5 Changes between 1.0.1l and 1.0.1m [xx XXX xxxx]
7 *) Removed the export ciphers from the DEFAULT ciphers
10 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
12 *) Build fixes for the Windows and OpenVMS platforms
13 [Matt Caswell and Richard Levitte]
15 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
17 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
18 message can cause a segmentation fault in OpenSSL due to a NULL pointer
19 dereference. This could lead to a Denial Of Service attack. Thanks to
20 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
24 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
25 dtls1_buffer_record function under certain conditions. In particular this
26 could occur if an attacker sent repeated DTLS records with the same
27 sequence number but for the next epoch. The memory leak could be exploited
28 by an attacker in a Denial of Service attack through memory exhaustion.
29 Thanks to Chris Mueller for reporting this issue.
33 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
34 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
35 method would be set to NULL which could later result in a NULL pointer
36 dereference. Thanks to Frank Schmirler for reporting this issue.
40 *) Abort handshake if server key exchange message is omitted for ephemeral
43 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
48 *) Remove non-export ephemeral RSA code on client and server. This code
49 violated the TLS standard by allowing the use of temporary RSA keys in
50 non-export ciphersuites and could be used by a server to effectively
51 downgrade the RSA key length used to a value smaller than the server
52 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
53 INRIA or reporting this issue.
57 *) Fixed issue where DH client certificates are accepted without verification.
58 An OpenSSL server will accept a DH certificate for client authentication
59 without the certificate verify message. This effectively allows a client to
60 authenticate without the use of a private key. This only affects servers
61 which trust a client certificate authority which issues certificates
62 containing DH keys: these are extremely rare and hardly ever encountered.
63 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
68 *) Ensure that the session ID context of an SSL is updated when its
69 SSL_CTX is updated via SSL_set_SSL_CTX.
71 The session ID context is typically set from the parent SSL_CTX,
72 and can vary with the CTX.
75 *) Fix various certificate fingerprint issues.
77 By using non-DER or invalid encodings outside the signed portion of a
78 certificate the fingerprint can be changed without breaking the signature.
79 Although no details of the signed portion of the certificate can be changed
80 this can cause problems with some applications: e.g. those using the
81 certificate fingerprint for blacklists.
83 1. Reject signatures with non zero unused bits.
85 If the BIT STRING containing the signature has non zero unused bits reject
86 the signature. All current signature algorithms require zero unused bits.
88 2. Check certificate algorithm consistency.
90 Check the AlgorithmIdentifier inside TBS matches the one in the
91 certificate signature. NB: this will result in signature failure
92 errors for some broken certificates.
94 Thanks to Konrad Kraszewski from Google for reporting this issue.
96 3. Check DSA/ECDSA signatures use DER.
98 Reencode DSA/ECDSA signatures and compare with the original received
99 signature. Return an error if there is a mismatch.
101 This will reject various cases including garbage after signature
102 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
103 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
104 (negative or with leading zeroes).
106 Further analysis was conducted and fixes were developed by Stephen Henson
107 of the OpenSSL core team.
112 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
113 results on some platforms, including x86_64. This bug occurs at random
114 with a very low probability, and is not known to be exploitable in any
115 way, though its exact impact is difficult to determine. Thanks to Pieter
116 Wuille (Blockstream) who reported this issue and also suggested an initial
117 fix. Further analysis was conducted by the OpenSSL development team and
118 Adam Langley of Google. The final fix was developed by Andy Polyakov of
119 the OpenSSL core team.
123 *) Do not resume sessions on the server if the negotiated protocol
124 version does not match the session's version. Resuming with a different
125 version, while not strictly forbidden by the RFC, is of questionable
126 sanity and breaks all known clients.
127 [David Benjamin, Emilia Käsper]
129 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
130 early CCS messages during renegotiation. (Note that because
131 renegotiation is encrypted, this early CCS was not exploitable.)
134 *) Tighten client-side session ticket handling during renegotiation:
135 ensure that the client only accepts a session ticket if the server sends
136 the extension anew in the ServerHello. Previously, a TLS client would
137 reuse the old extension state and thus accept a session ticket if one was
138 announced in the initial ServerHello.
140 Similarly, ensure that the client requires a session ticket if one
141 was advertised in the ServerHello. Previously, a TLS client would
142 ignore a missing NewSessionTicket message.
145 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
149 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
150 sends a carefully crafted handshake message, to cause OpenSSL to fail
151 to free up to 64k of memory causing a memory leak. This could be
152 exploited in a Denial Of Service attack. This issue affects OpenSSL
153 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
154 whether SRTP is used or configured. Implementations of OpenSSL that
155 have been compiled with OPENSSL_NO_SRTP defined are not affected.
157 The fix was developed by the OpenSSL team.
161 *) Session Ticket Memory Leak.
163 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
164 integrity of that ticket is first verified. In the event of a session
165 ticket integrity check failing, OpenSSL will fail to free memory
166 causing a memory leak. By sending a large number of invalid session
167 tickets an attacker could exploit this issue in a Denial Of Service
172 *) Build option no-ssl3 is incomplete.
174 When OpenSSL is configured with "no-ssl3" as a build option, servers
175 could accept and complete a SSL 3.0 handshake, and clients could be
176 configured to send them.
178 [Akamai and the OpenSSL team]
180 *) Add support for TLS_FALLBACK_SCSV.
181 Client applications doing fallback retries should call
182 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
184 [Adam Langley, Bodo Moeller]
186 *) Add additional DigestInfo checks.
188 Reencode DigestInto in DER and check against the original when
189 verifying RSA signature: this will reject any improperly encoded
190 DigestInfo structures.
192 Note: this is a precautionary measure and no attacks are currently known.
196 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
198 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
199 SRP code can be overrun an internal buffer. Add sanity check that
200 g, A, B < N to SRP code.
202 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
203 Group for discovering this issue.
207 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
208 TLS 1.0 instead of higher protocol versions when the ClientHello message
209 is badly fragmented. This allows a man-in-the-middle attacker to force a
210 downgrade to TLS 1.0 even if both the server and the client support a
211 higher protocol version, by modifying the client's TLS records.
213 Thanks to David Benjamin and Adam Langley (Google) for discovering and
214 researching this issue.
218 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
219 to a denial of service attack. A malicious server can crash the client
220 with a null pointer dereference (read) by specifying an anonymous (EC)DH
221 ciphersuite and sending carefully crafted handshake messages.
223 Thanks to Felix Gröbert (Google) for discovering and researching this
228 *) By sending carefully crafted DTLS packets an attacker could cause openssl
229 to leak memory. This can be exploited through a Denial of Service attack.
230 Thanks to Adam Langley for discovering and researching this issue.
234 *) An attacker can force openssl to consume large amounts of memory whilst
235 processing DTLS handshake messages. This can be exploited through a
236 Denial of Service attack.
237 Thanks to Adam Langley for discovering and researching this issue.
241 *) An attacker can force an error condition which causes openssl to crash
242 whilst processing DTLS packets due to memory being freed twice. This
243 can be exploited through a Denial of Service attack.
244 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
249 *) If a multithreaded client connects to a malicious server using a resumed
250 session and the server sends an ec point format extension it could write
251 up to 255 bytes to freed memory.
253 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
258 *) A malicious server can crash an OpenSSL client with a null pointer
259 dereference (read) by specifying an SRP ciphersuite even though it was not
260 properly negotiated with the client. This can be exploited through a
261 Denial of Service attack.
263 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
264 discovering and researching this issue.
268 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
269 X509_name_oneline, X509_name_print_ex et al. to leak some information
270 from the stack. Applications may be affected if they echo pretty printing
271 output to the attacker.
273 Thanks to Ivan Fratric (Google) for discovering this issue.
275 [Emilia Käsper, and Steve Henson]
277 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
278 for corner cases. (Certain input points at infinity could lead to
279 bogus results, with non-infinity inputs mapped to infinity too.)
282 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
284 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
285 handshake can force the use of weak keying material in OpenSSL
286 SSL/TLS clients and servers.
288 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
289 researching this issue. (CVE-2014-0224)
290 [KIKUCHI Masashi, Steve Henson]
292 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
293 OpenSSL DTLS client the code can be made to recurse eventually crashing
296 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
298 [Imre Rad, Steve Henson]
300 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
301 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
302 client or server. This is potentially exploitable to run arbitrary
303 code on a vulnerable client or server.
305 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
306 [Jüri Aedla, Steve Henson]
308 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
309 are subject to a denial of service attack.
311 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
312 this issue. (CVE-2014-3470)
313 [Felix Gröbert, Ivan Fratric, Steve Henson]
315 *) Harmonize version and its documentation. -f flag is used to display
317 [mancha <mancha1@zoho.com>]
319 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
321 [mancha <mancha1@zoho.com>]
323 *) Fix some double frees. These are not thought to be exploitable.
324 [mancha <mancha1@zoho.com>]
326 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
328 *) A missing bounds check in the handling of the TLS heartbeat extension
329 can be used to reveal up to 64k of memory to a connected client or
332 Thanks for Neel Mehta of Google Security for discovering this bug and to
333 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
334 preparing the fix (CVE-2014-0160)
335 [Adam Langley, Bodo Moeller]
337 *) Fix for the attack described in the paper "Recovering OpenSSL
338 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
339 by Yuval Yarom and Naomi Benger. Details can be obtained from:
340 http://eprint.iacr.org/2014/140
342 Thanks to Yuval Yarom and Naomi Benger for discovering this
343 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
344 [Yuval Yarom and Naomi Benger]
346 *) TLS pad extension: draft-agl-tls-padding-03
348 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
349 TLS client Hello record length value would otherwise be > 255 and
350 less that 512 pad with a dummy extension containing zeroes so it
351 is at least 512 bytes long.
353 [Adam Langley, Steve Henson]
355 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
357 *) Fix for TLS record tampering bug. A carefully crafted invalid
358 handshake could crash OpenSSL with a NULL pointer exception.
359 Thanks to Anton Johansson for reporting this issues.
362 *) Keep original DTLS digest and encryption contexts in retransmission
363 structures so we can use the previous session parameters if they need
364 to be resent. (CVE-2013-6450)
367 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
368 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
369 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
370 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
371 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
372 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
373 [Rob Stradling, Adam Langley]
375 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
377 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
378 supporting platforms or when small records were transferred.
379 [Andy Polyakov, Steve Henson]
381 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
383 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
385 This addresses the flaw in CBC record processing discovered by
386 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
387 at: http://www.isg.rhul.ac.uk/tls/
389 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
390 Security Group at Royal Holloway, University of London
391 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
392 Emilia Käsper for the initial patch.
394 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
396 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
397 ciphersuites which can be exploited in a denial of service attack.
398 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
399 and detecting this bug and to Wolfgang Ettlinger
400 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
404 *) Return an error when checking OCSP signatures when key is NULL.
405 This fixes a DoS attack. (CVE-2013-0166)
408 *) Make openssl verify return errors.
409 [Chris Palmer <palmer@google.com> and Ben Laurie]
411 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
412 the right response is stapled. Also change SSL_get_certificate()
413 so it returns the certificate actually sent.
414 See http://rt.openssl.org/Ticket/Display.html?id=2836.
415 [Rob Stradling <rob.stradling@comodo.com>]
417 *) Fix possible deadlock when decoding public keys.
420 *) Don't use TLS 1.0 record version number in initial client hello
424 Changes between 1.0.1b and 1.0.1c [10 May 2012]
426 *) Sanity check record length before skipping explicit IV in TLS
427 1.2, 1.1 and DTLS to fix DoS attack.
429 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
430 fuzzing as a service testing platform.
434 *) Initialise tkeylen properly when encrypting CMS messages.
435 Thanks to Solar Designer of Openwall for reporting this issue.
438 *) In FIPS mode don't try to use composite ciphers as they are not
442 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
444 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
445 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
446 mean any application compiled against OpenSSL 1.0.0 headers setting
447 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
448 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
449 0x10000000L Any application which was previously compiled against
450 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
451 will need to be recompiled as a result. Letting be results in
452 inability to disable specifically TLS 1.1 and in client context,
453 in unlike event, limit maximum offered version to TLS 1.0 [see below].
456 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
457 disable just protocol X, but all protocols above X *if* there are
458 protocols *below* X still enabled. In more practical terms it means
459 that if application wants to disable TLS1.0 in favor of TLS1.1 and
460 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
461 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
465 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
467 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
468 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
469 in CRYPTO_realloc_clean.
471 Thanks to Tavis Ormandy, Google Security Team, for discovering this
472 issue and to Adam Langley <agl@chromium.org> for fixing it.
474 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
476 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
479 *) Workarounds for some broken servers that "hang" if a client hello
480 record length exceeds 255 bytes.
482 1. Do not use record version number > TLS 1.0 in initial client
483 hello: some (but not all) hanging servers will now work.
484 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
485 the number of ciphers sent in the client hello. This should be
486 set to an even number, such as 50, for example by passing:
487 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
488 Most broken servers should now work.
489 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
490 TLS 1.2 client support entirely.
493 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
496 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
498 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
499 STRING form instead of a DigestInfo.
502 *) The format used for MDC2 RSA signatures is inconsistent between EVP
503 and the RSA_sign/RSA_verify functions. This was made more apparent when
504 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
505 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
506 the correct format in RSA_verify so both forms transparently work.
509 *) Some servers which support TLS 1.0 can choke if we initially indicate
510 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
511 encrypted premaster secret. As a workaround use the maximum pemitted
512 client version in client hello, this should keep such servers happy
513 and still work with previous versions of OpenSSL.
516 *) Add support for TLS/DTLS heartbeats.
517 [Robin Seggelmann <seggelmann@fh-muenster.de>]
519 *) Add support for SCTP.
520 [Robin Seggelmann <seggelmann@fh-muenster.de>]
522 *) Improved PRNG seeding for VOS.
523 [Paul Green <Paul.Green@stratus.com>]
525 *) Extensive assembler packs updates, most notably:
527 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
528 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
529 - x86_64: bit-sliced AES implementation;
530 - ARM: NEON support, contemporary platforms optimizations;
531 - s390x: z196 support;
532 - *: GHASH and GF(2^m) multiplication implementations;
536 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
537 (removal of unnecessary code)
538 [Peter Sylvester <peter.sylvester@edelweb.fr>]
540 *) Add TLS key material exporter from RFC 5705.
543 *) Add DTLS-SRTP negotiation from RFC 5764.
546 *) Add Next Protocol Negotiation,
547 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
548 disabled with a no-npn flag to config or Configure. Code donated
550 [Adam Langley <agl@google.com> and Ben Laurie]
552 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
553 NIST-P256, NIST-P521, with constant-time single point multiplication on
554 typical inputs. Compiler support for the nonstandard type __uint128_t is
555 required to use this (present in gcc 4.4 and later, for 64-bit builds).
556 Code made available under Apache License version 2.0.
558 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
559 line to include this in your build of OpenSSL, and run "make depend" (or
560 "make update"). This enables the following EC_METHODs:
562 EC_GFp_nistp224_method()
563 EC_GFp_nistp256_method()
564 EC_GFp_nistp521_method()
566 EC_GROUP_new_by_curve_name() will automatically use these (while
567 EC_GROUP_new_curve_GFp() currently prefers the more flexible
569 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
571 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
572 all platforms. Move ssize_t definition from e_os.h to the public
573 header file e_os2.h as it now appears in public header file cms.h
576 *) New -sigopt option to the ca, req and x509 utilities. Additional
577 signature parameters can be passed using this option and in
581 *) Add RSA PSS signing function. This will generate and set the
582 appropriate AlgorithmIdentifiers for PSS based on those in the
583 corresponding EVP_MD_CTX structure. No application support yet.
586 *) Support for companion algorithm specific ASN1 signing routines.
587 New function ASN1_item_sign_ctx() signs a pre-initialised
588 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
589 the appropriate parameters.
592 *) Add new algorithm specific ASN1 verification initialisation function
593 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
594 handling will be the same no matter what EVP_PKEY_METHOD is used.
595 Add a PSS handler to support verification of PSS signatures: checked
596 against a number of sample certificates.
599 *) Add signature printing for PSS. Add PSS OIDs.
600 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
602 *) Add algorithm specific signature printing. An individual ASN1 method
603 can now print out signatures instead of the standard hex dump.
605 More complex signatures (e.g. PSS) can print out more meaningful
606 information. Include DSA version that prints out the signature
610 *) Password based recipient info support for CMS library: implementing
614 *) Split password based encryption into PBES2 and PBKDF2 functions. This
615 neatly separates the code into cipher and PBE sections and is required
616 for some algorithms that split PBES2 into separate pieces (such as
620 *) Session-handling fixes:
621 - Fix handling of connections that are resuming with a session ID,
622 but also support Session Tickets.
623 - Fix a bug that suppressed issuing of a new ticket if the client
624 presented a ticket with an expired session.
625 - Try to set the ticket lifetime hint to something reasonable.
626 - Make tickets shorter by excluding irrelevant information.
627 - On the client side, don't ignore renewed tickets.
628 [Adam Langley, Bodo Moeller (Google)]
630 *) Fix PSK session representation.
633 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
635 This work was sponsored by Intel.
638 *) Add GCM support to TLS library. Some custom code is needed to split
639 the IV between the fixed (from PRF) and explicit (from TLS record)
640 portions. This adds all GCM ciphersuites supported by RFC5288 and
641 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
642 add a special AESGCM string for GCM only.
645 *) Expand range of ctrls for AES GCM. Permit setting invocation
646 field on decrypt and retrieval of invocation field only on encrypt.
649 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
650 As required by RFC5289 these ciphersuites cannot be used if for
651 versions of TLS earlier than 1.2.
654 *) For FIPS capable OpenSSL interpret a NULL default public key method
655 as unset and return the appopriate default but do *not* set the default.
656 This means we can return the appopriate method in applications that
657 swicth between FIPS and non-FIPS modes.
660 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
661 ENGINE is used then we cannot handle that in the FIPS module so we
662 keep original code iff non-FIPS operations are allowed.
665 *) Add -attime option to openssl utilities.
666 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
668 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
671 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
672 FIPS EC methods unconditionally for now.
675 *) New build option no-ec2m to disable characteristic 2 code.
678 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
679 all cases can be covered as some introduce binary incompatibilities.
682 *) Redirect RSA operations to FIPS module including keygen,
683 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
686 *) Add similar low level API blocking to ciphers.
689 *) Low level digest APIs are not approved in FIPS mode: any attempt
690 to use these will cause a fatal error. Applications that *really* want
691 to use them can use the private_* version instead.
694 *) Redirect cipher operations to FIPS module for FIPS builds.
697 *) Redirect digest operations to FIPS module for FIPS builds.
700 *) Update build system to add "fips" flag which will link in fipscanister.o
701 for static and shared library builds embedding a signature if needed.
704 *) Output TLS supported curves in preference order instead of numerical
705 order. This is currently hardcoded for the highest order curves first.
706 This should be configurable so applications can judge speed vs strength.
709 *) Add TLS v1.2 server support for client authentication.
712 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
716 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
717 FIPS modules versions.
720 *) Add TLS v1.2 client side support for client authentication. Keep cache
721 of handshake records longer as we don't know the hash algorithm to use
722 until after the certificate request message is received.
725 *) Initial TLS v1.2 client support. Add a default signature algorithms
726 extension including all the algorithms we support. Parse new signature
727 format in client key exchange. Relax some ECC signing restrictions for
728 TLS v1.2 as indicated in RFC5246.
731 *) Add server support for TLS v1.2 signature algorithms extension. Switch
732 to new signature format when needed using client digest preference.
733 All server ciphersuites should now work correctly in TLS v1.2. No client
734 support yet and no support for client certificates.
737 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
738 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
739 ciphersuites. At present only RSA key exchange ciphersuites work with
740 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
741 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
742 and version checking.
745 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
746 with this defined it will not be affected by any changes to ssl internal
747 structures. Add several utility functions to allow openssl application
748 to work with OPENSSL_NO_SSL_INTERN defined.
752 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
754 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
757 *) Permit abbreviated handshakes when renegotiating using the function
758 SSL_renegotiate_abbreviated().
759 [Robin Seggelmann <seggelmann@fh-muenster.de>]
761 *) Add call to ENGINE_register_all_complete() to
762 ENGINE_load_builtin_engines(), so some implementations get used
763 automatically instead of needing explicit application support.
766 *) Add support for TLS key exporter as described in RFC5705.
767 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
769 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
770 a few changes are required:
772 Add SSL_OP_NO_TLSv1_1 flag.
774 Update version checking logic to handle version 1.1.
775 Add explicit IV handling (ported from DTLS code).
776 Add command line options to s_client/s_server.
779 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
781 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
782 in CMS and PKCS7 code. When RSA decryption fails use a random key for
783 content decryption and always return the same error. Note: this attack
784 needs on average 2^20 messages so it only affects automated senders. The
785 old behaviour can be reenabled in the CMS code by setting the
786 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
787 an MMA defence is not necessary.
788 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
789 this issue. (CVE-2012-0884)
792 *) Fix CVE-2011-4619: make sure we really are receiving a
793 client hello before rejecting multiple SGC restarts. Thanks to
794 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
797 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
799 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
800 Thanks to Antonio Martin, Enterprise Secure Access Research and
801 Development, Cisco Systems, Inc. for discovering this bug and
802 preparing a fix. (CVE-2012-0050)
805 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
807 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
808 of the Vaudenay padding oracle attack on CBC mode encryption
809 which enables an efficient plaintext recovery attack against
810 the OpenSSL implementation of DTLS. Their attack exploits timing
811 differences arising during decryption processing. A research
812 paper describing this attack can be found at:
813 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
814 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
815 Security Group at Royal Holloway, University of London
816 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
817 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
818 for preparing the fix. (CVE-2011-4108)
819 [Robin Seggelmann, Michael Tuexen]
821 *) Clear bytes used for block padding of SSL 3.0 records.
823 [Adam Langley (Google)]
825 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
826 Kadianakis <desnacked@gmail.com> for discovering this issue and
827 Adam Langley for preparing the fix. (CVE-2011-4619)
828 [Adam Langley (Google)]
830 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
831 [Andrey Kulikov <amdeich@gmail.com>]
833 *) Prevent malformed RFC3779 data triggering an assertion failure.
834 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
835 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
836 [Rob Austein <sra@hactrn.net>]
838 *) Improved PRNG seeding for VOS.
839 [Paul Green <Paul.Green@stratus.com>]
841 *) Fix ssl_ciph.c set-up race.
842 [Adam Langley (Google)]
844 *) Fix spurious failures in ecdsatest.c.
845 [Emilia Käsper (Google)]
847 *) Fix the BIO_f_buffer() implementation (which was mixing different
848 interpretations of the '..._len' fields).
849 [Adam Langley (Google)]
851 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
852 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
853 threads won't reuse the same blinding coefficients.
855 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
856 lock to call BN_BLINDING_invert_ex, and avoids one use of
857 BN_BLINDING_update for each BN_BLINDING structure (previously,
858 the last update always remained unused).
859 [Emilia Käsper (Google)]
861 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
862 [Bob Buckholz (Google)]
864 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
866 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
867 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
868 [Kaspar Brand <ossl@velox.ch>]
870 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
871 for multi-threaded use of ECDH. (CVE-2011-3210)
872 [Adam Langley (Google)]
874 *) Fix x509_name_ex_d2i memory leak on bad inputs.
877 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
878 signature public key algorithm by using OID xref utilities instead.
879 Before this you could only use some ECC ciphersuites with SHA1 only.
882 *) Add protection against ECDSA timing attacks as mentioned in the paper
883 by Billy Bob Brumley and Nicola Tuveri, see:
885 http://eprint.iacr.org/2011/232.pdf
887 [Billy Bob Brumley and Nicola Tuveri]
889 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
891 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
892 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
894 *) Fix bug in string printing code: if *any* escaping is enabled we must
895 escape the escape character (backslash) or the resulting string is
899 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
901 *) Disable code workaround for ancient and obsolete Netscape browsers
902 and servers: an attacker can use it in a ciphersuite downgrade attack.
903 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
906 *) Fixed J-PAKE implementation error, originally discovered by
907 Sebastien Martini, further info and confirmation from Stefan
908 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
911 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
913 *) Fix extension code to avoid race conditions which can result in a buffer
914 overrun vulnerability: resumed sessions must not be modified as they can
915 be shared by multiple threads. CVE-2010-3864
918 *) Fix WIN32 build system to correctly link an ENGINE directory into
922 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
924 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
926 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
928 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
930 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
931 context. The operation can be customised via the ctrl mechanism in
932 case ENGINEs want to include additional functionality.
935 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
938 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
939 output hashes compatible with older versions of OpenSSL.
940 [Willy Weisz <weisz@vcpc.univie.ac.at>]
942 *) Fix compression algorithm handling: if resuming a session use the
943 compression algorithm of the resumed session instead of determining
944 it from client hello again. Don't allow server to change algorithm.
947 *) Add load_crls() function to apps tidying load_certs() too. Add option
948 to verify utility to allow additional CRLs to be included.
951 *) Update OCSP request code to permit adding custom headers to the request:
952 some responders need this.
955 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
957 [Julia Lawall <julia@diku.dk>]
959 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
960 needlessly dereferenced structures, used obsolete functions and
961 didn't handle all updated verify codes correctly.
964 *) Disable MD2 in the default configuration.
967 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
968 indicate the initial BIO being pushed or popped. This makes it possible
969 to determine whether the BIO is the one explicitly called or as a result
970 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
971 it handles reference counts correctly and doesn't zero out the I/O bio
972 when it is not being explicitly popped. WARNING: applications which
973 included workarounds for the old buggy behaviour will need to be modified
974 or they could free up already freed BIOs.
977 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
978 renaming to all platforms (within the 0.9.8 branch, this was
979 done conditionally on Netware platforms to avoid a name clash).
980 [Guenter <lists@gknw.net>]
982 *) Add ECDHE and PSK support to DTLS.
983 [Michael Tuexen <tuexen@fh-muenster.de>]
985 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
989 *) Add "missing" function EVP_MD_flags() (without this the only way to
990 retrieve a digest flags is by accessing the structure directly. Update
991 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
992 or cipher is registered as in the "from" argument. Print out all
993 registered digests in the dgst usage message instead of manually
994 attempting to work them out.
997 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
998 this allows the use of compression and extensions. Change default cipher
999 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
1000 by default unless an application cipher string requests it.
1003 *) Alter match criteria in PKCS12_parse(). It used to try to use local
1004 key ids to find matching certificates and keys but some PKCS#12 files
1005 don't follow the (somewhat unwritten) rules and this strategy fails.
1006 Now just gather all certificates together and the first private key
1007 then look for the first certificate that matches the key.
1010 *) Support use of registered digest and cipher names for dgst and cipher
1011 commands instead of having to add each one as a special case. So now
1018 openssl dgst -sha256 foo
1020 and this works for ENGINE based algorithms too.
1024 *) Update Gost ENGINE to support parameter files.
1025 [Victor B. Wagner <vitus@cryptocom.ru>]
1027 *) Support GeneralizedTime in ca utility.
1028 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
1030 *) Enhance the hash format used for certificate directory links. The new
1031 form uses the canonical encoding (meaning equivalent names will work
1032 even if they aren't identical) and uses SHA1 instead of MD5. This form
1033 is incompatible with the older format and as a result c_rehash should
1034 be used to rebuild symbolic links.
1037 *) Make PKCS#8 the default write format for private keys, replacing the
1038 traditional format. This form is standardised, more secure and doesn't
1039 include an implicit MD5 dependency.
1042 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
1043 committed to OpenSSL should pass this lot as a minimum.
1046 *) Add session ticket override functionality for use by EAP-FAST.
1047 [Jouni Malinen <j@w1.fi>]
1049 *) Modify HMAC functions to return a value. Since these can be implemented
1050 in an ENGINE errors can occur.
1053 *) Type-checked OBJ_bsearch_ex.
1056 *) Type-checked OBJ_bsearch. Also some constification necessitated
1057 by type-checking. Still to come: TXT_DB, bsearch(?),
1058 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
1062 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
1063 seconds to a tm structure directly, instead of going through OS
1064 specific date routines. This avoids any issues with OS routines such
1065 as the year 2038 bug. New *_adj() functions for ASN1 time structures
1066 and X509_time_adj_ex() to cover the extended range. The existing
1067 X509_time_adj() is still usable and will no longer have any date issues.
1070 *) Delta CRL support. New use deltas option which will attempt to locate
1071 and search any appropriate delta CRLs available.
1073 This work was sponsored by Google.
1076 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
1077 code and add additional score elements. Validate alternate CRL paths
1078 as part of the CRL checking and indicate a new error "CRL path validation
1079 error" in this case. Applications wanting additional details can use
1080 the verify callback and check the new "parent" field. If this is not
1081 NULL CRL path validation is taking place. Existing applications wont
1082 see this because it requires extended CRL support which is off by
1085 This work was sponsored by Google.
1088 *) Support for freshest CRL extension.
1090 This work was sponsored by Google.
1093 *) Initial indirect CRL support. Currently only supported in the CRLs
1094 passed directly and not via lookup. Process certificate issuer
1095 CRL entry extension and lookup CRL entries by bother issuer name
1096 and serial number. Check and process CRL issuer entry in IDP extension.
1098 This work was sponsored by Google.
1101 *) Add support for distinct certificate and CRL paths. The CRL issuer
1102 certificate is validated separately in this case. Only enabled if
1103 an extended CRL support flag is set: this flag will enable additional
1104 CRL functionality in future.
1106 This work was sponsored by Google.
1109 *) Add support for policy mappings extension.
1111 This work was sponsored by Google.
1114 *) Fixes to pathlength constraint, self issued certificate handling,
1115 policy processing to align with RFC3280 and PKITS tests.
1117 This work was sponsored by Google.
1120 *) Support for name constraints certificate extension. DN, email, DNS
1121 and URI types are currently supported.
1123 This work was sponsored by Google.
1126 *) To cater for systems that provide a pointer-based thread ID rather
1127 than numeric, deprecate the current numeric thread ID mechanism and
1128 replace it with a structure and associated callback type. This
1129 mechanism allows a numeric "hash" to be extracted from a thread ID in
1130 either case, and on platforms where pointers are larger than 'long',
1131 mixing is done to help ensure the numeric 'hash' is usable even if it
1132 can't be guaranteed unique. The default mechanism is to use "&errno"
1133 as a pointer-based thread ID to distinguish between threads.
1135 Applications that want to provide their own thread IDs should now use
1136 CRYPTO_THREADID_set_callback() to register a callback that will call
1137 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
1139 Note that ERR_remove_state() is now deprecated, because it is tied
1140 to the assumption that thread IDs are numeric. ERR_remove_state(0)
1141 to free the current thread's error state should be replaced by
1142 ERR_remove_thread_state(NULL).
1144 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
1145 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
1146 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
1147 application was previously providing a numeric thread callback that
1148 was inappropriate for distinguishing threads, then uniqueness might
1149 have been obtained with &errno that happened immediately in the
1150 intermediate development versions of OpenSSL; this is no longer the
1151 case, the numeric thread callback will now override the automatic use
1153 [Geoff Thorpe, with help from Bodo Moeller]
1155 *) Initial support for different CRL issuing certificates. This covers a
1156 simple case where the self issued certificates in the chain exist and
1157 the real CRL issuer is higher in the existing chain.
1159 This work was sponsored by Google.
1162 *) Removed effectively defunct crypto/store from the build.
1165 *) Revamp of STACK to provide stronger type-checking. Still to come:
1166 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
1167 ASN1_STRING, CONF_VALUE.
1170 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
1171 RAM on SSL connections. This option can save about 34k per idle SSL.
1174 *) Revamp of LHASH to provide stronger type-checking. Still to come:
1175 STACK, TXT_DB, bsearch, qsort.
1178 *) Initial support for Cryptographic Message Syntax (aka CMS) based
1179 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
1180 support for data, signedData, compressedData, digestedData and
1181 encryptedData, envelopedData types included. Scripts to check against
1182 RFC4134 examples draft and interop and consistency checks of many
1183 content types and variants.
1186 *) Add options to enc utility to support use of zlib compression BIO.
1189 *) Extend mk1mf to support importing of options and assembly language
1190 files from Configure script, currently only included in VC-WIN32.
1191 The assembly language rules can now optionally generate the source
1192 files from the associated perl scripts.
1195 *) Implement remaining functionality needed to support GOST ciphersuites.
1196 Interop testing has been performed using CryptoPro implementations.
1197 [Victor B. Wagner <vitus@cryptocom.ru>]
1199 *) s390x assembler pack.
1202 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
1206 *) Implement Opaque PRF Input TLS extension as specified in
1207 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
1208 official specification yet and no extension type assignment by
1209 IANA exists, this extension (for now) will have to be explicitly
1210 enabled when building OpenSSL by providing the extension number
1211 to use. For example, specify an option
1213 -DTLSEXT_TYPE_opaque_prf_input=0x9527
1215 to the "config" or "Configure" script to enable the extension,
1216 assuming extension number 0x9527 (which is a completely arbitrary
1217 and unofficial assignment based on the MD5 hash of the Internet
1218 Draft). Note that by doing so, you potentially lose
1219 interoperability with other TLS implementations since these might
1220 be using the same extension number for other purposes.
1222 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
1223 opaque PRF input value to use in the handshake. This will create
1224 an interal copy of the length-'len' string at 'src', and will
1225 return non-zero for success.
1227 To get more control and flexibility, provide a callback function
1230 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
1231 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
1235 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
1238 Callback function 'cb' will be called in handshakes, and is
1239 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
1240 Argument 'arg' is for application purposes (the value as given to
1241 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
1242 be provided to the callback function). The callback function
1243 has to return non-zero to report success: usually 1 to use opaque
1244 PRF input just if possible, or 2 to enforce use of the opaque PRF
1245 input. In the latter case, the library will abort the handshake
1246 if opaque PRF input is not successfully negotiated.
1248 Arguments 'peerinput' and 'len' given to the callback function
1249 will always be NULL and 0 in the case of a client. A server will
1250 see the client's opaque PRF input through these variables if
1251 available (NULL and 0 otherwise). Note that if the server
1252 provides an opaque PRF input, the length must be the same as the
1253 length of the client's opaque PRF input.
1255 Note that the callback function will only be called when creating
1256 a new session (session resumption can resume whatever was
1257 previously negotiated), and will not be called in SSL 2.0
1258 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
1259 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
1260 for applications that need to enforce opaque PRF input.
1264 *) Update ssl code to support digests other than SHA1+MD5 for handshake
1267 [Victor B. Wagner <vitus@cryptocom.ru>]
1269 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1270 RFC4507bis. The encrypted ticket format is an encrypted encoded
1271 SSL_SESSION structure, that way new session features are automatically
1274 If a client application caches session in an SSL_SESSION structure
1275 support is transparent because tickets are now stored in the encoded
1278 The SSL_CTX structure automatically generates keys for ticket
1279 protection in servers so again support should be possible
1280 with no application modification.
1282 If a client or server wishes to disable RFC4507 support then the option
1283 SSL_OP_NO_TICKET can be set.
1285 Add a TLS extension debugging callback to allow the contents of any client
1286 or server extensions to be examined.
1288 This work was sponsored by Google.
1291 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
1292 OpenSSL should now compile cleanly on gcc 4.2
1293 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
1295 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
1296 support including streaming MAC support: this is required for GOST
1297 ciphersuite support.
1298 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
1300 *) Add option -stream to use PKCS#7 streaming in smime utility. New
1301 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
1302 to output in BER and PEM format.
1305 *) Experimental support for use of HMAC via EVP_PKEY interface. This
1306 allows HMAC to be handled via the EVP_DigestSign*() interface. The
1307 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
1308 ENGINE support for HMAC keys which are unextractable. New -mac and
1309 -macopt options to dgst utility.
1312 *) New option -sigopt to dgst utility. Update dgst to use
1313 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
1314 alternative signing paramaters such as X9.31 or PSS in the dgst
1318 *) Change ssl_cipher_apply_rule(), the internal function that does
1319 the work each time a ciphersuite string requests enabling
1320 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
1321 removing ("!foo+bar") a class of ciphersuites: Now it maintains
1322 the order of disabled ciphersuites such that those ciphersuites
1323 that most recently went from enabled to disabled not only stay
1324 in order with respect to each other, but also have higher priority
1325 than other disabled ciphersuites the next time ciphersuites are
1328 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
1329 the same ciphersuites as with "HIGH" alone, but in a specific
1330 order where the PSK ciphersuites come first (since they are the
1331 most recently disabled ciphersuites when "HIGH" is parsed).
1333 Also, change ssl_create_cipher_list() (using this new
1334 funcionality) such that between otherwise identical
1335 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
1339 *) Change ssl_create_cipher_list() so that it automatically
1340 arranges the ciphersuites in reasonable order before starting
1341 to process the rule string. Thus, the definition for "DEFAULT"
1342 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
1343 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
1344 This makes it much easier to arrive at a reasonable default order
1345 in applications for which anonymous ciphers are OK (meaning
1346 that you can't actually use DEFAULT).
1347 [Bodo Moeller; suggested by Victor Duchovni]
1349 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
1350 processing) into multiple integers instead of setting
1351 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
1352 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
1353 (These masks as well as the individual bit definitions are hidden
1354 away into the non-exported interface ssl/ssl_locl.h, so this
1355 change to the definition of the SSL_CIPHER structure shouldn't
1356 affect applications.) This give us more bits for each of these
1357 categories, so there is no longer a need to coagulate AES128 and
1358 AES256 into a single algorithm bit, and to coagulate Camellia128
1359 and Camellia256 into a single algorithm bit, which has led to all
1362 Thus, among other things, the kludge introduced in 0.9.7m and
1363 0.9.8e for masking out AES256 independently of AES128 or masking
1364 out Camellia256 independently of AES256 is not needed here in 0.9.9.
1366 With the change, we also introduce new ciphersuite aliases that
1367 so far were missing: "AES128", "AES256", "CAMELLIA128", and
1371 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
1372 Use the leftmost N bytes of the signature input if the input is
1373 larger than the prime q (with N being the size in bytes of q).
1376 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
1377 it yet and it is largely untested.
1380 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
1383 *) Initial incomplete changes to avoid need for function casts in OpenSSL
1384 some compilers (gcc 4.2 and later) reject their use. Safestack is
1385 reimplemented. Update ASN1 to avoid use of legacy functions.
1388 *) Win32/64 targets are linked with Winsock2.
1391 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
1392 to external functions. This can be used to increase CRL handling
1393 efficiency especially when CRLs are very large by (for example) storing
1394 the CRL revoked certificates in a database.
1397 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
1398 new CRLs added to a directory can be used. New command line option
1399 -verify_return_error to s_client and s_server. This causes real errors
1400 to be returned by the verify callback instead of carrying on no matter
1401 what. This reflects the way a "real world" verify callback would behave.
1404 *) GOST engine, supporting several GOST algorithms and public key formats.
1405 Kindly donated by Cryptocom.
1408 *) Partial support for Issuing Distribution Point CRL extension. CRLs
1409 partitioned by DP are handled but no indirect CRL or reason partitioning
1410 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
1411 selected via a scoring technique which handles IDP and AKID in CRLs.
1414 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
1415 will ultimately be used for all verify operations: this will remove the
1416 X509_STORE dependency on certificate verification and allow alternative
1417 lookup methods. X509_STORE based implementations of these two callbacks.
1420 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
1421 Modify get_crl() to find a valid (unexpired) CRL if possible.
1424 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
1425 this would be called X509_CRL_cmp() but that name is already used by
1426 a function that just compares CRL issuer names. Cache several CRL
1427 extensions in X509_CRL structure and cache CRLDP in X509.
1430 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
1431 this maps equivalent X509_NAME structures into a consistent structure.
1432 Name comparison can then be performed rapidly using memcmp().
1435 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
1439 *) Allow digests to supply their own micalg string for S/MIME type using
1440 the ctrl EVP_MD_CTRL_MICALG.
1443 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
1444 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
1445 ctrl. It can then customise the structure before and/or after signing
1449 *) New function OBJ_add_sigid() to allow application defined signature OIDs
1450 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
1451 to free up any added signature OIDs.
1454 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
1455 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
1456 digest and cipher tables. New options added to openssl utility:
1457 list-message-digest-algorithms and list-cipher-algorithms.
1460 *) Change the array representation of binary polynomials: the list
1461 of degrees of non-zero coefficients is now terminated with -1.
1462 Previously it was terminated with 0, which was also part of the
1463 value; thus, the array representation was not applicable to
1464 polynomials where t^0 has coefficient zero. This change makes
1465 the array representation useful in a more general context.
1468 *) Various modifications and fixes to SSL/TLS cipher string
1469 handling. For ECC, the code now distinguishes between fixed ECDH
1470 with RSA certificates on the one hand and with ECDSA certificates
1471 on the other hand, since these are separate ciphersuites. The
1472 unused code for Fortezza ciphersuites has been removed.
1474 For consistency with EDH, ephemeral ECDH is now called "EECDH"
1475 (not "ECDHE"). For consistency with the code for DH
1476 certificates, use of ECDH certificates is now considered ECDH
1477 authentication, not RSA or ECDSA authentication (the latter is
1478 merely the CA's signing algorithm and not actively used in the
1481 The temporary ciphersuite alias "ECCdraft" is no longer
1482 available, and ECC ciphersuites are no longer excluded from "ALL"
1483 and "DEFAULT". The following aliases now exist for RFC 4492
1484 ciphersuites, most of these by analogy with the DH case:
1486 kECDHr - ECDH cert, signed with RSA
1487 kECDHe - ECDH cert, signed with ECDSA
1488 kECDH - ECDH cert (signed with either RSA or ECDSA)
1489 kEECDH - ephemeral ECDH
1490 ECDH - ECDH cert or ephemeral ECDH
1496 AECDH - anonymous ECDH
1497 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
1501 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
1502 Use correct micalg parameters depending on digest(s) in signed message.
1505 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
1506 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
1509 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
1510 an engine to register a method. Add ENGINE lookups for methods and
1511 functional reference processing.
1514 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
1515 EVP_{Sign,Verify}* which allow an application to customise the signature
1519 *) New -resign option to smime utility. This adds one or more signers
1520 to an existing PKCS#7 signedData structure. Also -md option to use an
1521 alternative message digest algorithm for signing.
1524 *) Tidy up PKCS#7 routines and add new functions to make it easier to
1525 create PKCS7 structures containing multiple signers. Update smime
1526 application to support multiple signers.
1529 *) New -macalg option to pkcs12 utility to allow setting of an alternative
1533 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
1534 Reorganize PBE internals to lookup from a static table using NIDs,
1535 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
1536 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
1537 PRF which will be automatically used with PBES2.
1540 *) Replace the algorithm specific calls to generate keys in "req" with the
1544 *) Update PKCS#7 enveloped data routines to use new API. This is now
1545 supported by any public key method supporting the encrypt operation. A
1546 ctrl is added to allow the public key algorithm to examine or modify
1547 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
1551 *) Add a ctrl to asn1 method to allow a public key algorithm to express
1552 a default digest type to use. In most cases this will be SHA1 but some
1553 algorithms (such as GOST) need to specify an alternative digest. The
1554 return value indicates how strong the prefernce is 1 means optional and
1555 2 is mandatory (that is it is the only supported type). Modify
1556 ASN1_item_sign() to accept a NULL digest argument to indicate it should
1557 use the default md. Update openssl utilities to use the default digest
1558 type for signing if it is not explicitly indicated.
1561 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
1562 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
1563 signing method from the key type. This effectively removes the link
1564 between digests and public key types.
1567 *) Add an OID cross reference table and utility functions. Its purpose is to
1568 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
1569 rsaEncryption. This will allow some of the algorithm specific hackery
1570 needed to use the correct OID to be removed.
1573 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
1574 structures for PKCS7_sign(). They are now set up by the relevant public
1578 *) Add provisional EC pkey method with support for ECDSA and ECDH.
1581 *) Add support for key derivation (agreement) in the API, DH method and
1585 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
1586 public and private key formats. As a side effect these add additional
1587 command line functionality not previously available: DSA signatures can be
1588 generated and verified using pkeyutl and DH key support and generation in
1593 [Oliver Tappe <zooey@hirschkaefer.de>]
1595 *) New make target "install_html_docs" installs HTML renditions of the
1597 [Oliver Tappe <zooey@hirschkaefer.de>]
1599 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
1600 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
1601 support key and parameter generation and add initial key generation
1602 functionality for RSA.
1605 *) Add functions for main EVP_PKEY_method operations. The undocumented
1606 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
1607 EVP_PKEY_{encrypt,decrypt}_old.
1610 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
1611 key API, doesn't do much yet.
1614 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
1615 public key algorithms. New option to openssl utility:
1616 "list-public-key-algorithms" to print out info.
1619 *) Implement the Supported Elliptic Curves Extension for
1620 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1623 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
1624 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
1627 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
1628 utilities such as rsa, dsa, dsaparam etc except they process any key
1632 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1633 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1634 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
1638 *) Initial support for pluggable public key ASN1.
1639 De-spaghettify the public key ASN1 handling. Move public and private
1640 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
1641 algorithm specific handling to a single module within the relevant
1642 algorithm directory. Add functions to allow (near) opaque processing
1643 of public and private key structures.
1646 *) Implement the Supported Point Formats Extension for
1647 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1650 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
1651 for the psk identity [hint] and the psk callback functions to the
1652 SSL_SESSION, SSL and SSL_CTX structure.
1655 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
1659 SSL_CTX_use_psk_identity_hint
1660 SSL_get_psk_identity_hint
1661 SSL_get_psk_identity
1662 SSL_use_psk_identity_hint
1664 [Mika Kousa and Pasi Eronen of Nokia Corporation]
1666 *) Add RFC 3161 compliant time stamp request creation, response generation
1667 and response verification functionality.
1668 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
1670 *) Add initial support for TLS extensions, specifically for the server_name
1671 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1672 have new members for a host name. The SSL data structure has an
1673 additional member SSL_CTX *initial_ctx so that new sessions can be
1674 stored in that context to allow for session resumption, even after the
1675 SSL has been switched to a new SSL_CTX in reaction to a client's
1676 server_name extension.
1678 New functions (subject to change):
1680 SSL_get_servername()
1681 SSL_get_servername_type()
1684 New CTRL codes and macros (subject to change):
1686 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1687 - SSL_CTX_set_tlsext_servername_callback()
1688 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1689 - SSL_CTX_set_tlsext_servername_arg()
1690 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1692 openssl s_client has a new '-servername ...' option.
1694 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1695 '-key2 ...', '-servername_fatal' (subject to change). This allows
1696 testing the HostName extension for a specific single host name ('-cert'
1697 and '-key' remain fallbacks for handshakes without HostName
1698 negotiation). If the unrecogninzed_name alert has to be sent, this by
1699 default is a warning; it becomes fatal with the '-servername_fatal'
1702 [Peter Sylvester, Remy Allais, Christophe Renou]
1704 *) Whirlpool hash implementation is added.
1707 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1708 bn(64,32). Because of instruction set limitations it doesn't have
1709 any negative impact on performance. This was done mostly in order
1710 to make it possible to share assembler modules, such as bn_mul_mont
1711 implementations, between 32- and 64-bit builds without hassle.
1714 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1715 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1719 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1720 dedicated Montgomery multiplication procedure, is introduced.
1721 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1722 "64-bit" performance on certain 32-bit targets.
1725 *) New option SSL_OP_NO_COMP to disable use of compression selectively
1726 in SSL structures. New SSL ctrl to set maximum send fragment size.
1727 Save memory by seeting the I/O buffer sizes dynamically instead of
1728 using the maximum available value.
1731 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
1732 in addition to the text details.
1735 *) Very, very preliminary EXPERIMENTAL support for printing of general
1736 ASN1 structures. This currently produces rather ugly output and doesn't
1737 handle several customised structures at all.
1740 *) Integrated support for PVK file format and some related formats such
1741 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
1742 these in the 'rsa' and 'dsa' utilities.
1745 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
1748 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
1749 place for the (very old) "NETSCAPE" format certificates which are now
1750 handled using new ASN1 code equivalents.
1753 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
1754 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
1755 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
1758 *) Modify CRL distribution points extension code to print out previously
1759 unsupported fields. Enhance extension setting code to allow setting of
1763 *) Add print and set support for Issuing Distribution Point CRL extension.
1766 *) Change 'Configure' script to enable Camellia by default.
1769 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
1771 *) When rejecting SSL/TLS records due to an incorrect version number, never
1772 update s->server with a new major version number. As of
1773 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
1774 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
1775 the previous behavior could result in a read attempt at NULL when
1776 receiving specific incorrect SSL/TLS records once record payload
1777 protection is active. (CVE-2010-0740)
1778 [Bodo Moeller, Adam Langley <agl@chromium.org>]
1780 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
1781 could be crashed if the relevant tables were not present (e.g. chrooted).
1782 [Tomas Hoger <thoger@redhat.com>]
1784 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
1786 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
1787 [Martin Olsson, Neel Mehta]
1789 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
1790 accommodate for stack sorting, always a write lock!).
1793 *) On some versions of WIN32 Heap32Next is very slow. This can cause
1794 excessive delays in the RAND_poll(): over a minute. As a workaround
1795 include a time check in the inner Heap32Next loop too.
1798 *) The code that handled flushing of data in SSL/TLS originally used the
1799 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
1800 the problem outlined in PR#1949. The fix suggested there however can
1801 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
1802 of Apache). So instead simplify the code to flush unconditionally.
1803 This should be fine since flushing with no data to flush is a no op.
1806 *) Handle TLS versions 2.0 and later properly and correctly use the
1807 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
1808 off ancient servers have a habit of sticking around for a while...
1811 *) Modify compression code so it frees up structures without using the
1812 ex_data callbacks. This works around a problem where some applications
1813 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
1814 restarting) then use compression (e.g. SSL with compression) later.
1815 This results in significant per-connection memory leaks and
1816 has caused some security issues including CVE-2008-1678 and
1820 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
1821 change when encrypting or decrypting.
1824 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
1825 connect and renegotiate with servers which do not support RI.
1826 Until RI is more widely deployed this option is enabled by default.
1829 *) Add "missing" ssl ctrls to clear options and mode.
1832 *) If client attempts to renegotiate and doesn't support RI respond with
1833 a no_renegotiation alert as required by RFC5746. Some renegotiating
1834 TLS clients will continue a connection gracefully when they receive
1835 the alert. Unfortunately OpenSSL mishandled this alert and would hang
1836 waiting for a server hello which it will never receive. Now we treat a
1837 received no_renegotiation alert as a fatal error. This is because
1838 applications requesting a renegotiation might well expect it to succeed
1839 and would have no code in place to handle the server denying it so the
1840 only safe thing to do is to terminate the connection.
1843 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
1844 peer supports secure renegotiation and 0 otherwise. Print out peer
1845 renegotiation support in s_client/s_server.
1848 *) Replace the highly broken and deprecated SPKAC certification method with
1849 the updated NID creation version. This should correctly handle UTF8.
1852 *) Implement RFC5746. Re-enable renegotiation but require the extension
1853 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
1854 turns out to be a bad idea. It has been replaced by
1855 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
1856 SSL_CTX_set_options(). This is really not recommended unless you
1857 know what you are doing.
1858 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
1860 *) Fixes to stateless session resumption handling. Use initial_ctx when
1861 issuing and attempting to decrypt tickets in case it has changed during
1862 servername handling. Use a non-zero length session ID when attempting
1863 stateless session resumption: this makes it possible to determine if
1864 a resumption has occurred immediately after receiving server hello
1865 (several places in OpenSSL subtly assume this) instead of later in
1869 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
1870 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
1871 fixes for a few places where the return code is not checked
1873 [Julia Lawall <julia@diku.dk>]
1875 *) Add --strict-warnings option to Configure script to include devteam
1876 warnings in other configurations.
1879 *) Add support for --libdir option and LIBDIR variable in makefiles. This
1880 makes it possible to install openssl libraries in locations which
1881 have names other than "lib", for example "/usr/lib64" which some
1883 [Steve Henson, based on patch from Jeremy Utley]
1885 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
1886 X690 8.9.12 and can produce some misleading textual output of OIDs.
1887 [Steve Henson, reported by Dan Kaminsky]
1889 *) Delete MD2 from algorithm tables. This follows the recommendation in
1890 several standards that it is not used in new applications due to
1891 several cryptographic weaknesses. For binary compatibility reasons
1892 the MD2 API is still compiled in by default.
1895 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
1899 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
1900 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
1902 [Guenter <lists@gknw.net>]
1904 *) Fix the server certificate chain building code to use X509_verify_cert(),
1905 it used to have an ad-hoc builder which was unable to cope with anything
1906 other than a simple chain.
1907 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
1909 *) Don't check self signed certificate signatures in X509_verify_cert()
1910 by default (a flag can override this): it just wastes time without
1911 adding any security. As a useful side effect self signed root CAs
1912 with non-FIPS digests are now usable in FIPS mode.
1915 *) In dtls1_process_out_of_seq_message() the check if the current message
1916 is already buffered was missing. For every new message was memory
1917 allocated, allowing an attacker to perform an denial of service attack
1918 with sending out of seq handshake messages until there is no memory
1919 left. Additionally every future messege was buffered, even if the
1920 sequence number made no sense and would be part of another handshake.
1921 So only messages with sequence numbers less than 10 in advance will be
1922 buffered. (CVE-2009-1378)
1923 [Robin Seggelmann, discovered by Daniel Mentz]
1925 *) Records are buffered if they arrive with a future epoch to be
1926 processed after finishing the corresponding handshake. There is
1927 currently no limitation to this buffer allowing an attacker to perform
1928 a DOS attack with sending records with future epochs until there is no
1929 memory left. This patch adds the pqueue_size() function to detemine
1930 the size of a buffer and limits the record buffer to 100 entries.
1932 [Robin Seggelmann, discovered by Daniel Mentz]
1934 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
1935 parent structure is freed. (CVE-2009-1379)
1938 *) Handle non-blocking I/O properly in SSL_shutdown() call.
1939 [Darryl Miles <darryl-mailinglists@netbauds.net>]
1942 [Ilya O. <vrghost@gmail.com>]
1944 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
1946 *) Disable renegotiation completely - this fixes a severe security
1947 problem (CVE-2009-3555) at the cost of breaking all
1948 renegotiation. Renegotiation can be re-enabled by setting
1949 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
1950 run-time. This is really not recommended unless you know what
1954 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
1956 *) Don't set val to NULL when freeing up structures, it is freed up by
1957 underlying code. If sizeof(void *) > sizeof(long) this can result in
1958 zeroing past the valid field. (CVE-2009-0789)
1959 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
1961 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
1962 checked correctly. This would allow some invalid signed attributes to
1963 appear to verify correctly. (CVE-2009-0591)
1964 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1966 *) Reject UniversalString and BMPString types with invalid lengths. This
1967 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
1968 a legal length. (CVE-2009-0590)
1971 *) Set S/MIME signing as the default purpose rather than setting it
1972 unconditionally. This allows applications to override it at the store
1976 *) Permit restricted recursion of ASN1 strings. This is needed in practice
1977 to handle some structures.
1980 *) Improve efficiency of mem_gets: don't search whole buffer each time
1982 [Jeremy Shapiro <jnshapir@us.ibm.com>]
1984 *) New -hex option for openssl rand.
1987 *) Print out UTF8String and NumericString when parsing ASN1.
1990 *) Support NumericString type for name components.
1993 *) Allow CC in the environment to override the automatically chosen
1994 compiler. Note that nothing is done to ensure flags work with the
1998 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
2000 *) Properly check EVP_VerifyFinal() and similar return values
2002 [Ben Laurie, Bodo Moeller, Google Security Team]
2004 *) Enable TLS extensions by default.
2007 *) Allow the CHIL engine to be loaded, whether the application is
2008 multithreaded or not. (This does not release the developer from the
2009 obligation to set up the dynamic locking callbacks.)
2010 [Sander Temme <sander@temme.net>]
2012 *) Use correct exit code if there is an error in dgst command.
2013 [Steve Henson; problem pointed out by Roland Dirlewanger]
2015 *) Tweak Configure so that you need to say "experimental-jpake" to enable
2016 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
2019 *) Add experimental JPAKE support, including demo authentication in
2020 s_client and s_server.
2023 *) Set the comparison function in v3_addr_canonize().
2024 [Rob Austein <sra@hactrn.net>]
2026 *) Add support for XMPP STARTTLS in s_client.
2027 [Philip Paeps <philip@freebsd.org>]
2029 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
2030 to ensure that even with this option, only ciphersuites in the
2031 server's preference list will be accepted. (Note that the option
2032 applies only when resuming a session, so the earlier behavior was
2033 just about the algorithm choice for symmetric cryptography.)
2036 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
2038 *) Fix NULL pointer dereference if a DTLS server received
2039 ChangeCipherSpec as first record (CVE-2009-1386).
2042 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
2043 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
2046 *) The fix in 0.9.8c that supposedly got rid of unsafe
2047 double-checked locking was incomplete for RSA blinding,
2048 addressing just one layer of what turns out to have been
2049 doubly unsafe triple-checked locking.
2051 So now fix this for real by retiring the MONT_HELPER macro
2052 in crypto/rsa/rsa_eay.c.
2054 [Bodo Moeller; problem pointed out by Marius Schilder]
2056 *) Various precautionary measures:
2058 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
2060 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
2061 (NB: This would require knowledge of the secret session ticket key
2062 to exploit, in which case you'd be SOL either way.)
2064 - Change bn_nist.c so that it will properly handle input BIGNUMs
2065 outside the expected range.
2067 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
2070 [Neel Mehta, Bodo Moeller]
2072 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
2073 the load fails. Useful for distros.
2074 [Ben Laurie and the FreeBSD team]
2076 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
2079 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
2082 *) Expand ENGINE to support engine supplied SSL client certificate functions.
2084 This work was sponsored by Logica.
2087 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
2088 keystores. Support for SSL/TLS client authentication too.
2089 Not compiled unless enable-capieng specified to Configure.
2091 This work was sponsored by Logica.
2094 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
2095 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
2096 attribute creation routines such as certifcate requests and PKCS#12
2100 Changes between 0.9.8g and 0.9.8h [28 May 2008]
2102 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
2103 handshake which could lead to a cilent crash as found using the
2104 Codenomicon TLS test suite (CVE-2008-1672)
2105 [Steve Henson, Mark Cox]
2107 *) Fix double free in TLS server name extensions which could lead to
2108 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
2111 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
2113 Clear the error queue to ensure that error entries left from
2114 older function calls do not interfere with the correct operation.
2115 [Lutz Jaenicke, Erik de Castro Lopo]
2117 *) Remove root CA certificates of commercial CAs:
2119 The OpenSSL project does not recommend any specific CA and does not
2120 have any policy with respect to including or excluding any CA.
2121 Therefore it does not make any sense to ship an arbitrary selection
2122 of root CA certificates with the OpenSSL software.
2125 *) RSA OAEP patches to fix two separate invalid memory reads.
2126 The first one involves inputs when 'lzero' is greater than
2127 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
2128 before the beginning of from). The second one involves inputs where
2129 the 'db' section contains nothing but zeroes (there is a one-byte
2130 invalid read after the end of 'db').
2131 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2133 *) Partial backport from 0.9.9-dev:
2135 Introduce bn_mul_mont (dedicated Montgomery multiplication
2136 procedure) as a candidate for BIGNUM assembler implementation.
2137 While 0.9.9-dev uses assembler for various architectures, only
2138 x86_64 is available by default here in the 0.9.8 branch, and
2139 32-bit x86 is available through a compile-time setting.
2141 To try the 32-bit x86 assembler implementation, use Configure
2142 option "enable-montasm" (which exists only for this backport).
2144 As "enable-montasm" for 32-bit x86 disclaims code stability
2145 anyway, in this constellation we activate additional code
2146 backported from 0.9.9-dev for further performance improvements,
2147 namely BN_from_montgomery_word. (To enable this otherwise,
2148 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
2150 [Andy Polyakov (backport partially by Bodo Moeller)]
2152 *) Add TLS session ticket callback. This allows an application to set
2153 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
2154 values. This is useful for key rollover for example where several key
2155 sets may exist with different names.
2158 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
2159 This was broken until now in 0.9.8 releases, such that the only way
2160 a registered ENGINE could be used (assuming it initialises
2161 successfully on the host) was to explicitly set it as the default
2162 for the relevant algorithms. This is in contradiction with 0.9.7
2163 behaviour and the documentation. With this fix, when an ENGINE is
2164 registered into a given algorithm's table of implementations, the
2165 'uptodate' flag is reset so that auto-discovery will be used next
2166 time a new context for that algorithm attempts to select an
2168 [Ian Lister (tweaked by Geoff Thorpe)]
2170 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
2171 implemention in the following ways:
2173 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
2176 Lack of BER streaming support means one pass streaming processing is
2177 only supported if data is detached: setting the streaming flag is
2178 ignored for embedded content.
2180 CMS support is disabled by default and must be explicitly enabled
2181 with the enable-cms configuration option.
2184 *) Update the GMP engine glue to do direct copies between BIGNUM and
2185 mpz_t when openssl and GMP use the same limb size. Otherwise the
2186 existing "conversion via a text string export" trick is still used.
2187 [Paul Sheer <paulsheer@gmail.com>]
2189 *) Zlib compression BIO. This is a filter BIO which compressed and
2190 uncompresses any data passed through it.
2193 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
2194 RFC3394 compatible AES key wrapping.
2197 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
2198 sets string data without copying. X509_ALGOR_set0() and
2199 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
2200 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
2201 from an X509_ATTRIBUTE structure optionally checking it occurs only
2202 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
2206 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
2207 to get the expected BN_FLG_CONSTTIME behavior.
2208 [Bodo Moeller (Google)]
2212 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
2213 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
2214 - added some more tests to do_tests.pl
2215 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
2216 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
2217 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
2218 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
2219 - various changes to netware.pl to enable gcc-cross builds on Win32
2221 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
2222 - various changes to fix missing prototype warnings
2223 - fixed x86nasm.pl to create correct asm files for NASM COFF output
2224 - added AES, WHIRLPOOL and CPUID assembler code to build files
2225 - added missing AES assembler make rules to mk1mf.pl
2226 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
2227 [Guenter Knauf <eflash@gmx.net>]
2229 *) Implement certificate status request TLS extension defined in RFC3546.
2230 A client can set the appropriate parameters and receive the encoded
2231 OCSP response via a callback. A server can query the supplied parameters
2232 and set the encoded OCSP response in the callback. Add simplified examples
2233 to s_client and s_server.
2236 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
2238 *) Fix various bugs:
2239 + Binary incompatibility of ssl_ctx_st structure
2240 + DTLS interoperation with non-compliant servers
2241 + Don't call get_session_cb() without proposed session
2242 + Fix ia64 assembler code
2243 [Andy Polyakov, Steve Henson]
2245 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
2247 *) DTLS Handshake overhaul. There were longstanding issues with
2248 OpenSSL DTLS implementation, which were making it impossible for
2249 RFC 4347 compliant client to communicate with OpenSSL server.
2250 Unfortunately just fixing these incompatibilities would "cut off"
2251 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
2252 server keeps tolerating non RFC compliant syntax. The opposite is
2253 not true, 0.9.8f client can not communicate with earlier server.
2254 This update even addresses CVE-2007-4995.
2257 *) Changes to avoid need for function casts in OpenSSL: some compilers
2258 (gcc 4.2 and later) reject their use.
2259 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
2262 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2263 RFC4507bis. The encrypted ticket format is an encrypted encoded
2264 SSL_SESSION structure, that way new session features are automatically
2267 If a client application caches session in an SSL_SESSION structure
2268 support is transparent because tickets are now stored in the encoded
2271 The SSL_CTX structure automatically generates keys for ticket
2272 protection in servers so again support should be possible
2273 with no application modification.
2275 If a client or server wishes to disable RFC4507 support then the option
2276 SSL_OP_NO_TICKET can be set.
2278 Add a TLS extension debugging callback to allow the contents of any client
2279 or server extensions to be examined.
2281 This work was sponsored by Google.
2284 *) Add initial support for TLS extensions, specifically for the server_name
2285 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2286 have new members for a host name. The SSL data structure has an
2287 additional member SSL_CTX *initial_ctx so that new sessions can be
2288 stored in that context to allow for session resumption, even after the
2289 SSL has been switched to a new SSL_CTX in reaction to a client's
2290 server_name extension.
2292 New functions (subject to change):
2294 SSL_get_servername()
2295 SSL_get_servername_type()
2298 New CTRL codes and macros (subject to change):
2300 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2301 - SSL_CTX_set_tlsext_servername_callback()
2302 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2303 - SSL_CTX_set_tlsext_servername_arg()
2304 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2306 openssl s_client has a new '-servername ...' option.
2308 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2309 '-key2 ...', '-servername_fatal' (subject to change). This allows
2310 testing the HostName extension for a specific single host name ('-cert'
2311 and '-key' remain fallbacks for handshakes without HostName
2312 negotiation). If the unrecogninzed_name alert has to be sent, this by
2313 default is a warning; it becomes fatal with the '-servername_fatal'
2316 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
2318 *) Add AES and SSE2 assembly language support to VC++ build.
2321 *) Mitigate attack on final subtraction in Montgomery reduction.
2324 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
2325 (which previously caused an internal error).
2328 *) Squeeze another 10% out of IGE mode when in != out.
2331 *) AES IGE mode speedup.
2332 [Dean Gaudet (Google)]
2334 *) Add the Korean symmetric 128-bit cipher SEED (see
2335 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
2336 add SEED ciphersuites from RFC 4162:
2338 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
2339 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
2340 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
2341 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
2343 To minimize changes between patchlevels in the OpenSSL 0.9.8
2344 series, SEED remains excluded from compilation unless OpenSSL
2345 is configured with 'enable-seed'.
2346 [KISA, Bodo Moeller]
2348 *) Mitigate branch prediction attacks, which can be practical if a
2349 single processor is shared, allowing a spy process to extract
2350 information. For detailed background information, see
2351 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
2352 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
2353 and Necessary Software Countermeasures"). The core of the change
2354 are new versions BN_div_no_branch() and
2355 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
2356 respectively, which are slower, but avoid the security-relevant
2357 conditional branches. These are automatically called by BN_div()
2358 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
2359 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
2360 remove a conditional branch.
2362 BN_FLG_CONSTTIME is the new name for the previous
2363 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
2364 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
2365 in the exponent causes BN_mod_exp_mont() to use the alternative
2366 implementation in BN_mod_exp_mont_consttime().) The old name
2367 remains as a deprecated alias.
2369 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
2370 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
2371 constant-time implementations for more than just exponentiation.
2372 Here too the old name is kept as a deprecated alias.
2374 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2375 the BN_BLINDING structure gets an independent copy of the
2376 modulus. This means that the previous "BIGNUM *m" argument to
2377 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2378 essentially becomes "const BIGNUM *m", although we can't actually
2379 change this in the header file before 0.9.9. It allows
2380 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2381 enable BN_FLG_CONSTTIME.
2383 [Matthew D Wood (Intel Corp)]
2385 *) In the SSL/TLS server implementation, be strict about session ID
2386 context matching (which matters if an application uses a single
2387 external cache for different purposes). Previously,
2388 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2389 set. This did ensure strict client verification, but meant that,
2390 with applications using a single external cache for quite
2391 different requirements, clients could circumvent ciphersuite
2392 restrictions for a given session ID context by starting a session
2393 in a different context.
2396 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2397 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2398 authentication-only ciphersuites.
2401 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
2402 not complete and could lead to a possible single byte overflow
2403 (CVE-2007-5135) [Ben Laurie]
2405 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
2407 *) Since AES128 and AES256 (and similarly Camellia128 and
2408 Camellia256) share a single mask bit in the logic of
2409 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2410 kludge to work properly if AES128 is available and AES256 isn't
2411 (or if Camellia128 is available and Camellia256 isn't).
2414 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
2415 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
2416 When a point or a seed is encoded in a BIT STRING, we need to
2417 prevent the removal of trailing zero bits to get the proper DER
2418 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
2419 of a NamedBitList, for which trailing 0 bits need to be removed.)
2422 *) Have SSL/TLS server implementation tolerate "mismatched" record
2423 protocol version while receiving ClientHello even if the
2424 ClientHello is fragmented. (The server can't insist on the
2425 particular protocol version it has chosen before the ServerHello
2426 message has informed the client about his choice.)
2429 *) Add RFC 3779 support.
2430 [Rob Austein for ARIN, Ben Laurie]
2432 *) Load error codes if they are not already present instead of using a
2433 static variable. This allows them to be cleanly unloaded and reloaded.
2434 Improve header file function name parsing.
2437 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
2438 or CAPABILITY handshake as required by RFCs.
2441 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
2443 *) Introduce limits to prevent malicious keys being able to
2444 cause a denial of service. (CVE-2006-2940)
2445 [Steve Henson, Bodo Moeller]
2447 *) Fix ASN.1 parsing of certain invalid structures that can result
2448 in a denial of service. (CVE-2006-2937) [Steve Henson]
2450 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2451 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2453 *) Fix SSL client code which could crash if connecting to a
2454 malicious SSLv2 server. (CVE-2006-4343)
2455 [Tavis Ormandy and Will Drewry, Google Security Team]
2457 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
2458 match only those. Before that, "AES256-SHA" would be interpreted
2459 as a pattern and match "AES128-SHA" too (since AES128-SHA got
2460 the same strength classification in 0.9.7h) as we currently only
2461 have a single AES bit in the ciphersuite description bitmap.
2462 That change, however, also applied to ciphersuite strings such as
2463 "RC4-MD5" that intentionally matched multiple ciphersuites --
2464 namely, SSL 2.0 ciphersuites in addition to the more common ones
2465 from SSL 3.0/TLS 1.0.
2467 So we change the selection algorithm again: Naming an explicit
2468 ciphersuite selects this one ciphersuite, and any other similar
2469 ciphersuite (same bitmap) from *other* protocol versions.
2470 Thus, "RC4-MD5" again will properly select both the SSL 2.0
2471 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
2473 Since SSL 2.0 does not have any ciphersuites for which the
2474 128/256 bit distinction would be relevant, this works for now.
2475 The proper fix will be to use different bits for AES128 and
2476 AES256, which would have avoided the problems from the beginning;
2477 however, bits are scarce, so we can only do this in a new release
2478 (not just a patchlevel) when we can change the SSL_CIPHER
2479 definition to split the single 'unsigned long mask' bitmap into
2480 multiple values to extend the available space.
2484 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2486 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2487 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2489 *) Add AES IGE and biIGE modes.
2492 *) Change the Unix randomness entropy gathering to use poll() when
2493 possible instead of select(), since the latter has some
2494 undesirable limitations.
2495 [Darryl Miles via Richard Levitte and Bodo Moeller]
2497 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2498 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2499 cannot be implicitly activated as part of, e.g., the "AES" alias.
2500 However, please upgrade to OpenSSL 0.9.9[-dev] for
2501 non-experimental use of the ECC ciphersuites to get TLS extension
2502 support, which is required for curve and point format negotiation
2503 to avoid potential handshake problems.
2506 *) Disable rogue ciphersuites:
2508 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2509 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2510 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2512 The latter two were purportedly from
2513 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2516 Also deactivate the remaining ciphersuites from
2517 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2518 unofficial, and the ID has long expired.
2521 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2522 dual-core machines) and other potential thread-safety issues.
2525 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
2526 versions), which is now available for royalty-free use
2527 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
2528 Also, add Camellia TLS ciphersuites from RFC 4132.
2530 To minimize changes between patchlevels in the OpenSSL 0.9.8
2531 series, Camellia remains excluded from compilation unless OpenSSL
2532 is configured with 'enable-camellia'.
2535 *) Disable the padding bug check when compression is in use. The padding
2536 bug check assumes the first packet is of even length, this is not
2537 necessarily true if compresssion is enabled and can result in false
2538 positives causing handshake failure. The actual bug test is ancient
2539 code so it is hoped that implementations will either have fixed it by
2540 now or any which still have the bug do not support compression.
2543 Changes between 0.9.8a and 0.9.8b [04 May 2006]
2545 *) When applying a cipher rule check to see if string match is an explicit
2546 cipher suite and only match that one cipher suite if it is.
2549 *) Link in manifests for VC++ if needed.
2550 [Austin Ziegler <halostatue@gmail.com>]
2552 *) Update support for ECC-based TLS ciphersuites according to
2553 draft-ietf-tls-ecc-12.txt with proposed changes (but without
2554 TLS extensions, which are supported starting with the 0.9.9
2555 branch, not in the OpenSSL 0.9.8 branch).
2558 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
2559 opaque EVP_CIPHER_CTX handling.
2562 *) Fixes and enhancements to zlib compression code. We now only use
2563 "zlib1.dll" and use the default __cdecl calling convention on Win32
2564 to conform with the standards mentioned here:
2565 http://www.zlib.net/DLL_FAQ.txt
2566 Static zlib linking now works on Windows and the new --with-zlib-include
2567 --with-zlib-lib options to Configure can be used to supply the location
2568 of the headers and library. Gracefully handle case where zlib library
2572 *) Several fixes and enhancements to the OID generation code. The old code
2573 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
2574 handle numbers larger than ULONG_MAX, truncated printing and had a
2575 non standard OBJ_obj2txt() behaviour.
2578 *) Add support for building of engines under engine/ as shared libraries
2579 under VC++ build system.
2582 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
2583 Hopefully, we will not see any false combination of paths any more.
2586 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
2588 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2589 (part of SSL_OP_ALL). This option used to disable the
2590 countermeasure against man-in-the-middle protocol-version
2591 rollback in the SSL 2.0 server implementation, which is a bad
2592 idea. (CVE-2005-2969)
2594 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2595 for Information Security, National Institute of Advanced Industrial
2596 Science and Technology [AIST], Japan)]
2598 *) Add two function to clear and return the verify parameter flags.
2601 *) Keep cipherlists sorted in the source instead of sorting them at
2602 runtime, thus removing the need for a lock.
2605 *) Avoid some small subgroup attacks in Diffie-Hellman.
2606 [Nick Mathewson and Ben Laurie]
2608 *) Add functions for well-known primes.
2611 *) Extended Windows CE support.
2612 [Satoshi Nakamura and Andy Polyakov]
2614 *) Initialize SSL_METHOD structures at compile time instead of during
2615 runtime, thus removing the need for a lock.
2618 *) Make PKCS7_decrypt() work even if no certificate is supplied by
2619 attempting to decrypt each encrypted key in turn. Add support to
2623 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
2625 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
2628 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
2631 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
2632 key into the same file any more.
2635 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
2638 *) Add -utf8 command line and config file option to 'ca'.
2639 [Stefan <stf@udoma.org]
2641 *) Removed the macro des_crypt(), as it seems to conflict with some
2642 libraries. Use DES_crypt().
2645 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
2646 involves renaming the source and generated shared-libs for
2647 both. The engines will accept the corrected or legacy ids
2648 ('ncipher' and '4758_cca' respectively) when binding. NB,
2649 this only applies when building 'shared'.
2650 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
2652 *) Add attribute functions to EVP_PKEY structure. Modify
2653 PKCS12_create() to recognize a CSP name attribute and
2654 use it. Make -CSP option work again in pkcs12 utility.
2657 *) Add new functionality to the bn blinding code:
2658 - automatic re-creation of the BN_BLINDING parameters after
2659 a fixed number of uses (currently 32)
2660 - add new function for parameter creation
2661 - introduce flags to control the update behaviour of the
2662 BN_BLINDING parameters
2663 - hide BN_BLINDING structure
2664 Add a second BN_BLINDING slot to the RSA structure to improve
2665 performance when a single RSA object is shared among several
2669 *) Add support for DTLS.
2670 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
2672 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
2673 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
2676 *) Remove buggy and incompletet DH cert support from
2677 ssl/ssl_rsa.c and ssl/s3_both.c
2680 *) Use SHA-1 instead of MD5 as the default digest algorithm for
2681 the apps/openssl applications.
2684 *) Compile clean with "-Wall -Wmissing-prototypes
2685 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
2686 DEBUG_SAFESTACK must also be set.
2689 *) Change ./Configure so that certain algorithms can be disabled by default.
2690 The new counterpiece to "no-xxx" is "enable-xxx".
2692 The patented RC5 and MDC2 algorithms will now be disabled unless
2693 "enable-rc5" and "enable-mdc2", respectively, are specified.
2695 (IDEA remains enabled despite being patented. This is because IDEA
2696 is frequently required for interoperability, and there is no license
2697 fee for non-commercial use. As before, "no-idea" can be used to
2698 avoid this algorithm.)
2702 *) Add processing of proxy certificates (see RFC 3820). This work was
2703 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
2704 EGEE (Enabling Grids for E-science in Europe).
2707 *) RC4 performance overhaul on modern architectures/implementations, such
2708 as Intel P4, IA-64 and AMD64.
2711 *) New utility extract-section.pl. This can be used specify an alternative
2712 section number in a pod file instead of having to treat each file as
2713 a separate case in Makefile. This can be done by adding two lines to the
2716 =for comment openssl_section:XXX
2718 The blank line is mandatory.
2722 *) New arguments -certform, -keyform and -pass for s_client and s_server
2723 to allow alternative format key and certificate files and passphrase
2727 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
2728 update associated structures and add various utility functions.
2730 Add new policy related verify parameters, include policy checking in
2731 standard verify code. Enhance 'smime' application with extra parameters
2732 to support policy checking and print out.
2735 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
2736 Nehemiah processors. These extensions support AES encryption in hardware
2737 as well as RNG (though RNG support is currently disabled).
2738 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
2740 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
2743 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
2744 [Andy Polyakov and a number of other people]
2746 *) Improved PowerPC platform support. Most notably BIGNUM assembler
2747 implementation contributed by IBM.
2748 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
2750 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
2751 exponent rather than 'unsigned long'. There is a corresponding change to
2752 the new 'rsa_keygen' element of the RSA_METHOD structure.
2753 [Jelte Jansen, Geoff Thorpe]
2755 *) Functionality for creating the initial serial number file is now
2756 moved from CA.pl to the 'ca' utility with a new option -create_serial.
2758 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
2759 number file to 1, which is bound to cause problems. To avoid
2760 the problems while respecting compatibility between different 0.9.7
2761 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
2762 CA.pl for serial number initialization. With the new release 0.9.8,
2763 we can fix the problem directly in the 'ca' utility.)
2766 *) Reduced header interdepencies by declaring more opaque objects in
2767 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
2768 give fewer recursive includes, which could break lazy source code - so
2769 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
2770 developers should define this symbol when building and using openssl to
2771 ensure they track the recommended behaviour, interfaces, [etc], but
2772 backwards-compatible behaviour prevails when this isn't defined.
2775 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
2778 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
2779 This will generate a random key of the appropriate length based on the
2780 cipher context. The EVP_CIPHER can provide its own random key generation
2781 routine to support keys of a specific form. This is used in the des and
2782 3des routines to generate a key of the correct parity. Update S/MIME
2783 code to use new functions and hence generate correct parity DES keys.
2784 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
2785 valid (weak or incorrect parity).
2788 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
2789 as looking them up. This is useful when the verified structure may contain
2790 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
2791 present unless the new PKCS7_NO_CRL flag is asserted.
2794 *) Extend ASN1 oid configuration module. It now additionally accepts the
2797 shortName = some long name, 1.2.3.4
2800 *) Reimplemented the BN_CTX implementation. There is now no more static
2801 limitation on the number of variables it can handle nor the depth of the
2802 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
2803 information can now expand as required, and rather than having a single
2804 static array of bignums, BN_CTX now uses a linked-list of such arrays
2805 allowing it to expand on demand whilst maintaining the usefulness of
2806 BN_CTX's "bundling".
2809 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
2810 to allow all RSA operations to function using a single BN_CTX.
2813 *) Preliminary support for certificate policy evaluation and checking. This
2814 is initially intended to pass the tests outlined in "Conformance Testing
2815 of Relying Party Client Certificate Path Processing Logic" v1.07.
2818 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
2819 remained unused and not that useful. A variety of other little bignum
2820 tweaks and fixes have also been made continuing on from the audit (see
2824 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
2825 associated ASN1, EVP and SSL functions and old ASN1 macros.
2828 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
2829 and this should never fail. So the return value from the use of
2830 BN_set_word() (which can fail due to needless expansion) is now deprecated;
2831 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
2834 *) BN_CTX_get() should return zero-valued bignums, providing the same
2835 initialised value as BN_new().
2836 [Geoff Thorpe, suggested by Ulf Möller]
2838 *) Support for inhibitAnyPolicy certificate extension.
2841 *) An audit of the BIGNUM code is underway, for which debugging code is
2842 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
2843 is considered valid when processing BIGNUMs, and causes execution to
2844 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
2845 further steps are taken to deliberately pollute unused data in BIGNUM
2846 structures to try and expose faulty code further on. For now, openssl will
2847 (in its default mode of operation) continue to tolerate the inconsistent
2848 forms that it has tolerated in the past, but authors and packagers should
2849 consider trying openssl and their own applications when compiled with
2850 these debugging symbols defined. It will help highlight potential bugs in
2851 their own code, and will improve the test coverage for OpenSSL itself. At
2852 some point, these tighter rules will become openssl's default to improve
2853 maintainability, though the assert()s and other overheads will remain only
2854 in debugging configurations. See bn.h for more details.
2855 [Geoff Thorpe, Nils Larsch, Ulf Möller]
2857 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
2858 that can only be obtained through BN_CTX_new() (which implicitly
2859 initialises it). The presence of this function only made it possible
2860 to overwrite an existing structure (and cause memory leaks).
2863 *) Because of the callback-based approach for implementing LHASH as a
2864 template type, lh_insert() adds opaque objects to hash-tables and
2865 lh_doall() or lh_doall_arg() are typically used with a destructor callback
2866 to clean up those corresponding objects before destroying the hash table
2867 (and losing the object pointers). So some over-zealous constifications in
2868 LHASH have been relaxed so that lh_insert() does not take (nor store) the
2869 objects as "const" and the lh_doall[_arg] callback wrappers are not
2870 prototyped to have "const" restrictions on the object pointers they are
2871 given (and so aren't required to cast them away any more).
2874 *) The tmdiff.h API was so ugly and minimal that our own timing utility
2875 (speed) prefers to use its own implementation. The two implementations
2876 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
2877 its object type properly exposed (MS_TM) instead of casting to/from "char
2878 *". This may still change yet if someone realises MS_TM and "ms_time_***"
2879 aren't necessarily the greatest nomenclatures - but this is what was used
2880 internally to the implementation so I've used that for now.
2883 *) Ensure that deprecated functions do not get compiled when
2884 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
2885 the self-tests were still using deprecated key-generation functions so
2886 these have been updated also.
2889 *) Reorganise PKCS#7 code to separate the digest location functionality
2890 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
2891 New function PKCS7_set_digest() to set the digest type for PKCS#7
2892 digestedData type. Add additional code to correctly generate the
2893 digestedData type and add support for this type in PKCS7 initialization
2897 *) New function PKCS7_set0_type_other() this initializes a PKCS7
2898 structure of type "other".
2901 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
2902 sure the loop does correctly stop and breaking ("division by zero")
2903 modulus operations are not performed. The (pre-generated) prime
2904 table crypto/bn/bn_prime.h was already correct, but it could not be
2905 re-generated on some platforms because of the "division by zero"
2906 situation in the script.
2907 [Ralf S. Engelschall]
2909 *) Update support for ECC-based TLS ciphersuites according to
2910 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
2911 SHA-1 now is only used for "small" curves (where the
2912 representation of a field element takes up to 24 bytes); for
2913 larger curves, the field element resulting from ECDH is directly
2914 used as premaster secret.
2915 [Douglas Stebila (Sun Microsystems Laboratories)]
2917 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
2918 curve secp160r1 to the tests.
2919 [Douglas Stebila (Sun Microsystems Laboratories)]
2921 *) Add the possibility to load symbols globally with DSO.
2922 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
2924 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
2925 control of the error stack.
2928 *) Add support for STORE in ENGINE.
2931 *) Add the STORE type. The intention is to provide a common interface
2932 to certificate and key stores, be they simple file-based stores, or
2933 HSM-type store, or LDAP stores, or...
2934 NOTE: The code is currently UNTESTED and isn't really used anywhere.
2937 *) Add a generic structure called OPENSSL_ITEM. This can be used to
2938 pass a list of arguments to any function as well as provide a way
2939 for a function to pass data back to the caller.
2942 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
2943 works like BUF_strdup() but can be used to duplicate a portion of
2944 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
2948 *) Add the function sk_find_ex() which works like sk_find(), but will
2949 return an index to an element even if an exact match couldn't be
2950 found. The index is guaranteed to point at the element where the
2951 searched-for key would be inserted to preserve sorting order.
2954 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
2955 takes an extra flags argument for optional functionality. Currently,
2956 the following flags are defined:
2958 OBJ_BSEARCH_VALUE_ON_NOMATCH
2959 This one gets OBJ_bsearch_ex() to return a pointer to the first
2960 element where the comparing function returns a negative or zero
2963 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
2964 This one gets OBJ_bsearch_ex() to return a pointer to the first
2965 element where the comparing function returns zero. This is useful
2966 if there are more than one element where the comparing function
2970 *) Make it possible to create self-signed certificates with 'openssl ca'
2971 in such a way that the self-signed certificate becomes part of the
2972 CA database and uses the same mechanisms for serial number generation
2973 as all other certificate signing. The new flag '-selfsign' enables
2974 this functionality. Adapt CA.sh and CA.pl.in.
2977 *) Add functionality to check the public key of a certificate request
2978 against a given private. This is useful to check that a certificate
2979 request can be signed by that key (self-signing).
2982 *) Make it possible to have multiple active certificates with the same
2983 subject in the CA index file. This is done only if the keyword
2984 'unique_subject' is set to 'no' in the main CA section (default
2985 if 'CA_default') of the configuration file. The value is saved
2986 with the database itself in a separate index attribute file,
2987 named like the index file with '.attr' appended to the name.
2990 *) Generate muti valued AVAs using '+' notation in config files for
2994 *) Support for nameConstraints certificate extension.
2997 *) Support for policyConstraints certificate extension.
3000 *) Support for policyMappings certificate extension.
3003 *) Make sure the default DSA_METHOD implementation only uses its
3004 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
3005 and change its own handlers to be NULL so as to remove unnecessary
3006 indirection. This lets alternative implementations fallback to the
3007 default implementation more easily.
3010 *) Support for directoryName in GeneralName related extensions
3014 *) Make it possible to link applications using Makefile.shared.
3015 Make that possible even when linking against static libraries!
3018 *) Support for single pass processing for S/MIME signing. This now
3019 means that S/MIME signing can be done from a pipe, in addition
3020 cleartext signing (multipart/signed type) is effectively streaming
3021 and the signed data does not need to be all held in memory.
3023 This is done with a new flag PKCS7_STREAM. When this flag is set
3024 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
3025 is done after the data is output (and digests calculated) in
3026 SMIME_write_PKCS7().
3029 *) Add full support for -rpath/-R, both in shared libraries and
3030 applications, at least on the platforms where it's known how
3034 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
3035 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
3036 will now compute a table of multiples of the generator that
3037 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
3038 faster (notably in the case of a single point multiplication,
3039 scalar * generator).
3040 [Nils Larsch, Bodo Moeller]
3042 *) IPv6 support for certificate extensions. The various extensions
3043 which use the IP:a.b.c.d can now take IPv6 addresses using the
3044 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
3048 *) Added an ENGINE that implements RSA by performing private key
3049 exponentiations with the GMP library. The conversions to and from
3050 GMP's mpz_t format aren't optimised nor are any montgomery forms
3051 cached, and on x86 it appears OpenSSL's own performance has caught up.
3052 However there are likely to be other architectures where GMP could
3053 provide a boost. This ENGINE is not built in by default, but it can be
3054 specified at Configure time and should be accompanied by the necessary
3055 linker additions, eg;
3056 ./config -DOPENSSL_USE_GMP -lgmp
3059 *) "openssl engine" will not display ENGINE/DSO load failure errors when
3060 testing availability of engines with "-t" - the old behaviour is
3061 produced by increasing the feature's verbosity with "-tt".
3064 *) ECDSA routines: under certain error conditions uninitialized BN objects
3065 could be freed. Solution: make sure initialization is performed early
3066 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
3070 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
3071 and DH_METHOD (eg. by ENGINE implementations) to override the normal
3072 software implementations. For DSA and DH, parameter generation can
3073 also be overriden by providing the appropriate method callbacks.
3076 *) Change the "progress" mechanism used in key-generation and
3077 primality testing to functions that take a new BN_GENCB pointer in
3078 place of callback/argument pairs. The new API functions have "_ex"
3079 postfixes and the older functions are reimplemented as wrappers for
3080 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
3081 declarations of the old functions to help (graceful) attempts to
3082 migrate to the new functions. Also, the new key-generation API
3083 functions operate on a caller-supplied key-structure and return
3084 success/failure rather than returning a key or NULL - this is to
3085 help make "keygen" another member function of RSA_METHOD etc.
3087 Example for using the new callback interface:
3089 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
3093 BN_GENCB_set(&my_cb, my_callback, my_arg);
3095 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
3096 /* For the meaning of a, b in calls to my_callback(), see the
3097 * documentation of the function that calls the callback.
3098 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
3099 * my_callback should return 1 if it wants BN_is_prime_ex()
3100 * to continue, or 0 to stop.
3105 *) Change the ZLIB compression method to be stateful, and make it
3106 available to TLS with the number defined in
3107 draft-ietf-tls-compression-04.txt.
3110 *) Add the ASN.1 structures and functions for CertificatePair, which
3111 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
3113 CertificatePair ::= SEQUENCE {
3114 forward [0] Certificate OPTIONAL,
3115 reverse [1] Certificate OPTIONAL,
3116 -- at least one of the pair shall be present -- }
3118 Also implement the PEM functions to read and write certificate
3119 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
3121 This needed to be defined, mostly for the sake of the LDAP
3122 attribute crossCertificatePair, but may prove useful elsewhere as
3126 *) Make it possible to inhibit symlinking of shared libraries in
3127 Makefile.shared, for Cygwin's sake.
3130 *) Extend the BIGNUM API by creating a function
3131 void BN_set_negative(BIGNUM *a, int neg);
3132 and a macro that behave like
3133 int BN_is_negative(const BIGNUM *a);
3135 to avoid the need to access 'a->neg' directly in applications.
3138 *) Implement fast modular reduction for pseudo-Mersenne primes
3139 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
3140 EC_GROUP_new_curve_GFp() will now automatically use this
3142 [Nils Larsch <nla@trustcenter.de>]
3144 *) Add new lock type (CRYPTO_LOCK_BN).
3147 *) Change the ENGINE framework to automatically load engines
3148 dynamically from specific directories unless they could be
3149 found to already be built in or loaded. Move all the
3150 current engines except for the cryptodev one to a new
3152 The engines in engines/ are built as shared libraries if
3153 the "shared" options was given to ./Configure or ./config.
3154 Otherwise, they are inserted in libcrypto.a.
3155 /usr/local/ssl/engines is the default directory for dynamic
3156 engines, but that can be overriden at configure time through
3157 the usual use of --prefix and/or --openssldir, and at run
3158 time with the environment variable OPENSSL_ENGINES.
3159 [Geoff Thorpe and Richard Levitte]
3161 *) Add Makefile.shared, a helper makefile to build shared
3162 libraries. Addapt Makefile.org.
3165 *) Add version info to Win32 DLLs.
3166 [Peter 'Luna' Runestig" <peter@runestig.com>]
3168 *) Add new 'medium level' PKCS#12 API. Certificates and keys
3169 can be added using this API to created arbitrary PKCS#12
3170 files while avoiding the low level API.
3172 New options to PKCS12_create(), key or cert can be NULL and
3173 will then be omitted from the output file. The encryption
3174 algorithm NIDs can be set to -1 for no encryption, the mac
3175 iteration count can be set to 0 to omit the mac.
3177 Enhance pkcs12 utility by making the -nokeys and -nocerts
3178 options work when creating a PKCS#12 file. New option -nomac
3179 to omit the mac, NONE can be set for an encryption algorithm.
3180 New code is modified to use the enhanced PKCS12_create()
3181 instead of the low level API.
3184 *) Extend ASN1 encoder to support indefinite length constructed
3185 encoding. This can output sequences tags and octet strings in
3186 this form. Modify pk7_asn1.c to support indefinite length
3187 encoding. This is experimental and needs additional code to
3188 be useful, such as an ASN1 bio and some enhanced streaming
3191 Extend template encode functionality so that tagging is passed
3192 down to the template encoder.
3195 *) Let 'openssl req' fail if an argument to '-newkey' is not
3196 recognized instead of using RSA as a default.
3199 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
3200 As these are not official, they are not included in "ALL";
3201 the "ECCdraft" ciphersuite group alias can be used to select them.
3202 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
3204 *) Add ECDH engine support.
3205 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
3207 *) Add ECDH in new directory crypto/ecdh/.
3208 [Douglas Stebila (Sun Microsystems Laboratories)]
3210 *) Let BN_rand_range() abort with an error after 100 iterations
3211 without success (which indicates a broken PRNG).
3214 *) Change BN_mod_sqrt() so that it verifies that the input value
3215 is really the square of the return value. (Previously,
3216 BN_mod_sqrt would show GIGO behaviour.)
3219 *) Add named elliptic curves over binary fields from X9.62, SECG,
3220 and WAP/WTLS; add OIDs that were still missing.
3222 [Sheueling Chang Shantz and Douglas Stebila
3223 (Sun Microsystems Laboratories)]
3225 *) Extend the EC library for elliptic curves over binary fields
3226 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
3229 EC_GF2m_simple_method
3233 EC_GROUP_new_curve_GF2m
3234 EC_GROUP_set_curve_GF2m
3235 EC_GROUP_get_curve_GF2m
3236 EC_POINT_set_affine_coordinates_GF2m
3237 EC_POINT_get_affine_coordinates_GF2m
3238 EC_POINT_set_compressed_coordinates_GF2m
3240 Point compression for binary fields is disabled by default for
3241 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
3244 As binary polynomials are represented as BIGNUMs, various members
3245 of the EC_GROUP and EC_POINT data structures can be shared
3246 between the implementations for prime fields and binary fields;
3247 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
3248 are essentially identical to their ..._GFp counterparts.
3249 (For simplicity, the '..._GFp' prefix has been dropped from
3250 various internal method names.)
3252 An internal 'field_div' method (similar to 'field_mul' and
3253 'field_sqr') has been added; this is used only for binary fields.
3255 [Sheueling Chang Shantz and Douglas Stebila
3256 (Sun Microsystems Laboratories)]
3258 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
3259 through methods ('mul', 'precompute_mult').
3261 The generic implementations (now internally called 'ec_wNAF_mul'
3262 and 'ec_wNAF_precomputed_mult') remain the default if these
3263 methods are undefined.
3265 [Sheueling Chang Shantz and Douglas Stebila
3266 (Sun Microsystems Laboratories)]
3268 *) New function EC_GROUP_get_degree, which is defined through
3269 EC_METHOD. For curves over prime fields, this returns the bit
3270 length of the modulus.
3272 [Sheueling Chang Shantz and Douglas Stebila
3273 (Sun Microsystems Laboratories)]
3275 *) New functions EC_GROUP_dup, EC_POINT_dup.
3276 (These simply call ..._new and ..._copy).
3278 [Sheueling Chang Shantz and Douglas Stebila
3279 (Sun Microsystems Laboratories)]
3281 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
3282 Polynomials are represented as BIGNUMs (where the sign bit is not
3283 used) in the following functions [macros]:
3286 BN_GF2m_sub [= BN_GF2m_add]
3287 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
3288 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
3289 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
3291 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
3292 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
3293 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
3294 BN_GF2m_cmp [= BN_ucmp]
3296 (Note that only the 'mod' functions are actually for fields GF(2^m).
3297 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
3299 For some functions, an the irreducible polynomial defining a
3300 field can be given as an 'unsigned int[]' with strictly
3301 decreasing elements giving the indices of those bits that are set;
3302 i.e., p[] represents the polynomial
3303 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
3305 p[0] > p[1] > ... > p[k] = 0.
3306 This applies to the following functions:
3311 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
3312 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
3314 BN_GF2m_mod_sqrt_arr
3315 BN_GF2m_mod_solve_quad_arr
3319 Conversion can be performed by the following functions:
3324 bntest.c has additional tests for binary polynomial arithmetic.
3326 Two implementations for BN_GF2m_mod_div() are available.
3327 The default algorithm simply uses BN_GF2m_mod_inv() and
3328 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
3329 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
3330 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
3332 [Sheueling Chang Shantz and Douglas Stebila
3333 (Sun Microsystems Laboratories)]
3335 *) Add new error code 'ERR_R_DISABLED' that can be used when some
3336 functionality is disabled at compile-time.
3337 [Douglas Stebila <douglas.stebila@sun.com>]
3339 *) Change default behaviour of 'openssl asn1parse' so that more
3340 information is visible when viewing, e.g., a certificate:
3342 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
3343 mode the content of non-printable OCTET STRINGs is output in a
3344 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
3345 avoid the appearance of a printable string.
3346 [Nils Larsch <nla@trustcenter.de>]
3348 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
3350 EC_GROUP_set_asn1_flag()
3351 EC_GROUP_get_asn1_flag()
3352 EC_GROUP_set_point_conversion_form()
3353 EC_GROUP_get_point_conversion_form()
3354 These control ASN1 encoding details:
3355 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
3356 has been set to OPENSSL_EC_NAMED_CURVE.
3357 - Points are encoded in uncompressed form by default; options for
3358 asn1_for are as for point2oct, namely
3359 POINT_CONVERSION_COMPRESSED
3360 POINT_CONVERSION_UNCOMPRESSED
3361 POINT_CONVERSION_HYBRID
3363 Also add 'seed' and 'seed_len' members to EC_GROUP with access
3366 EC_GROUP_get0_seed()
3367 EC_GROUP_get_seed_len()
3368 This is used only for ASN1 purposes (so far).
3369 [Nils Larsch <nla@trustcenter.de>]
3371 *) Add 'field_type' member to EC_METHOD, which holds the NID
3372 of the appropriate field type OID. The new function
3373 EC_METHOD_get_field_type() returns this value.
3374 [Nils Larsch <nla@trustcenter.de>]
3379 EC_POINT_point2hex()
3380 EC_POINT_hex2point()
3381 providing useful interfaces to EC_POINT_point2oct() and
3382 EC_POINT_oct2point().
3383 [Nils Larsch <nla@trustcenter.de>]
3385 *) Change internals of the EC library so that the functions
3386 EC_GROUP_set_generator()
3387 EC_GROUP_get_generator()
3388 EC_GROUP_get_order()
3389 EC_GROUP_get_cofactor()
3390 are implemented directly in crypto/ec/ec_lib.c and not dispatched
3391 to methods, which would lead to unnecessary code duplication when
3392 adding different types of curves.
3393 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
3395 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
3396 arithmetic, and such that modified wNAFs are generated
3397 (which avoid length expansion in many cases).
3400 *) Add a function EC_GROUP_check_discriminant() (defined via
3401 EC_METHOD) that verifies that the curve discriminant is non-zero.
3403 Add a function EC_GROUP_check() that makes some sanity tests
3404 on a EC_GROUP, its generator and order. This includes
3405 EC_GROUP_check_discriminant().
3406 [Nils Larsch <nla@trustcenter.de>]
3408 *) Add ECDSA in new directory crypto/ecdsa/.
3410 Add applications 'openssl ecparam' and 'openssl ecdsa'
3411 (these are based on 'openssl dsaparam' and 'openssl dsa').
3413 ECDSA support is also included in various other files across the
3414 library. Most notably,
3415 - 'openssl req' now has a '-newkey ecdsa:file' option;
3416 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
3417 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
3418 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
3419 them suitable for ECDSA where domain parameters must be
3420 extracted before the specific public key;
3421 - ECDSA engine support has been added.
3422 [Nils Larsch <nla@trustcenter.de>]
3424 *) Include some named elliptic curves, and add OIDs from X9.62,
3425 SECG, and WAP/WTLS. Each curve can be obtained from the new
3427 EC_GROUP_new_by_curve_name(),
3428 and the list of available named curves can be obtained with
3429 EC_get_builtin_curves().
3430 Also add a 'curve_name' member to EC_GROUP objects, which can be
3432 EC_GROUP_set_curve_name()
3433 EC_GROUP_get_curve_name()
3434 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
3436 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3437 was actually never needed) and in BN_mul(). The removal in BN_mul()
3438 required a small change in bn_mul_part_recursive() and the addition
3439 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3440 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3441 bn_sub_words() and bn_add_words() except they take arrays with
3445 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
3447 *) Cleanse PEM buffers before freeing them since they may contain
3449 [Benjamin Bennett <ben@psc.edu>]
3451 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3452 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3453 authentication-only ciphersuites.
3456 *) Since AES128 and AES256 share a single mask bit in the logic of
3457 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3458 kludge to work properly if AES128 is available and AES256 isn't.
3461 *) Expand security boundary to match 1.1.1 module.
3464 *) Remove redundant features: hash file source, editing of test vectors
3465 modify fipsld to use external fips_premain.c signature.
3468 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
3469 run algorithm test programs.
3472 *) Make algorithm test programs more tolerant of whitespace.
3475 *) Have SSL/TLS server implementation tolerate "mismatched" record
3476 protocol version while receiving ClientHello even if the
3477 ClientHello is fragmented. (The server can't insist on the
3478 particular protocol version it has chosen before the ServerHello
3479 message has informed the client about his choice.)
3482 *) Load error codes if they are not already present instead of using a
3483 static variable. This allows them to be cleanly unloaded and reloaded.
3486 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3488 *) Introduce limits to prevent malicious keys being able to
3489 cause a denial of service. (CVE-2006-2940)
3490 [Steve Henson, Bodo Moeller]
3492 *) Fix ASN.1 parsing of certain invalid structures that can result
3493 in a denial of service. (CVE-2006-2937) [Steve Henson]
3495 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3496 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3498 *) Fix SSL client code which could crash if connecting to a
3499 malicious SSLv2 server. (CVE-2006-4343)
3500 [Tavis Ormandy and Will Drewry, Google Security Team]
3502 *) Change ciphersuite string processing so that an explicit
3503 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3504 will no longer include "AES128-SHA"), and any other similar
3505 ciphersuite (same bitmap) from *other* protocol versions (so that
3506 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3507 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3508 changes from 0.9.8b and 0.9.8d.
3511 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
3513 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3514 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3516 *) Change the Unix randomness entropy gathering to use poll() when
3517 possible instead of select(), since the latter has some
3518 undesirable limitations.
3519 [Darryl Miles via Richard Levitte and Bodo Moeller]
3521 *) Disable rogue ciphersuites:
3523 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3524 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3525 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3527 The latter two were purportedly from
3528 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3531 Also deactive the remaining ciphersuites from
3532 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3533 unofficial, and the ID has long expired.
3536 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3537 dual-core machines) and other potential thread-safety issues.
3540 Changes between 0.9.7i and 0.9.7j [04 May 2006]
3542 *) Adapt fipsld and the build system to link against the validated FIPS
3543 module in FIPS mode.
3546 *) Fixes for VC++ 2005 build under Windows.
3549 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
3550 from a Windows bash shell such as MSYS. It is autodetected from the
3551 "config" script when run from a VC++ environment. Modify standard VC++
3552 build to use fipscanister.o from the GNU make build.
3555 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
3557 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
3558 The value now differs depending on if you build for FIPS or not.
3559 BEWARE! A program linked with a shared FIPSed libcrypto can't be
3560 safely run with a non-FIPSed libcrypto, as it may crash because of
3561 the difference induced by this change.
3564 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
3566 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3567 (part of SSL_OP_ALL). This option used to disable the
3568 countermeasure against man-in-the-middle protocol-version
3569 rollback in the SSL 2.0 server implementation, which is a bad
3570 idea. (CVE-2005-2969)
3572 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3573 for Information Security, National Institute of Advanced Industrial
3574 Science and Technology [AIST], Japan)]
3576 *) Minimal support for X9.31 signatures and PSS padding modes. This is
3577 mainly for FIPS compliance and not fully integrated at this stage.
3580 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
3581 the exponentiation using a fixed-length exponent. (Otherwise,
3582 the information leaked through timing could expose the secret key
3583 after many signatures; cf. Bleichenbacher's attack on DSA with
3587 *) Make a new fixed-window mod_exp implementation the default for
3588 RSA, DSA, and DH private-key operations so that the sequence of
3589 squares and multiplies and the memory access pattern are
3590 independent of the particular secret key. This will mitigate
3591 cache-timing and potential related attacks.
3593 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
3594 and this is automatically used by BN_mod_exp_mont() if the new flag
3595 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
3596 will use this BN flag for private exponents unless the flag
3597 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
3598 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
3600 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
3602 *) Change the client implementation for SSLv23_method() and
3603 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
3604 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
3605 (Previously, the SSL 2.0 backwards compatible Client Hello
3606 message format would be used even with SSL_OP_NO_SSLv2.)
3609 *) Add support for smime-type MIME parameter in S/MIME messages which some
3613 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
3614 a threadsafe manner. Modify rsa code to use new function and add calls
3615 to dsa and dh code (which had race conditions before).
3618 *) Include the fixed error library code in the C error file definitions
3619 instead of fixing them up at runtime. This keeps the error code
3620 structures constant.
3623 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
3625 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
3628 *) Fixes for newer kerberos headers. NB: the casts are needed because
3629 the 'length' field is signed on one version and unsigned on another
3630 with no (?) obvious way to tell the difference, without these VC++
3631 complains. Also the "definition" of FAR (blank) is no longer included
3632 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
3633 some needed definitions.
3636 *) Undo Cygwin change.
3639 *) Added support for proxy certificates according to RFC 3820.
3640 Because they may be a security thread to unaware applications,
3641 they must be explicitely allowed in run-time. See
3642 docs/HOWTO/proxy_certificates.txt for further information.
3645 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
3647 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
3648 server and client random values. Previously
3649 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
3650 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
3652 This change has negligible security impact because:
3654 1. Server and client random values still have 24 bytes of pseudo random
3657 2. Server and client random values are sent in the clear in the initial
3660 3. The master secret is derived using the premaster secret (48 bytes in
3661 size for static RSA ciphersuites) as well as client server and random
3664 The OpenSSL team would like to thank the UK NISCC for bringing this issue
3667 [Stephen Henson, reported by UK NISCC]
3669 *) Use Windows randomness collection on Cygwin.
3672 *) Fix hang in EGD/PRNGD query when communication socket is closed
3673 prematurely by EGD/PRNGD.
3674 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
3676 *) Prompt for pass phrases when appropriate for PKCS12 input format.
3679 *) Back-port of selected performance improvements from development
3680 branch, as well as improved support for PowerPC platforms.
3683 *) Add lots of checks for memory allocation failure, error codes to indicate
3684 failure and freeing up memory if a failure occurs.
3685 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
3687 *) Add new -passin argument to dgst.
3690 *) Perform some character comparisons of different types in X509_NAME_cmp:
3691 this is needed for some certificates that reencode DNs into UTF8Strings
3692 (in violation of RFC3280) and can't or wont issue name rollover
3696 *) Make an explicit check during certificate validation to see that
3697 the CA setting in each certificate on the chain is correct. As a
3698 side effect always do the following basic checks on extensions,
3699 not just when there's an associated purpose to the check:
3701 - if there is an unhandled critical extension (unless the user
3702 has chosen to ignore this fault)
3703 - if the path length has been exceeded (if one is set at all)
3704 - that certain extensions fit the associated purpose (if one has
3708 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
3710 *) Avoid a race condition when CRLs are checked in a multi threaded
3711 environment. This would happen due to the reordering of the revoked
3712 entries during signature checking and serial number lookup. Now the
3713 encoding is cached and the serial number sort performed under a lock.
3714 Add new STACK function sk_is_sorted().
3717 *) Add Delta CRL to the extension code.
3720 *) Various fixes to s3_pkt.c so alerts are sent properly.
3721 [David Holmes <d.holmes@f5.com>]
3723 *) Reduce the chances of duplicate issuer name and serial numbers (in
3724 violation of RFC3280) using the OpenSSL certificate creation utilities.
3725 This is done by creating a random 64 bit value for the initial serial
3726 number when a serial number file is created or when a self signed
3727 certificate is created using 'openssl req -x509'. The initial serial
3728 number file is created using 'openssl x509 -next_serial' in CA.pl
3729 rather than being initialized to 1.
3732 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
3734 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3735 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3736 [Joe Orton, Steve Henson]
3738 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
3740 [Joe Orton, Steve Henson]
3742 *) Make it possible to have multiple active certificates with the same
3743 subject in the CA index file. This is done only if the keyword
3744 'unique_subject' is set to 'no' in the main CA section (default
3745 if 'CA_default') of the configuration file. The value is saved
3746 with the database itself in a separate index attribute file,
3747 named like the index file with '.attr' appended to the name.
3750 *) X509 verify fixes. Disable broken certificate workarounds when
3751 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
3752 keyUsage extension present. Don't accept CRLs with unhandled critical
3753 extensions: since verify currently doesn't process CRL extensions this
3754 rejects a CRL with *any* critical extensions. Add new verify error codes
3758 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
3759 A clarification of RFC2560 will require the use of OCTET STRINGs and
3760 some implementations cannot handle the current raw format. Since OpenSSL
3761 copies and compares OCSP nonces as opaque blobs without any attempt at
3762 parsing them this should not create any compatibility issues.
3765 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
3766 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
3767 this HMAC (and other) operations are several times slower than OpenSSL
3771 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
3772 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3774 *) Use the correct content when signing type "other".
3777 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
3779 *) Fix various bugs revealed by running the NISCC test suite:
3781 Stop out of bounds reads in the ASN1 code when presented with
3782 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3784 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
3786 If verify callback ignores invalid public key errors don't try to check
3787 certificate signature with the NULL public key.
3791 *) New -ignore_err option in ocsp application to stop the server
3792 exiting on the first error in a request.
3795 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3796 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3800 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3801 extra data after the compression methods not only for TLS 1.0
3802 but also for SSL 3.0 (as required by the specification).
3803 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3805 *) Change X509_certificate_type() to mark the key as exported/exportable
3806 when it's 512 *bits* long, not 512 bytes.
3809 *) Change AES_cbc_encrypt() so it outputs exact multiple of
3810 blocks during encryption.
3813 *) Various fixes to base64 BIO and non blocking I/O. On write
3814 flushes were not handled properly if the BIO retried. On read
3815 data was not being buffered properly and had various logic bugs.
3816 This also affects blocking I/O when the data being decoded is a
3820 *) Various S/MIME bugfixes and compatibility changes:
3821 output correct application/pkcs7 MIME type if
3822 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
3823 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
3824 of files as .eml work). Correctly handle very long lines in MIME
3828 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
3830 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3831 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3832 a protocol version number mismatch like a decryption error
3833 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3836 *) Turn on RSA blinding by default in the default implementation
3837 to avoid a timing attack. Applications that don't want it can call
3838 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3839 They would be ill-advised to do so in most cases.
3840 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3842 *) Change RSA blinding code so that it works when the PRNG is not
3843 seeded (in this case, the secret RSA exponent is abused as
3844 an unpredictable seed -- if it is not unpredictable, there
3845 is no point in blinding anyway). Make RSA blinding thread-safe
3846 by remembering the creator's thread ID in rsa->blinding and
3847 having all other threads use local one-time blinding factors
3848 (this requires more computation than sharing rsa->blinding, but
3849 avoids excessive locking; and if an RSA object is not shared
3850 between threads, blinding will still be very fast).
3853 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
3854 ENGINE as defaults for all supported algorithms irrespective of
3855 the 'flags' parameter. 'flags' is now honoured, so applications
3856 should make sure they are passing it correctly.
3859 *) Target "mingw" now allows native Windows code to be generated in
3860 the Cygwin environment as well as with the MinGW compiler.
3863 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
3865 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3866 via timing by performing a MAC computation even if incorrrect
3867 block cipher padding has been found. This is a countermeasure
3868 against active attacks where the attacker has to distinguish
3869 between bad padding and a MAC verification error. (CVE-2003-0078)
3871 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3872 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3873 Martin Vuagnoux (EPFL, Ilion)]
3875 *) Make the no-err option work as intended. The intention with no-err
3876 is not to have the whole error stack handling routines removed from
3877 libcrypto, it's only intended to remove all the function name and
3878 reason texts, thereby removing some of the footprint that may not
3879 be interesting if those errors aren't displayed anyway.
3881 NOTE: it's still possible for any application or module to have it's
3882 own set of error texts inserted. The routines are there, just not
3883 used by default when no-err is given.
3886 *) Add support for FreeBSD on IA64.
3887 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
3889 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
3890 Kerberos function mit_des_cbc_cksum(). Before this change,
3891 the value returned by DES_cbc_cksum() was like the one from
3892 mit_des_cbc_cksum(), except the bytes were swapped.
3893 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
3895 *) Allow an application to disable the automatic SSL chain building.
3896 Before this a rather primitive chain build was always performed in
3897 ssl3_output_cert_chain(): an application had no way to send the
3898 correct chain if the automatic operation produced an incorrect result.
3900 Now the chain builder is disabled if either:
3902 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
3904 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
3906 The reasoning behind this is that an application would not want the
3907 auto chain building to take place if extra chain certificates are
3908 present and it might also want a means of sending no additional
3909 certificates (for example the chain has two certificates and the
3913 *) Add the possibility to build without the ENGINE framework.
3914 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3916 *) Under Win32 gmtime() can return NULL: check return value in
3917 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
3920 *) DSA routines: under certain error conditions uninitialized BN objects
3921 could be freed. Solution: make sure initialization is performed early
3922 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
3923 Nils Larsch <nla@trustcenter.de> via PR#459)
3926 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
3927 checked on reconnect on the client side, therefore session resumption
3928 could still fail with a "ssl session id is different" error. This
3929 behaviour is masked when SSL_OP_ALL is used due to
3930 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
3931 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3932 followup to PR #377.
3935 *) IA-32 assembler support enhancements: unified ELF targets, support
3936 for SCO/Caldera platforms, fix for Cygwin shared build.
3939 *) Add support for FreeBSD on sparc64. As a consequence, support for
3940 FreeBSD on non-x86 processors is separate from x86 processors on
3941 the config script, much like the NetBSD support.
3942 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
3944 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
3946 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
3949 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
3950 code (06) was taken as the first octet of the session ID and the last
3951 octet was ignored consequently. As a result SSLv2 client side session
3952 caching could not have worked due to the session ID mismatch between
3954 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3958 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
3959 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
3963 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
3964 seems that in spite of existing for more than a year, many application
3965 author have done nothing to provide the necessary callbacks, which
3966 means that this particular engine will not work properly anywhere.
3967 This is a very unfortunate situation which forces us, in the name
3968 of usability, to give the hw_ncipher.c a static lock, which is part
3970 NOTE: This is for the 0.9.7 series ONLY. This hack will never
3971 appear in 0.9.8 or later. We EXPECT application authors to have
3972 dealt properly with this when 0.9.8 is released (unless we actually
3973 make such changes in the libcrypto locking code that changes will
3974 have to be made anyway).
3977 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
3978 octets have been read, EOF or an error occurs. Without this change
3979 some truncated ASN1 structures will not produce an error.
3982 *) Disable Heimdal support, since it hasn't been fully implemented.
3983 Still give the possibility to force the use of Heimdal, but with
3984 warnings and a request that patches get sent to openssl-dev.
3987 *) Add the VC-CE target, introduce the WINCE sysname, and add
3988 INSTALL.WCE and appropriate conditionals to make it build.
3989 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3991 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
3992 cygssl-x.y.z.dll, where x, y and z are the major, minor and
3993 edit numbers of the version.
3994 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3996 *) Introduce safe string copy and catenation functions
3997 (BUF_strlcpy() and BUF_strlcat()).
3998 [Ben Laurie (CHATS) and Richard Levitte]
4000 *) Avoid using fixed-size buffers for one-line DNs.
4001 [Ben Laurie (CHATS)]
4003 *) Add BUF_MEM_grow_clean() to avoid information leakage when
4004 resizing buffers containing secrets, and use where appropriate.
4005 [Ben Laurie (CHATS)]
4007 *) Avoid using fixed size buffers for configuration file location.
4008 [Ben Laurie (CHATS)]
4010 *) Avoid filename truncation for various CA files.
4011 [Ben Laurie (CHATS)]
4013 *) Use sizeof in preference to magic numbers.
4014 [Ben Laurie (CHATS)]
4016 *) Avoid filename truncation in cert requests.
4017 [Ben Laurie (CHATS)]
4019 *) Add assertions to check for (supposedly impossible) buffer
4021 [Ben Laurie (CHATS)]
4023 *) Don't cache truncated DNS entries in the local cache (this could
4024 potentially lead to a spoofing attack).
4025 [Ben Laurie (CHATS)]
4027 *) Fix various buffers to be large enough for hex/decimal
4028 representations in a platform independent manner.
4029 [Ben Laurie (CHATS)]
4031 *) Add CRYPTO_realloc_clean() to avoid information leakage when
4032 resizing buffers containing secrets, and use where appropriate.
4033 [Ben Laurie (CHATS)]
4035 *) Add BIO_indent() to avoid much slightly worrying code to do
4037 [Ben Laurie (CHATS)]
4039 *) Convert sprintf()/BIO_puts() to BIO_printf().
4040 [Ben Laurie (CHATS)]
4042 *) buffer_gets() could terminate with the buffer only half
4044 [Ben Laurie (CHATS)]
4046 *) Add assertions to prevent user-supplied crypto functions from
4047 overflowing internal buffers by having large block sizes, etc.
4048 [Ben Laurie (CHATS)]
4050 *) New OPENSSL_assert() macro (similar to assert(), but enabled
4052 [Ben Laurie (CHATS)]
4054 *) Eliminate unused copy of key in RC4.
4055 [Ben Laurie (CHATS)]
4057 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
4058 [Ben Laurie (CHATS)]
4060 *) Fix off-by-one error in EGD path.
4061 [Ben Laurie (CHATS)]
4063 *) If RANDFILE path is too long, ignore instead of truncating.
4064 [Ben Laurie (CHATS)]
4066 *) Eliminate unused and incorrectly sized X.509 structure
4068 [Ben Laurie (CHATS)]
4070 *) Eliminate unused and dangerous function knumber().
4071 [Ben Laurie (CHATS)]
4073 *) Eliminate unused and dangerous structure, KSSL_ERR.
4074 [Ben Laurie (CHATS)]
4076 *) Protect against overlong session ID context length in an encoded
4077 session object. Since these are local, this does not appear to be
4079 [Ben Laurie (CHATS)]
4081 *) Change from security patch (see 0.9.6e below) that did not affect
4082 the 0.9.6 release series:
4084 Remote buffer overflow in SSL3 protocol - an attacker could
4085 supply an oversized master key in Kerberos-enabled versions.
4087 [Ben Laurie (CHATS)]
4089 *) Change the SSL kerb5 codes to match RFC 2712.
4092 *) Make -nameopt work fully for req and add -reqopt switch.
4093 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
4095 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
4096 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
4098 *) Make sure tests can be performed even if the corresponding algorithms
4099 have been removed entirely. This was also the last step to make
4100 OpenSSL compilable with DJGPP under all reasonable conditions.
4101 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
4103 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
4104 to allow version independent disabling of normally unselected ciphers,
4105 which may be activated as a side-effect of selecting a single cipher.
4107 (E.g., cipher list string "RSA" enables ciphersuites that are left
4108 out of "ALL" because they do not provide symmetric encryption.
4109 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
4110 [Lutz Jaenicke, Bodo Moeller]
4112 *) Add appropriate support for separate platform-dependent build
4113 directories. The recommended way to make a platform-dependent
4114 build directory is the following (tested on Linux), maybe with
4117 # Place yourself outside of the OpenSSL source tree. In
4118 # this example, the environment variable OPENSSL_SOURCE
4119 # is assumed to contain the absolute OpenSSL source directory.
4120 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
4121 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
4122 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
4123 mkdir -p `dirname $F`
4124 ln -s $OPENSSL_SOURCE/$F $F
4127 To be absolutely sure not to disturb the source tree, a "make clean"
4128 is a good thing. If it isn't successfull, don't worry about it,
4129 it probably means the source directory is very clean.
4132 *) Make sure any ENGINE control commands make local copies of string
4133 pointers passed to them whenever necessary. Otherwise it is possible
4134 the caller may have overwritten (or deallocated) the original string
4135 data when a later ENGINE operation tries to use the stored values.
4136 [Götz Babin-Ebell <babinebell@trustcenter.de>]
4138 *) Improve diagnostics in file reading and command-line digests.
4139 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
4141 *) Add AES modes CFB and OFB to the object database. Correct an
4142 error in AES-CFB decryption.
4145 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
4146 allows existing EVP_CIPHER_CTX structures to be reused after
4147 calling EVP_*Final(). This behaviour is used by encryption
4148 BIOs and some applications. This has the side effect that
4149 applications must explicitly clean up cipher contexts with
4150 EVP_CIPHER_CTX_cleanup() or they will leak memory.
4153 *) Check the values of dna and dnb in bn_mul_recursive before calling
4154 bn_mul_comba (a non zero value means the a or b arrays do not contain
4155 n2 elements) and fallback to bn_mul_normal if either is not zero.
4158 *) Fix escaping of non-ASCII characters when using the -subj option
4159 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
4162 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
4163 form for "surname", serialNumber has no short form.
4164 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
4165 therefore remove "mail" short name for "internet 7".
4166 The OID for unique identifiers in X509 certificates is
4167 x500UniqueIdentifier, not uniqueIdentifier.
4168 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
4171 *) Add an "init" command to the ENGINE config module and auto initialize
4172 ENGINEs. Without any "init" command the ENGINE will be initialized
4173 after all ctrl commands have been executed on it. If init=1 the
4174 ENGINE is initailized at that point (ctrls before that point are run
4175 on the uninitialized ENGINE and after on the initialized one). If
4176 init=0 then the ENGINE will not be iniatialized at all.
4179 *) Fix the 'app_verify_callback' interface so that the user-defined
4180 argument is actually passed to the callback: In the
4181 SSL_CTX_set_cert_verify_callback() prototype, the callback
4182 declaration has been changed from
4185 int (*cb)(X509_STORE_CTX *,void *);
4186 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
4187 i=s->ctx->app_verify_callback(&ctx)
4188 has been changed into
4189 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
4191 To update applications using SSL_CTX_set_cert_verify_callback(),
4192 a dummy argument can be added to their callback functions.
4193 [D. K. Smetters <smetters@parc.xerox.com>]
4195 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
4196 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
4198 *) Add and OPENSSL_LOAD_CONF define which will cause
4199 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
4200 This allows older applications to transparently support certain
4201 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
4202 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
4203 load the config file and OPENSSL_add_all_algorithms_conf() which will
4204 always load it have also been added.
4207 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
4208 Adjust NIDs and EVP layer.
4209 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4211 *) Config modules support in openssl utility.
4213 Most commands now load modules from the config file,
4214 though in a few (such as version) this isn't done
4215 because it couldn't be used for anything.
4217 In the case of ca and req the config file used is
4218 the same as the utility itself: that is the -config
4219 command line option can be used to specify an
4223 *) Move default behaviour from OPENSSL_config(). If appname is NULL
4224 use "openssl_conf" if filename is NULL use default openssl config file.
4227 *) Add an argument to OPENSSL_config() to allow the use of an alternative
4228 config section name. Add a new flag to tolerate a missing config file
4229 and move code to CONF_modules_load_file().
4232 *) Support for crypto accelerator cards from Accelerated Encryption
4233 Processing, www.aep.ie. (Use engine 'aep')
4234 The support was copied from 0.9.6c [engine] and adapted/corrected
4235 to work with the new engine framework.
4236 [AEP Inc. and Richard Levitte]
4238 *) Support for SureWare crypto accelerator cards from Baltimore
4239 Technologies. (Use engine 'sureware')
4240 The support was copied from 0.9.6c [engine] and adapted
4241 to work with the new engine framework.
4244 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
4245 make the newer ENGINE framework commands for the CHIL engine work.
4246 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
4248 *) Make it possible to produce shared libraries on ReliantUNIX.
4249 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
4251 *) Add the configuration target debug-linux-ppro.
4252 Make 'openssl rsa' use the general key loading routines
4253 implemented in apps.c, and make those routines able to
4254 handle the key format FORMAT_NETSCAPE and the variant
4256 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4258 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4259 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4261 *) Add -keyform to rsautl, and document -engine.
4262 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
4264 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
4265 BIO_R_NO_SUCH_FILE error code rather than the generic
4266 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
4269 *) Add new functions
4271 ERR_peek_last_error_line
4272 ERR_peek_last_error_line_data.
4273 These are similar to
4276 ERR_peek_error_line_data,
4277 but report on the latest error recorded rather than the first one
4278 still in the error queue.
4279 [Ben Laurie, Bodo Moeller]
4281 *) default_algorithms option in ENGINE config module. This allows things
4283 default_algorithms = ALL
4284 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
4287 *) Prelminary ENGINE config module.
4290 *) New experimental application configuration code.
4293 *) Change the AES code to follow the same name structure as all other
4294 symmetric ciphers, and behave the same way. Move everything to
4295 the directory crypto/aes, thereby obsoleting crypto/rijndael.
4296 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4298 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
4299 [Ben Laurie and Theo de Raadt]
4301 *) Add option to output public keys in req command.
4302 [Massimiliano Pala madwolf@openca.org]
4304 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
4305 (up to about 10% better than before for P-192 and P-224).
4308 *) New functions/macros
4310 SSL_CTX_set_msg_callback(ctx, cb)
4311 SSL_CTX_set_msg_callback_arg(ctx, arg)
4312 SSL_set_msg_callback(ssl, cb)
4313 SSL_set_msg_callback_arg(ssl, arg)
4315 to request calling a callback function
4317 void cb(int write_p, int version, int content_type,
4318 const void *buf, size_t len, SSL *ssl, void *arg)
4320 whenever a protocol message has been completely received
4321 (write_p == 0) or sent (write_p == 1). Here 'version' is the
4322 protocol version according to which the SSL library interprets
4323 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
4324 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
4325 the content type as defined in the SSL 3.0/TLS 1.0 protocol
4326 specification (change_cipher_spec(20), alert(21), handshake(22)).
4327 'buf' and 'len' point to the actual message, 'ssl' to the
4328 SSL object, and 'arg' is the application-defined value set by
4329 SSL[_CTX]_set_msg_callback_arg().
4331 'openssl s_client' and 'openssl s_server' have new '-msg' options
4332 to enable a callback that displays all protocol messages.
4335 *) Change the shared library support so shared libraries are built as
4336 soon as the corresponding static library is finished, and thereby get
4337 openssl and the test programs linked against the shared library.
4338 This still only happens when the keyword "shard" has been given to
4339 the configuration scripts.
4341 NOTE: shared library support is still an experimental thing, and
4342 backward binary compatibility is still not guaranteed.
4343 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
4345 *) Add support for Subject Information Access extension.
4346 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4348 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
4349 additional bytes when new memory had to be allocated, not just
4350 when reusing an existing buffer.
4353 *) New command line and configuration option 'utf8' for the req command.
4354 This allows field values to be specified as UTF8 strings.
4357 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
4358 runs for the former and machine-readable output for the latter.
4361 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
4362 of the e-mail address in the DN (i.e., it will go into a certificate
4363 extension only). The new configuration file option 'email_in_dn = no'
4364 has the same effect.
4365 [Massimiliano Pala madwolf@openca.org]
4367 *) Change all functions with names starting with des_ to be starting
4368 with DES_ instead. Add wrappers that are compatible with libdes,
4369 but are named _ossl_old_des_*. Finally, add macros that map the
4370 des_* symbols to the corresponding _ossl_old_des_* if libdes
4371 compatibility is desired. If OpenSSL 0.9.6c compatibility is
4372 desired, the des_* symbols will be mapped to DES_*, with one
4375 Since we provide two compatibility mappings, the user needs to
4376 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
4377 compatibility is desired. The default (i.e., when that macro
4378 isn't defined) is OpenSSL 0.9.6c compatibility.
4380 There are also macros that enable and disable the support of old
4381 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
4382 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
4383 are defined, the default will apply: to support the old des routines.
4385 In either case, one must include openssl/des.h to get the correct
4386 definitions. Do not try to just include openssl/des_old.h, that
4389 NOTE: This is a major break of an old API into a new one. Software
4390 authors are encouraged to switch to the DES_ style functions. Some
4391 time in the future, des_old.h and the libdes compatibility functions
4392 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
4393 default), and then completely removed.
4396 *) Test for certificates which contain unsupported critical extensions.
4397 If such a certificate is found during a verify operation it is
4398 rejected by default: this behaviour can be overridden by either
4399 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
4400 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
4401 X509_supported_extension() has also been added which returns 1 if a
4402 particular extension is supported.
4405 *) Modify the behaviour of EVP cipher functions in similar way to digests
4406 to retain compatibility with existing code.
4409 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
4410 compatibility with existing code. In particular the 'ctx' parameter does
4411 not have to be to be initialized before the call to EVP_DigestInit() and
4412 it is tidied up after a call to EVP_DigestFinal(). New function
4413 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
4414 EVP_MD_CTX_copy() changed to not require the destination to be
4415 initialized valid and new function EVP_MD_CTX_copy_ex() added which
4416 requires the destination to be valid.
4418 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
4419 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
4422 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
4423 so that complete 'Handshake' protocol structures are kept in memory
4424 instead of overwriting 'msg_type' and 'length' with 'body' data.
4427 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
4428 [Massimo Santin via Richard Levitte]
4430 *) Major restructuring to the underlying ENGINE code. This includes
4431 reduction of linker bloat, separation of pure "ENGINE" manipulation
4432 (initialisation, etc) from functionality dealing with implementations
4433 of specific crypto iterfaces. This change also introduces integrated
4434 support for symmetric ciphers and digest implementations - so ENGINEs
4435 can now accelerate these by providing EVP_CIPHER and EVP_MD
4436 implementations of their own. This is detailed in crypto/engine/README
4437 as it couldn't be adequately described here. However, there are a few
4438 API changes worth noting - some RSA, DSA, DH, and RAND functions that
4439 were changed in the original introduction of ENGINE code have now
4440 reverted back - the hooking from this code to ENGINE is now a good
4441 deal more passive and at run-time, operations deal directly with
4442 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
4443 dereferencing through an ENGINE pointer any more. Also, the ENGINE
4444 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
4445 they were not being used by the framework as there is no concept of a
4446 BIGNUM_METHOD and they could not be generalised to the new
4447 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
4448 ENGINE_cpy() has been removed as it cannot be consistently defined in
4452 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
4455 *) Change mkdef.pl to sort symbols that get the same entry number,
4456 and make sure the automatically generated functions ERR_load_*
4457 become part of libeay.num as well.
4460 *) New function SSL_renegotiate_pending(). This returns true once
4461 renegotiation has been requested (either SSL_renegotiate() call
4462 or HelloRequest/ClientHello receveived from the peer) and becomes
4463 false once a handshake has been completed.
4464 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
4465 sends a HelloRequest, but does not ensure that a handshake takes
4466 place. SSL_renegotiate_pending() is useful for checking if the
4467 client has followed the request.)
4470 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
4471 By default, clients may request session resumption even during
4472 renegotiation (if session ID contexts permit); with this option,
4473 session resumption is possible only in the first handshake.
4475 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4476 more bits available for options that should not be part of
4477 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4480 *) Add some demos for certificate and certificate request creation.
4483 *) Make maximum certificate chain size accepted from the peer application
4484 settable (SSL*_get/set_max_cert_list()), as proposed by
4485 "Douglas E. Engert" <deengert@anl.gov>.
4488 *) Add support for shared libraries for Unixware-7
4489 (Boyd Lynn Gerber <gerberb@zenez.com>).
4492 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4493 be done prior to destruction. Use this to unload error strings from
4494 ENGINEs that load their own error strings. NB: This adds two new API
4495 functions to "get" and "set" this destroy handler in an ENGINE.
4498 *) Alter all existing ENGINE implementations (except "openssl" and
4499 "openbsd") to dynamically instantiate their own error strings. This
4500 makes them more flexible to be built both as statically-linked ENGINEs
4501 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4502 Also, add stub code to each that makes building them as self-contained
4503 shared-libraries easier (see README.ENGINE).
4506 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4507 implementations into applications that are completely implemented in
4508 self-contained shared-libraries. The "dynamic" ENGINE exposes control
4509 commands that can be used to configure what shared-library to load and
4510 to control aspects of the way it is handled. Also, made an update to
4511 the README.ENGINE file that brings its information up-to-date and
4512 provides some information and instructions on the "dynamic" ENGINE
4513 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
4516 *) Make it possible to unload ranges of ERR strings with a new
4517 "ERR_unload_strings" function.
4520 *) Add a copy() function to EVP_MD.
4523 *) Make EVP_MD routines take a context pointer instead of just the
4524 md_data void pointer.
4527 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
4528 that the digest can only process a single chunk of data
4529 (typically because it is provided by a piece of
4530 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
4531 is only going to provide a single chunk of data, and hence the
4532 framework needn't accumulate the data for oneshot drivers.
4535 *) As with "ERR", make it possible to replace the underlying "ex_data"
4536 functions. This change also alters the storage and management of global
4537 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
4538 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
4539 index counters. The API functions that use this state have been changed
4540 to take a "class_index" rather than pointers to the class's local STACK
4541 and counter, and there is now an API function to dynamically create new
4542 classes. This centralisation allows us to (a) plug a lot of the
4543 thread-safety problems that existed, and (b) makes it possible to clean
4544 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
4545 such data would previously have always leaked in application code and
4546 workarounds were in place to make the memory debugging turn a blind eye
4547 to it. Application code that doesn't use this new function will still
4548 leak as before, but their memory debugging output will announce it now
4549 rather than letting it slide.
4551 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
4552 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
4553 has a return value to indicate success or failure.
4556 *) Make it possible to replace the underlying "ERR" functions such that the
4557 global state (2 LHASH tables and 2 locks) is only used by the "default"
4558 implementation. This change also adds two functions to "get" and "set"
4559 the implementation prior to it being automatically set the first time
4560 any other ERR function takes place. Ie. an application can call "get",
4561 pass the return value to a module it has just loaded, and that module
4562 can call its own "set" function using that value. This means the
4563 module's "ERR" operations will use (and modify) the error state in the
4564 application and not in its own statically linked copy of OpenSSL code.
4567 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
4568 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
4569 the operation, and provides a more encapsulated way for external code
4570 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
4571 to use these functions rather than manually incrementing the counts.
4573 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
4576 *) Add EVP test program.
4579 *) Add symmetric cipher support to ENGINE. Expect the API to change!
4582 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
4583 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
4584 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
4585 These allow a CRL to be built without having to access X509_CRL fields
4586 directly. Modify 'ca' application to use new functions.
4589 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
4590 bug workarounds. Rollback attack detection is a security feature.
4591 The problem will only arise on OpenSSL servers when TLSv1 is not
4592 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
4593 Software authors not wanting to support TLSv1 will have special reasons
4594 for their choice and can explicitly enable this option.
4595 [Bodo Moeller, Lutz Jaenicke]
4597 *) Rationalise EVP so it can be extended: don't include a union of
4598 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
4599 (similar to those existing for EVP_CIPHER_CTX).
4604 EVP_MD_CTX_init(&md); /* new function call */
4605 EVP_DigestInit(&md, EVP_sha1());
4606 EVP_DigestUpdate(&md, in, len);
4607 EVP_DigestFinal(&md, out, NULL);
4608 EVP_MD_CTX_cleanup(&md); /* new function call */
4612 *) Make DES key schedule conform to the usual scheme, as well as
4613 correcting its structure. This means that calls to DES functions
4614 now have to pass a pointer to a des_key_schedule instead of a
4615 plain des_key_schedule (which was actually always a pointer
4618 des_key_schedule ks;
4620 des_set_key_checked(..., &ks);
4621 des_ncbc_encrypt(..., &ks, ...);
4623 (Note that a later change renames 'des_...' into 'DES_...'.)
4626 *) Initial reduction of linker bloat: the use of some functions, such as
4627 PEM causes large amounts of unused functions to be linked in due to
4628 poor organisation. For example pem_all.c contains every PEM function
4629 which has a knock on effect of linking in large amounts of (unused)
4630 ASN1 code. Grouping together similar functions and splitting unrelated
4631 functions prevents this.
4634 *) Cleanup of EVP macros.
4637 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
4638 correct _ecb suffix.
4641 *) Add initial OCSP responder support to ocsp application. The
4642 revocation information is handled using the text based index
4643 use by the ca application. The responder can either handle
4644 requests generated internally, supplied in files (for example
4645 via a CGI script) or using an internal minimal server.
4648 *) Add configuration choices to get zlib compression for TLS.
4651 *) Changes to Kerberos SSL for RFC 2712 compliance:
4652 1. Implemented real KerberosWrapper, instead of just using
4653 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
4654 2. Implemented optional authenticator field of KerberosWrapper.
4656 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
4657 and authenticator structs; see crypto/krb5/.
4659 Generalized Kerberos calls to support multiple Kerberos libraries.
4660 [Vern Staats <staatsvr@asc.hpc.mil>,
4661 Jeffrey Altman <jaltman@columbia.edu>
4662 via Richard Levitte]
4664 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
4665 already does with RSA. testdsa.h now has 'priv_key/pub_key'
4666 values for each of the key sizes rather than having just
4667 parameters (and 'speed' generating keys each time).
4670 *) Speed up EVP routines.
4673 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
4674 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
4675 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
4676 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
4678 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
4679 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
4680 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
4683 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
4685 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
4688 *) Added the OS2-EMX target.
4689 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
4691 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
4692 to support NCONF routines in extension code. New function CONF_set_nconf()
4693 to allow functions which take an NCONF to also handle the old LHASH
4694 structure: this means that the old CONF compatible routines can be
4695 retained (in particular wrt extensions) without having to duplicate the
4696 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
4699 *) Enhance the general user interface with mechanisms for inner control
4700 and with possibilities to have yes/no kind of prompts.
4703 *) Change all calls to low level digest routines in the library and
4704 applications to use EVP. Add missing calls to HMAC_cleanup() and
4705 don't assume HMAC_CTX can be copied using memcpy().
4706 [Verdon Walker <VWalker@novell.com>, Steve Henson]
4708 *) Add the possibility to control engines through control names but with
4709 arbitrary arguments instead of just a string.
4710 Change the key loaders to take a UI_METHOD instead of a callback
4711 function pointer. NOTE: this breaks binary compatibility with earlier
4712 versions of OpenSSL [engine].
4713 Adapt the nCipher code for these new conditions and add a card insertion
4717 *) Enhance the general user interface with mechanisms to better support
4718 dialog box interfaces, application-defined prompts, the possibility
4719 to use defaults (for example default passwords from somewhere else)
4720 and interrupts/cancellations.
4723 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
4724 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
4727 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
4728 tidy up some unnecessarily weird code in 'sk_new()').
4729 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
4731 *) Change the key loading routines for ENGINEs to use the same kind
4732 callback (pem_password_cb) as all other routines that need this
4736 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
4737 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
4738 than this minimum value is recommended.
4741 *) New random seeder for OpenVMS, using the system process statistics
4742 that are easily reachable.
4745 *) Windows apparently can't transparently handle global
4746 variables defined in DLLs. Initialisations such as:
4748 const ASN1_ITEM *it = &ASN1_INTEGER_it;
4750 wont compile. This is used by the any applications that need to
4751 declare their own ASN1 modules. This was fixed by adding the option
4752 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
4753 needed for static libraries under Win32.
4756 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
4757 setting of purpose and trust fields. New X509_STORE trust and
4758 purpose functions and tidy up setting in other SSL functions.
4761 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
4762 structure. These are inherited by X509_STORE_CTX when it is
4763 initialised. This allows various defaults to be set in the
4764 X509_STORE structure (such as flags for CRL checking and custom
4765 purpose or trust settings) for functions which only use X509_STORE_CTX
4766 internally such as S/MIME.
4768 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
4769 trust settings if they are not set in X509_STORE. This allows X509_STORE
4770 purposes and trust (in S/MIME for example) to override any set by default.
4772 Add command line options for CRL checking to smime, s_client and s_server
4776 *) Initial CRL based revocation checking. If the CRL checking flag(s)
4777 are set then the CRL is looked up in the X509_STORE structure and
4778 its validity and signature checked, then if the certificate is found
4779 in the CRL the verify fails with a revoked error.
4781 Various new CRL related callbacks added to X509_STORE_CTX structure.
4783 Command line options added to 'verify' application to support this.
4785 This needs some additional work, such as being able to handle multiple
4786 CRLs with different times, extension based lookup (rather than just
4787 by subject name) and ultimately more complete V2 CRL extension
4791 *) Add a general user interface API (crypto/ui/). This is designed
4792 to replace things like des_read_password and friends (backward
4793 compatibility functions using this new API are provided).
4794 The purpose is to remove prompting functions from the DES code
4795 section as well as provide for prompting through dialog boxes in
4796 a window system and the like.
4799 *) Add "ex_data" support to ENGINE so implementations can add state at a
4800 per-structure level rather than having to store it globally.
4803 *) Make it possible for ENGINE structures to be copied when retrieved by
4804 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
4805 This causes the "original" ENGINE structure to act like a template,
4806 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
4807 operational state can be localised to each ENGINE structure, despite the
4808 fact they all share the same "methods". New ENGINE structures returned in
4809 this case have no functional references and the return value is the single
4810 structural reference. This matches the single structural reference returned
4811 by ENGINE_by_id() normally, when it is incremented on the pre-existing
4815 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
4816 needs to match any other type at all we need to manually clear the
4820 *) Changes to the "openssl engine" utility to include;
4821 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
4822 about an ENGINE's available control commands.
4823 - executing control commands from command line arguments using the
4824 '-pre' and '-post' switches. '-post' is only used if '-t' is
4825 specified and the ENGINE is successfully initialised. The syntax for
4826 the individual commands are colon-separated, for example;
4827 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
4830 *) New dynamic control command support for ENGINEs. ENGINEs can now
4831 declare their own commands (numbers), names (strings), descriptions,
4832 and input types for run-time discovery by calling applications. A
4833 subset of these commands are implicitly classed as "executable"
4834 depending on their input type, and only these can be invoked through
4835 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
4836 can be based on user input, config files, etc). The distinction is
4837 that "executable" commands cannot return anything other than a boolean
4838 result and can only support numeric or string input, whereas some
4839 discoverable commands may only be for direct use through
4840 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
4841 pointers, or other custom uses. The "executable" commands are to
4842 support parameterisations of ENGINE behaviour that can be
4843 unambiguously defined by ENGINEs and used consistently across any
4844 OpenSSL-based application. Commands have been added to all the
4845 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
4846 control over shared-library paths without source code alterations.
4849 *) Changed all ENGINE implementations to dynamically allocate their
4850 ENGINEs rather than declaring them statically. Apart from this being
4851 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
4852 this also allows the implementations to compile without using the
4853 internal engine_int.h header.
4856 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
4857 'const' value. Any code that should be able to modify a RAND_METHOD
4858 should already have non-const pointers to it (ie. they should only
4859 modify their own ones).
4862 *) Made a variety of little tweaks to the ENGINE code.
4863 - "atalla" and "ubsec" string definitions were moved from header files
4864 to C code. "nuron" string definitions were placed in variables
4865 rather than hard-coded - allowing parameterisation of these values
4866 later on via ctrl() commands.
4867 - Removed unused "#if 0"'d code.
4868 - Fixed engine list iteration code so it uses ENGINE_free() to release
4869 structural references.
4870 - Constified the RAND_METHOD element of ENGINE structures.
4871 - Constified various get/set functions as appropriate and added
4872 missing functions (including a catch-all ENGINE_cpy that duplicates
4873 all ENGINE values onto a new ENGINE except reference counts/state).
4874 - Removed NULL parameter checks in get/set functions. Setting a method
4875 or function to NULL is a way of cancelling out a previously set
4876 value. Passing a NULL ENGINE parameter is just plain stupid anyway
4877 and doesn't justify the extra error symbols and code.
4878 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
4879 flags from engine_int.h to engine.h.
4880 - Changed prototypes for ENGINE handler functions (init(), finish(),
4881 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
4884 *) Implement binary inversion algorithm for BN_mod_inverse in addition
4885 to the algorithm using long division. The binary algorithm can be
4886 used only if the modulus is odd. On 32-bit systems, it is faster
4887 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
4888 roughly 5-15% for 256-bit moduli), so we use it only for moduli
4889 up to 450 bits. In 64-bit environments, the binary algorithm
4890 appears to be advantageous for much longer moduli; here we use it
4891 for moduli up to 2048 bits.
4894 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
4895 could not support the combine flag in choice fields.
4898 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
4899 extensions from a certificate request to the certificate.
4902 *) Allow multiple 'certopt' and 'nameopt' options to be separated
4903 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
4904 file: this allows the display of the certificate about to be
4905 signed to be customised, to allow certain fields to be included
4906 or excluded and extension details. The old system didn't display
4907 multicharacter strings properly, omitted fields not in the policy
4908 and couldn't display additional details such as extensions.
4911 *) Function EC_POINTs_mul for multiple scalar multiplication
4912 of an arbitrary number of elliptic curve points
4913 \sum scalars[i]*points[i],
4914 optionally including the generator defined for the EC_GROUP:
4915 scalar*generator + \sum scalars[i]*points[i].
4917 EC_POINT_mul is a simple wrapper function for the typical case
4918 that the point list has just one item (besides the optional
4922 *) First EC_METHODs for curves over GF(p):
4924 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
4925 operations and provides various method functions that can also
4926 operate with faster implementations of modular arithmetic.
4928 EC_GFp_mont_method() reuses most functions that are part of
4929 EC_GFp_simple_method, but uses Montgomery arithmetic.
4931 [Bodo Moeller; point addition and point doubling
4932 implementation directly derived from source code provided by
4933 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
4935 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
4936 crypto/ec/ec_lib.c):
4938 Curves are EC_GROUP objects (with an optional group generator)
4939 based on EC_METHODs that are built into the library.
4941 Points are EC_POINT objects based on EC_GROUP objects.
4943 Most of the framework would be able to handle curves over arbitrary
4944 finite fields, but as there are no obvious types for fields other
4945 than GF(p), some functions are limited to that for now.
4948 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
4949 that the file contains a complete HTTP response.
4952 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
4953 change the def and num file printf format specifier from "%-40sXXX"
4954 to "%-39s XXX". The latter will always guarantee a space after the
4955 field while the former will cause them to run together if the field
4956 is 40 of more characters long.
4959 *) Constify the cipher and digest 'method' functions and structures
4960 and modify related functions to take constant EVP_MD and EVP_CIPHER
4964 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
4965 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
4968 *) Modify EVP_Digest*() routines so they now return values. Although the
4969 internal software routines can never fail additional hardware versions
4973 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
4975 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
4976 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
4979 ERR_R_NESTED_ASN1_ERROR
4981 ERR_R_MISSING_ASN1_EOS
4982 were 4 .. 9, conflicting with
4983 ERR_LIB_RSA (= ERR_R_RSA_LIB)
4985 ERR_LIB_PEM (= ERR_R_PEM_LIB).
4986 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
4988 Add new error code 'ERR_R_INTERNAL_ERROR'.
4991 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
4995 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
4996 sets the subject name for a new request or supersedes the
4997 subject name in a given request. Formats that can be parsed are
4998 'CN=Some Name, OU=myOU, C=IT'
5000 'CN=Some Name/OU=myOU/C=IT'.
5002 Add options '-batch' and '-verbose' to 'openssl req'.
5003 [Massimiliano Pala <madwolf@hackmasters.net>]
5005 *) Introduce the possibility to access global variables through
5006 functions on platform were that's the best way to handle exporting
5007 global variables in shared libraries. To enable this functionality,
5008 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
5009 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
5010 is normally done by Configure or something similar).
5012 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
5013 in the source file (foo.c) like this:
5015 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
5016 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
5018 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
5019 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
5021 OPENSSL_DECLARE_GLOBAL(int,foo);
5022 #define foo OPENSSL_GLOBAL_REF(foo)
5023 OPENSSL_DECLARE_GLOBAL(double,bar);
5024 #define bar OPENSSL_GLOBAL_REF(bar)
5026 The #defines are very important, and therefore so is including the
5027 header file everywhere where the defined globals are used.
5029 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
5030 of ASN.1 items, but that structure is a bit different.
5032 The largest change is in util/mkdef.pl which has been enhanced with
5033 better and easier to understand logic to choose which symbols should
5034 go into the Windows .def files as well as a number of fixes and code
5035 cleanup (among others, algorithm keywords are now sorted
5036 lexicographically to avoid constant rewrites).
5039 *) In BN_div() keep a copy of the sign of 'num' before writing the
5040 result to 'rm' because if rm==num the value will be overwritten
5041 and produce the wrong result if 'num' is negative: this caused
5042 problems with BN_mod() and BN_nnmod().
5045 *) Function OCSP_request_verify(). This checks the signature on an
5046 OCSP request and verifies the signer certificate. The signer
5047 certificate is just checked for a generic purpose and OCSP request
5051 *) Add OCSP_check_validity() function to check the validity of OCSP
5052 responses. OCSP responses are prepared in real time and may only
5053 be a few seconds old. Simply checking that the current time lies
5054 between thisUpdate and nextUpdate max reject otherwise valid responses
5055 caused by either OCSP responder or client clock inaccuracy. Instead
5056 we allow thisUpdate and nextUpdate to fall within a certain period of
5057 the current time. The age of the response can also optionally be
5058 checked. Two new options -validity_period and -status_age added to
5062 *) If signature or public key algorithm is unrecognized print out its
5063 OID rather that just UNKNOWN.
5066 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
5067 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
5068 ID to be generated from the issuer certificate alone which can then be
5069 passed to OCSP_id_issuer_cmp().
5072 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
5073 ASN1 modules to export functions returning ASN1_ITEM pointers
5074 instead of the ASN1_ITEM structures themselves. This adds several
5075 new macros which allow the underlying ASN1 function/structure to
5076 be accessed transparently. As a result code should not use ASN1_ITEM
5077 references directly (such as &X509_it) but instead use the relevant
5078 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
5079 use of the new ASN1 code on platforms where exporting structures
5080 is problematical (for example in shared libraries) but exporting
5081 functions returning pointers to structures is not.
5084 *) Add support for overriding the generation of SSL/TLS session IDs.
5085 These callbacks can be registered either in an SSL_CTX or per SSL.
5086 The purpose of this is to allow applications to control, if they wish,
5087 the arbitrary values chosen for use as session IDs, particularly as it
5088 can be useful for session caching in multiple-server environments. A
5089 command-line switch for testing this (and any client code that wishes
5090 to use such a feature) has been added to "s_server".
5091 [Geoff Thorpe, Lutz Jaenicke]
5093 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
5094 of the form '#if defined(...) || defined(...) || ...' and
5095 '#if !defined(...) && !defined(...) && ...'. This also avoids
5096 the growing number of special cases it was previously handling.
5099 *) Make all configuration macros available for application by making
5100 sure they are available in opensslconf.h, by giving them names starting
5101 with "OPENSSL_" to avoid conflicts with other packages and by making
5102 sure e_os2.h will cover all platform-specific cases together with
5104 Additionally, it is now possible to define configuration/platform-
5105 specific names (called "system identities"). In the C code, these
5106 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
5107 macro with the name beginning with "OPENSSL_SYS_", which is determined
5108 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
5112 *) New option -set_serial to 'req' and 'x509' this allows the serial
5113 number to use to be specified on the command line. Previously self
5114 signed certificates were hard coded with serial number 0 and the
5115 CA options of 'x509' had to use a serial number in a file which was
5119 *) New options to 'ca' utility to support V2 CRL entry extensions.
5120 Currently CRL reason, invalidity date and hold instruction are
5121 supported. Add new CRL extensions to V3 code and some new objects.
5124 *) New function EVP_CIPHER_CTX_set_padding() this is used to
5125 disable standard block padding (aka PKCS#5 padding) in the EVP
5126 API, which was previously mandatory. This means that the data is
5127 not padded in any way and so the total length much be a multiple
5128 of the block size, otherwise an error occurs.
5131 *) Initial (incomplete) OCSP SSL support.
5134 *) New function OCSP_parse_url(). This splits up a URL into its host,
5135 port and path components: primarily to parse OCSP URLs. New -url
5136 option to ocsp utility.
5139 *) New nonce behavior. The return value of OCSP_check_nonce() now
5140 reflects the various checks performed. Applications can decide
5141 whether to tolerate certain situations such as an absent nonce
5142 in a response when one was present in a request: the ocsp application
5143 just prints out a warning. New function OCSP_add1_basic_nonce()
5144 this is to allow responders to include a nonce in a response even if
5145 the request is nonce-less.
5148 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
5149 skipped when using openssl x509 multiple times on a single input file,
5150 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
5153 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
5154 set string type: to handle setting ASN1_TIME structures. Fix ca
5155 utility to correctly initialize revocation date of CRLs.
5158 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
5159 the clients preferred ciphersuites and rather use its own preferences.
5160 Should help to work around M$ SGC (Server Gated Cryptography) bug in
5161 Internet Explorer by ensuring unchanged hash method during stepup.
5162 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
5165 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
5166 to aes and add a new 'exist' option to print out symbols that don't
5170 *) Additional options to ocsp utility to allow flags to be set and
5171 additional certificates supplied.
5174 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
5175 OCSP client a number of certificate to only verify the response
5179 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
5180 handle the new API. Currently only ECB, CBC modes supported. Add new
5183 Add TLS AES ciphersuites as described in RFC3268, "Advanced
5184 Encryption Standard (AES) Ciphersuites for Transport Layer
5185 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
5186 not enabled by default and were not part of the "ALL" ciphersuite
5187 alias because they were not yet official; they could be
5188 explicitly requested by specifying the "AESdraft" ciphersuite
5189 group alias. In the final release of OpenSSL 0.9.7, the group
5190 alias is called "AES" and is part of "ALL".)
5191 [Ben Laurie, Steve Henson, Bodo Moeller]
5193 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
5194 request to response.
5197 *) Functions for OCSP responders. OCSP_request_onereq_count(),
5198 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
5199 extract information from a certificate request. OCSP_response_create()
5200 creates a response and optionally adds a basic response structure.
5201 OCSP_basic_add1_status() adds a complete single response to a basic
5202 response and returns the OCSP_SINGLERESP structure just added (to allow
5203 extensions to be included for example). OCSP_basic_add1_cert() adds a
5204 certificate to a basic response and OCSP_basic_sign() signs a basic
5205 response with various flags. New helper functions ASN1_TIME_check()
5206 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
5207 (converts ASN1_TIME to GeneralizedTime).
5210 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
5211 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
5212 structure from a certificate. X509_pubkey_digest() digests the public_key
5213 contents: this is used in various key identifiers.
5216 *) Make sk_sort() tolerate a NULL argument.
5217 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
5219 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
5220 passed by the function are trusted implicitly. If any of them signed the
5221 response then it is assumed to be valid and is not verified.
5224 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
5225 to data. This was previously part of the PKCS7 ASN1 code. This
5226 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
5227 [Steve Henson, reported by Kenneth R. Robinette
5228 <support@securenetterm.com>]
5230 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
5231 routines: without these tracing memory leaks is very painful.
5232 Fix leaks in PKCS12 and PKCS7 routines.
5235 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
5236 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
5237 effectively meant GeneralizedTime would never be used. Now it
5238 is initialised to -1 but X509_time_adj() now has to check the value
5239 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
5240 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
5241 [Steve Henson, reported by Kenneth R. Robinette
5242 <support@securenetterm.com>]
5244 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
5245 result in a zero length in the ASN1_INTEGER structure which was
5246 not consistent with the structure when d2i_ASN1_INTEGER() was used
5247 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
5248 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
5249 where it did not print out a minus for negative ASN1_INTEGER.
5252 *) Add summary printout to ocsp utility. The various functions which
5253 convert status values to strings have been renamed to:
5254 OCSP_response_status_str(), OCSP_cert_status_str() and
5255 OCSP_crl_reason_str() and are no longer static. New options
5256 to verify nonce values and to disable verification. OCSP response
5257 printout format cleaned up.
5260 *) Add additional OCSP certificate checks. These are those specified
5261 in RFC2560. This consists of two separate checks: the CA of the
5262 certificate being checked must either be the OCSP signer certificate
5263 or the issuer of the OCSP signer certificate. In the latter case the
5264 OCSP signer certificate must contain the OCSP signing extended key
5265 usage. This check is performed by attempting to match the OCSP
5266 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
5267 in the OCSP_CERTID structures of the response.
5270 *) Initial OCSP certificate verification added to OCSP_basic_verify()
5271 and related routines. This uses the standard OpenSSL certificate
5272 verify routines to perform initial checks (just CA validity) and
5273 to obtain the certificate chain. Then additional checks will be
5274 performed on the chain. Currently the root CA is checked to see
5275 if it is explicitly trusted for OCSP signing. This is used to set
5276 a root CA as a global signing root: that is any certificate that
5277 chains to that CA is an acceptable OCSP signing certificate.
5280 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
5281 extensions from a separate configuration file.
5282 As when reading extensions from the main configuration file,
5283 the '-extensions ...' option may be used for specifying the
5285 [Massimiliano Pala <madwolf@comune.modena.it>]
5287 *) New OCSP utility. Allows OCSP requests to be generated or
5288 read. The request can be sent to a responder and the output
5289 parsed, outputed or printed in text form. Not complete yet:
5290 still needs to check the OCSP response validity.
5293 *) New subcommands for 'openssl ca':
5294 'openssl ca -status <serial>' prints the status of the cert with
5295 the given serial number (according to the index file).
5296 'openssl ca -updatedb' updates the expiry status of certificates
5298 [Massimiliano Pala <madwolf@comune.modena.it>]
5300 *) New '-newreq-nodes' command option to CA.pl. This is like
5301 '-newreq', but calls 'openssl req' with the '-nodes' option
5302 so that the resulting key is not encrypted.
5303 [Damien Miller <djm@mindrot.org>]
5305 *) New configuration for the GNU Hurd.
5306 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
5308 *) Initial code to implement OCSP basic response verify. This
5309 is currently incomplete. Currently just finds the signer's
5310 certificate and verifies the signature on the response.
5313 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
5314 value of OPENSSLDIR. This is available via the new '-d' option
5315 to 'openssl version', and is also included in 'openssl version -a'.
5318 *) Allowing defining memory allocation callbacks that will be given
5319 file name and line number information in additional arguments
5320 (a const char* and an int). The basic functionality remains, as
5321 well as the original possibility to just replace malloc(),
5322 realloc() and free() by functions that do not know about these
5323 additional arguments. To register and find out the current
5324 settings for extended allocation functions, the following
5325 functions are provided:
5327 CRYPTO_set_mem_ex_functions
5328 CRYPTO_set_locked_mem_ex_functions
5329 CRYPTO_get_mem_ex_functions
5330 CRYPTO_get_locked_mem_ex_functions
5332 These work the same way as CRYPTO_set_mem_functions and friends.
5333 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
5334 extended allocation function is enabled.
5335 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
5336 a conventional allocation function is enabled.
5337 [Richard Levitte, Bodo Moeller]
5339 *) Finish off removing the remaining LHASH function pointer casts.
5340 There should no longer be any prototype-casting required when using
5341 the LHASH abstraction, and any casts that remain are "bugs". See
5342 the callback types and macros at the head of lhash.h for details
5343 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
5346 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
5347 If /dev/[u]random devices are not available or do not return enough
5348 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
5350 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
5351 /etc/entropy will be queried once each in this sequence, quering stops
5352 when enough entropy was collected without querying more sockets.
5355 *) Change the Unix RAND_poll() variant to be able to poll several
5356 random devices, as specified by DEVRANDOM, until a sufficient amount
5357 of data has been collected. We spend at most 10 ms on each file
5358 (select timeout) and read in non-blocking mode. DEVRANDOM now
5359 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
5360 (previously it was just the string "/dev/urandom"), so on typical
5361 platforms the 10 ms delay will never occur.
5362 Also separate out the Unix variant to its own file, rand_unix.c.
5363 For VMS, there's a currently-empty rand_vms.c.
5366 *) Move OCSP client related routines to ocsp_cl.c. These
5367 provide utility functions which an application needing
5368 to issue a request to an OCSP responder and analyse the
5369 response will typically need: as opposed to those which an
5370 OCSP responder itself would need which will be added later.
5372 OCSP_request_sign() signs an OCSP request with an API similar
5373 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
5374 response. OCSP_response_get1_basic() extracts basic response
5375 from response. OCSP_resp_find_status(): finds and extracts status
5376 information from an OCSP_CERTID structure (which will be created
5377 when the request structure is built). These are built from lower
5378 level functions which work on OCSP_SINGLERESP structures but
5379 wont normally be used unless the application wishes to examine
5380 extensions in the OCSP response for example.
5382 Replace nonce routines with a pair of functions.
5383 OCSP_request_add1_nonce() adds a nonce value and optionally
5384 generates a random value. OCSP_check_nonce() checks the
5385 validity of the nonce in an OCSP response.
5388 *) Change function OCSP_request_add() to OCSP_request_add0_id().
5389 This doesn't copy the supplied OCSP_CERTID and avoids the
5390 need to free up the newly created id. Change return type
5391 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
5392 This can then be used to add extensions to the request.
5393 Deleted OCSP_request_new(), since most of its functionality
5394 is now in OCSP_REQUEST_new() (and the case insensitive name
5395 clash) apart from the ability to set the request name which
5396 will be added elsewhere.
5399 *) Update OCSP API. Remove obsolete extensions argument from
5400 various functions. Extensions are now handled using the new
5401 OCSP extension code. New simple OCSP HTTP function which
5402 can be used to send requests and parse the response.
5405 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
5406 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
5407 uses the special reorder version of SET OF to sort the attributes
5408 and reorder them to match the encoded order. This resolves a long
5409 standing problem: a verify on a PKCS7 structure just after signing
5410 it used to fail because the attribute order did not match the
5411 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
5412 it uses the received order. This is necessary to tolerate some broken
5413 software that does not order SET OF. This is handled by encoding
5414 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
5415 to produce the required SET OF.
5418 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
5419 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
5420 files to get correct declarations of the ASN.1 item variables.
5423 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
5424 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
5425 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
5426 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
5427 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
5428 ASN1_ITEM and no wrapper functions.
5431 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
5432 replace the old function pointer based I/O routines. Change most of
5433 the *_d2i_bio() and *_d2i_fp() functions to use these.
5436 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
5437 lines, recognice more "algorithms" that can be deselected, and make
5438 it complain about algorithm deselection that isn't recognised.
5441 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
5442 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
5443 to use new functions. Add NO_ASN1_OLD which can be set to remove
5444 some old style ASN1 functions: this can be used to determine if old
5445 code will still work when these eventually go away.
5448 *) New extension functions for OCSP structures, these follow the
5449 same conventions as certificates and CRLs.
5452 *) New function X509V3_add1_i2d(). This automatically encodes and
5453 adds an extension. Its behaviour can be customised with various
5454 flags to append, replace or delete. Various wrappers added for
5455 certifcates and CRLs.
5458 *) Fix to avoid calling the underlying ASN1 print routine when
5459 an extension cannot be parsed. Correct a typo in the
5460 OCSP_SERVICELOC extension. Tidy up print OCSP format.
5463 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
5464 entries for variables.
5467 *) Add functionality to apps/openssl.c for detecting locking
5468 problems: As the program is single-threaded, all we have
5469 to do is register a locking callback using an array for
5470 storing which locks are currently held by the program.
5473 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
5474 SSL_get_ex_data_X509_STORE_idx(), which is used in
5475 ssl_verify_cert_chain() and thus can be called at any time
5476 during TLS/SSL handshakes so that thread-safety is essential.
5477 Unfortunately, the ex_data design is not at all suited
5478 for multi-threaded use, so it probably should be abolished.
5481 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5482 [Broadcom, tweaked and integrated by Geoff Thorpe]
5484 *) Move common extension printing code to new function
5485 X509V3_print_extensions(). Reorganise OCSP print routines and
5486 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5489 *) New function X509_signature_print() to remove duplication in some
5493 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5494 set (this was treated exactly the same as SET OF previously). This
5495 is used to reorder the STACK representing the structure to match the
5496 encoding. This will be used to get round a problem where a PKCS7
5497 structure which was signed could not be verified because the STACK
5498 order did not reflect the encoded order.
5501 *) Reimplement the OCSP ASN1 module using the new code.
5504 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5505 for its ASN1 operations. The old style function pointers still exist
5506 for now but they will eventually go away.
5509 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
5510 completely replaces the old ASN1 functionality with a table driven
5511 encoder and decoder which interprets an ASN1_ITEM structure describing
5512 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
5513 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
5514 has also been converted to the new form.
5517 *) Change BN_mod_exp_recp so that negative moduli are tolerated
5518 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
5519 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
5520 for negative moduli.
5523 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
5524 of not touching the result's sign bit.
5527 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
5531 *) Changed the LHASH code to use prototypes for callbacks, and created
5532 macros to declare and implement thin (optionally static) functions
5533 that provide type-safety and avoid function pointer casting for the
5534 type-specific callbacks.
5537 *) Added Kerberos Cipher Suites to be used with TLS, as written in
5539 [Veers Staats <staatsvr@asc.hpc.mil>,
5540 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
5542 *) Reformat the FAQ so the different questions and answers can be divided
5543 in sections depending on the subject.
5546 *) Have the zlib compression code load ZLIB.DLL dynamically under
5550 *) New function BN_mod_sqrt for computing square roots modulo a prime
5551 (using the probabilistic Tonelli-Shanks algorithm unless
5552 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
5553 be handled deterministically).
5554 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5556 *) Make BN_mod_inverse faster by explicitly handling small quotients
5557 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
5558 512 bits], about 30% for larger ones [1024 or 2048 bits].)
5561 *) New function BN_kronecker.
5564 *) Fix BN_gcd so that it works on negative inputs; the result is
5565 positive unless both parameters are zero.
5566 Previously something reasonably close to an infinite loop was
5567 possible because numbers could be growing instead of shrinking
5568 in the implementation of Euclid's algorithm.
5571 *) Fix BN_is_word() and BN_is_one() macros to take into account the
5572 sign of the number in question.
5574 Fix BN_is_word(a,w) to work correctly for w == 0.
5576 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
5577 because its test if the absolute value of 'a' equals 'w'.
5578 Note that BN_abs_is_word does *not* handle w == 0 reliably;
5579 it exists mostly for use in the implementations of BN_is_zero(),
5580 BN_is_one(), and BN_is_word().
5583 *) New function BN_swap.
5586 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
5587 the exponentiation functions are more likely to produce reasonable
5588 results on negative inputs.
5591 *) Change BN_mod_mul so that the result is always non-negative.
5592 Previously, it could be negative if one of the factors was negative;
5593 I don't think anyone really wanted that behaviour.
5596 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
5597 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
5598 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
5599 and add new functions:
5608 BN_mod_lshift1_quick
5612 These functions always generate non-negative results.
5614 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
5615 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
5617 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
5618 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
5619 be reduced modulo m.
5620 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5623 The following entry accidentily appeared in the CHANGES file
5624 distributed with OpenSSL 0.9.7. The modifications described in
5625 it do *not* apply to OpenSSL 0.9.7.
5627 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5628 was actually never needed) and in BN_mul(). The removal in BN_mul()
5629 required a small change in bn_mul_part_recursive() and the addition
5630 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5631 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5632 bn_sub_words() and bn_add_words() except they take arrays with
5637 *) In 'openssl passwd', verify passwords read from the terminal
5638 unless the '-salt' option is used (which usually means that
5639 verification would just waste user's time since the resulting
5640 hash is going to be compared with some given password hash)
5641 or the new '-noverify' option is used.
5643 This is an incompatible change, but it does not affect
5644 non-interactive use of 'openssl passwd' (passwords on the command
5645 line, '-stdin' option, '-in ...' option) and thus should not
5649 *) Remove all references to RSAref, since there's no more need for it.
5652 *) Make DSO load along a path given through an environment variable
5653 (SHLIB_PATH) with shl_load().
5656 *) Constify the ENGINE code as a result of BIGNUM constification.
5657 Also constify the RSA code and most things related to it. In a
5658 few places, most notable in the depth of the ASN.1 code, ugly
5659 casts back to non-const were required (to be solved at a later
5663 *) Make it so the openssl application has all engines loaded by default.
5666 *) Constify the BIGNUM routines a little more.
5669 *) Add the following functions:
5671 ENGINE_load_cswift()
5673 ENGINE_load_atalla()
5675 ENGINE_load_builtin_engines()
5677 That way, an application can itself choose if external engines that
5678 are built-in in OpenSSL shall ever be used or not. The benefit is
5679 that applications won't have to be linked with libdl or other dso
5680 libraries unless it's really needed.
5682 Changed 'openssl engine' to load all engines on demand.
5683 Changed the engine header files to avoid the duplication of some
5684 declarations (they differed!).
5687 *) 'openssl engine' can now list capabilities.
5690 *) Better error reporting in 'openssl engine'.
5693 *) Never call load_dh_param(NULL) in s_server.
5696 *) Add engine application. It can currently list engines by name and
5697 identity, and test if they are actually available.
5700 *) Improve RPM specification file by forcing symbolic linking and making
5701 sure the installed documentation is also owned by root.root.
5702 [Damien Miller <djm@mindrot.org>]
5704 *) Give the OpenSSL applications more possibilities to make use of
5705 keys (public as well as private) handled by engines.
5708 *) Add OCSP code that comes from CertCo.
5711 *) Add VMS support for the Rijndael code.
5714 *) Added untested support for Nuron crypto accelerator.
5717 *) Add support for external cryptographic devices. This code was
5718 previously distributed separately as the "engine" branch.
5719 [Geoff Thorpe, Richard Levitte]
5721 *) Rework the filename-translation in the DSO code. It is now possible to
5722 have far greater control over how a "name" is turned into a filename
5723 depending on the operating environment and any oddities about the
5724 different shared library filenames on each system.
5727 *) Support threads on FreeBSD-elf in Configure.
5730 *) Fix for SHA1 assembly problem with MASM: it produces
5731 warnings about corrupt line number information when assembling
5732 with debugging information. This is caused by the overlapping
5734 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
5737 NCONF_get_number() has no error checking at all. As a replacement,
5738 NCONF_get_number_e() is defined (_e for "error checking") and is
5739 promoted strongly. The old NCONF_get_number is kept around for
5740 binary backward compatibility.
5741 Make it possible for methods to load from something other than a BIO,
5742 by providing a function pointer that is given a name instead of a BIO.
5743 For example, this could be used to load configuration data from an
5747 *) Fix for non blocking accept BIOs. Added new I/O special reason
5748 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
5749 with non blocking I/O was not possible because no retry code was
5750 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
5754 *) Added the beginnings of Rijndael support.
5757 *) Fix for bug in DirectoryString mask setting. Add support for
5758 X509_NAME_print_ex() in 'req' and X509_print_ex() function
5759 to allow certificate printing to more controllable, additional
5760 'certopt' option to 'x509' to allow new printing options to be
5764 *) Clean old EAY MD5 hack from e_os.h.
5767 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
5769 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5770 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5771 [Joe Orton, Steve Henson]
5773 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
5775 *) Fix additional bug revealed by the NISCC test suite:
5777 Stop bug triggering large recursion when presented with
5778 certain ASN.1 tags (CVE-2003-0851)
5781 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
5783 *) Fix various bugs revealed by running the NISCC test suite:
5785 Stop out of bounds reads in the ASN1 code when presented with
5786 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5788 If verify callback ignores invalid public key errors don't try to check
5789 certificate signature with the NULL public key.
5793 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5794 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5798 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5799 extra data after the compression methods not only for TLS 1.0
5800 but also for SSL 3.0 (as required by the specification).
5801 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5803 *) Change X509_certificate_type() to mark the key as exported/exportable
5804 when it's 512 *bits* long, not 512 bytes.
5807 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
5809 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5810 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5811 a protocol version number mismatch like a decryption error
5812 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5815 *) Turn on RSA blinding by default in the default implementation
5816 to avoid a timing attack. Applications that don't want it can call
5817 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5818 They would be ill-advised to do so in most cases.
5819 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5821 *) Change RSA blinding code so that it works when the PRNG is not
5822 seeded (in this case, the secret RSA exponent is abused as
5823 an unpredictable seed -- if it is not unpredictable, there
5824 is no point in blinding anyway). Make RSA blinding thread-safe
5825 by remembering the creator's thread ID in rsa->blinding and
5826 having all other threads use local one-time blinding factors
5827 (this requires more computation than sharing rsa->blinding, but
5828 avoids excessive locking; and if an RSA object is not shared
5829 between threads, blinding will still be very fast).
5832 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
5834 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5835 via timing by performing a MAC computation even if incorrrect
5836 block cipher padding has been found. This is a countermeasure
5837 against active attacks where the attacker has to distinguish
5838 between bad padding and a MAC verification error. (CVE-2003-0078)
5840 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5841 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5842 Martin Vuagnoux (EPFL, Ilion)]
5844 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
5846 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
5847 memory from it's contents. This is done with a counter that will
5848 place alternating values in each byte. This can be used to solve
5849 two issues: 1) the removal of calls to memset() by highly optimizing
5850 compilers, and 2) cleansing with other values than 0, since those can
5851 be read through on certain media, for example a swap space on disk.
5854 *) Bugfix: client side session caching did not work with external caching,
5855 because the session->cipher setting was not restored when reloading
5856 from the external cache. This problem was masked, when
5857 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
5858 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
5861 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
5862 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
5863 [Zeev Lieber <zeev-l@yahoo.com>]
5865 *) Undo an undocumented change introduced in 0.9.6e which caused
5866 repeated calls to OpenSSL_add_all_ciphers() and
5867 OpenSSL_add_all_digests() to be ignored, even after calling
5871 *) Change the default configuration reader to deal with last line not
5872 being properly terminated.
5875 *) Change X509_NAME_cmp() so it applies the special rules on handling
5876 DN values that are of type PrintableString, as well as RDNs of type
5877 emailAddress where the value has the type ia5String.
5878 [stefank@valicert.com via Richard Levitte]
5880 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
5881 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
5882 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
5883 the bitwise-OR of the two for use by the majority of applications
5884 wanting this behaviour, and update the docs. The documented
5885 behaviour and actual behaviour were inconsistent and had been
5886 changing anyway, so this is more a bug-fix than a behavioural
5888 [Geoff Thorpe, diagnosed by Nadav Har'El]
5890 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
5891 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
5894 *) Fix initialization code race conditions in
5895 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
5896 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
5897 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
5898 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
5899 ssl2_get_cipher_by_char(),
5900 ssl3_get_cipher_by_char().
5901 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
5903 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
5904 the cached sessions are flushed, as the remove_cb() might use ex_data
5905 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
5906 (see [openssl.org #212]).
5907 [Geoff Thorpe, Lutz Jaenicke]
5909 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
5910 length, instead of the encoding length to d2i_ASN1_OBJECT.
5913 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
5915 *) [In 0.9.6g-engine release:]
5916 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
5917 [Lynn Gazis <lgazis@rainbow.com>]
5919 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
5921 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
5922 and get fix the header length calculation.
5923 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
5924 Alon Kantor <alonk@checkpoint.com> (and others),
5927 *) Use proper error handling instead of 'assertions' in buffer
5928 overflow checks added in 0.9.6e. This prevents DoS (the
5929 assertions could call abort()).
5930 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
5932 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
5934 *) Add various sanity checks to asn1_get_length() to reject
5935 the ASN1 length bytes if they exceed sizeof(long), will appear
5936 negative or the content length exceeds the length of the
5938 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5940 *) Fix cipher selection routines: ciphers without encryption had no flags
5941 for the cipher strength set and where therefore not handled correctly
5942 by the selection routines (PR #130).
5945 *) Fix EVP_dsa_sha macro.
5949 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
5950 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
5951 that was added in OpenSSL 0.9.6d.
5953 As the countermeasure turned out to be incompatible with some
5954 broken SSL implementations, the new option is part of SSL_OP_ALL.
5955 SSL_OP_ALL is usually employed when compatibility with weird SSL
5956 implementations is desired (e.g. '-bugs' option to 's_client' and
5957 's_server'), so the new option is automatically set in many
5961 *) Changes in security patch:
5963 Changes marked "(CHATS)" were sponsored by the Defense Advanced
5964 Research Projects Agency (DARPA) and Air Force Research Laboratory,
5965 Air Force Materiel Command, USAF, under agreement number
5968 *) Add various sanity checks to asn1_get_length() to reject
5969 the ASN1 length bytes if they exceed sizeof(long), will appear
5970 negative or the content length exceeds the length of the
5971 supplied buffer. (CVE-2002-0659)
5972 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5974 *) Assertions for various potential buffer overflows, not known to
5976 [Ben Laurie (CHATS)]
5978 *) Various temporary buffers to hold ASCII versions of integers were
5979 too small for 64 bit platforms. (CVE-2002-0655)
5980 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
5982 *) Remote buffer overflow in SSL3 protocol - an attacker could
5983 supply an oversized session ID to a client. (CVE-2002-0656)
5984 [Ben Laurie (CHATS)]
5986 *) Remote buffer overflow in SSL2 protocol - an attacker could
5987 supply an oversized client master key. (CVE-2002-0656)
5988 [Ben Laurie (CHATS)]
5990 Changes between 0.9.6c and 0.9.6d [9 May 2002]
5992 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
5993 encoded as NULL) with id-dsa-with-sha1.
5994 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
5996 *) Check various X509_...() return values in apps/req.c.
5997 [Nils Larsch <nla@trustcenter.de>]
5999 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
6000 an end-of-file condition would erronously be flagged, when the CRLF
6001 was just at the end of a processed block. The bug was discovered when
6002 processing data through a buffering memory BIO handing the data to a
6003 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
6004 <ptsekov@syntrex.com> and Nedelcho Stanev.
6007 *) Implement a countermeasure against a vulnerability recently found
6008 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
6009 before application data chunks to avoid the use of known IVs
6010 with data potentially chosen by the attacker.
6013 *) Fix length checks in ssl3_get_client_hello().
6016 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
6017 to prevent ssl3_read_internal() from incorrectly assuming that
6018 ssl3_read_bytes() found application data while handshake
6019 processing was enabled when in fact s->s3->in_read_app_data was
6020 merely automatically cleared during the initial handshake.
6021 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
6023 *) Fix object definitions for Private and Enterprise: they were not
6024 recognized in their shortname (=lowercase) representation. Extend
6025 obj_dat.pl to issue an error when using undefined keywords instead
6026 of silently ignoring the problem (Svenning Sorensen
6027 <sss@sss.dnsalias.net>).
6030 *) Fix DH_generate_parameters() so that it works for 'non-standard'
6031 generators, i.e. generators other than 2 and 5. (Previously, the
6032 code did not properly initialise the 'add' and 'rem' values to
6033 BN_generate_prime().)
6035 In the new general case, we do not insist that 'generator' is
6036 actually a primitive root: This requirement is rather pointless;
6037 a generator of the order-q subgroup is just as good, if not
6041 *) Map new X509 verification errors to alerts. Discovered and submitted by
6042 Tom Wu <tom@arcot.com>.
6045 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
6046 returning non-zero before the data has been completely received
6047 when using non-blocking I/O.
6048 [Bodo Moeller; problem pointed out by John Hughes]
6050 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
6051 [Ben Laurie, Lutz Jaenicke]
6053 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
6054 Yoram Zahavi <YoramZ@gilian.com>).
6057 *) Add information about CygWin 1.3 and on, and preserve proper
6058 configuration for the versions before that.
6059 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6061 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
6062 check whether we deal with a copy of a session and do not delete from
6063 the cache in this case. Problem reported by "Izhar Shoshani Levi"
6064 <izhar@checkpoint.com>.
6067 *) Do not store session data into the internal session cache, if it
6068 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
6069 flag is set). Proposed by Aslam <aslam@funk.com>.
6072 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
6076 *) [In 0.9.6d-engine release:]
6077 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6078 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6080 *) Add the configuration target linux-s390x.
6081 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
6083 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
6084 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
6085 variable as an indication that a ClientHello message has been
6086 received. As the flag value will be lost between multiple
6087 invocations of ssl3_accept when using non-blocking I/O, the
6088 function may not be aware that a handshake has actually taken
6089 place, thus preventing a new session from being added to the
6092 To avoid this problem, we now set s->new_session to 2 instead of
6093 using a local variable.
6094 [Lutz Jaenicke, Bodo Moeller]
6096 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
6097 if the SSL_R_LENGTH_MISMATCH error is detected.
6098 [Geoff Thorpe, Bodo Moeller]
6100 *) New 'shared_ldflag' column in Configure platform table.
6103 *) Fix EVP_CIPHER_mode macro.
6104 ["Dan S. Camper" <dan@bti.net>]
6106 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
6107 type, we must throw them away by setting rr->length to 0.
6108 [D P Chang <dpc@qualys.com>]
6110 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
6112 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
6113 <Dominikus.Scherkl@biodata.com>. (The previous implementation
6114 worked incorrectly for those cases where range = 10..._2 and
6115 3*range is two bits longer than range.)
6118 *) Only add signing time to PKCS7 structures if it is not already
6122 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
6123 OBJ_ld_ce should be OBJ_id_ce.
6124 Also some ip-pda OIDs in crypto/objects/objects.txt were
6125 incorrect (cf. RFC 3039).
6126 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
6128 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
6129 returns early because it has nothing to do.
6130 [Andy Schneider <andy.schneider@bjss.co.uk>]
6132 *) [In 0.9.6c-engine release:]
6133 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
6134 [Andy Schneider <andy.schneider@bjss.co.uk>]
6136 *) [In 0.9.6c-engine release:]
6137 Add support for Cryptographic Appliance's keyserver technology.
6138 (Use engine 'keyclient')
6139 [Cryptographic Appliances and Geoff Thorpe]
6141 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
6142 is called via tools/c89.sh because arguments have to be
6143 rearranged (all '-L' options must appear before the first object
6145 [Richard Shapiro <rshapiro@abinitio.com>]
6147 *) [In 0.9.6c-engine release:]
6148 Add support for Broadcom crypto accelerator cards, backported
6150 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
6152 *) [In 0.9.6c-engine release:]
6153 Add support for SureWare crypto accelerator cards from
6154 Baltimore Technologies. (Use engine 'sureware')
6155 [Baltimore Technologies and Mark Cox]
6157 *) [In 0.9.6c-engine release:]
6158 Add support for crypto accelerator cards from Accelerated
6159 Encryption Processing, www.aep.ie. (Use engine 'aep')
6160 [AEP Inc. and Mark Cox]
6162 *) Add a configuration entry for gcc on UnixWare.
6163 [Gary Benson <gbenson@redhat.com>]
6165 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
6166 messages are stored in a single piece (fixed-length part and
6167 variable-length part combined) and fix various bugs found on the way.
6170 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
6171 instead. BIO_gethostbyname() does not know what timeouts are
6172 appropriate, so entries would stay in cache even when they have
6174 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
6176 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
6177 faced with a pathologically small ClientHello fragment that does
6178 not contain client_version: Instead of aborting with an error,
6179 simply choose the highest available protocol version (i.e.,
6180 TLS 1.0 unless it is disabled). In practice, ClientHello
6181 messages are never sent like this, but this change gives us
6182 strictly correct behaviour at least for TLS.
6185 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
6186 never resets s->method to s->ctx->method when called from within
6187 one of the SSL handshake functions.
6188 [Bodo Moeller; problem pointed out by Niko Baric]
6190 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
6191 (sent using the client's version number) if client_version is
6192 smaller than the protocol version in use. Also change
6193 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
6194 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
6195 the client will at least see that alert.
6198 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
6202 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
6203 client receives HelloRequest while in a handshake.
6204 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
6206 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
6207 should end in 'break', not 'goto end' which circuments various
6208 cleanups done in state SSL_ST_OK. But session related stuff
6209 must be disabled for SSL_ST_OK in the case that we just sent a
6212 Also avoid some overhead by not calling ssl_init_wbio_buffer()
6213 before just sending a HelloRequest.
6214 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
6216 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
6217 reveal whether illegal block cipher padding was found or a MAC
6218 verification error occured. (Neither SSLerr() codes nor alerts
6219 are directly visible to potential attackers, but the information
6220 may leak via logfiles.)
6222 Similar changes are not required for the SSL 2.0 implementation
6223 because the number of padding bytes is sent in clear for SSL 2.0,
6224 and the extra bytes are just ignored. However ssl/s2_pkt.c
6225 failed to verify that the purported number of padding bytes is in
6229 *) Add OpenUNIX-8 support including shared libraries
6230 (Boyd Lynn Gerber <gerberb@zenez.com>).
6233 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
6234 'wristwatch attack' using huge encoding parameters (cf.
6235 James H. Manger's CRYPTO 2001 paper). Note that the
6236 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
6237 encoding parameters and hence was not vulnerable.
6240 *) BN_sqr() bug fix.
6241 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
6243 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
6244 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
6245 followed by modular reduction.
6246 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
6248 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
6249 equivalent based on BN_pseudo_rand() instead of BN_rand().
6252 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
6253 This function was broken, as the check for a new client hello message
6254 to handle SGC did not allow these large messages.
6255 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
6258 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
6261 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
6262 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
6265 *) Rework the configuration and shared library support for Tru64 Unix.
6266 The configuration part makes use of modern compiler features and
6267 still retains old compiler behavior for those that run older versions
6268 of the OS. The shared library support part includes a variant that
6269 uses the RPATH feature, and is available through the special
6270 configuration target "alpha-cc-rpath", which will never be selected
6272 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
6274 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
6275 with the same message size as in ssl3_get_certificate_request().
6276 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
6277 messages might inadvertently be reject as too long.
6278 [Petr Lampa <lampa@fee.vutbr.cz>]
6280 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
6283 *) Modified SSL library such that the verify_callback that has been set
6284 specificly for an SSL object with SSL_set_verify() is actually being
6285 used. Before the change, a verify_callback set with this function was
6286 ignored and the verify_callback() set in the SSL_CTX at the time of
6287 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
6288 to allow the necessary settings.
6291 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
6292 explicitly to NULL, as at least on Solaris 8 this seems not always to be
6293 done automatically (in contradiction to the requirements of the C
6294 standard). This made problems when used from OpenSSH.
6297 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
6298 dh->length and always used
6300 BN_rand_range(priv_key, dh->p).
6302 BN_rand_range() is not necessary for Diffie-Hellman, and this
6303 specific range makes Diffie-Hellman unnecessarily inefficient if
6304 dh->length (recommended exponent length) is much smaller than the
6305 length of dh->p. We could use BN_rand_range() if the order of
6306 the subgroup was stored in the DH structure, but we only have
6311 BN_rand(priv_key, l, ...)
6313 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
6319 RSA_eay_public_encrypt
6320 RSA_eay_private_decrypt
6321 RSA_eay_private_encrypt (signing)
6322 RSA_eay_public_decrypt (signature verification)
6324 (default implementations for RSA_public_encrypt,
6325 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
6326 always reject numbers >= n.
6329 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
6330 to synchronize access to 'locking_thread'. This is necessary on
6331 systems where access to 'locking_thread' (an 'unsigned long'
6332 variable) is not atomic.
6335 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
6336 *before* setting the 'crypto_lock_rand' flag. The previous code had
6337 a race condition if 0 is a valid thread ID.
6338 [Travis Vitek <vitek@roguewave.com>]
6340 *) Add support for shared libraries under Irix.
6341 [Albert Chin-A-Young <china@thewrittenword.com>]
6343 *) Add configuration option to build on Linux on both big-endian and
6345 [Ralf Baechle <ralf@uni-koblenz.de>]
6347 *) Add the possibility to create shared libraries on HP-UX.
6350 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
6352 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
6353 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
6354 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
6355 PRNG state recovery was possible based on the output of
6356 one PRNG request appropriately sized to gain knowledge on
6357 'md' followed by enough consecutive 1-byte PRNG requests
6358 to traverse all of 'state'.
6360 1. When updating 'md_local' (the current thread's copy of 'md')
6361 during PRNG output generation, hash all of the previous
6362 'md_local' value, not just the half used for PRNG output.
6364 2. Make the number of bytes from 'state' included into the hash
6365 independent from the number of PRNG bytes requested.
6367 The first measure alone would be sufficient to avoid
6368 Markku-Juhani's attack. (Actually it had never occurred
6369 to me that the half of 'md_local' used for chaining was the
6370 half from which PRNG output bytes were taken -- I had always
6371 assumed that the secret half would be used.) The second
6372 measure makes sure that additional data from 'state' is never
6373 mixed into 'md_local' in small portions; this heuristically
6374 further strengthens the PRNG.
6377 *) Fix crypto/bn/asm/mips3.s.
6380 *) When only the key is given to "enc", the IV is undefined. Print out
6381 an error message in this case.
6384 *) Handle special case when X509_NAME is empty in X509 printing routines.
6387 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
6388 positive and less than q.
6391 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
6392 used: it isn't thread safe and the add_lock_callback should handle
6394 [Paul Rose <Paul.Rose@bridge.com>]
6396 *) Verify that incoming data obeys the block size in
6397 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
6401 [Ulf Möller, Bodo Möller]
6403 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
6404 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
6405 when fixing the server behaviour for backwards-compatible 'client
6406 hello' messages. (Note that the attack is impractical against
6407 SSL 3.0 and TLS 1.0 anyway because length and version checking
6408 means that the probability of guessing a valid ciphertext is
6409 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
6412 Before 0.9.5, the countermeasure (hide the error by generating a
6413 random 'decryption result') did not work properly because
6414 ERR_clear_error() was missing, meaning that SSL_get_error() would
6415 detect the supposedly ignored error.
6417 Both problems are now fixed.
6420 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
6421 (previously it was 1024).
6424 *) Fix for compatibility mode trust settings: ignore trust settings
6425 unless some valid trust or reject settings are present.
6428 *) Fix for blowfish EVP: its a variable length cipher.
6431 *) Fix various bugs related to DSA S/MIME verification. Handle missing
6432 parameters in DSA public key structures and return an error in the
6433 DSA routines if parameters are absent.
6436 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
6437 in the current directory if neither $RANDFILE nor $HOME was set.
6438 RAND_file_name() in 0.9.6a returned NULL in this case. This has
6439 caused some confusion to Windows users who haven't defined $HOME.
6440 Thus RAND_file_name() is changed again: e_os.h can define a
6441 DEFAULT_HOME, which will be used if $HOME is not set.
6442 For Windows, we use "C:"; on other platforms, we still require
6443 environment variables.
6445 *) Move 'if (!initialized) RAND_poll()' into regions protected by
6446 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
6447 having multiple threads call RAND_poll() concurrently.
6450 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
6451 combination of a flag and a thread ID variable.
6452 Otherwise while one thread is in ssleay_rand_bytes (which sets the
6453 flag), *other* threads can enter ssleay_add_bytes without obeying
6454 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
6455 that they do not hold after the first thread unsets add_do_not_lock).
6458 *) Change bctest again: '-x' expressions are not available in all
6462 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
6464 *) Fix a couple of memory leaks in PKCS7_dataDecode()
6465 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
6467 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
6468 the default extension for executables, if any. Also, make the perl
6469 scripts that use symlink() to test if it really exists and use "cp"
6470 if it doesn't. All this made OpenSSL compilable and installable in
6474 *) Fix for asn1_GetSequence() for indefinite length constructed data.
6475 If SEQUENCE is length is indefinite just set c->slen to the total
6476 amount of data available.
6477 [Steve Henson, reported by shige@FreeBSD.org]
6478 [This change does not apply to 0.9.7.]
6480 *) Change bctest to avoid here-documents inside command substitution
6481 (workaround for FreeBSD /bin/sh bug).
6482 For compatibility with Ultrix, avoid shell functions (introduced
6483 in the bctest version that searches along $PATH).
6486 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6487 with des_encrypt() defined on some operating systems, like Solaris
6491 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6492 On the Importance of Eliminating Errors in Cryptographic
6493 Computations, J. Cryptology 14 (2001) 2, 101-119,
6494 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6497 *) MIPS assembler BIGNUM division bug fix.
6500 *) Disabled incorrect Alpha assembler code.
6503 *) Fix PKCS#7 decode routines so they correctly update the length
6504 after reading an EOC for the EXPLICIT tag.
6506 [This change does not apply to 0.9.7.]
6508 *) Fix bug in PKCS#12 key generation routines. This was triggered
6509 if a 3DES key was generated with a 0 initial byte. Include
6510 PKCS12_BROKEN_KEYGEN compilation option to retain the old
6511 (but broken) behaviour.
6514 *) Enhance bctest to search for a working bc along $PATH and print
6516 [Tim Rice <tim@multitalents.net> via Richard Levitte]
6518 *) Fix memory leaks in err.c: free err_data string if necessary;
6519 don't write to the wrong index in ERR_set_error_data.
6522 *) Implement ssl23_peek (analogous to ssl23_read), which previously
6526 *) Replace rdtsc with _emit statements for VC++ version 5.
6527 [Jeremy Cooper <jeremy@baymoo.org>]
6529 *) Make it possible to reuse SSLv2 sessions.
6532 *) In copy_email() check for >= 0 as a return value for
6533 X509_NAME_get_index_by_NID() since 0 is a valid index.
6534 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
6536 *) Avoid coredump with unsupported or invalid public keys by checking if
6537 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
6538 PKCS7_verify() fails with non detached data.
6541 *) Don't use getenv in library functions when run as setuid/setgid.
6542 New function OPENSSL_issetugid().
6545 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
6546 due to incorrect handling of multi-threading:
6548 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
6550 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
6552 3. Count how many times MemCheck_off() has been called so that
6553 nested use can be treated correctly. This also avoids
6554 inband-signalling in the previous code (which relied on the
6555 assumption that thread ID 0 is impossible).
6558 *) Add "-rand" option also to s_client and s_server.
6561 *) Fix CPU detection on Irix 6.x.
6562 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
6563 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6565 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
6568 [This change does not apply to 0.9.7.]
6570 *) Use the cached encoding of an X509_NAME structure rather than
6571 copying it. This is apparently the reason for the libsafe "errors"
6572 but the code is actually correct.
6575 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
6576 Bleichenbacher's DSA attack.
6577 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
6578 to be set and top=0 forces the highest bit to be set; top=-1 is new
6579 and leaves the highest bit random.
6580 [Ulf Moeller, Bodo Moeller]
6582 *) In the NCONF_...-based implementations for CONF_... queries
6583 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
6584 a temporary CONF structure with the data component set to NULL
6585 (which gives segmentation faults in lh_retrieve).
6586 Instead, use NULL for the CONF pointer in CONF_get_string and
6587 CONF_get_number (which may use environment variables) and directly
6588 return NULL from CONF_get_section.
6591 *) Fix potential buffer overrun for EBCDIC.
6594 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
6595 keyUsage if basicConstraints absent for a CA.
6598 *) Make SMIME_write_PKCS7() write mail header values with a format that
6599 is more generally accepted (no spaces before the semicolon), since
6600 some programs can't parse those values properly otherwise. Also make
6601 sure BIO's that break lines after each write do not create invalid
6605 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
6606 macros previously used would not encode an empty SEQUENCE OF
6607 and break the signature.
6609 [This change does not apply to 0.9.7.]
6611 *) Zero the premaster secret after deriving the master secret in
6615 *) Add some EVP_add_digest_alias registrations (as found in
6616 OpenSSL_add_all_digests()) to SSL_library_init()
6617 aka OpenSSL_add_ssl_algorithms(). This provides improved
6618 compatibility with peers using X.509 certificates
6619 with unconventional AlgorithmIdentifier OIDs.
6622 *) Fix for Irix with NO_ASM.
6623 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6625 *) ./config script fixes.
6626 [Ulf Moeller, Richard Levitte]
6628 *) Fix 'openssl passwd -1'.
6631 *) Change PKCS12_key_gen_asc() so it can cope with non null
6632 terminated strings whose length is passed in the passlen
6633 parameter, for example from PEM callbacks. This was done
6634 by adding an extra length parameter to asc2uni().
6635 [Steve Henson, reported by <oddissey@samsung.co.kr>]
6637 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
6638 call failed, free the DSA structure.
6641 *) Fix to uni2asc() to cope with zero length Unicode strings.
6642 These are present in some PKCS#12 files.
6645 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
6646 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
6647 when writing a 32767 byte record.
6648 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
6650 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
6651 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
6653 (RSA objects have a reference count access to which is protected
6654 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
6655 so they are meant to be shared between threads.)
6656 [Bodo Moeller, Geoff Thorpe; original patch submitted by
6657 "Reddie, Steven" <Steven.Reddie@ca.com>]
6659 *) Fix a deadlock in CRYPTO_mem_leaks().
6662 *) Use better test patterns in bntest.
6665 *) rand_win.c fix for Borland C.
6668 *) BN_rshift bugfix for n == 0.
6671 *) Add a 'bctest' script that checks for some known 'bc' bugs
6672 so that 'make test' does not abort just because 'bc' is broken.
6675 *) Store verify_result within SSL_SESSION also for client side to
6676 avoid potential security hole. (Re-used sessions on the client side
6677 always resulted in verify_result==X509_V_OK, not using the original
6678 result of the server certificate verification.)
6681 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
6682 SSL3_RT_APPLICATION_DATA, return 0.
6683 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
6687 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
6688 releases, have been re-implemented by renaming the previous
6689 implementations of ssl2_read and ssl3_read to ssl2_read_internal
6690 and ssl3_read_internal, respectively, and adding 'peek' parameters
6691 to them. The new ssl[23]_{read,peek} functions are calls to
6692 ssl[23]_read_internal with the 'peek' flag set appropriately.
6693 A 'peek' parameter has also been added to ssl3_read_bytes, which
6694 does the actual work for ssl3_read_internal.
6697 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
6698 the method-specific "init()" handler. Also clean up ex_data after
6699 calling the method-specific "finish()" handler. Previously, this was
6700 happening the other way round.
6703 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
6704 The previous value, 12, was not always sufficient for BN_mod_exp().
6707 *) Make sure that shared libraries get the internal name engine with
6708 the full version number and not just 0. This should mark the
6709 shared libraries as not backward compatible. Of course, this should
6710 be changed again when we can guarantee backward binary compatibility.
6713 *) Fix typo in get_cert_by_subject() in by_dir.c
6714 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
6716 *) Rework the system to generate shared libraries:
6718 - Make note of the expected extension for the shared libraries and
6719 if there is a need for symbolic links from for example libcrypto.so.0
6720 to libcrypto.so.0.9.7. There is extended info in Configure for
6723 - Make as few rebuilds of the shared libraries as possible.
6725 - Still avoid linking the OpenSSL programs with the shared libraries.
6727 - When installing, install the shared libraries separately from the
6731 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
6733 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
6734 and not in SSL_clear because the latter is also used by the
6735 accept/connect functions; previously, the settings made by
6736 SSL_set_read_ahead would be lost during the handshake.
6737 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
6739 *) Correct util/mkdef.pl to be selective about disabled algorithms.
6740 Previously, it would create entries for disableed algorithms no
6744 *) Added several new manual pages for SSL_* function.
6747 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
6749 *) In ssl23_get_client_hello, generate an error message when faced
6750 with an initial SSL 3.0/TLS record that is too small to contain the
6751 first two bytes of the ClientHello message, i.e. client_version.
6752 (Note that this is a pathologic case that probably has never happened
6753 in real life.) The previous approach was to use the version number
6754 from the record header as a substitute; but our protocol choice
6755 should not depend on that one because it is not authenticated
6756 by the Finished messages.
6759 *) More robust randomness gathering functions for Windows.
6760 [Jeffrey Altman <jaltman@columbia.edu>]
6762 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
6763 not set then we don't setup the error code for issuer check errors
6764 to avoid possibly overwriting other errors which the callback does
6765 handle. If an application does set the flag then we assume it knows
6766 what it is doing and can handle the new informational codes
6770 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
6771 a general "ANY" type, as such it should be able to decode anything
6772 including tagged types. However it didn't check the class so it would
6773 wrongly interpret tagged types in the same way as their universal
6774 counterpart and unknown types were just rejected. Changed so that the
6775 tagged and unknown types are handled in the same way as a SEQUENCE:
6776 that is the encoding is stored intact. There is also a new type
6777 "V_ASN1_OTHER" which is used when the class is not universal, in this
6778 case we have no idea what the actual type is so we just lump them all
6782 *) On VMS, stdout may very well lead to a file that is written to
6783 in a record-oriented fashion. That means that every write() will
6784 write a separate record, which will be read separately by the
6785 programs trying to read from it. This can be very confusing.
6787 The solution is to put a BIO filter in the way that will buffer
6788 text until a linefeed is reached, and then write everything a
6789 line at a time, so every record written will be an actual line,
6790 not chunks of lines and not (usually doesn't happen, but I've
6791 seen it once) several lines in one record. BIO_f_linebuffer() is
6794 Currently, it's a VMS-only method, because that's where it has
6795 been tested well enough.
6798 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
6799 it can return incorrect results.
6800 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
6801 but it was in 0.9.6-beta[12].)
6804 *) Disable the check for content being present when verifying detached
6805 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
6806 include zero length content when signing messages.
6809 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
6810 BIO_ctrl (for BIO pairs).
6813 *) Add DSO method for VMS.
6816 *) Bug fix: Montgomery multiplication could produce results with the
6820 *) Add RPM specification openssl.spec and modify it to build three
6821 packages. The default package contains applications, application
6822 documentation and run-time libraries. The devel package contains
6823 include files, static libraries and function documentation. The
6824 doc package contains the contents of the doc directory. The original
6825 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
6828 *) Add a large number of documentation files for many SSL routines.
6829 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6831 *) Add a configuration entry for Sony News 4.
6832 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
6834 *) Don't set the two most significant bits to one when generating a
6835 random number < q in the DSA library.
6838 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
6839 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
6840 the underlying transport is blocking) if a handshake took place.
6841 (The default behaviour is needed by applications such as s_client
6842 and s_server that use select() to determine when to use SSL_read;
6843 but for applications that know in advance when to expect data, it
6844 just makes things more complicated.)
6847 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
6851 *) Add a few more EBCDIC conditionals that make `req' and `x509'
6852 work better on such systems.
6853 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6855 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
6856 Update PKCS12_parse() so it copies the friendlyName and the
6857 keyid to the certificates aux info.
6860 *) Fix bug in PKCS7_verify() which caused an infinite loop
6861 if there was more than one signature.
6862 [Sven Uszpelkat <su@celocom.de>]
6864 *) Major change in util/mkdef.pl to include extra information
6865 about each symbol, as well as presentig variables as well
6866 as functions. This change means that there's n more need
6867 to rebuild the .num files when some algorithms are excluded.
6870 *) Allow the verify time to be set by an application,
6871 rather than always using the current time.
6874 *) Phase 2 verify code reorganisation. The certificate
6875 verify code now looks up an issuer certificate by a
6876 number of criteria: subject name, authority key id
6877 and key usage. It also verifies self signed certificates
6878 by the same criteria. The main comparison function is
6879 X509_check_issued() which performs these checks.
6881 Lot of changes were necessary in order to support this
6882 without completely rewriting the lookup code.
6884 Authority and subject key identifier are now cached.
6886 The LHASH 'certs' is X509_STORE has now been replaced
6887 by a STACK_OF(X509_OBJECT). This is mainly because an
6888 LHASH can't store or retrieve multiple objects with
6889 the same hash value.
6891 As a result various functions (which were all internal
6892 use only) have changed to handle the new X509_STORE
6893 structure. This will break anything that messed round
6894 with X509_STORE internally.
6896 The functions X509_STORE_add_cert() now checks for an
6897 exact match, rather than just subject name.
6899 The X509_STORE API doesn't directly support the retrieval
6900 of multiple certificates matching a given criteria, however
6901 this can be worked round by performing a lookup first
6902 (which will fill the cache with candidate certificates)
6903 and then examining the cache for matches. This is probably
6904 the best we can do without throwing out X509_LOOKUP
6905 entirely (maybe later...).
6907 The X509_VERIFY_CTX structure has been enhanced considerably.
6909 All certificate lookup operations now go via a get_issuer()
6910 callback. Although this currently uses an X509_STORE it
6911 can be replaced by custom lookups. This is a simple way
6912 to bypass the X509_STORE hackery necessary to make this
6913 work and makes it possible to use more efficient techniques
6914 in future. A very simple version which uses a simple
6915 STACK for its trusted certificate store is also provided
6916 using X509_STORE_CTX_trusted_stack().
6918 The verify_cb() and verify() callbacks now have equivalents
6919 in the X509_STORE_CTX structure.
6921 X509_STORE_CTX also has a 'flags' field which can be used
6922 to customise the verify behaviour.
6925 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
6926 excludes S/MIME capabilities.
6929 *) When a certificate request is read in keep a copy of the
6930 original encoding of the signed data and use it when outputing
6931 again. Signatures then use the original encoding rather than
6932 a decoded, encoded version which may cause problems if the
6933 request is improperly encoded.
6936 *) For consistency with other BIO_puts implementations, call
6937 buffer_write(b, ...) directly in buffer_puts instead of calling
6940 In BIO_puts, increment b->num_write as in BIO_write.
6941 [Peter.Sylvester@EdelWeb.fr]
6943 *) Fix BN_mul_word for the case where the word is 0. (We have to use
6944 BN_zero, we may not return a BIGNUM with an array consisting of
6948 *) Avoid calling abort() from within the library when problems are
6949 detected, except if preprocessor symbols have been defined
6950 (such as REF_CHECK, BN_DEBUG etc.).
6953 *) New openssl application 'rsautl'. This utility can be
6954 used for low level RSA operations. DER public key
6955 BIO/fp routines also added.
6958 *) New Configure entry and patches for compiling on QNX 4.
6959 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
6961 *) A demo state-machine implementation was sponsored by
6962 Nuron (http://www.nuron.com/) and is now available in
6963 demos/state_machine.
6966 *) New options added to the 'dgst' utility for signature
6967 generation and verification.
6970 *) Unrecognized PKCS#7 content types are now handled via a
6971 catch all ASN1_TYPE structure. This allows unsupported
6972 types to be stored as a "blob" and an application can
6973 encode and decode it manually.
6976 *) Fix various signed/unsigned issues to make a_strex.c
6978 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
6980 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
6981 length if passed a buffer. ASN1_INTEGER_to_BN failed
6982 if passed a NULL BN and its argument was negative.
6983 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
6985 *) Modification to PKCS#7 encoding routines to output definite
6986 length encoding. Since currently the whole structures are in
6987 memory there's not real point in using indefinite length
6988 constructed encoding. However if OpenSSL is compiled with
6989 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
6992 *) Added BIO_vprintf() and BIO_vsnprintf().
6995 *) Added more prefixes to parse for in the the strings written
6996 through a logging bio, to cover all the levels that are available
6997 through syslog. The prefixes are now:
6999 PANIC, EMERG, EMR => LOG_EMERG
7000 ALERT, ALR => LOG_ALERT
7001 CRIT, CRI => LOG_CRIT
7002 ERROR, ERR => LOG_ERR
7003 WARNING, WARN, WAR => LOG_WARNING
7004 NOTICE, NOTE, NOT => LOG_NOTICE
7005 INFO, INF => LOG_INFO
7006 DEBUG, DBG => LOG_DEBUG
7008 and as before, if none of those prefixes are present at the
7009 beginning of the string, LOG_ERR is chosen.
7011 On Win32, the LOG_* levels are mapped according to this:
7013 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
7014 LOG_WARNING => EVENTLOG_WARNING_TYPE
7015 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
7019 *) Made it possible to reconfigure with just the configuration
7020 argument "reconf" or "reconfigure". The command line arguments
7021 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
7022 and are retrieved from there when reconfiguring.
7026 [Assar Westerlund <assar@sics.se>, Richard Levitte]
7028 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
7031 *) The obj_dat.pl script was messing up the sorting of object
7032 names. The reason was that it compared the quoted version
7033 of strings as a result "OCSP" > "OCSP Signing" because
7034 " > SPACE. Changed script to store unquoted versions of
7035 names and add quotes on output. It was also omitting some
7036 names from the lookup table if they were given a default
7037 value (that is if SN is missing it is given the same
7038 value as LN and vice versa), these are now added on the
7039 grounds that if an object has a name we should be able to
7040 look it up. Finally added warning output when duplicate
7041 short or long names are found.
7044 *) Changes needed for Tandem NSK.
7045 [Scott Uroff <scott@xypro.com>]
7047 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
7048 RSA_padding_check_SSLv23(), special padding was never detected
7049 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
7050 version rollback attacks was not effective.
7052 In s23_clnt.c, don't use special rollback-attack detection padding
7053 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
7054 client; similarly, in s23_srvr.c, don't do the rollback check if
7055 SSL 2.0 is the only protocol enabled in the server.
7058 *) Make it possible to get hexdumps of unprintable data with 'openssl
7059 asn1parse'. By implication, the functions ASN1_parse_dump() and
7060 BIO_dump_indent() are added.
7063 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
7064 these print out strings and name structures based on various
7065 flags including RFC2253 support and proper handling of
7066 multibyte characters. Added options to the 'x509' utility
7067 to allow the various flags to be set.
7070 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
7071 Also change the functions X509_cmp_current_time() and
7072 X509_gmtime_adj() work with an ASN1_TIME structure,
7073 this will enable certificates using GeneralizedTime in validity
7074 dates to be checked.
7077 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
7078 negative public key encodings) on by default,
7079 NO_NEG_PUBKEY_BUG can be set to disable it.
7082 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
7083 content octets. An i2c_ASN1_OBJECT is unnecessary because
7084 the encoding can be trivially obtained from the structure.
7087 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
7088 not read locks (CRYPTO_r_[un]lock).
7091 *) A first attempt at creating official support for shared
7092 libraries through configuration. I've kept it so the
7093 default is static libraries only, and the OpenSSL programs
7094 are always statically linked for now, but there are
7095 preparations for dynamic linking in place.
7096 This has been tested on Linux and Tru64.
7099 *) Randomness polling function for Win9x, as described in:
7100 Peter Gutmann, Software Generation of Practically Strong
7104 *) Fix so PRNG is seeded in req if using an already existing
7108 *) New options to smime application. -inform and -outform
7109 allow alternative formats for the S/MIME message including
7110 PEM and DER. The -content option allows the content to be
7111 specified separately. This should allow things like Netscape
7112 form signing output easier to verify.
7115 *) Fix the ASN1 encoding of tags using the 'long form'.
7118 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
7119 STRING types. These convert content octets to and from the
7120 underlying type. The actual tag and length octets are
7121 already assumed to have been read in and checked. These
7122 are needed because all other string types have virtually
7123 identical handling apart from the tag. By having versions
7124 of the ASN1 functions that just operate on content octets
7125 IMPLICIT tagging can be handled properly. It also allows
7126 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
7127 and ASN1_INTEGER are identical apart from the tag.
7130 *) Change the handling of OID objects as follows:
7132 - New object identifiers are inserted in objects.txt, following
7133 the syntax given in objects.README.
7134 - objects.pl is used to process obj_mac.num and create a new
7136 - obj_dat.pl is used to create a new obj_dat.h, using the data in
7139 This is currently kind of a hack, and the perl code in objects.pl
7140 isn't very elegant, but it works as I intended. The simplest way
7141 to check that it worked correctly is to look in obj_dat.h and
7142 check the array nid_objs and make sure the objects haven't moved
7143 around (this is important!). Additions are OK, as well as
7144 consistent name changes.
7147 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
7150 *) Addition of the command line parameter '-rand file' to 'openssl req'.
7151 The given file adds to whatever has already been seeded into the
7152 random pool through the RANDFILE configuration file option or
7153 environment variable, or the default random state file.
7156 *) mkstack.pl now sorts each macro group into lexical order.
7157 Previously the output order depended on the order the files
7158 appeared in the directory, resulting in needless rewriting
7162 *) Patches to make OpenSSL compile under Win32 again. Mostly
7163 work arounds for the VC++ problem that it treats func() as
7164 func(void). Also stripped out the parts of mkdef.pl that
7165 added extra typesafe functions: these no longer exist.
7168 *) Reorganisation of the stack code. The macros are now all
7169 collected in safestack.h . Each macro is defined in terms of
7170 a "stack macro" of the form SKM_<name>(type, a, b). The
7171 DEBUG_SAFESTACK is now handled in terms of function casts,
7172 this has the advantage of retaining type safety without the
7173 use of additional functions. If DEBUG_SAFESTACK is not defined
7174 then the non typesafe macros are used instead. Also modified the
7175 mkstack.pl script to handle the new form. Needs testing to see
7176 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
7177 the default if no major problems. Similar behaviour for ASN1_SET_OF
7178 and PKCS12_STACK_OF.
7181 *) When some versions of IIS use the 'NET' form of private key the
7182 key derivation algorithm is different. Normally MD5(password) is
7183 used as a 128 bit RC4 key. In the modified case
7184 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
7185 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
7186 as the old Netscape_RSA functions except they have an additional
7187 'sgckey' parameter which uses the modified algorithm. Also added
7188 an -sgckey command line option to the rsa utility. Thanks to
7189 Adrian Peck <bertie@ncipher.com> for posting details of the modified
7190 algorithm to openssl-dev.
7193 *) The evp_local.h macros were using 'c.##kname' which resulted in
7194 invalid expansion on some systems (SCO 5.0.5 for example).
7195 Corrected to 'c.kname'.
7196 [Phillip Porch <root@theporch.com>]
7198 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
7199 a STACK of email addresses from a certificate or request, these look
7200 in the subject name and the subject alternative name extensions and
7201 omit any duplicate addresses.
7204 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
7205 This makes DSA verification about 2 % faster.
7208 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
7209 (meaning that now 2^5 values will be precomputed, which is only 4 KB
7210 plus overhead for 1024 bit moduli).
7211 This makes exponentiations about 0.5 % faster for 1024 bit
7212 exponents (as measured by "openssl speed rsa2048").
7215 *) Rename memory handling macros to avoid conflicts with other
7217 Malloc => OPENSSL_malloc
7218 Malloc_locked => OPENSSL_malloc_locked
7219 Realloc => OPENSSL_realloc
7220 Free => OPENSSL_free
7223 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
7224 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
7227 *) CygWin32 support.
7228 [John Jarvie <jjarvie@newsguy.com>]
7230 *) The type-safe stack code has been rejigged. It is now only compiled
7231 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
7232 by default all type-specific stack functions are "#define"d back to
7233 standard stack functions. This results in more streamlined output
7234 but retains the type-safety checking possibilities of the original
7238 *) The STACK code has been cleaned up, and certain type declarations
7239 that didn't make a lot of sense have been brought in line. This has
7240 also involved a cleanup of sorts in safestack.h to more correctly
7241 map type-safe stack functions onto their plain stack counterparts.
7242 This work has also resulted in a variety of "const"ifications of
7243 lots of the code, especially "_cmp" operations which should normally
7244 be prototyped with "const" parameters anyway.
7247 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
7248 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
7249 (The PRNG state consists of two parts, the large pool 'state' and 'md',
7250 where all of 'md' is used each time the PRNG is used, but 'state'
7251 is used only indexed by a cyclic counter. As entropy may not be
7252 well distributed from the beginning, 'md' is important as a
7253 chaining variable. However, the output function chains only half
7254 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
7255 all of 'md', and seeding with STATE_SIZE dummy bytes will result
7256 in all of 'state' being rewritten, with the new values depending
7257 on virtually all of 'md'. This overcomes the 80 bit limitation.)
7260 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
7261 the handshake is continued after ssl_verify_cert_chain();
7262 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
7263 can lead to 'unexplainable' connection aborts later.
7264 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
7266 *) Major EVP API cipher revision.
7267 Add hooks for extra EVP features. This allows various cipher
7268 parameters to be set in the EVP interface. Support added for variable
7269 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
7270 setting of RC2 and RC5 parameters.
7272 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
7275 Remove lots of duplicated code from the EVP library. For example *every*
7276 cipher init() function handles the 'iv' in the same way according to the
7277 cipher mode. They also all do nothing if the 'key' parameter is NULL and
7278 for CFB and OFB modes they zero ctx->num.
7280 New functionality allows removal of S/MIME code RC2 hack.
7282 Most of the routines have the same form and so can be declared in terms
7285 By shifting this to the top level EVP_CipherInit() it can be removed from
7286 all individual ciphers. If the cipher wants to handle IVs or keys
7287 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
7290 Change lots of functions like EVP_EncryptUpdate() to now return a
7291 value: although software versions of the algorithms cannot fail
7292 any installed hardware versions can.
7295 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
7296 this option is set, tolerate broken clients that send the negotiated
7297 protocol version number instead of the requested protocol version
7301 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
7302 i.e. non-zero for export ciphersuites, zero otherwise.
7303 Previous versions had this flag inverted, inconsistent with
7304 rsa_tmp_cb (..._TMP_RSA_CB).
7305 [Bodo Moeller; problem reported by Amit Chopra]
7307 *) Add missing DSA library text string. Work around for some IIS
7308 key files with invalid SEQUENCE encoding.
7311 *) Add a document (doc/standards.txt) that list all kinds of standards
7312 and so on that are implemented in OpenSSL.
7315 *) Enhance c_rehash script. Old version would mishandle certificates
7316 with the same subject name hash and wouldn't handle CRLs at all.
7317 Added -fingerprint option to crl utility, to support new c_rehash
7321 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
7324 *) Fix for SSL server purpose checking. Server checking was
7325 rejecting certificates which had extended key usage present
7326 but no ssl client purpose.
7327 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
7329 *) Make PKCS#12 code work with no password. The PKCS#12 spec
7330 is a little unclear about how a blank password is handled.
7331 Since the password in encoded as a BMPString with terminating
7332 double NULL a zero length password would end up as just the
7333 double NULL. However no password at all is different and is
7334 handled differently in the PKCS#12 key generation code. NS
7335 treats a blank password as zero length. MSIE treats it as no
7336 password on export: but it will try both on import. We now do
7337 the same: PKCS12_parse() tries zero length and no password if
7338 the password is set to "" or NULL (NULL is now a valid password:
7339 it wasn't before) as does the pkcs12 application.
7342 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
7343 perror when PEM_read_bio_X509_REQ fails, the error message must
7344 be obtained from the error queue.
7347 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
7348 it in ERR_remove_state if appropriate, and change ERR_get_state
7349 accordingly to avoid race conditions (this is necessary because
7350 thread_hash is no longer constant once set).
7353 *) Bugfix for linux-elf makefile.one.
7356 *) RSA_get_default_method() will now cause a default
7357 RSA_METHOD to be chosen if one doesn't exist already.
7358 Previously this was only set during a call to RSA_new()
7359 or RSA_new_method(NULL) meaning it was possible for
7360 RSA_get_default_method() to return NULL.
7363 *) Added native name translation to the existing DSO code
7364 that will convert (if the flag to do so is set) filenames
7365 that are sufficiently small and have no path information
7366 into a canonical native form. Eg. "blah" converted to
7367 "libblah.so" or "blah.dll" etc.
7370 *) New function ERR_error_string_n(e, buf, len) which is like
7371 ERR_error_string(e, buf), but writes at most 'len' bytes
7372 including the 0 terminator. For ERR_error_string_n, 'buf'
7374 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
7376 *) CONF library reworked to become more general. A new CONF
7377 configuration file reader "class" is implemented as well as a
7378 new functions (NCONF_*, for "New CONF") to handle it. The now
7379 old CONF_* functions are still there, but are reimplemented to
7380 work in terms of the new functions. Also, a set of functions
7381 to handle the internal storage of the configuration data is
7382 provided to make it easier to write new configuration file
7383 reader "classes" (I can definitely see something reading a
7384 configuration file in XML format, for example), called _CONF_*,
7385 or "the configuration storage API"...
7387 The new configuration file reading functions are:
7389 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
7390 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
7392 NCONF_default, NCONF_WIN32
7394 NCONF_dump_fp, NCONF_dump_bio
7396 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
7397 NCONF_new creates a new CONF object. This works in the same way
7398 as other interfaces in OpenSSL, like the BIO interface.
7399 NCONF_dump_* dump the internal storage of the configuration file,
7400 which is useful for debugging. All other functions take the same
7401 arguments as the old CONF_* functions wth the exception of the
7402 first that must be a `CONF *' instead of a `LHASH *'.
7404 To make it easer to use the new classes with the old CONF_* functions,
7405 the function CONF_set_default_method is provided.
7408 *) Add '-tls1' option to 'openssl ciphers', which was already
7409 mentioned in the documentation but had not been implemented.
7410 (This option is not yet really useful because even the additional
7411 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
7414 *) Initial DSO code added into libcrypto for letting OpenSSL (and
7415 OpenSSL-based applications) load shared libraries and bind to
7416 them in a portable way.
7417 [Geoff Thorpe, with contributions from Richard Levitte]
7419 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
7421 *) Make sure _lrotl and _lrotr are only used with MSVC.
7423 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
7424 (the default implementation of RAND_status).
7426 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
7427 to '-clrext' (= clear extensions), as intended and documented.
7428 [Bodo Moeller; inconsistency pointed out by Michael Attili
7431 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
7432 was larger than the MD block size.
7433 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
7435 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
7436 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
7437 using the passed key: if the passed key was a private key the result
7438 of X509_print(), for example, would be to print out all the private key
7442 *) des_quad_cksum() byte order bug fix.
7443 [Ulf Möller, using the problem description in krb4-0.9.7, where
7444 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
7446 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
7448 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
7450 *) For easily testing in shell scripts whether some command
7451 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
7452 returns with exit code 0 iff no command of the given name is available.
7453 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
7454 the output goes to stdout and nothing is printed to stderr.
7455 Additional arguments are always ignored.
7457 Since for each cipher there is a command of the same name,
7458 the 'no-cipher' compilation switches can be tested this way.
7460 ('openssl no-XXX' is not able to detect pseudo-commands such
7461 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
7464 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
7467 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
7468 is set; it will be thrown away anyway because each handshake creates
7470 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
7471 to parameters -- in previous versions (since OpenSSL 0.9.3) the
7472 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
7473 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7476 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7477 'Q' and 'R' lose their special meanings (quit/renegotiate).
7478 This is part of what -quiet does; unlike -quiet, -ign_eof
7479 does not suppress any output.
7482 *) Add compatibility options to the purpose and trust code. The
7483 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7484 accepts a certificate or CA, this was the previous behaviour,
7485 with all the associated security issues.
7487 X509_TRUST_COMPAT is the old trust behaviour: only and
7488 automatically trust self signed roots in certificate store. A
7489 new trust setting X509_TRUST_DEFAULT is used to specify that
7490 a purpose has no associated trust setting and it should instead
7491 use the value in the default purpose.
7494 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7495 and fix a memory leak.
7498 *) In util/mkerr.pl (which implements 'make errors'), preserve
7499 reason strings from the previous version of the .c file, as
7500 the default to have only downcase letters (and digits) in
7501 automatically generated reasons codes is not always appropriate.
7504 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7505 using strerror. Previously, ERR_reason_error_string() returned
7506 library names as reason strings for SYSerr; but SYSerr is a special
7507 case where small numbers are errno values, not library numbers.
7510 *) Add '-dsaparam' option to 'openssl dhparam' application. This
7511 converts DSA parameters into DH parameters. (When creating parameters,
7512 DSA_generate_parameters is used.)
7515 *) Include 'length' (recommended exponent length) in C code generated
7516 by 'openssl dhparam -C'.
7519 *) The second argument to set_label in perlasm was already being used
7520 so couldn't be used as a "file scope" flag. Moved to third argument
7524 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
7525 instead of RAND_bytes for encryption IVs and salts.
7528 *) Include RAND_status() into RAND_METHOD instead of implementing
7529 it only for md_rand.c Otherwise replacing the PRNG by calling
7530 RAND_set_rand_method would be impossible.
7533 *) Don't let DSA_generate_key() enter an infinite loop if the random
7534 number generation fails.
7537 *) New 'rand' application for creating pseudo-random output.
7540 *) Added configuration support for Linux/IA64
7541 [Rolf Haberrecker <rolf@suse.de>]
7543 *) Assembler module support for Mingw32.
7546 *) Shared library support for HPUX (in shlib/).
7547 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
7549 *) Shared library support for Solaris gcc.
7550 [Lutz Behnke <behnke@trustcenter.de>]
7552 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
7554 *) PKCS7_encrypt() was adding text MIME headers twice because they
7555 were added manually and by SMIME_crlf_copy().
7558 *) In bntest.c don't call BN_rand with zero bits argument.
7559 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
7561 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
7562 case was implemented. This caused BN_div_recp() to fail occasionally.
7565 *) Add an optional second argument to the set_label() in the perl
7566 assembly language builder. If this argument exists and is set
7567 to 1 it signals that the assembler should use a symbol whose
7568 scope is the entire file, not just the current function. This
7569 is needed with MASM which uses the format label:: for this scope.
7570 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
7572 *) Change the ASN1 types so they are typedefs by default. Before
7573 almost all types were #define'd to ASN1_STRING which was causing
7574 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
7578 *) Change names of new functions to the new get1/get0 naming
7579 convention: After 'get1', the caller owns a reference count
7580 and has to call ..._free; 'get0' returns a pointer to some
7581 data structure without incrementing reference counters.
7582 (Some of the existing 'get' functions increment a reference
7583 counter, some don't.)
7584 Similarly, 'set1' and 'add1' functions increase reference
7585 counters or duplicate objects.
7588 *) Allow for the possibility of temp RSA key generation failure:
7589 the code used to assume it always worked and crashed on failure.
7592 *) Fix potential buffer overrun problem in BIO_printf().
7593 [Ulf Möller, using public domain code by Patrick Powell; problem
7594 pointed out by David Sacerdote <das33@cornell.edu>]
7596 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
7597 RAND_egd() and RAND_status(). In the command line application,
7598 the EGD socket can be specified like a seed file using RANDFILE
7602 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
7603 Some CAs (e.g. Verisign) distribute certificates in this form.
7606 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
7607 list to exclude them. This means that no special compilation option
7608 is needed to use anonymous DH: it just needs to be included in the
7612 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
7613 EVP_MD_type. The old functionality is available in a new macro called
7614 EVP_MD_md(). Change code that uses it and update docs.
7617 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
7618 where the 'void *' argument is replaced by a function pointer argument.
7619 Previously 'void *' was abused to point to functions, which works on
7620 many platforms, but is not correct. As these functions are usually
7621 called by macros defined in OpenSSL header files, most source code
7622 should work without changes.
7625 *) <openssl/opensslconf.h> (which is created by Configure) now contains
7626 sections with information on -D... compiler switches used for
7627 compiling the library so that applications can see them. To enable
7628 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
7629 must be defined. E.g.,
7630 #define OPENSSL_ALGORITHM_DEFINES
7631 #include <openssl/opensslconf.h>
7632 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
7633 [Richard Levitte, Ulf and Bodo Möller]
7635 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
7639 *) Change the 'other' type in certificate aux info to a STACK_OF
7640 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
7641 the required ASN1 format: arbitrary types determined by an OID.
7644 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
7645 argument to 'req'. This is not because the function is newer or
7646 better than others it just uses the work 'NEW' in the certificate
7647 request header lines. Some software needs this.
7650 *) Reorganise password command line arguments: now passwords can be
7651 obtained from various sources. Delete the PEM_cb function and make
7652 it the default behaviour: i.e. if the callback is NULL and the
7653 usrdata argument is not NULL interpret it as a null terminated pass
7654 phrase. If usrdata and the callback are NULL then the pass phrase
7655 is prompted for as usual.
7658 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
7659 the support is automatically enabled. The resulting binaries will
7660 autodetect the card and use it if present.
7661 [Ben Laurie and Compaq Inc.]
7663 *) Work around for Netscape hang bug. This sends certificate request
7664 and server done in one record. Since this is perfectly legal in the
7665 SSL/TLS protocol it isn't a "bug" option and is on by default. See
7666 the bugs/SSLv3 entry for more info.
7669 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
7672 *) Add -rand argument to smime and pkcs12 applications and read/write
7676 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
7679 *) Add command line password options to the remaining applications.
7682 *) Bug fix for BN_div_recp() for numerators with an even number of
7686 *) More tests in bntest.c, and changed test_bn output.
7689 *) ./config recognizes MacOS X now.
7692 *) Bug fix for BN_div() when the first words of num and divsor are
7693 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
7696 *) Add support for various broken PKCS#8 formats, and command line
7697 options to produce them.
7700 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
7701 get temporary BIGNUMs from a BN_CTX.
7704 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
7708 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
7709 include a #define from the old name to the new. The original intent
7710 was that statically linked binaries could for example just call
7711 SSLeay_add_all_ciphers() to just add ciphers to the table and not
7712 link with digests. This never worked becayse SSLeay_add_all_digests()
7713 and SSLeay_add_all_ciphers() were in the same source file so calling
7714 one would link with the other. They are now in separate source files.
7717 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
7720 *) Use a less unusual form of the Miller-Rabin primality test (it used
7721 a binary algorithm for exponentiation integrated into the Miller-Rabin
7722 loop, our standard modexp algorithms are faster).
7725 *) Support for the EBCDIC character set completed.
7726 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
7728 *) Source code cleanups: use const where appropriate, eliminate casts,
7729 use void * instead of char * in lhash.
7732 *) Bugfix: ssl3_send_server_key_exchange was not restartable
7733 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
7734 this the server could overwrite ephemeral keys that the client
7738 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
7739 using 50 iterations of the Rabin-Miller test.
7741 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
7742 iterations of the Rabin-Miller test as required by the appendix
7743 to FIPS PUB 186[-1]) instead of DSA_is_prime.
7744 As BN_is_prime_fasttest includes trial division, DSA parameter
7745 generation becomes much faster.
7747 This implies a change for the callback functions in DSA_is_prime
7748 and DSA_generate_parameters: The callback function is called once
7749 for each positive witness in the Rabin-Miller test, not just
7750 occasionally in the inner loop; and the parameters to the
7751 callback function now provide an iteration count for the outer
7752 loop rather than for the current invocation of the inner loop.
7753 DSA_generate_parameters additionally can call the callback
7754 function with an 'iteration count' of -1, meaning that a
7755 candidate has passed the trial division test (when q is generated
7756 from an application-provided seed, trial division is skipped).
7759 *) New function BN_is_prime_fasttest that optionally does trial
7760 division before starting the Rabin-Miller test and has
7761 an additional BN_CTX * argument (whereas BN_is_prime always
7762 has to allocate at least one BN_CTX).
7763 'callback(1, -1, cb_arg)' is called when a number has passed the
7764 trial division stage.
7767 *) Fix for bug in CRL encoding. The validity dates weren't being handled
7771 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
7774 *) New function BN_pseudo_rand().
7777 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
7778 bignum version of BN_from_montgomery() with the working code from
7779 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
7783 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
7784 made it impossible to use the same SSL_SESSION data structure in
7785 SSL2 clients in multiple threads.
7788 *) The return value of RAND_load_file() no longer counts bytes obtained
7789 by stat(). RAND_load_file(..., -1) is new and uses the complete file
7790 to seed the PRNG (previously an explicit byte count was required).
7791 [Ulf Möller, Bodo Möller]
7793 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
7794 used (char *) instead of (void *) and had casts all over the place.
7797 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
7800 *) Retain source code compatibility for BN_prime_checks macro:
7801 BN_is_prime(..., BN_prime_checks, ...) now uses
7802 BN_prime_checks_for_size to determine the appropriate number of
7803 Rabin-Miller iterations.
7806 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
7807 DH_CHECK_P_NOT_SAFE_PRIME.
7808 (Check if this is true? OpenPGP calls them "strong".)
7811 *) Merge the functionality of "dh" and "gendh" programs into a new program
7812 "dhparam". The old programs are retained for now but will handle DH keys
7813 (instead of parameters) in future.
7816 *) Make the ciphers, s_server and s_client programs check the return values
7817 when a new cipher list is set.
7820 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
7821 ciphers. Before when the 56bit ciphers were enabled the sorting was
7824 The syntax for the cipher sorting has been extended to support sorting by
7825 cipher-strength (using the strength_bits hard coded in the tables).
7826 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
7828 Fix a bug in the cipher-command parser: when supplying a cipher command
7829 string with an "undefined" symbol (neither command nor alphanumeric
7830 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
7831 an error is flagged.
7833 Due to the strength-sorting extension, the code of the
7834 ssl_create_cipher_list() function was completely rearranged. I hope that
7835 the readability was also increased :-)
7836 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7838 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
7839 for the first serial number and places 2 in the serial number file. This
7840 avoids problems when the root CA is created with serial number zero and
7841 the first user certificate has the same issuer name and serial number
7845 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
7846 the new code. Add documentation for this stuff.
7849 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
7850 X509_*() to X509at_*() on the grounds that they don't handle X509
7851 structures and behave in an analagous way to the X509v3 functions:
7852 they shouldn't be called directly but wrapper functions should be used
7855 So we also now have some wrapper functions that call the X509at functions
7856 when passed certificate requests. (TO DO: similar things can be done with
7857 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
7858 things. Some of these need some d2i or i2d and print functionality
7859 because they handle more complex structures.)
7862 *) Add missing #ifndefs that caused missing symbols when building libssl
7863 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
7864 NO_RSA in ssl/s2*.c.
7865 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
7867 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
7868 has a return value which indicates the quality of the random data
7869 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
7870 error queue. New function RAND_pseudo_bytes() generates output that is
7871 guaranteed to be unique but not unpredictable. RAND_add is like
7872 RAND_seed, but takes an extra argument for an entropy estimate
7873 (RAND_seed always assumes full entropy).
7876 *) Do more iterations of Rabin-Miller probable prime test (specifically,
7877 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
7878 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
7879 in crypto/bn/bn_prime.c for the complete table). This guarantees a
7880 false-positive rate of at most 2^-80 for random input.
7883 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
7886 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
7887 in the 0.9.5 release), this returns the chain
7888 from an X509_CTX structure with a dup of the stack and all
7889 the X509 reference counts upped: so the stack will exist
7890 after X509_CTX_cleanup() has been called. Modify pkcs12.c
7893 Also make SSL_SESSION_print() print out the verify return
7897 *) Add manpage for the pkcs12 command. Also change the default
7898 behaviour so MAC iteration counts are used unless the new
7899 -nomaciter option is used. This improves file security and
7900 only older versions of MSIE (4.0 for example) need it.
7903 *) Honor the no-xxx Configure options when creating .DEF files.
7906 *) Add PKCS#10 attributes to field table: challengePassword,
7907 unstructuredName and unstructuredAddress. These are taken from
7908 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
7909 international characters are used.
7911 More changes to X509_ATTRIBUTE code: allow the setting of types
7912 based on strings. Remove the 'loc' parameter when adding
7913 attributes because these will be a SET OF encoding which is sorted
7917 *) Initial changes to the 'req' utility to allow request generation
7918 automation. This will allow an application to just generate a template
7919 file containing all the field values and have req construct the
7922 Initial support for X509_ATTRIBUTE handling. Stacks of these are
7923 used all over the place including certificate requests and PKCS#7
7924 structures. They are currently handled manually where necessary with
7925 some primitive wrappers for PKCS#7. The new functions behave in a
7926 manner analogous to the X509 extension functions: they allow
7927 attributes to be looked up by NID and added.
7929 Later something similar to the X509V3 code would be desirable to
7930 automatically handle the encoding, decoding and printing of the
7931 more complex types. The string types like challengePassword can
7932 be handled by the string table functions.
7934 Also modified the multi byte string table handling. Now there is
7935 a 'global mask' which masks out certain types. The table itself
7936 can use the flag STABLE_NO_MASK to ignore the mask setting: this
7937 is useful when for example there is only one permissible type
7938 (as in countryName) and using the mask might result in no valid
7942 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
7943 SSL_get_peer_finished to allow applications to obtain the latest
7944 Finished messages sent to the peer or expected from the peer,
7945 respectively. (SSL_get_peer_finished is usually the Finished message
7946 actually received from the peer, otherwise the protocol will be aborted.)
7948 As the Finished message are message digests of the complete handshake
7949 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
7950 be used for external authentication procedures when the authentication
7951 provided by SSL/TLS is not desired or is not enough.
7954 *) Enhanced support for Alpha Linux is added. Now ./config checks if
7955 the host supports BWX extension and if Compaq C is present on the
7956 $PATH. Just exploiting of the BWX extension results in 20-30%
7957 performance kick for some algorithms, e.g. DES and RC4 to mention
7958 a couple. Compaq C in turn generates ~20% faster code for MD5 and
7962 *) Add support for MS "fast SGC". This is arguably a violation of the
7963 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
7964 weak crypto and after checking the certificate is SGC a second one
7965 with strong crypto. MS SGC stops the first handshake after receiving
7966 the server certificate message and sends a second client hello. Since
7967 a server will typically do all the time consuming operations before
7968 expecting any further messages from the client (server key exchange
7969 is the most expensive) there is little difference between the two.
7971 To get OpenSSL to support MS SGC we have to permit a second client
7972 hello message after we have sent server done. In addition we have to
7973 reset the MAC if we do get this second client hello.
7976 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
7977 if a DER encoded private key is RSA or DSA traditional format. Changed
7978 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
7979 format DER encoded private key. Newer code should use PKCS#8 format which
7980 has the key type encoded in the ASN1 structure. Added DER private key
7981 support to pkcs8 application.
7984 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
7985 ciphersuites has been selected (as required by the SSL 3/TLS 1
7986 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
7987 is set, we interpret this as a request to violate the specification
7988 (the worst that can happen is a handshake failure, and 'correct'
7989 behaviour would result in a handshake failure anyway).
7992 *) In SSL_CTX_add_session, take into account that there might be multiple
7993 SSL_SESSION structures with the same session ID (e.g. when two threads
7994 concurrently obtain them from an external cache).
7995 The internal cache can handle only one SSL_SESSION with a given ID,
7996 so if there's a conflict, we now throw out the old one to achieve
8000 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
8001 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
8002 some routines that use cipher OIDs: some ciphers do not have OIDs
8003 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
8007 *) Simplify the trust setting structure and code. Now we just have
8008 two sequences of OIDs for trusted and rejected settings. These will
8009 typically have values the same as the extended key usage extension
8010 and any application specific purposes.
8012 The trust checking code now has a default behaviour: it will just
8013 check for an object with the same NID as the passed id. Functions can
8014 be provided to override either the default behaviour or the behaviour
8015 for a given id. SSL client, server and email already have functions
8016 in place for compatibility: they check the NID and also return "trusted"
8017 if the certificate is self signed.
8020 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
8021 traditional format into an EVP_PKEY structure.
8024 *) Add a password callback function PEM_cb() which either prompts for
8025 a password if usr_data is NULL or otherwise assumes it is a null
8026 terminated password. Allow passwords to be passed on command line
8027 environment or config files in a few more utilities.
8030 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
8031 keys. Add some short names for PKCS#8 PBE algorithms and allow them
8032 to be specified on the command line for the pkcs8 and pkcs12 utilities.
8033 Update documentation.
8036 *) Support for ASN1 "NULL" type. This could be handled before by using
8037 ASN1_TYPE but there wasn't any function that would try to read a NULL
8038 and produce an error if it couldn't. For compatibility we also have
8039 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
8040 don't allocate anything because they don't need to.
8043 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
8045 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
8047 *) Rebuild of the memory allocation routines used by OpenSSL code and
8048 possibly others as well. The purpose is to make an interface that
8049 provide hooks so anyone can build a separate set of allocation and
8050 deallocation routines to be used by OpenSSL, for example memory
8051 pool implementations, or something else, which was previously hard
8052 since Malloc(), Realloc() and Free() were defined as macros having
8053 the values malloc, realloc and free, respectively (except for Win32
8054 compilations). The same is provided for memory debugging code.
8055 OpenSSL already comes with functionality to find memory leaks, but
8056 this gives people a chance to debug other memory problems.
8058 With these changes, a new set of functions and macros have appeared:
8060 CRYPTO_set_mem_debug_functions() [F]
8061 CRYPTO_get_mem_debug_functions() [F]
8062 CRYPTO_dbg_set_options() [F]
8063 CRYPTO_dbg_get_options() [F]
8064 CRYPTO_malloc_debug_init() [M]
8066 The memory debug functions are NULL by default, unless the library
8067 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
8068 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
8069 gives the standard debugging functions that come with OpenSSL) or
8070 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
8071 provided by the library user) must be used. When the standard
8072 debugging functions are used, CRYPTO_dbg_set_options can be used to
8073 request additional information:
8074 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
8075 the CRYPTO_MDEBUG_xxx macro when compiling the library.
8077 Also, things like CRYPTO_set_mem_functions will always give the
8078 expected result (the new set of functions is used for allocation
8079 and deallocation) at all times, regardless of platform and compiler
8082 To finish it up, some functions that were never use in any other
8083 way than through macros have a new API and new semantic:
8086 CRYPTO_dbg_realloc()
8089 All macros of value have retained their old syntax.
8090 [Richard Levitte and Bodo Moeller]
8092 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
8093 ordering of SMIMECapabilities wasn't in "strength order" and there
8094 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
8098 *) Some ASN1 types with illegal zero length encoding (INTEGER,
8099 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
8100 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
8102 *) Merge in my S/MIME library for OpenSSL. This provides a simple
8103 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
8104 functionality to handle multipart/signed properly) and a utility
8105 called 'smime' to call all this stuff. This is based on code I
8106 originally wrote for Celo who have kindly allowed it to be
8107 included in OpenSSL.
8110 *) Add variants des_set_key_checked and des_set_key_unchecked of
8111 des_set_key (aka des_key_sched). Global variable des_check_key
8112 decides which of these is called by des_set_key; this way
8113 des_check_key behaves as it always did, but applications and
8114 the library itself, which was buggy for des_check_key == 1,
8115 have a cleaner way to pick the version they need.
8118 *) New function PKCS12_newpass() which changes the password of a
8122 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
8123 dynamic mix. In both cases the ids can be used as an index into the
8124 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
8125 functions so they accept a list of the field values and the
8126 application doesn't need to directly manipulate the X509_TRUST
8130 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
8134 *) Modify the way the V3 extension code looks up extensions. This now
8135 works in a similar way to the object code: we have some "standard"
8136 extensions in a static table which is searched with OBJ_bsearch()
8137 and the application can add dynamic ones if needed. The file
8138 crypto/x509v3/ext_dat.h now has the info: this file needs to be
8139 updated whenever a new extension is added to the core code and kept
8140 in ext_nid order. There is a simple program 'tabtest.c' which checks
8141 this. New extensions are not added too often so this file can readily
8142 be maintained manually.
8144 There are two big advantages in doing things this way. The extensions
8145 can be looked up immediately and no longer need to be "added" using
8146 X509V3_add_standard_extensions(): this function now does nothing.
8147 [Side note: I get *lots* of email saying the extension code doesn't
8148 work because people forget to call this function]
8149 Also no dynamic allocation is done unless new extensions are added:
8150 so if we don't add custom extensions there is no need to call
8151 X509V3_EXT_cleanup().
8154 *) Modify enc utility's salting as follows: make salting the default. Add a
8155 magic header, so unsalted files fail gracefully instead of just decrypting
8156 to garbage. This is because not salting is a big security hole, so people
8157 should be discouraged from doing it.
8160 *) Fixes and enhancements to the 'x509' utility. It allowed a message
8161 digest to be passed on the command line but it only used this
8162 parameter when signing a certificate. Modified so all relevant
8163 operations are affected by the digest parameter including the
8164 -fingerprint and -x509toreq options. Also -x509toreq choked if a
8165 DSA key was used because it didn't fix the digest.
8168 *) Initial certificate chain verify code. Currently tests the untrusted
8169 certificates for consistency with the verify purpose (which is set
8170 when the X509_STORE_CTX structure is set up) and checks the pathlength.
8172 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
8173 this is because it will reject chains with invalid extensions whereas
8174 every previous version of OpenSSL and SSLeay made no checks at all.
8176 Trust code: checks the root CA for the relevant trust settings. Trust
8177 settings have an initial value consistent with the verify purpose: e.g.
8178 if the verify purpose is for SSL client use it expects the CA to be
8179 trusted for SSL client use. However the default value can be changed to
8180 permit custom trust settings: one example of this would be to only trust
8181 certificates from a specific "secure" set of CAs.
8183 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
8184 which should be used for version portability: especially since the
8185 verify structure is likely to change more often now.
8187 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
8188 to set them. If not set then assume SSL clients will verify SSL servers
8191 Two new options to the verify program: -untrusted allows a set of
8192 untrusted certificates to be passed in and -purpose which sets the
8193 intended purpose of the certificate. If a purpose is set then the
8194 new chain verify code is used to check extension consistency.
8197 *) Support for the authority information access extension.
8200 *) Modify RSA and DSA PEM read routines to transparently handle
8201 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
8202 public keys in a format compatible with certificate
8203 SubjectPublicKeyInfo structures. Unfortunately there were already
8204 functions called *_PublicKey_* which used various odd formats so
8205 these are retained for compatibility: however the DSA variants were
8206 never in a public release so they have been deleted. Changed dsa/rsa
8207 utilities to handle the new format: note no releases ever handled public
8208 keys so we should be OK.
8210 The primary motivation for this change is to avoid the same fiasco
8211 that dogs private keys: there are several incompatible private key
8212 formats some of which are standard and some OpenSSL specific and
8213 require various evil hacks to allow partial transparent handling and
8214 even then it doesn't work with DER formats. Given the option anything
8215 other than PKCS#8 should be dumped: but the other formats have to
8216 stay in the name of compatibility.
8218 With public keys and the benefit of hindsight one standard format
8219 is used which works with EVP_PKEY, RSA or DSA structures: though
8220 it clearly returns an error if you try to read the wrong kind of key.
8222 Added a -pubkey option to the 'x509' utility to output the public key.
8223 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
8224 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
8225 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
8226 that do the same as the EVP_PKEY_assign_*() except they up the
8227 reference count of the added key (they don't "swallow" the
8231 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
8232 CRLs would fail if the file contained no certificates or no CRLs:
8233 added a new function to read in both types and return the number
8234 read: this means that if none are read it will be an error. The
8235 DER versions of the certificate and CRL reader would always fail
8236 because it isn't possible to mix certificates and CRLs in DER format
8237 without choking one or the other routine. Changed this to just read
8238 a certificate: this is the best we can do. Also modified the code
8239 in apps/verify.c to take notice of return codes: it was previously
8240 attempting to read in certificates from NULL pointers and ignoring
8241 any errors: this is one reason why the cert and CRL reader seemed
8242 to work. It doesn't check return codes from the default certificate
8243 routines: these may well fail if the certificates aren't installed.
8246 *) Code to support otherName option in GeneralName.
8249 *) First update to verify code. Change the verify utility
8250 so it warns if it is passed a self signed certificate:
8251 for consistency with the normal behaviour. X509_verify
8252 has been modified to it will now verify a self signed
8253 certificate if *exactly* the same certificate appears
8254 in the store: it was previously impossible to trust a
8255 single self signed certificate. This means that:
8256 openssl verify ss.pem
8257 now gives a warning about a self signed certificate but
8258 openssl verify -CAfile ss.pem ss.pem
8262 *) For servers, store verify_result in SSL_SESSION data structure
8263 (and add it to external session representation).
8264 This is needed when client certificate verifications fails,
8265 but an application-provided verification callback (set by
8266 SSL_CTX_set_cert_verify_callback) allows accepting the session
8267 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
8268 but returns 1): When the session is reused, we have to set
8269 ssl->verify_result to the appropriate error code to avoid
8271 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
8273 *) Fix a bug in the new PKCS#7 code: it didn't consider the
8274 case in PKCS7_dataInit() where the signed PKCS7 structure
8275 didn't contain any existing data because it was being created.
8276 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
8278 *) Add a salt to the key derivation routines in enc.c. This
8279 forms the first 8 bytes of the encrypted file. Also add a
8280 -S option to allow a salt to be input on the command line.
8283 *) New function X509_cmp(). Oddly enough there wasn't a function
8284 to compare two certificates. We do this by working out the SHA1
8285 hash and comparing that. X509_cmp() will be needed by the trust
8289 *) SSL_get1_session() is like SSL_get_session(), but increments
8290 the reference count in the SSL_SESSION returned.
8291 [Geoff Thorpe <geoff@eu.c2.net>]
8293 *) Fix for 'req': it was adding a null to request attributes.
8294 Also change the X509_LOOKUP and X509_INFO code to handle
8295 certificate auxiliary information.
8298 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
8302 *) Add the possibility to add extra information to the memory leak
8303 detecting output, to form tracebacks, showing from where each
8304 allocation was originated: CRYPTO_push_info("constant string") adds
8305 the string plus current file name and line number to a per-thread
8306 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
8307 is like calling CYRPTO_pop_info() until the stack is empty.
8308 Also updated memory leak detection code to be multi-thread-safe.
8311 *) Add options -text and -noout to pkcs7 utility and delete the
8312 encryption options which never did anything. Update docs.
8315 *) Add options to some of the utilities to allow the pass phrase
8316 to be included on either the command line (not recommended on
8317 OSes like Unix) or read from the environment. Update the
8318 manpages and fix a few bugs.
8321 *) Add a few manpages for some of the openssl commands.
8324 *) Fix the -revoke option in ca. It was freeing up memory twice,
8325 leaking and not finding already revoked certificates.
8328 *) Extensive changes to support certificate auxiliary information.
8329 This involves the use of X509_CERT_AUX structure and X509_AUX
8330 functions. An X509_AUX function such as PEM_read_X509_AUX()
8331 can still read in a certificate file in the usual way but it
8332 will also read in any additional "auxiliary information". By
8333 doing things this way a fair degree of compatibility can be
8334 retained: existing certificates can have this information added
8335 using the new 'x509' options.
8337 Current auxiliary information includes an "alias" and some trust
8338 settings. The trust settings will ultimately be used in enhanced
8339 certificate chain verification routines: currently a certificate
8340 can only be trusted if it is self signed and then it is trusted
8344 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
8345 The problem was that one of the replacement routines had not been working
8346 since SSLeay releases. For now the offending routine has been replaced
8347 with non-optimised assembler. Even so, this now gives around 95%
8348 performance improvement for 1024 bit RSA signs.
8351 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
8352 handling. Most clients have the effective key size in bits equal to
8353 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
8354 A few however don't do this and instead use the size of the decrypted key
8355 to determine the RC2 key length and the AlgorithmIdentifier to determine
8356 the effective key length. In this case the effective key length can still
8357 be 40 bits but the key length can be 168 bits for example. This is fixed
8358 by manually forcing an RC2 key into the EVP_PKEY structure because the
8359 EVP code can't currently handle unusual RC2 key sizes: it always assumes
8360 the key length and effective key length are equal.
8363 *) Add a bunch of functions that should simplify the creation of
8364 X509_NAME structures. Now you should be able to do:
8365 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
8366 and have it automatically work out the correct field type and fill in
8367 the structures. The more adventurous can try:
8368 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
8369 and it will (hopefully) work out the correct multibyte encoding.
8372 *) Change the 'req' utility to use the new field handling and multibyte
8373 copy routines. Before the DN field creation was handled in an ad hoc
8374 way in req, ca, and x509 which was rather broken and didn't support
8375 BMPStrings or UTF8Strings. Since some software doesn't implement
8376 BMPStrings or UTF8Strings yet, they can be enabled using the config file
8377 using the dirstring_type option. See the new comment in the default
8378 openssl.cnf for more info.
8381 *) Make crypto/rand/md_rand.c more robust:
8382 - Assure unique random numbers after fork().
8383 - Make sure that concurrent threads access the global counter and
8384 md serializably so that we never lose entropy in them
8385 or use exactly the same state in multiple threads.
8386 Access to the large state is not always serializable because
8387 the additional locking could be a performance killer, and
8388 md should be large enough anyway.
8391 *) New file apps/app_rand.c with commonly needed functionality
8392 for handling the random seed file.
8394 Use the random seed file in some applications that previously did not:
8396 dsaparam -genkey (which also ignored its '-rand' option),
8399 x509 (when signing).
8400 Except on systems with /dev/urandom, it is crucial to have a random
8401 seed file at least for key creation, DSA signing, and for DH exchanges;
8402 for RSA signatures we could do without one.
8404 gendh and gendsa (unlike genrsa) used to read only the first byte
8405 of each file listed in the '-rand' option. The function as previously
8406 found in genrsa is now in app_rand.c and is used by all programs
8407 that support '-rand'.
8410 *) In RAND_write_file, use mode 0600 for creating files;
8411 don't just chmod when it may be too late.
8414 *) Report an error from X509_STORE_load_locations
8415 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
8418 *) New function ASN1_mbstring_copy() this copies a string in either
8419 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
8420 into an ASN1_STRING type. A mask of permissible types is passed
8421 and it chooses the "minimal" type to use or an error if not type
8425 *) Add function equivalents to the various macros in asn1.h. The old
8426 macros are retained with an M_ prefix. Code inside the library can
8427 use the M_ macros. External code (including the openssl utility)
8428 should *NOT* in order to be "shared library friendly".
8431 *) Add various functions that can check a certificate's extensions
8432 to see if it usable for various purposes such as SSL client,
8433 server or S/MIME and CAs of these types. This is currently
8434 VERY EXPERIMENTAL but will ultimately be used for certificate chain
8435 verification. Also added a -purpose flag to x509 utility to
8436 print out all the purposes.
8439 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
8443 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
8444 for, obtain and decode and extension and obtain its critical flag.
8445 This allows all the necessary extension code to be handled in a
8446 single function call.
8449 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
8450 platforms. See crypto/rc4/rc4_enc.c for further details.
8453 *) New -noout option to asn1parse. This causes no output to be produced
8454 its main use is when combined with -strparse and -out to extract data
8455 from a file (which may not be in ASN.1 format).
8458 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
8459 when producing the local key id.
8460 [Richard Levitte <levitte@stacken.kth.se>]
8462 *) New option -dhparam in s_server. This allows a DH parameter file to be
8463 stated explicitly. If it is not stated then it tries the first server
8464 certificate file. The previous behaviour hard coded the filename
8468 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
8469 a public key to be input or output. For example:
8470 openssl rsa -in key.pem -pubout -out pubkey.pem
8471 Also added necessary DSA public key functions to handle this.
8474 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
8475 in the message. This was handled by allowing
8476 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8477 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8479 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8480 to the end of the strings whereas this didn't. This would cause problems
8481 if strings read with d2i_ASN1_bytes() were later modified.
8482 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8484 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8485 data and it contains EOF it will end up returning an error. This is
8486 caused by input 46 bytes long. The cause is due to the way base64
8487 BIOs find the start of base64 encoded data. They do this by trying a
8488 trial decode on each line until they find one that works. When they
8489 do a flag is set and it starts again knowing it can pass all the
8490 data directly through the decoder. Unfortunately it doesn't reset
8491 the context it uses. This means that if EOF is reached an attempt
8492 is made to pass two EOFs through the context and this causes the
8493 resulting error. This can also cause other problems as well. As is
8494 usual with these problems it takes *ages* to find and the fix is
8495 trivial: move one line.
8496 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8498 *) Ugly workaround to get s_client and s_server working under Windows. The
8499 old code wouldn't work because it needed to select() on sockets and the
8500 tty (for keypresses and to see if data could be written). Win32 only
8501 supports select() on sockets so we select() with a 1s timeout on the
8502 sockets and then see if any characters are waiting to be read, if none
8503 are present then we retry, we also assume we can always write data to
8504 the tty. This isn't nice because the code then blocks until we've
8505 received a complete line of data and it is effectively polling the
8506 keyboard at 1s intervals: however it's quite a bit better than not
8507 working at all :-) A dedicated Windows application might handle this
8508 with an event loop for example.
8511 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
8512 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
8513 will be called when RSA_sign() and RSA_verify() are used. This is useful
8514 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
8515 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
8516 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
8517 This necessitated the support of an extra signature type NID_md5_sha1
8518 for SSL signatures and modifications to the SSL library to use it instead
8519 of calling RSA_public_decrypt() and RSA_private_encrypt().
8522 *) Add new -verify -CAfile and -CApath options to the crl program, these
8523 will lookup a CRL issuers certificate and verify the signature in a
8524 similar way to the verify program. Tidy up the crl program so it
8525 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
8526 less strict. It will now permit CRL extensions even if it is not
8527 a V2 CRL: this will allow it to tolerate some broken CRLs.
8530 *) Initialize all non-automatic variables each time one of the openssl
8531 sub-programs is started (this is necessary as they may be started
8532 multiple times from the "OpenSSL>" prompt).
8533 [Lennart Bang, Bodo Moeller]
8535 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
8536 removing all other RSA functionality (this is what NO_RSA does). This
8537 is so (for example) those in the US can disable those operations covered
8538 by the RSA patent while allowing storage and parsing of RSA keys and RSA
8542 *) Non-copying interface to BIO pairs.
8543 (still largely untested)
8546 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
8547 ASCII string. This was handled independently in various places before.
8550 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
8551 UTF8 strings a character at a time.
8554 *) Use client_version from client hello to select the protocol
8555 (s23_srvr.c) and for RSA client key exchange verification
8556 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
8559 *) Add various utility functions to handle SPKACs, these were previously
8560 handled by poking round in the structure internals. Added new function
8561 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
8562 print, verify and generate SPKACs. Based on an original idea from
8563 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
8566 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
8569 *) Allow the config file extension section to be overwritten on the
8570 command line. Based on an original idea from Massimiliano Pala
8571 <madwolf@comune.modena.it>. The new option is called -extensions
8572 and can be applied to ca, req and x509. Also -reqexts to override
8573 the request extensions in req and -crlexts to override the crl extensions
8577 *) Add new feature to the SPKAC handling in ca. Now you can include
8578 the same field multiple times by preceding it by "XXXX." for example:
8581 this is the same syntax as used in the req config file.
8584 *) Allow certificate extensions to be added to certificate requests. These
8585 are specified in a 'req_extensions' option of the req section of the
8586 config file. They can be printed out with the -text option to req but
8587 are otherwise ignored at present.
8590 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
8591 data read consists of only the final block it would not decrypted because
8592 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
8593 A misplaced 'break' also meant the decrypted final block might not be
8594 copied until the next read.
8597 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
8598 a few extra parameters to the DH structure: these will be useful if
8599 for example we want the value of 'q' or implement X9.42 DH.
8602 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
8603 provides hooks that allow the default DSA functions or functions on a
8604 "per key" basis to be replaced. This allows hardware acceleration and
8605 hardware key storage to be handled without major modification to the
8606 library. Also added low level modexp hooks and CRYPTO_EX structure and
8607 associated functions.
8610 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
8611 as "read only": it can't be written to and the buffer it points to will
8612 not be freed. Reading from a read only BIO is much more efficient than
8613 a normal memory BIO. This was added because there are several times when
8614 an area of memory needs to be read from a BIO. The previous method was
8615 to create a memory BIO and write the data to it, this results in two
8616 copies of the data and an O(n^2) reading algorithm. There is a new
8617 function BIO_new_mem_buf() which creates a read only memory BIO from
8618 an area of memory. Also modified the PKCS#7 routines to use read only
8622 *) Bugfix: ssl23_get_client_hello did not work properly when called in
8623 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
8624 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
8625 but a retry condition occured while trying to read the rest.
8628 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
8629 NID_pkcs7_encrypted by default: this was wrong since this should almost
8630 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
8631 the encrypted data type: this is a more sensible place to put it and it
8632 allows the PKCS#12 code to be tidied up that duplicated this
8636 *) Changed obj_dat.pl script so it takes its input and output files on
8637 the command line. This should avoid shell escape redirection problems
8641 *) Initial support for certificate extension requests, these are included
8642 in things like Xenroll certificate requests. Included functions to allow
8643 extensions to be obtained and added.
8646 *) -crlf option to s_client and s_server for sending newlines as
8647 CRLF (as required by many protocols).
8650 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
8652 *) Install libRSAglue.a when OpenSSL is built with RSAref.
8653 [Ralf S. Engelschall]
8655 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
8656 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
8658 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
8662 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
8663 DH parameters/keys (q is lost during that conversion, but the resulting
8664 DH parameters contain its length).
8666 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
8667 much faster than DH_generate_parameters (which creates parameters
8668 where p = 2*q + 1), and also the smaller q makes DH computations
8669 much more efficient (160-bit exponentiation instead of 1024-bit
8670 exponentiation); so this provides a convenient way to support DHE
8671 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
8672 utter importance to use
8673 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8675 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8676 when such DH parameters are used, because otherwise small subgroup
8677 attacks may become possible!
8680 *) Avoid memory leak in i2d_DHparams.
8683 *) Allow the -k option to be used more than once in the enc program:
8684 this allows the same encrypted message to be read by multiple recipients.
8687 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
8688 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
8689 it will always use the numerical form of the OID, even if it has a short
8693 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
8694 method only got called if p,q,dmp1,dmq1,iqmp components were present,
8695 otherwise bn_mod_exp was called. In the case of hardware keys for example
8696 no private key components need be present and it might store extra data
8697 in the RSA structure, which cannot be accessed from bn_mod_exp.
8698 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
8699 private key operations.
8702 *) Added support for SPARC Linux.
8705 *) pem_password_cb function type incompatibly changed from
8706 typedef int pem_password_cb(char *buf, int size, int rwflag);
8708 ....(char *buf, int size, int rwflag, void *userdata);
8709 so that applications can pass data to their callbacks:
8710 The PEM[_ASN1]_{read,write}... functions and macros now take an
8711 additional void * argument, which is just handed through whenever
8712 the password callback is called.
8713 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
8715 New function SSL_CTX_set_default_passwd_cb_userdata.
8717 Compatibility note: As many C implementations push function arguments
8718 onto the stack in reverse order, the new library version is likely to
8719 interoperate with programs that have been compiled with the old
8720 pem_password_cb definition (PEM_whatever takes some data that
8721 happens to be on the stack as its last argument, and the callback
8722 just ignores this garbage); but there is no guarantee whatsoever that
8725 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
8726 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
8727 problems not only on Windows, but also on some Unix platforms.
8728 To avoid problematic command lines, these definitions are now in an
8729 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
8730 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
8733 *) MIPS III/IV assembler module is reimplemented.
8736 *) More DES library cleanups: remove references to srand/rand and
8737 delete an unused file.
8740 *) Add support for the the free Netwide assembler (NASM) under Win32,
8741 since not many people have MASM (ml) and it can be hard to obtain.
8742 This is currently experimental but it seems to work OK and pass all
8743 the tests. Check out INSTALL.W32 for info.
8746 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
8747 without temporary keys kept an extra copy of the server key,
8748 and connections with temporary keys did not free everything in case
8752 *) New function RSA_check_key and new openssl rsa option -check
8753 for verifying the consistency of RSA keys.
8754 [Ulf Moeller, Bodo Moeller]
8756 *) Various changes to make Win32 compile work:
8757 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
8758 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
8759 comparison" warnings.
8760 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
8763 *) Add a debugging option to PKCS#5 v2 key generation function: when
8764 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
8765 derived keys are printed to stderr.
8768 *) Copy the flags in ASN1_STRING_dup().
8769 [Roman E. Pavlov <pre@mo.msk.ru>]
8771 *) The x509 application mishandled signing requests containing DSA
8772 keys when the signing key was also DSA and the parameters didn't match.
8774 It was supposed to omit the parameters when they matched the signing key:
8775 the verifying software was then supposed to automatically use the CA's
8776 parameters if they were absent from the end user certificate.
8778 Omitting parameters is no longer recommended. The test was also
8779 the wrong way round! This was probably due to unusual behaviour in
8780 EVP_cmp_parameters() which returns 1 if the parameters match.
8781 This meant that parameters were omitted when they *didn't* match and
8782 the certificate was useless. Certificates signed with 'ca' didn't have
8784 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
8786 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
8787 The interface is as follows:
8788 Applications can use
8789 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
8790 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
8791 "off" is now the default.
8792 The library internally uses
8793 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
8794 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
8795 to disable memory-checking temporarily.
8797 Some inconsistent states that previously were possible (and were
8798 even the default) are now avoided.
8800 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
8801 with each memory chunk allocated; this is occasionally more helpful
8802 than just having a counter.
8804 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
8806 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
8810 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
8811 which largely parallels "options", but is for changing API behaviour,
8812 whereas "options" are about protocol behaviour.
8813 Initial "mode" flags are:
8815 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
8816 a single record has been written.
8817 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
8818 retries use the same buffer location.
8819 (But all of the contents must be
8823 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
8826 *) Fix problems with no-hmac etc.
8827 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
8829 *) New functions RSA_get_default_method(), RSA_set_method() and
8830 RSA_get_method(). These allows replacement of RSA_METHODs without having
8831 to mess around with the internals of an RSA structure.
8834 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
8835 Also really enable memory leak checks in openssl.c and in some
8837 [Chad C. Mulligan, Bodo Moeller]
8839 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
8840 up the length of negative integers. This has now been simplified to just
8841 store the length when it is first determined and use it later, rather
8842 than trying to keep track of where data is copied and updating it to
8844 [Steve Henson, reported by Brien Wheeler
8845 <bwheeler@authentica-security.com>]
8847 *) Add a new function PKCS7_signatureVerify. This allows the verification
8848 of a PKCS#7 signature but with the signing certificate passed to the
8849 function itself. This contrasts with PKCS7_dataVerify which assumes the
8850 certificate is present in the PKCS#7 structure. This isn't always the
8851 case: certificates can be omitted from a PKCS#7 structure and be
8852 distributed by "out of band" means (such as a certificate database).
8855 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
8856 function prototypes in pem.h, also change util/mkdef.pl to add the
8857 necessary function names.
8860 *) mk1mf.pl (used by Windows builds) did not properly read the
8861 options set by Configure in the top level Makefile, and Configure
8862 was not even able to write more than one option correctly.
8863 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
8866 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
8867 file to be loaded from a BIO or FILE pointer. The BIO version will
8868 for example allow memory BIOs to contain config info.
8871 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
8872 Whoever hopes to achieve shared-library compatibility across versions
8873 must use this, not the compile-time macro.
8874 (Exercise 0.9.4: Which is the minimum library version required by
8876 Note: All this applies only to multi-threaded programs, others don't
8880 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
8881 through a BIO pair triggered the default case, i.e.
8882 SSLerr(...,SSL_R_UNKNOWN_STATE).
8885 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
8886 can use the SSL library even if none of the specific BIOs is
8890 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
8891 for the encoded length.
8892 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
8894 *) Add initial documentation of the X509V3 functions.
8897 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
8898 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
8899 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
8900 secure PKCS#8 private key format with a high iteration count.
8903 *) Fix determination of Perl interpreter: A perl or perl5
8904 _directory_ in $PATH was also accepted as the interpreter.
8905 [Ralf S. Engelschall]
8907 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
8908 wrong with it but it was very old and did things like calling
8909 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
8913 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
8914 to use the new extension code.
8917 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
8918 with macros. This should make it easier to change their form, add extra
8919 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
8923 *) Add to configuration table a new entry that can specify an alternative
8924 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
8925 according to Mark Crispin <MRC@Panda.COM>.
8929 *) DES CBC did not update the IV. Weird.
8932 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
8933 Changing the behaviour of the former might break existing programs --
8934 where IV updating is needed, des_ncbc_encrypt can be used.
8937 *) When bntest is run from "make test" it drives bc to check its
8938 calculations, as well as internally checking them. If an internal check
8939 fails, it needs to cause bc to give a non-zero result or make test carries
8940 on without noticing the failure. Fixed.
8943 *) DES library cleanups.
8946 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
8947 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
8948 ciphers. NOTE: although the key derivation function has been verified
8949 against some published test vectors it has not been extensively tested
8950 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
8954 *) Instead of "mkdir -p", which is not fully portable, use new
8955 Perl script "util/mkdir-p.pl".
8958 *) Rewrite the way password based encryption (PBE) is handled. It used to
8959 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
8960 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
8961 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
8962 the 'parameter' field of the AlgorithmIdentifier is passed to the
8963 underlying key generation function so it must do its own ASN1 parsing.
8964 This has also changed the EVP_PBE_CipherInit() function which now has a
8965 'parameter' argument instead of literal salt and iteration count values
8966 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
8969 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
8970 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
8971 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
8972 KEY" because this clashed with PKCS#8 unencrypted string. Since this
8973 value was just used as a "magic string" and not used directly its
8974 value doesn't matter.
8977 *) Introduce some semblance of const correctness to BN. Shame C doesn't
8981 *) "linux-sparc64" configuration (ultrapenguin).
8982 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
8983 "linux-sparc" configuration.
8984 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
8986 *) config now generates no-xxx options for missing ciphers.
8989 *) Support the EBCDIC character set (work in progress).
8990 File ebcdic.c not yet included because it has a different license.
8991 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8993 *) Support BS2000/OSD-POSIX.
8994 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8996 *) Make callbacks for key generation use void * instead of char *.
8999 *) Make S/MIME samples compile (not yet tested).
9002 *) Additional typesafe stacks.
9005 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
9009 Changes between 0.9.3 and 0.9.3a [29 May 1999]
9011 *) New configuration variant "sco5-gcc".
9013 *) Updated some demos.
9014 [Sean O Riordain, Wade Scholine]
9016 *) Add missing BIO_free at exit of pkcs12 application.
9019 *) Fix memory leak in conf.c.
9022 *) Updates for Win32 to assembler version of MD5.
9025 *) Set #! path to perl in apps/der_chop to where we found it
9026 instead of using a fixed path.
9029 *) SHA library changes for irix64-mips4-cc.
9032 *) Improvements for VMS support.
9036 Changes between 0.9.2b and 0.9.3 [24 May 1999]
9038 *) Bignum library bug fix. IRIX 6 passes "make test" now!
9039 This also avoids the problems with SC4.2 and unpatched SC5.
9040 [Andy Polyakov <appro@fy.chalmers.se>]
9042 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
9043 These are required because of the typesafe stack would otherwise break
9044 existing code. If old code used a structure member which used to be STACK
9045 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
9046 sk_num or sk_value it would produce an error because the num, data members
9047 are not present in STACK_OF. Now it just produces a warning. sk_set
9048 replaces the old method of assigning a value to sk_value
9049 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
9050 that does this will no longer work (and should use sk_set instead) but
9051 this could be regarded as a "questionable" behaviour anyway.
9054 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
9055 correctly handle encrypted S/MIME data.
9058 *) Change type of various DES function arguments from des_cblock
9059 (which means, in function argument declarations, pointer to char)
9060 to des_cblock * (meaning pointer to array with 8 char elements),
9061 which allows the compiler to do more typechecking; it was like
9062 that back in SSLeay, but with lots of ugly casts.
9064 Introduce new type const_des_cblock.
9067 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
9068 problems: find RecipientInfo structure that matches recipient certificate
9069 and initialise the ASN1 structures properly based on passed cipher.
9072 *) Belatedly make the BN tests actually check the results.
9075 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
9076 to and from BNs: it was completely broken. New compilation option
9077 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
9078 key elements as negative integers.
9081 *) Reorganize and speed up MD5.
9082 [Andy Polyakov <appro@fy.chalmers.se>]
9085 [Richard Levitte <richard@levitte.org>]
9087 *) New option -out to asn1parse to allow the parsed structure to be
9088 output to a file. This is most useful when combined with the -strparse
9089 option to examine the output of things like OCTET STRINGS.
9092 *) Make SSL library a little more fool-proof by not requiring any longer
9093 that SSL_set_{accept,connect}_state be called before
9094 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
9095 in many applications because usually everything *appeared* to work as
9096 intended anyway -- now it really works as intended).
9099 *) Move openssl.cnf out of lib/.
9102 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
9103 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
9104 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
9105 [Ralf S. Engelschall]
9107 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
9108 handle PKCS#7 enveloped data properly.
9109 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
9111 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
9112 copying pointers. The cert_st handling is changed by this in
9113 various ways (and thus what used to be known as ctx->default_cert
9114 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
9115 any longer when s->cert does not give us what we need).
9116 ssl_cert_instantiate becomes obsolete by this change.
9117 As soon as we've got the new code right (possibly it already is?),
9118 we have solved a couple of bugs of the earlier code where s->cert
9119 was used as if it could not have been shared with other SSL structures.
9121 Note that using the SSL API in certain dirty ways now will result
9122 in different behaviour than observed with earlier library versions:
9123 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
9124 does not influence s as it used to.
9126 In order to clean up things more thoroughly, inside SSL_SESSION
9127 we don't use CERT any longer, but a new structure SESS_CERT
9128 that holds per-session data (if available); currently, this is
9129 the peer's certificate chain and, for clients, the server's certificate
9130 and temporary key. CERT holds only those values that can have
9131 meaningful defaults in an SSL_CTX.
9134 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
9135 from the internal representation. Various PKCS#7 fixes: remove some
9136 evil casts and set the enc_dig_alg field properly based on the signing
9140 *) Allow PKCS#12 password to be set from the command line or the
9141 environment. Let 'ca' get its config file name from the environment
9142 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
9146 *) Allow certificate policies extension to use an IA5STRING for the
9147 organization field. This is contrary to the PKIX definition but
9148 VeriSign uses it and IE5 only recognises this form. Document 'x509'
9152 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
9153 without disallowing inline assembler and the like for non-pedantic builds.
9156 *) Support Borland C++ builder.
9157 [Janez Jere <jj@void.si>, modified by Ulf Möller]
9162 *) SHA-1 cleanups and performance enhancements.
9163 [Andy Polyakov <appro@fy.chalmers.se>]
9165 *) Sparc v8plus assembler for the bignum library.
9166 [Andy Polyakov <appro@fy.chalmers.se>]
9168 *) Accept any -xxx and +xxx compiler options in Configure.
9171 *) Update HPUX configuration.
9174 *) Add missing sk_<type>_unshift() function to safestack.h
9175 [Ralf S. Engelschall]
9177 *) New function SSL_CTX_use_certificate_chain_file that sets the
9178 "extra_cert"s in addition to the certificate. (This makes sense
9179 only for "PEM" format files, as chains as a whole are not
9183 *) Support verify_depth from the SSL API.
9184 x509_vfy.c had what can be considered an off-by-one-error:
9185 Its depth (which was not part of the external interface)
9186 was actually counting the number of certificates in a chain;
9187 now it really counts the depth.
9190 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
9191 instead of X509err, which often resulted in confusing error
9192 messages since the error codes are not globally unique
9193 (e.g. an alleged error in ssl3_accept when a certificate
9194 didn't match the private key).
9196 *) New function SSL_CTX_set_session_id_context that allows to set a default
9197 value (so that you don't need SSL_set_session_id_context for each
9198 connection using the SSL_CTX).
9201 *) OAEP decoding bug fix.
9204 *) Support INSTALL_PREFIX for package builders, as proposed by
9208 *) New Configure options "threads" and "no-threads". For systems
9209 where the proper compiler options are known (currently Solaris
9210 and Linux), "threads" is the default.
9213 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
9216 *) Install various scripts to $(OPENSSLDIR)/misc, not to
9217 $(INSTALLTOP)/bin -- they shouldn't clutter directories
9218 such as /usr/local/bin.
9221 *) "make linux-shared" to build shared libraries.
9222 [Niels Poppe <niels@netbox.org>]
9224 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
9227 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
9228 extension adding in x509 utility.
9231 *) Remove NOPROTO sections and error code comments.
9234 *) Partial rewrite of the DEF file generator to now parse the ANSI
9238 *) New Configure options --prefix=DIR and --openssldir=DIR.
9241 *) Complete rewrite of the error code script(s). It is all now handled
9242 by one script at the top level which handles error code gathering,
9243 header rewriting and C source file generation. It should be much better
9244 than the old method: it now uses a modified version of Ulf's parser to
9245 read the ANSI prototypes in all header files (thus the old K&R definitions
9246 aren't needed for error creation any more) and do a better job of
9247 translating function codes into names. The old 'ASN1 error code imbedded
9248 in a comment' is no longer necessary and it doesn't use .err files which
9249 have now been deleted. Also the error code call doesn't have to appear all
9250 on one line (which resulted in some large lines...).
9253 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
9256 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
9257 0 (which usually indicates a closed connection), but continue reading.
9260 *) Fix some race conditions.
9263 *) Add support for CRL distribution points extension. Add Certificate
9264 Policies and CRL distribution points documentation.
9267 *) Move the autogenerated header file parts to crypto/opensslconf.h.
9270 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
9271 8 of keying material. Merlin has also confirmed interop with this fix
9272 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
9273 [Merlin Hughes <merlin@baltimore.ie>]
9275 *) Fix lots of warnings.
9276 [Richard Levitte <levitte@stacken.kth.se>]
9278 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
9279 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
9280 [Richard Levitte <levitte@stacken.kth.se>]
9282 *) Fix problems with sizeof(long) == 8.
9283 [Andy Polyakov <appro@fy.chalmers.se>]
9285 *) Change functions to ANSI C.
9288 *) Fix typos in error codes.
9289 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
9291 *) Remove defunct assembler files from Configure.
9294 *) SPARC v8 assembler BIGNUM implementation.
9295 [Andy Polyakov <appro@fy.chalmers.se>]
9297 *) Support for Certificate Policies extension: both print and set.
9298 Various additions to support the r2i method this uses.
9301 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
9302 return a const string when you are expecting an allocated buffer.
9305 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
9306 types DirectoryString and DisplayText.
9309 *) Add code to allow r2i extensions to access the configuration database,
9310 add an LHASH database driver and add several ctx helper functions.
9313 *) Fix an evil bug in bn_expand2() which caused various BN functions to
9314 fail when they extended the size of a BIGNUM.
9317 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
9318 support typesafe stack.
9321 *) Fix typo in SSL_[gs]et_options().
9322 [Nils Frostberg <nils@medcom.se>]
9324 *) Delete various functions and files that belonged to the (now obsolete)
9325 old X509V3 handling code.
9328 *) New Configure option "rsaref".
9331 *) Don't auto-generate pem.h.
9334 *) Introduce type-safe ASN.1 SETs.
9337 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
9338 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
9340 *) Introduce type-safe STACKs. This will almost certainly break lots of code
9341 that links with OpenSSL (well at least cause lots of warnings), but fear
9342 not: the conversion is trivial, and it eliminates loads of evil casts. A
9343 few STACKed things have been converted already. Feel free to convert more.
9344 In the fullness of time, I'll do away with the STACK type altogether.
9347 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
9348 specified in <certfile> by updating the entry in the index.txt file.
9349 This way one no longer has to edit the index.txt file manually for
9350 revoking a certificate. The -revoke option does the gory details now.
9351 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
9353 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
9354 `-text' option at all and this way the `-noout -text' combination was
9355 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
9356 [Ralf S. Engelschall]
9358 *) Make sure a corresponding plain text error message exists for the
9359 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
9360 verify callback function determined that a certificate was revoked.
9361 [Ralf S. Engelschall]
9363 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
9364 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
9365 all available cipers including rc5, which was forgotten until now.
9366 In order to let the testing shell script know which algorithms
9367 are available, a new (up to now undocumented) command
9368 "openssl list-cipher-commands" is used.
9371 *) Bugfix: s_client occasionally would sleep in select() when
9372 it should have checked SSL_pending() first.
9375 *) New functions DSA_do_sign and DSA_do_verify to provide access to
9376 the raw DSA values prior to ASN.1 encoding.
9379 *) Tweaks to Configure
9380 [Niels Poppe <niels@netbox.org>]
9382 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
9386 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
9389 *) New config option to avoid instructions that are illegal on the 80386.
9390 The default code is faster, but requires at least a 486.
9393 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
9394 SSL2_SERVER_VERSION (not used at all) macros, which are now the
9395 same as SSL2_VERSION anyway.
9398 *) New "-showcerts" option for s_client.
9401 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
9402 application. Various cleanups and fixes.
9405 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
9406 modify error routines to work internally. Add error codes and PBE init
9407 to library startup routines.
9410 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
9411 packing functions to asn1 and evp. Changed function names and error
9412 codes along the way.
9415 *) PKCS12 integration: and so it begins... First of several patches to
9416 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
9417 objects to objects.h
9420 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
9421 and display support for Thawte strong extranet extension.
9424 *) Add LinuxPPC support.
9425 [Jeff Dubrule <igor@pobox.org>]
9427 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
9428 bn_div_words in alpha.s.
9429 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
9431 *) Make sure the RSA OAEP test is skipped under -DRSAref because
9432 OAEP isn't supported when OpenSSL is built with RSAref.
9433 [Ulf Moeller <ulf@fitug.de>]
9435 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
9436 so they no longer are missing under -DNOPROTO.
9437 [Soren S. Jorvang <soren@t.dk>]
9440 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
9442 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
9443 doesn't work when the session is reused. Coming soon!
9446 *) Fix a security hole, that allows sessions to be reused in the wrong
9447 context thus bypassing client cert protection! All software that uses
9448 client certs and session caches in multiple contexts NEEDS PATCHING to
9449 allow session reuse! A fuller solution is in the works.
9450 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
9452 *) Some more source tree cleanups (removed obsolete files
9453 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
9454 permission on "config" script to be executable) and a fix for the INSTALL
9456 [Ulf Moeller <ulf@fitug.de>]
9458 *) Remove some legacy and erroneous uses of malloc, free instead of
9460 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
9462 *) Make rsa_oaep_test return non-zero on error.
9463 [Ulf Moeller <ulf@fitug.de>]
9465 *) Add support for native Solaris shared libraries. Configure
9466 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
9467 if someone would make that last step automatic.
9468 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
9470 *) ctx_size was not built with the right compiler during "make links". Fixed.
9473 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
9474 except NULL ciphers". This means the default cipher list will no longer
9475 enable NULL ciphers. They need to be specifically enabled e.g. with
9476 the string "DEFAULT:eNULL".
9479 *) Fix to RSA private encryption routines: if p < q then it would
9480 occasionally produce an invalid result. This will only happen with
9481 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9484 *) Be less restrictive and allow also `perl util/perlpath.pl
9485 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9486 because this way one can also use an interpreter named `perl5' (which is
9487 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9488 installed as `perl').
9489 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9491 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9492 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9494 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9495 advapi32.lib to Win32 build and change the pem test comparision
9496 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9497 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9498 and crypto/des/ede_cbcm_enc.c.
9501 *) DES quad checksum was broken on big-endian architectures. Fixed.
9504 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9505 Win32 test batch file so it (might) work again. The Win32 test batch file
9506 is horrible: I feel ill....
9509 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
9510 in e_os.h. Audit of header files to check ANSI and non ANSI
9511 sections: 10 functions were absent from non ANSI section and not exported
9512 from Windows DLLs. Fixed up libeay.num for new functions.
9515 *) Make `openssl version' output lines consistent.
9516 [Ralf S. Engelschall]
9518 *) Fix Win32 symbol export lists for BIO functions: Added
9519 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
9520 to ms/libeay{16,32}.def.
9521 [Ralf S. Engelschall]
9523 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
9524 fine under Unix and passes some trivial tests I've now added. But the
9525 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
9526 added to make sure no one expects that this stuff really works in the
9527 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
9528 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
9530 [Ralf S. Engelschall]
9532 *) Fix the generation of two part addresses in perl.
9533 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
9535 *) Add config entry for Linux on MIPS.
9536 [John Tobey <jtobey@channel1.com>]
9538 *) Make links whenever Configure is run, unless we are on Windoze.
9541 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
9542 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
9546 *) Add a useful kludge to allow package maintainers to specify compiler and
9547 other platforms details on the command line without having to patch the
9548 Configure script everytime: One now can use ``perl Configure
9549 <id>:<details>'', i.e. platform ids are allowed to have details appended
9550 to them (seperated by colons). This is treated as there would be a static
9551 pre-configured entry in Configure's %table under key <id> with value
9552 <details> and ``perl Configure <id>'' is called. So, when you want to
9553 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
9554 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
9555 now, which overrides the FreeBSD-elf entry on-the-fly.
9556 [Ralf S. Engelschall]
9558 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
9561 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
9562 on the `perl Configure ...' command line. This way one can compile
9563 OpenSSL libraries with Position Independent Code (PIC) which is needed
9564 for linking it into DSOs.
9565 [Ralf S. Engelschall]
9567 *) Remarkably, export ciphers were totally broken and no-one had noticed!
9571 *) Cleaned up the LICENSE document: The official contact for any license
9572 questions now is the OpenSSL core team under openssl-core@openssl.org.
9573 And add a paragraph about the dual-license situation to make sure people
9574 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
9575 to the OpenSSL toolkit.
9576 [Ralf S. Engelschall]
9578 *) General source tree makefile cleanups: Made `making xxx in yyy...'
9579 display consistent in the source tree and replaced `/bin/rm' by `rm'.
9580 Additonally cleaned up the `make links' target: Remove unnecessary
9581 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
9582 to speed processing and no longer clutter the display with confusing
9583 stuff. Instead only the actually done links are displayed.
9584 [Ralf S. Engelschall]
9586 *) Permit null encryption ciphersuites, used for authentication only. It used
9587 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
9588 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
9592 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
9593 signed attributes when verifying signatures (this would break them),
9594 the detached data encoding was wrong and public keys obtained using
9595 X509_get_pubkey() weren't freed.
9598 *) Add text documentation for the BUFFER functions. Also added a work around
9599 to a Win95 console bug. This was triggered by the password read stuff: the
9600 last character typed gets carried over to the next fread(). If you were
9601 generating a new cert request using 'req' for example then the last
9602 character of the passphrase would be CR which would then enter the first
9606 *) Added the new `Includes OpenSSL Cryptography Software' button as
9607 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
9608 button and can be used by applications based on OpenSSL to show the
9609 relationship to the OpenSSL project.
9610 [Ralf S. Engelschall]
9612 *) Remove confusing variables in function signatures in files
9613 ssl/ssl_lib.c and ssl/ssl.h.
9614 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9616 *) Don't install bss_file.c under PREFIX/include/
9617 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9619 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
9620 functions that return function pointers and has support for NT specific
9621 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
9622 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
9623 unsigned to signed types: this was killing the Win32 compile.
9626 *) Add new certificate file to stack functions,
9627 SSL_add_dir_cert_subjects_to_stack() and
9628 SSL_add_file_cert_subjects_to_stack(). These largely supplant
9629 SSL_load_client_CA_file(), and can be used to add multiple certs easily
9630 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
9631 This means that Apache-SSL and similar packages don't have to mess around
9632 to add as many CAs as they want to the preferred list.
9635 *) Experiment with doxygen documentation. Currently only partially applied to
9637 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
9638 openssl.doxy as the configuration file.
9641 *) Get rid of remaining C++-style comments which strict C compilers hate.
9642 [Ralf S. Engelschall, pointed out by Carlos Amengual]
9644 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
9645 compiled in by default: it has problems with large keys.
9648 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
9649 DH private keys and/or callback functions which directly correspond to
9650 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
9651 is needed for applications which have to configure certificates on a
9652 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
9654 For the RSA certificate situation is makes no difference, but
9655 for the DSA certificate situation this fixes the "no shared cipher"
9656 problem where the OpenSSL cipher selection procedure failed because the
9657 temporary keys were not overtaken from the context and the API provided
9658 no way to reconfigure them.
9659 The new functions now let applications reconfigure the stuff and they
9660 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
9661 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
9662 non-public-API function ssl_cert_instantiate() is used as a helper
9663 function and also to reduce code redundancy inside ssl_rsa.c.
9664 [Ralf S. Engelschall]
9666 *) Move s_server -dcert and -dkey options out of the undocumented feature
9667 area because they are useful for the DSA situation and should be
9668 recognized by the users.
9669 [Ralf S. Engelschall]
9671 *) Fix the cipher decision scheme for export ciphers: the export bits are
9672 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
9673 SSL_EXP_MASK. So, the original variable has to be used instead of the
9674 already masked variable.
9675 [Richard Levitte <levitte@stacken.kth.se>]
9677 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
9678 [Richard Levitte <levitte@stacken.kth.se>]
9680 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
9681 from `int' to `unsigned int' because it's a length and initialized by
9682 EVP_DigestFinal() which expects an `unsigned int *'.
9683 [Richard Levitte <levitte@stacken.kth.se>]
9685 *) Don't hard-code path to Perl interpreter on shebang line of Configure
9686 script. Instead use the usual Shell->Perl transition trick.
9687 [Ralf S. Engelschall]
9689 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
9690 (in addition to RSA certificates) to match the behaviour of `openssl dsa
9691 -noout -modulus' as it's already the case for `openssl rsa -noout
9692 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
9693 currently the public key is printed (a decision which was already done by
9694 `openssl dsa -modulus' in the past) which serves a similar purpose.
9695 Additionally the NO_RSA no longer completely removes the whole -modulus
9696 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
9698 [Ralf S. Engelschall]
9700 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
9701 BIO. See the source (crypto/evp/bio_ok.c) for more info.
9702 [Arne Ansper <arne@ats.cyber.ee>]
9704 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
9705 to be added. Now both 'req' and 'ca' can use new objects defined in the
9709 *) Add cool BIO that does syslog (or event log on NT).
9710 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
9712 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
9713 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
9714 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
9715 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
9718 *) Add preliminary config info for new extension code.
9721 *) Make RSA_NO_PADDING really use no padding.
9722 [Ulf Moeller <ulf@fitug.de>]
9724 *) Generate errors when private/public key check is done.
9727 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
9728 for some CRL extensions and new objects added.
9731 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
9732 key usage extension and fuller support for authority key id.
9735 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
9736 padding method for RSA, which is recommended for new applications in PKCS
9737 #1 v2.0 (RFC 2437, October 1998).
9738 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
9739 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
9740 against Bleichbacher's attack on RSA.
9741 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
9744 *) Updates to the new SSL compression code
9745 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9747 *) Fix so that the version number in the master secret, when passed
9748 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
9749 (because the server will not accept higher), that the version number
9750 is 0x03,0x01, not 0x03,0x00
9751 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9753 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
9754 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
9755 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
9758 *) Support for RAW extensions where an arbitrary extension can be
9759 created by including its DER encoding. See apps/openssl.cnf for
9763 *) Make sure latest Perl versions don't interpret some generated C array
9764 code as Perl array code in the crypto/err/err_genc.pl script.
9765 [Lars Weber <3weber@informatik.uni-hamburg.de>]
9767 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
9768 not many people have the assembler. Various Win32 compilation fixes and
9769 update to the INSTALL.W32 file with (hopefully) more accurate Win32
9773 *) Modify configure script 'Configure' to automatically create crypto/date.h
9774 file under Win32 and also build pem.h from pem.org. New script
9775 util/mkfiles.pl to create the MINFO file on environments that can't do a
9776 'make files': perl util/mkfiles.pl >MINFO should work.
9779 *) Major rework of DES function declarations, in the pursuit of correctness
9780 and purity. As a result, many evil casts evaporated, and some weirdness,
9781 too. You may find this causes warnings in your code. Zapping your evil
9782 casts will probably fix them. Mostly.
9785 *) Fix for a typo in asn1.h. Bug fix to object creation script
9786 obj_dat.pl. It considered a zero in an object definition to mean
9787 "end of object": none of the objects in objects.h have any zeros
9788 so it wasn't spotted.
9789 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
9791 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
9792 Masking (CBCM). In the absence of test vectors, the best I have been able
9793 to do is check that the decrypt undoes the encrypt, so far. Send me test
9794 vectors if you have them.
9797 *) Correct calculation of key length for export ciphers (too much space was
9798 allocated for null ciphers). This has not been tested!
9801 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
9802 message is now correct (it understands "crypto" and "ssl" on its
9803 command line). There is also now an "update" option. This will update
9804 the util/ssleay.num and util/libeay.num files with any new functions.
9806 perl util/mkdef.pl crypto ssl update
9807 it will update them.
9810 *) Overhauled the Perl interface (perl/*):
9811 - ported BN stuff to OpenSSL's different BN library
9812 - made the perl/ source tree CVS-aware
9813 - renamed the package from SSLeay to OpenSSL (the files still contain
9814 their history because I've copied them in the repository)
9815 - removed obsolete files (the test scripts will be replaced
9816 by better Test::Harness variants in the future)
9817 [Ralf S. Engelschall]
9819 *) First cut for a very conservative source tree cleanup:
9820 1. merge various obsolete readme texts into doc/ssleay.txt
9821 where we collect the old documents and readme texts.
9822 2. remove the first part of files where I'm already sure that we no
9823 longer need them because of three reasons: either they are just temporary
9824 files which were left by Eric or they are preserved original files where
9825 I've verified that the diff is also available in the CVS via "cvs diff
9826 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
9827 the crypto/md/ stuff).
9828 [Ralf S. Engelschall]
9830 *) More extension code. Incomplete support for subject and issuer alt
9831 name, issuer and authority key id. Change the i2v function parameters
9832 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
9833 what that's for :-) Fix to ASN1 macro which messed up
9834 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
9837 *) Preliminary support for ENUMERATED type. This is largely copied from the
9841 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
9842 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9844 *) Make sure `make rehash' target really finds the `openssl' program.
9845 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9847 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
9848 like to hear about it if this slows down other processors.
9851 *) Add CygWin32 platform information to Configure script.
9852 [Alan Batie <batie@aahz.jf.intel.com>]
9854 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
9855 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
9857 *) New program nseq to manipulate netscape certificate sequences
9860 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
9864 *) Fixes to BN code. Previously the default was to define BN_RECURSION
9865 but the BN code had some problems that would cause failures when
9866 doing certificate verification and some other functions.
9867 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9869 *) Add ASN1 and PEM code to support netscape certificate sequences.
9872 *) Add ASN1 and PEM code to support netscape certificate sequences.
9875 *) Add several PKIX and private extended key usage OIDs.
9878 *) Modify the 'ca' program to handle the new extension code. Modify
9879 openssl.cnf for new extension format, add comments.
9882 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
9883 and add a sample to openssl.cnf so req -x509 now adds appropriate
9887 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
9888 error code, add initial support to X509_print() and x509 application.
9891 *) Takes a deep breath and start addding X509 V3 extension support code. Add
9892 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
9893 stuff is currently isolated and isn't even compiled yet.
9896 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
9897 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
9898 Removed the versions check from X509 routines when loading extensions:
9899 this allows certain broken certificates that don't set the version
9900 properly to be processed.
9903 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
9904 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
9905 can still be regenerated with "make depend".
9908 *) Spelling mistake in C version of CAST-128.
9909 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
9911 *) Changes to the error generation code. The perl script err-code.pl
9912 now reads in the old error codes and retains the old numbers, only
9913 adding new ones if necessary. It also only changes the .err files if new
9914 codes are added. The makefiles have been modified to only insert errors
9915 when needed (to avoid needlessly modifying header files). This is done
9916 by only inserting errors if the .err file is newer than the auto generated
9917 C file. To rebuild all the error codes from scratch (the old behaviour)
9918 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
9919 or delete all the .err files.
9922 *) CAST-128 was incorrectly implemented for short keys. The C version has
9923 been fixed, but is untested. The assembler versions are also fixed, but
9924 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
9925 to regenerate it if needed.
9926 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
9927 Hagino <itojun@kame.net>]
9929 *) File was opened incorrectly in randfile.c.
9930 [Ulf Möller <ulf@fitug.de>]
9932 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
9933 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
9934 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
9935 al: it's just almost always a UTCTime. Note this patch adds new error
9936 codes so do a "make errors" if there are problems.
9939 *) Correct Linux 1 recognition in config.
9940 [Ulf Möller <ulf@fitug.de>]
9942 *) Remove pointless MD5 hash when using DSA keys in ca.
9943 [Anonymous <nobody@replay.com>]
9945 *) Generate an error if given an empty string as a cert directory. Also
9946 generate an error if handed NULL (previously returned 0 to indicate an
9947 error, but didn't set one).
9948 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
9950 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
9953 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
9954 parameters. This was causing a warning which killed off the Win32 compile.
9957 *) Remove C++ style comments from crypto/bn/bn_local.h.
9958 [Neil Costigan <neil.costigan@celocom.com>]
9960 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
9961 based on a text string, looking up short and long names and finally
9962 "dot" format. The "dot" format stuff didn't work. Added new function
9963 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
9964 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
9965 OID is not part of the table.
9968 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
9969 X509_LOOKUP_by_alias().
9972 *) Sort openssl functions by name.
9975 *) Get the gendsa program working (hopefully) and add it to app list. Remove
9976 encryption from sample DSA keys (in case anyone is interested the password
9980 *) Make _all_ *_free functions accept a NULL pointer.
9981 [Frans Heymans <fheymans@isaserver.be>]
9983 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
9985 [Anonymous <nobody@replay.com>]
9987 *) s_server should send the CAfile as acceptable CAs, not its own cert.
9988 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9990 *) Don't blow it for numeric -newkey arguments to apps/req.
9991 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9993 *) Temp key "for export" tests were wrong in s3_srvr.c.
9994 [Anonymous <nobody@replay.com>]
9996 *) Add prototype for temp key callback functions
9997 SSL_CTX_set_tmp_{rsa,dh}_callback().
10000 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
10001 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
10004 *) X509_name_add_entry() freed the wrong thing after an error.
10005 [Arne Ansper <arne@ats.cyber.ee>]
10007 *) rsa_eay.c would attempt to free a NULL context.
10008 [Arne Ansper <arne@ats.cyber.ee>]
10010 *) BIO_s_socket() had a broken should_retry() on Windoze.
10011 [Arne Ansper <arne@ats.cyber.ee>]
10013 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
10014 [Arne Ansper <arne@ats.cyber.ee>]
10016 *) Make sure the already existing X509_STORE->depth variable is initialized
10017 in X509_STORE_new(), but document the fact that this variable is still
10018 unused in the certificate verification process.
10019 [Ralf S. Engelschall]
10021 *) Fix the various library and apps files to free up pkeys obtained from
10022 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
10025 *) Fix reference counting in X509_PUBKEY_get(). This makes
10026 demos/maurice/example2.c work, amongst others, probably.
10027 [Steve Henson and Ben Laurie]
10029 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
10030 `openssl' and second, the shortcut symlinks for the `openssl <command>'
10031 are no longer created. This way we have a single and consistent command
10032 line interface `openssl <command>', similar to `cvs <command>'.
10033 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
10035 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
10036 BIT STRING wrapper always have zero unused bits.
10039 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
10042 *) Make the top-level INSTALL documentation easier to understand.
10045 *) Makefiles updated to exit if an error occurs in a sub-directory
10046 make (including if user presses ^C) [Paul Sutton]
10048 *) Make Montgomery context stuff explicit in RSA data structure.
10051 *) Fix build order of pem and err to allow for generated pem.h.
10054 *) Fix renumbering bug in X509_NAME_delete_entry().
10057 *) Enhanced the err-ins.pl script so it makes the error library number
10058 global and can add a library name. This is needed for external ASN1 and
10059 other error libraries.
10062 *) Fixed sk_insert which never worked properly.
10065 *) Fix ASN1 macros so they can handle indefinite length construted
10066 EXPLICIT tags. Some non standard certificates use these: they can now
10070 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
10071 into a single doc/ssleay.txt bundle. This way the information is still
10072 preserved but no longer messes up this directory. Now it's new room for
10073 the new set of documenation files.
10074 [Ralf S. Engelschall]
10076 *) SETs were incorrectly DER encoded. This was a major pain, because they
10077 shared code with SEQUENCEs, which aren't coded the same. This means that
10078 almost everything to do with SETs or SEQUENCEs has either changed name or
10079 number of arguments.
10080 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
10082 *) Fix test data to work with the above.
10085 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
10086 was already fixed by Eric for 0.9.1 it seems.
10087 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
10089 *) Autodetect FreeBSD3.
10092 *) Fix various bugs in Configure. This affects the following platforms:
10096 unixware-2.0-pentium
10100 *) Eliminate generated files from CVS. Reorder tests to regenerate files
10101 before they are needed.
10104 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
10108 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
10110 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
10111 changed SSLeay to OpenSSL in version strings.
10112 [Ralf S. Engelschall]
10114 *) Some fixups to the top-level documents.
10117 *) Fixed the nasty bug where rsaref.h was not found under compile-time
10118 because the symlink to include/ was missing.
10119 [Ralf S. Engelschall]
10121 *) Incorporated the popular no-RSA/DSA-only patches
10122 which allow to compile a RSA-free SSLeay.
10123 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
10125 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
10126 when "ssleay" is still not found.
10127 [Ralf S. Engelschall]
10129 *) Added more platforms to Configure: Cray T3E, HPUX 11,
10130 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
10132 *) Updated the README file.
10133 [Ralf S. Engelschall]
10135 *) Added various .cvsignore files in the CVS repository subdirs
10136 to make a "cvs update" really silent.
10137 [Ralf S. Engelschall]
10139 *) Recompiled the error-definition header files and added
10140 missing symbols to the Win32 linker tables.
10141 [Ralf S. Engelschall]
10143 *) Cleaned up the top-level documents;
10144 o new files: CHANGES and LICENSE
10145 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
10146 o merged COPYRIGHT into LICENSE
10147 o removed obsolete TODO file
10148 o renamed MICROSOFT to INSTALL.W32
10149 [Ralf S. Engelschall]
10151 *) Removed dummy files from the 0.9.1b source tree:
10152 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
10153 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
10154 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
10155 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
10156 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
10157 [Ralf S. Engelschall]
10159 *) Added various platform portability fixes.
10162 *) The Genesis of the OpenSSL rpject:
10163 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
10164 Young and Tim J. Hudson created while they were working for C2Net until
10166 [The OpenSSL Project]
10169 Changes between 0.9.0b and 0.9.1b [not released]
10171 *) Updated a few CA certificates under certs/
10174 *) Changed some BIGNUM api stuff.
10177 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
10178 DGUX x86, Linux Alpha, etc.
10181 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
10182 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
10186 *) Add -strparse option to asn1pars program which parses nested
10188 [Dr Stephen Henson <shenson@bigfoot.com>]
10190 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
10193 *) DSA fix for "ca" program.
10196 *) Added "-genkey" option to "dsaparam" program.
10199 *) Added RIPE MD160 (rmd160) message digest.
10202 *) Added -a (all) option to "ssleay version" command.
10205 *) Added PLATFORM define which is the id given to Configure.
10208 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
10211 *) Extended the ASN.1 parser routines.
10214 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
10217 *) Added a BN_CTX to the BN library.
10220 *) Fixed the weak key values in DES library
10223 *) Changed API in EVP library for cipher aliases.
10226 *) Added support for RC2/64bit cipher.
10229 *) Converted the lhash library to the crypto/mem.c functions.
10232 *) Added more recognized ASN.1 object ids.
10235 *) Added more RSA padding checks for SSL/TLS.
10238 *) Added BIO proxy/filter functionality.
10241 *) Added extra_certs to SSL_CTX which can be used
10242 send extra CA certificates to the client in the CA cert chain sending
10243 process. It can be configured with SSL_CTX_add_extra_chain_cert().
10246 *) Now Fortezza is denied in the authentication phase because
10247 this is key exchange mechanism is not supported by SSLeay at all.
10250 *) Additional PKCS1 checks.
10253 *) Support the string "TLSv1" for all TLS v1 ciphers.
10256 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
10257 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
10260 *) Fixed a few memory leaks.
10263 *) Fixed various code and comment typos.
10266 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
10267 bytes sent in the client random.
10268 [Edward Bishop <ebishop@spyglass.com>]