security hole, that allows sessions to be reused in the wrong
9690 context thus bypassing client cert protection! All software that uses
9691 client certs and session caches in multiple contexts NEEDS PATCHING to
9692 allow session reuse! A fuller solution is in the works.
9693 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
9695 *) Some more source tree cleanups (removed obsolete files
9696 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
9697 permission on "config" script to be executable) and a fix for the INSTALL
9699 [Ulf Moeller <ulf@fitug.de>]
9701 *) Remove some legacy and erroneous uses of malloc, free instead of
9703 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
9705 *) Make rsa_oaep_test return non-zero on error.
9706 [Ulf Moeller <ulf@fitug.de>]
9708 *) Add support for native Solaris shared libraries. Configure
9709 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
9710 if someone would make that last step automatic.
9711 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
9713 *) ctx_size was not built with the right compiler during "make links". Fixed.
9716 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
9717 except NULL ciphers". This means the default cipher list will no longer
9718 enable NULL ciphers. They need to be specifically enabled e.g. with
9719 the string "DEFAULT:eNULL".
9722 *) Fix to RSA private encryption routines: if p < q then it would
9723 occasionally produce an invalid result. This will only happen with
9724 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9727 *) Be less restrictive and allow also `perl util/perlpath.pl
9728 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9729 because this way one can also use an interpreter named `perl5' (which is
9730 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9731 installed as `perl').
9732 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9734 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9735 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9737 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9738 advapi32.lib to Win32 build and change the pem test comparision
9739 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9740 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9741 and crypto/des/ede_cbcm_enc.c.
9744 *) DES quad checksum was broken on big-endian architectures. Fixed.
9747 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9748 Win32 test batch file so it (might) work again. The Win32 test batch file
9749 is horrible: I feel ill....
9752 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
9753 in e_os.h. Audit of header files to check ANSI and non ANSI
9754 sections: 10 functions were absent from non ANSI section and not exported
9755 from Windows DLLs. Fixed up libeay.num for new functions.
9758 *) Make `openssl version' output lines consistent.
9759 [Ralf S. Engelschall]
9761 *) Fix Win32 symbol export lists for BIO functions: Added
9762 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
9763 to ms/libeay{16,32}.def.
9764 [Ralf S. Engelschall]
9766 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
9767 fine under Unix and passes some trivial tests I've now added. But the
9768 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
9769 added to make sure no one expects that this stuff really works in the
9770 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
9771 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
9773 [Ralf S. Engelschall]
9775 *) Fix the generation of two part addresses in perl.
9776 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
9778 *) Add config entry for Linux on MIPS.
9779 [John Tobey <jtobey@channel1.com>]
9781 *) Make links whenever Configure is run, unless we are on Windoze.
9784 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
9785 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
9789 *) Add a useful kludge to allow package maintainers to specify compiler and
9790 other platforms details on the command line without having to patch the
9791 Configure script everytime: One now can use ``perl Configure
9792 <id>:<details>'', i.e. platform ids are allowed to have details appended
9793 to them (seperated by colons). This is treated as there would be a static
9794 pre-configured entry in Configure's %table under key <id> with value
9795 <details> and ``perl Configure <id>'' is called. So, when you want to
9796 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
9797 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
9798 now, which overrides the FreeBSD-elf entry on-the-fly.
9799 [Ralf S. Engelschall]
9801 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
9804 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
9805 on the `perl Configure ...' command line. This way one can compile
9806 OpenSSL libraries with Position Independent Code (PIC) which is needed
9807 for linking it into DSOs.
9808 [Ralf S. Engelschall]
9810 *) Remarkably, export ciphers were totally broken and no-one had noticed!
9814 *) Cleaned up the LICENSE document: The official contact for any license
9815 questions now is the OpenSSL core team under openssl-core@openssl.org.
9816 And add a paragraph about the dual-license situation to make sure people
9817 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
9818 to the OpenSSL toolkit.
9819 [Ralf S. Engelschall]
9821 *) General source tree makefile cleanups: Made `making xxx in yyy...'
9822 display consistent in the source tree and replaced `/bin/rm' by `rm'.
9823 Additonally cleaned up the `make links' target: Remove unnecessary
9824 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
9825 to speed processing and no longer clutter the display with confusing
9826 stuff. Instead only the actually done links are displayed.
9827 [Ralf S. Engelschall]
9829 *) Permit null encryption ciphersuites, used for authentication only. It used
9830 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
9831 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
9835 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
9836 signed attributes when verifying signatures (this would break them),
9837 the detached data encoding was wrong and public keys obtained using
9838 X509_get_pubkey() weren't freed.
9841 *) Add text documentation for the BUFFER functions. Also added a work around
9842 to a Win95 console bug. This was triggered by the password read stuff: the
9843 last character typed gets carried over to the next fread(). If you were
9844 generating a new cert request using 'req' for example then the last
9845 character of the passphrase would be CR which would then enter the first
9849 *) Added the new `Includes OpenSSL Cryptography Software' button as
9850 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
9851 button and can be used by applications based on OpenSSL to show the
9852 relationship to the OpenSSL project.
9853 [Ralf S. Engelschall]
9855 *) Remove confusing variables in function signatures in files
9856 ssl/ssl_lib.c and ssl/ssl.h.
9857 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9859 *) Don't install bss_file.c under PREFIX/include/
9860 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9862 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
9863 functions that return function pointers and has support for NT specific
9864 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
9865 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
9866 unsigned to signed types: this was killing the Win32 compile.
9869 *) Add new certificate file to stack functions,
9870 SSL_add_dir_cert_subjects_to_stack() and
9871 SSL_add_file_cert_subjects_to_stack(). These largely supplant
9872 SSL_load_client_CA_file(), and can be used to add multiple certs easily
9873 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
9874 This means that Apache-SSL and similar packages don't have to mess around
9875 to add as many CAs as they want to the preferred list.
9878 *) Experiment with doxygen documentation. Currently only partially applied to
9880 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
9881 openssl.doxy as the configuration file.
9884 *) Get rid of remaining C++-style comments which strict C compilers hate.
9885 [Ralf S. Engelschall, pointed out by Carlos Amengual]
9887 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
9888 compiled in by default: it has problems with large keys.
9891 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
9892 DH private keys and/or callback functions which directly correspond to
9893 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
9894 is needed for applications which have to configure certificates on a
9895 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
9897 For the RSA certificate situation is makes no difference, but
9898 for the DSA certificate situation this fixes the "no shared cipher"
9899 problem where the OpenSSL cipher selection procedure failed because the
9900 temporary keys were not overtaken from the context and the API provided
9901 no way to reconfigure them.
9902 The new functions now let applications reconfigure the stuff and they
9903 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
9904 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
9905 non-public-API function ssl_cert_instantiate() is used as a helper
9906 function and also to reduce code redundancy inside ssl_rsa.c.
9907 [Ralf S. Engelschall]
9909 *) Move s_server -dcert and -dkey options out of the undocumented feature
9910 area because they are useful for the DSA situation and should be
9911 recognized by the users.
9912 [Ralf S. Engelschall]
9914 *) Fix the cipher decision scheme for export ciphers: the export bits are
9915 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
9916 SSL_EXP_MASK. So, the original variable has to be used instead of the
9917 already masked variable.
9918 [Richard Levitte <levitte@stacken.kth.se>]
9920 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
9921 [Richard Levitte <levitte@stacken.kth.se>]
9923 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
9924 from `int' to `unsigned int' because it's a length and initialized by
9925 EVP_DigestFinal() which expects an `unsigned int *'.
9926 [Richard Levitte <levitte@stacken.kth.se>]
9928 *) Don't hard-code path to Perl interpreter on shebang line of Configure
9929 script. Instead use the usual Shell->Perl transition trick.
9930 [Ralf S. Engelschall]
9932 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
9933 (in addition to RSA certificates) to match the behaviour of `openssl dsa
9934 -noout -modulus' as it's already the case for `openssl rsa -noout
9935 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
9936 currently the public key is printed (a decision which was already done by
9937 `openssl dsa -modulus' in the past) which serves a similar purpose.
9938 Additionally the NO_RSA no longer completely removes the whole -modulus
9939 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
9941 [Ralf S. Engelschall]
9943 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
9944 BIO. See the source (crypto/evp/bio_ok.c) for more info.
9945 [Arne Ansper <arne@ats.cyber.ee>]
9947 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
9948 to be added. Now both 'req' and 'ca' can use new objects defined in the
9952 *) Add cool BIO that does syslog (or event log on NT).
9953 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
9955 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
9956 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
9957 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
9958 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
9961 *) Add preliminary config info for new extension code.
9964 *) Make RSA_NO_PADDING really use no padding.
9965 [Ulf Moeller <ulf@fitug.de>]
9967 *) Generate errors when private/public key check is done.
9970 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
9971 for some CRL extensions and new objects added.
9974 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
9975 key usage extension and fuller support for authority key id.
9978 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
9979 padding method for RSA, which is recommended for new applications in PKCS
9980 #1 v2.0 (RFC 2437, October 1998).
9981 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
9982 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
9983 against Bleichbacher's attack on RSA.
9984 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
9987 *) Updates to the new SSL compression code
9988 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9990 *) Fix so that the version number in the master secret, when passed
9991 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
9992 (because the server will not accept higher), that the version number
9993 is 0x03,0x01, not 0x03,0x00
9994 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9996 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
9997 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
9998 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
10001 *) Support for RAW extensions where an arbitrary extension can be
10002 created by including its DER encoding. See apps/openssl.cnf for
10006 *) Make sure latest Perl versions don't interpret some generated C array
10007 code as Perl array code in the crypto/err/err_genc.pl script.
10008 [Lars Weber <3weber@informatik.uni-hamburg.de>]
10010 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
10011 not many people have the assembler. Various Win32 compilation fixes and
10012 update to the INSTALL.W32 file with (hopefully) more accurate Win32
10013 build instructions.
10016 *) Modify configure script 'Configure' to automatically create crypto/date.h
10017 file under Win32 and also build pem.h from pem.org. New script
10018 util/mkfiles.pl to create the MINFO file on environments that can't do a
10019 'make files': perl util/mkfiles.pl >MINFO should work.
10022 *) Major rework of DES function declarations, in the pursuit of correctness
10023 and purity. As a result, many evil casts evaporated, and some weirdness,
10024 too. You may find this causes warnings in your code. Zapping your evil
10025 casts will probably fix them. Mostly.
10028 *) Fix for a typo in asn1.h. Bug fix to object creation script
10029 obj_dat.pl. It considered a zero in an object definition to mean
10030 "end of object": none of the objects in objects.h have any zeros
10031 so it wasn't spotted.
10032 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
10034 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
10035 Masking (CBCM). In the absence of test vectors, the best I have been able
10036 to do is check that the decrypt undoes the encrypt, so far. Send me test
10037 vectors if you have them.
10040 *) Correct calculation of key length for export ciphers (too much space was
10041 allocated for null ciphers). This has not been tested!
10044 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
10045 message is now correct (it understands "crypto" and "ssl" on its
10046 command line). There is also now an "update" option. This will update
10047 the util/ssleay.num and util/libeay.num files with any new functions.
10049 perl util/mkdef.pl crypto ssl update
10050 it will update them.
10053 *) Overhauled the Perl interface (perl/*):
10054 - ported BN stuff to OpenSSL's different BN library
10055 - made the perl/ source tree CVS-aware
10056 - renamed the package from SSLeay to OpenSSL (the files still contain
10057 their history because I've copied them in the repository)
10058 - removed obsolete files (the test scripts will be replaced
10059 by better Test::Harness variants in the future)
10060 [Ralf S. Engelschall]
10062 *) First cut for a very conservative source tree cleanup:
10063 1. merge various obsolete readme texts into doc/ssleay.txt
10064 where we collect the old documents and readme texts.
10065 2. remove the first part of files where I'm already sure that we no
10066 longer need them because of three reasons: either they are just temporary
10067 files which were left by Eric or they are preserved original files where
10068 I've verified that the diff is also available in the CVS via "cvs diff
10069 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
10070 the crypto/md/ stuff).
10071 [Ralf S. Engelschall]
10073 *) More extension code. Incomplete support for subject and issuer alt
10074 name, issuer and authority key id. Change the i2v function parameters
10075 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
10076 what that's for :-) Fix to ASN1 macro which messed up
10077 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
10080 *) Preliminary support for ENUMERATED type. This is largely copied from the
10084 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
10085 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10087 *) Make sure `make rehash' target really finds the `openssl' program.
10088 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10090 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
10091 like to hear about it if this slows down other processors.
10094 *) Add CygWin32 platform information to Configure script.
10095 [Alan Batie <batie@aahz.jf.intel.com>]
10097 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
10098 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
10100 *) New program nseq to manipulate netscape certificate sequences
10103 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
10107 *) Fixes to BN code. Previously the default was to define BN_RECURSION
10108 but the BN code had some problems that would cause failures when
10109 doing certificate verification and some other functions.
10110 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10112 *) Add ASN1 and PEM code to support netscape certificate sequences.
10115 *) Add ASN1 and PEM code to support netscape certificate sequences.
10118 *) Add several PKIX and private extended key usage OIDs.
10121 *) Modify the 'ca' program to handle the new extension code. Modify
10122 openssl.cnf for new extension format, add comments.
10125 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
10126 and add a sample to openssl.cnf so req -x509 now adds appropriate
10130 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
10131 error code, add initial support to X509_print() and x509 application.
10134 *) Takes a deep breath and start addding X509 V3 extension support code. Add
10135 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
10136 stuff is currently isolated and isn't even compiled yet.
10139 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
10140 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
10141 Removed the versions check from X509 routines when loading extensions:
10142 this allows certain broken certificates that don't set the version
10143 properly to be processed.
10146 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
10147 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
10148 can still be regenerated with "make depend".
10151 *) Spelling mistake in C version of CAST-128.
10152 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
10154 *) Changes to the error generation code. The perl script err-code.pl
10155 now reads in the old error codes and retains the old numbers, only
10156 adding new ones if necessary. It also only changes the .err files if new
10157 codes are added. The makefiles have been modified to only insert errors
10158 when needed (to avoid needlessly modifying header files). This is done
10159 by only inserting errors if the .err file is newer than the auto generated
10160 C file. To rebuild all the error codes from scratch (the old behaviour)
10161 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
10162 or delete all the .err files.
10165 *) CAST-128 was incorrectly implemented for short keys. The C version has
10166 been fixed, but is untested. The assembler versions are also fixed, but
10167 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
10168 to regenerate it if needed.
10169 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
10170 Hagino <itojun@kame.net>]
10172 *) File was opened incorrectly in randfile.c.
10173 [Ulf Möller <ulf@fitug.de>]
10175 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
10176 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
10177 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
10178 al: it's just almost always a UTCTime. Note this patch adds new error
10179 codes so do a "make errors" if there are problems.
10182 *) Correct Linux 1 recognition in config.
10183 [Ulf Möller <ulf@fitug.de>]
10185 *) Remove pointless MD5 hash when using DSA keys in ca.
10186 [Anonymous <nobody@replay.com>]
10188 *) Generate an error if given an empty string as a cert directory. Also
10189 generate an error if handed NULL (previously returned 0 to indicate an
10190 error, but didn't set one).
10191 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
10193 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
10196 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
10197 parameters. This was causing a warning which killed off the Win32 compile.
10200 *) Remove C++ style comments from crypto/bn/bn_local.h.
10201 [Neil Costigan <neil.costigan@celocom.com>]
10203 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
10204 based on a text string, looking up short and long names and finally
10205 "dot" format. The "dot" format stuff didn't work. Added new function
10206 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
10207 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
10208 OID is not part of the table.
10211 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
10212 X509_LOOKUP_by_alias().
10215 *) Sort openssl functions by name.
10218 *) Get the gendsa program working (hopefully) and add it to app list. Remove
10219 encryption from sample DSA keys (in case anyone is interested the password
10223 *) Make _all_ *_free functions accept a NULL pointer.
10224 [Frans Heymans <fheymans@isaserver.be>]
10226 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
10228 [Anonymous <nobody@replay.com>]
10230 *) s_server should send the CAfile as acceptable CAs, not its own cert.
10231 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10233 *) Don't blow it for numeric -newkey arguments to apps/req.
10234 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10236 *) Temp key "for export" tests were wrong in s3_srvr.c.
10237 [Anonymous <nobody@replay.com>]
10239 *) Add prototype for temp key callback functions
10240 SSL_CTX_set_tmp_{rsa,dh}_callback().
10243 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
10244 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
10247 *) X509_name_add_entry() freed the wrong thing after an error.
10248 [Arne Ansper <arne@ats.cyber.ee>]
10250 *) rsa_eay.c would attempt to free a NULL context.
10251 [Arne Ansper <arne@ats.cyber.ee>]
10253 *) BIO_s_socket() had a broken should_retry() on Windoze.
10254 [Arne Ansper <arne@ats.cyber.ee>]
10256 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
10257 [Arne Ansper <arne@ats.cyber.ee>]
10259 *) Make sure the already existing X509_STORE->depth variable is initialized
10260 in X509_STORE_new(), but document the fact that this variable is still
10261 unused in the certificate verification process.
10262 [Ralf S. Engelschall]
10264 *) Fix the various library and apps files to free up pkeys obtained from
10265 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
10268 *) Fix reference counting in X509_PUBKEY_get(). This makes
10269 demos/maurice/example2.c work, amongst others, probably.
10270 [Steve Henson and Ben Laurie]
10272 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
10273 `openssl' and second, the shortcut symlinks for the `openssl <command>'
10274 are no longer created. This way we have a single and consistent command
10275 line interface `openssl <command>', similar to `cvs <command>'.
10276 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
10278 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
10279 BIT STRING wrapper always have zero unused bits.
10282 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
10285 *) Make the top-level INSTALL documentation easier to understand.
10288 *) Makefiles updated to exit if an error occurs in a sub-directory
10289 make (including if user presses ^C) [Paul Sutton]
10291 *) Make Montgomery context stuff explicit in RSA data structure.
10294 *) Fix build order of pem and err to allow for generated pem.h.
10297 *) Fix renumbering bug in X509_NAME_delete_entry().
10300 *) Enhanced the err-ins.pl script so it makes the error library number
10301 global and can add a library name. This is needed for external ASN1 and
10302 other error libraries.
10305 *) Fixed sk_insert which never worked properly.
10308 *) Fix ASN1 macros so they can handle indefinite length construted
10309 EXPLICIT tags. Some non standard certificates use these: they can now
10313 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
10314 into a single doc/ssleay.txt bundle. This way the information is still
10315 preserved but no longer messes up this directory. Now it's new room for
10316 the new set of documenation files.
10317 [Ralf S. Engelschall]
10319 *) SETs were incorrectly DER encoded. This was a major pain, because they
10320 shared code with SEQUENCEs, which aren't coded the same. This means that
10321 almost everything to do with SETs or SEQUENCEs has either changed name or
10322 number of arguments.
10323 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
10325 *) Fix test data to work with the above.
10328 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
10329 was already fixed by Eric for 0.9.1 it seems.
10330 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
10332 *) Autodetect FreeBSD3.
10335 *) Fix various bugs in Configure. This affects the following platforms:
10339 unixware-2.0-pentium
10343 *) Eliminate generated files from CVS. Reorder tests to regenerate files
10344 before they are needed.
10347 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
10351 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
10353 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
10354 changed SSLeay to OpenSSL in version strings.
10355 [Ralf S. Engelschall]
10357 *) Some fixups to the top-level documents.
10360 *) Fixed the nasty bug where rsaref.h was not found under compile-time
10361 because the symlink to include/ was missing.
10362 [Ralf S. Engelschall]
10364 *) Incorporated the popular no-RSA/DSA-only patches
10365 which allow to compile a RSA-free SSLeay.
10366 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
10368 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
10369 when "ssleay" is still not found.
10370 [Ralf S. Engelschall]
10372 *) Added more platforms to Configure: Cray T3E, HPUX 11,
10373 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
10375 *) Updated the README file.
10376 [Ralf S. Engelschall]
10378 *) Added various .cvsignore files in the CVS repository subdirs
10379 to make a "cvs update" really silent.
10380 [Ralf S. Engelschall]
10382 *) Recompiled the error-definition header files and added
10383 missing symbols to the Win32 linker tables.
10384 [Ralf S. Engelschall]
10386 *) Cleaned up the top-level documents;
10387 o new files: CHANGES and LICENSE
10388 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
10389 o merged COPYRIGHT into LICENSE
10390 o removed obsolete TODO file
10391 o renamed MICROSOFT to INSTALL.W32
10392 [Ralf S. Engelschall]
10394 *) Removed dummy files from the 0.9.1b source tree:
10395 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
10396 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
10397 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
10398 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
10399 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
10400 [Ralf S. Engelschall]
10402 *) Added various platform portability fixes.
10405 *) The Genesis of the OpenSSL rpject:
10406 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
10407 Young and Tim J. Hudson created while they were working for C2Net until
10409 [The OpenSSL Project]
10412 Changes between 0.9.0b and 0.9.1b [not released]
10414 *) Updated a few CA certificates under certs/
10417 *) Changed some BIGNUM api stuff.
10420 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
10421 DGUX x86, Linux Alpha, etc.
10424 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
10425 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
10429 *) Add -strparse option to asn1pars program which parses nested
10431 [Dr Stephen Henson <shenson@bigfoot.com>]
10433 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
10436 *) DSA fix for "ca" program.
10439 *) Added "-genkey" option to "dsaparam" program.
10442 *) Added RIPE MD160 (rmd160) message digest.
10445 *) Added -a (all) option to "ssleay version" command.
10448 *) Added PLATFORM define which is the id given to Configure.
10451 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
10454 *) Extended the ASN.1 parser routines.
10457 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
10460 *) Added a BN_CTX to the BN library.
10463 *) Fixed the weak key values in DES library
10466 *) Changed API in EVP library for cipher aliases.
10469 *) Added support for RC2/64bit cipher.
10472 *) Converted the lhash library to the crypto/mem.c functions.
10475 *) Added more recognized ASN.1 object ids.
10478 *) Added more RSA padding checks for SSL/TLS.
10481 *) Added BIO proxy/filter functionality.
10484 *) Added extra_certs to SSL_CTX which can be used
10485 send extra CA certificates to the client in the CA cert chain sending
10486 process. It can be configured with SSL_CTX_add_extra_chain_cert().
10489 *) Now Fortezza is denied in the authentication phase because
10490 this is key exchange mechanism is not supported by SSLeay at all.
10493 *) Additional PKCS1 checks.
10496 *) Support the string "TLSv1" for all TLS v1 ciphers.
10499 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
10500 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
10503 *) Fixed a few memory leaks.
10506 *) Fixed various code and comment typos.
10509 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
10510 bytes sent in the client random.
10511 [Edward Bishop <ebishop@spyglass.com>]
projects / openssl.git / blob