5 Changes between 1.0.1j and 1.0.1k [xx XXX xxxx]
7 *) Tighten client-side session ticket handling during renegotiation:
8 ensure that the client only accepts a session ticket if the server sends
9 the extension anew in the ServerHello. Previously, a TLS client would
10 reuse the old extension state and thus accept a session ticket if one was
11 announced in the initial ServerHello.
14 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
18 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
19 sends a carefully crafted handshake message, to cause OpenSSL to fail
20 to free up to 64k of memory causing a memory leak. This could be
21 exploited in a Denial Of Service attack. This issue affects OpenSSL
22 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
23 whether SRTP is used or configured. Implementations of OpenSSL that
24 have been compiled with OPENSSL_NO_SRTP defined are not affected.
26 The fix was developed by the OpenSSL team.
30 *) Session Ticket Memory Leak.
32 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
33 integrity of that ticket is first verified. In the event of a session
34 ticket integrity check failing, OpenSSL will fail to free memory
35 causing a memory leak. By sending a large number of invalid session
36 tickets an attacker could exploit this issue in a Denial Of Service
41 *) Build option no-ssl3 is incomplete.
43 When OpenSSL is configured with "no-ssl3" as a build option, servers
44 could accept and complete a SSL 3.0 handshake, and clients could be
45 configured to send them.
47 [Akamai and the OpenSSL team]
49 *) Add support for TLS_FALLBACK_SCSV.
50 Client applications doing fallback retries should call
51 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
53 [Adam Langley, Bodo Moeller]
55 *) Add additional DigestInfo checks.
57 Reencode DigestInto in DER and check against the original when
58 verifying RSA signature: this will reject any improperly encoded
59 DigestInfo structures.
61 Note: this is a precautionary measure and no attacks are currently known.
65 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
67 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
68 SRP code can be overrun an internal buffer. Add sanity check that
69 g, A, B < N to SRP code.
71 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
72 Group for discovering this issue.
76 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
77 TLS 1.0 instead of higher protocol versions when the ClientHello message
78 is badly fragmented. This allows a man-in-the-middle attacker to force a
79 downgrade to TLS 1.0 even if both the server and the client support a
80 higher protocol version, by modifying the client's TLS records.
82 Thanks to David Benjamin and Adam Langley (Google) for discovering and
83 researching this issue.
87 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
88 to a denial of service attack. A malicious server can crash the client
89 with a null pointer dereference (read) by specifying an anonymous (EC)DH
90 ciphersuite and sending carefully crafted handshake messages.
92 Thanks to Felix Gröbert (Google) for discovering and researching this
97 *) By sending carefully crafted DTLS packets an attacker could cause openssl
98 to leak memory. This can be exploited through a Denial of Service attack.
99 Thanks to Adam Langley for discovering and researching this issue.
103 *) An attacker can force openssl to consume large amounts of memory whilst
104 processing DTLS handshake messages. This can be exploited through a
105 Denial of Service attack.
106 Thanks to Adam Langley for discovering and researching this issue.
110 *) An attacker can force an error condition which causes openssl to crash
111 whilst processing DTLS packets due to memory being freed twice. This
112 can be exploited through a Denial of Service attack.
113 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
118 *) If a multithreaded client connects to a malicious server using a resumed
119 session and the server sends an ec point format extension it could write
120 up to 255 bytes to freed memory.
122 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
127 *) A malicious server can crash an OpenSSL client with a null pointer
128 dereference (read) by specifying an SRP ciphersuite even though it was not
129 properly negotiated with the client. This can be exploited through a
130 Denial of Service attack.
132 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
133 discovering and researching this issue.
137 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
138 X509_name_oneline, X509_name_print_ex et al. to leak some information
139 from the stack. Applications may be affected if they echo pretty printing
140 output to the attacker.
142 Thanks to Ivan Fratric (Google) for discovering this issue.
144 [Emilia Käsper, and Steve Henson]
146 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
147 for corner cases. (Certain input points at infinity could lead to
148 bogus results, with non-infinity inputs mapped to infinity too.)
151 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
153 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
154 handshake can force the use of weak keying material in OpenSSL
155 SSL/TLS clients and servers.
157 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
158 researching this issue. (CVE-2014-0224)
159 [KIKUCHI Masashi, Steve Henson]
161 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
162 OpenSSL DTLS client the code can be made to recurse eventually crashing
165 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
167 [Imre Rad, Steve Henson]
169 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
170 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
171 client or server. This is potentially exploitable to run arbitrary
172 code on a vulnerable client or server.
174 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
175 [Jüri Aedla, Steve Henson]
177 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
178 are subject to a denial of service attack.
180 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
181 this issue. (CVE-2014-3470)
182 [Felix Gröbert, Ivan Fratric, Steve Henson]
184 *) Harmonize version and its documentation. -f flag is used to display
186 [mancha <mancha1@zoho.com>]
188 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
190 [mancha <mancha1@zoho.com>]
192 *) Fix some double frees. These are not thought to be exploitable.
193 [mancha <mancha1@zoho.com>]
195 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
197 *) A missing bounds check in the handling of the TLS heartbeat extension
198 can be used to reveal up to 64k of memory to a connected client or
201 Thanks for Neel Mehta of Google Security for discovering this bug and to
202 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
203 preparing the fix (CVE-2014-0160)
204 [Adam Langley, Bodo Moeller]
206 *) Fix for the attack described in the paper "Recovering OpenSSL
207 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
208 by Yuval Yarom and Naomi Benger. Details can be obtained from:
209 http://eprint.iacr.org/2014/140
211 Thanks to Yuval Yarom and Naomi Benger for discovering this
212 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
213 [Yuval Yarom and Naomi Benger]
215 *) TLS pad extension: draft-agl-tls-padding-03
217 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
218 TLS client Hello record length value would otherwise be > 255 and
219 less that 512 pad with a dummy extension containing zeroes so it
220 is at least 512 bytes long.
222 [Adam Langley, Steve Henson]
224 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
226 *) Fix for TLS record tampering bug. A carefully crafted invalid
227 handshake could crash OpenSSL with a NULL pointer exception.
228 Thanks to Anton Johansson for reporting this issues.
231 *) Keep original DTLS digest and encryption contexts in retransmission
232 structures so we can use the previous session parameters if they need
233 to be resent. (CVE-2013-6450)
236 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
237 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
238 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
239 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
240 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
241 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
242 [Rob Stradling, Adam Langley]
244 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
246 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
247 supporting platforms or when small records were transferred.
248 [Andy Polyakov, Steve Henson]
250 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
252 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
254 This addresses the flaw in CBC record processing discovered by
255 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
256 at: http://www.isg.rhul.ac.uk/tls/
258 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
259 Security Group at Royal Holloway, University of London
260 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
261 Emilia Käsper for the initial patch.
263 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
265 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
266 ciphersuites which can be exploited in a denial of service attack.
267 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
268 and detecting this bug and to Wolfgang Ettlinger
269 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
273 *) Return an error when checking OCSP signatures when key is NULL.
274 This fixes a DoS attack. (CVE-2013-0166)
277 *) Make openssl verify return errors.
278 [Chris Palmer <palmer@google.com> and Ben Laurie]
280 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
281 the right response is stapled. Also change SSL_get_certificate()
282 so it returns the certificate actually sent.
283 See http://rt.openssl.org/Ticket/Display.html?id=2836.
284 [Rob Stradling <rob.stradling@comodo.com>]
286 *) Fix possible deadlock when decoding public keys.
289 *) Don't use TLS 1.0 record version number in initial client hello
293 Changes between 1.0.1b and 1.0.1c [10 May 2012]
295 *) Sanity check record length before skipping explicit IV in TLS
296 1.2, 1.1 and DTLS to fix DoS attack.
298 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
299 fuzzing as a service testing platform.
303 *) Initialise tkeylen properly when encrypting CMS messages.
304 Thanks to Solar Designer of Openwall for reporting this issue.
307 *) In FIPS mode don't try to use composite ciphers as they are not
311 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
313 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
314 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
315 mean any application compiled against OpenSSL 1.0.0 headers setting
316 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
317 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
318 0x10000000L Any application which was previously compiled against
319 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
320 will need to be recompiled as a result. Letting be results in
321 inability to disable specifically TLS 1.1 and in client context,
322 in unlike event, limit maximum offered version to TLS 1.0 [see below].
325 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
326 disable just protocol X, but all protocols above X *if* there are
327 protocols *below* X still enabled. In more practical terms it means
328 that if application wants to disable TLS1.0 in favor of TLS1.1 and
329 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
330 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
334 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
336 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
337 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
338 in CRYPTO_realloc_clean.
340 Thanks to Tavis Ormandy, Google Security Team, for discovering this
341 issue and to Adam Langley <agl@chromium.org> for fixing it.
343 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
345 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
348 *) Workarounds for some broken servers that "hang" if a client hello
349 record length exceeds 255 bytes.
351 1. Do not use record version number > TLS 1.0 in initial client
352 hello: some (but not all) hanging servers will now work.
353 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
354 the number of ciphers sent in the client hello. This should be
355 set to an even number, such as 50, for example by passing:
356 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
357 Most broken servers should now work.
358 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
359 TLS 1.2 client support entirely.
362 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
365 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
367 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
368 STRING form instead of a DigestInfo.
371 *) The format used for MDC2 RSA signatures is inconsistent between EVP
372 and the RSA_sign/RSA_verify functions. This was made more apparent when
373 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
374 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
375 the correct format in RSA_verify so both forms transparently work.
378 *) Some servers which support TLS 1.0 can choke if we initially indicate
379 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
380 encrypted premaster secret. As a workaround use the maximum pemitted
381 client version in client hello, this should keep such servers happy
382 and still work with previous versions of OpenSSL.
385 *) Add support for TLS/DTLS heartbeats.
386 [Robin Seggelmann <seggelmann@fh-muenster.de>]
388 *) Add support for SCTP.
389 [Robin Seggelmann <seggelmann@fh-muenster.de>]
391 *) Improved PRNG seeding for VOS.
392 [Paul Green <Paul.Green@stratus.com>]
394 *) Extensive assembler packs updates, most notably:
396 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
397 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
398 - x86_64: bit-sliced AES implementation;
399 - ARM: NEON support, contemporary platforms optimizations;
400 - s390x: z196 support;
401 - *: GHASH and GF(2^m) multiplication implementations;
405 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
406 (removal of unnecessary code)
407 [Peter Sylvester <peter.sylvester@edelweb.fr>]
409 *) Add TLS key material exporter from RFC 5705.
412 *) Add DTLS-SRTP negotiation from RFC 5764.
415 *) Add Next Protocol Negotiation,
416 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
417 disabled with a no-npn flag to config or Configure. Code donated
419 [Adam Langley <agl@google.com> and Ben Laurie]
421 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
422 NIST-P256, NIST-P521, with constant-time single point multiplication on
423 typical inputs. Compiler support for the nonstandard type __uint128_t is
424 required to use this (present in gcc 4.4 and later, for 64-bit builds).
425 Code made available under Apache License version 2.0.
427 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
428 line to include this in your build of OpenSSL, and run "make depend" (or
429 "make update"). This enables the following EC_METHODs:
431 EC_GFp_nistp224_method()
432 EC_GFp_nistp256_method()
433 EC_GFp_nistp521_method()
435 EC_GROUP_new_by_curve_name() will automatically use these (while
436 EC_GROUP_new_curve_GFp() currently prefers the more flexible
438 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
440 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
441 all platforms. Move ssize_t definition from e_os.h to the public
442 header file e_os2.h as it now appears in public header file cms.h
445 *) New -sigopt option to the ca, req and x509 utilities. Additional
446 signature parameters can be passed using this option and in
450 *) Add RSA PSS signing function. This will generate and set the
451 appropriate AlgorithmIdentifiers for PSS based on those in the
452 corresponding EVP_MD_CTX structure. No application support yet.
455 *) Support for companion algorithm specific ASN1 signing routines.
456 New function ASN1_item_sign_ctx() signs a pre-initialised
457 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
458 the appropriate parameters.
461 *) Add new algorithm specific ASN1 verification initialisation function
462 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
463 handling will be the same no matter what EVP_PKEY_METHOD is used.
464 Add a PSS handler to support verification of PSS signatures: checked
465 against a number of sample certificates.
468 *) Add signature printing for PSS. Add PSS OIDs.
469 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
471 *) Add algorithm specific signature printing. An individual ASN1 method
472 can now print out signatures instead of the standard hex dump.
474 More complex signatures (e.g. PSS) can print out more meaningful
475 information. Include DSA version that prints out the signature
479 *) Password based recipient info support for CMS library: implementing
483 *) Split password based encryption into PBES2 and PBKDF2 functions. This
484 neatly separates the code into cipher and PBE sections and is required
485 for some algorithms that split PBES2 into separate pieces (such as
489 *) Session-handling fixes:
490 - Fix handling of connections that are resuming with a session ID,
491 but also support Session Tickets.
492 - Fix a bug that suppressed issuing of a new ticket if the client
493 presented a ticket with an expired session.
494 - Try to set the ticket lifetime hint to something reasonable.
495 - Make tickets shorter by excluding irrelevant information.
496 - On the client side, don't ignore renewed tickets.
497 [Adam Langley, Bodo Moeller (Google)]
499 *) Fix PSK session representation.
502 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
504 This work was sponsored by Intel.
507 *) Add GCM support to TLS library. Some custom code is needed to split
508 the IV between the fixed (from PRF) and explicit (from TLS record)
509 portions. This adds all GCM ciphersuites supported by RFC5288 and
510 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
511 add a special AESGCM string for GCM only.
514 *) Expand range of ctrls for AES GCM. Permit setting invocation
515 field on decrypt and retrieval of invocation field only on encrypt.
518 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
519 As required by RFC5289 these ciphersuites cannot be used if for
520 versions of TLS earlier than 1.2.
523 *) For FIPS capable OpenSSL interpret a NULL default public key method
524 as unset and return the appopriate default but do *not* set the default.
525 This means we can return the appopriate method in applications that
526 swicth between FIPS and non-FIPS modes.
529 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
530 ENGINE is used then we cannot handle that in the FIPS module so we
531 keep original code iff non-FIPS operations are allowed.
534 *) Add -attime option to openssl utilities.
535 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
537 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
540 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
541 FIPS EC methods unconditionally for now.
544 *) New build option no-ec2m to disable characteristic 2 code.
547 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
548 all cases can be covered as some introduce binary incompatibilities.
551 *) Redirect RSA operations to FIPS module including keygen,
552 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
555 *) Add similar low level API blocking to ciphers.
558 *) Low level digest APIs are not approved in FIPS mode: any attempt
559 to use these will cause a fatal error. Applications that *really* want
560 to use them can use the private_* version instead.
563 *) Redirect cipher operations to FIPS module for FIPS builds.
566 *) Redirect digest operations to FIPS module for FIPS builds.
569 *) Update build system to add "fips" flag which will link in fipscanister.o
570 for static and shared library builds embedding a signature if needed.
573 *) Output TLS supported curves in preference order instead of numerical
574 order. This is currently hardcoded for the highest order curves first.
575 This should be configurable so applications can judge speed vs strength.
578 *) Add TLS v1.2 server support for client authentication.
581 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
585 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
586 FIPS modules versions.
589 *) Add TLS v1.2 client side support for client authentication. Keep cache
590 of handshake records longer as we don't know the hash algorithm to use
591 until after the certificate request message is received.
594 *) Initial TLS v1.2 client support. Add a default signature algorithms
595 extension including all the algorithms we support. Parse new signature
596 format in client key exchange. Relax some ECC signing restrictions for
597 TLS v1.2 as indicated in RFC5246.
600 *) Add server support for TLS v1.2 signature algorithms extension. Switch
601 to new signature format when needed using client digest preference.
602 All server ciphersuites should now work correctly in TLS v1.2. No client
603 support yet and no support for client certificates.
606 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
607 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
608 ciphersuites. At present only RSA key exchange ciphersuites work with
609 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
610 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
611 and version checking.
614 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
615 with this defined it will not be affected by any changes to ssl internal
616 structures. Add several utility functions to allow openssl application
617 to work with OPENSSL_NO_SSL_INTERN defined.
621 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
623 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
626 *) Permit abbreviated handshakes when renegotiating using the function
627 SSL_renegotiate_abbreviated().
628 [Robin Seggelmann <seggelmann@fh-muenster.de>]
630 *) Add call to ENGINE_register_all_complete() to
631 ENGINE_load_builtin_engines(), so some implementations get used
632 automatically instead of needing explicit application support.
635 *) Add support for TLS key exporter as described in RFC5705.
636 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
638 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
639 a few changes are required:
641 Add SSL_OP_NO_TLSv1_1 flag.
643 Update version checking logic to handle version 1.1.
644 Add explicit IV handling (ported from DTLS code).
645 Add command line options to s_client/s_server.
648 Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
650 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
652 This addresses the flaw in CBC record processing discovered by
653 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
654 at: http://www.isg.rhul.ac.uk/tls/
656 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
657 Security Group at Royal Holloway, University of London
658 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
659 Emilia Käsper for the initial patch.
661 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
663 *) Return an error when checking OCSP signatures when key is NULL.
664 This fixes a DoS attack. (CVE-2013-0166)
667 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
668 the right response is stapled. Also change SSL_get_certificate()
669 so it returns the certificate actually sent.
670 See http://rt.openssl.org/Ticket/Display.html?id=2836.
672 [Rob Stradling <rob.stradling@comodo.com>]
674 *) Fix possible deadlock when decoding public keys.
677 Changes between 1.0.0i and 1.0.0j [10 May 2012]
679 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
682 *) Sanity check record length before skipping explicit IV in DTLS
685 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
686 fuzzing as a service testing platform.
690 *) Initialise tkeylen properly when encrypting CMS messages.
691 Thanks to Solar Designer of Openwall for reporting this issue.
694 Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
696 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
697 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
698 in CRYPTO_realloc_clean.
700 Thanks to Tavis Ormandy, Google Security Team, for discovering this
701 issue and to Adam Langley <agl@chromium.org> for fixing it.
703 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
705 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
707 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
708 in CMS and PKCS7 code. When RSA decryption fails use a random key for
709 content decryption and always return the same error. Note: this attack
710 needs on average 2^20 messages so it only affects automated senders. The
711 old behaviour can be reenabled in the CMS code by setting the
712 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
713 an MMA defence is not necessary.
714 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
715 this issue. (CVE-2012-0884)
718 *) Fix CVE-2011-4619: make sure we really are receiving a
719 client hello before rejecting multiple SGC restarts. Thanks to
720 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
723 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
725 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
726 Thanks to Antonio Martin, Enterprise Secure Access Research and
727 Development, Cisco Systems, Inc. for discovering this bug and
728 preparing a fix. (CVE-2012-0050)
731 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
733 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
734 of the Vaudenay padding oracle attack on CBC mode encryption
735 which enables an efficient plaintext recovery attack against
736 the OpenSSL implementation of DTLS. Their attack exploits timing
737 differences arising during decryption processing. A research
738 paper describing this attack can be found at:
739 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
740 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
741 Security Group at Royal Holloway, University of London
742 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
743 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
744 for preparing the fix. (CVE-2011-4108)
745 [Robin Seggelmann, Michael Tuexen]
747 *) Clear bytes used for block padding of SSL 3.0 records.
749 [Adam Langley (Google)]
751 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
752 Kadianakis <desnacked@gmail.com> for discovering this issue and
753 Adam Langley for preparing the fix. (CVE-2011-4619)
754 [Adam Langley (Google)]
756 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
757 [Andrey Kulikov <amdeich@gmail.com>]
759 *) Prevent malformed RFC3779 data triggering an assertion failure.
760 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
761 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
762 [Rob Austein <sra@hactrn.net>]
764 *) Improved PRNG seeding for VOS.
765 [Paul Green <Paul.Green@stratus.com>]
767 *) Fix ssl_ciph.c set-up race.
768 [Adam Langley (Google)]
770 *) Fix spurious failures in ecdsatest.c.
771 [Emilia Käsper (Google)]
773 *) Fix the BIO_f_buffer() implementation (which was mixing different
774 interpretations of the '..._len' fields).
775 [Adam Langley (Google)]
777 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
778 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
779 threads won't reuse the same blinding coefficients.
781 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
782 lock to call BN_BLINDING_invert_ex, and avoids one use of
783 BN_BLINDING_update for each BN_BLINDING structure (previously,
784 the last update always remained unused).
785 [Emilia Käsper (Google)]
787 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
788 [Bob Buckholz (Google)]
790 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
792 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
793 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
794 [Kaspar Brand <ossl@velox.ch>]
796 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
797 for multi-threaded use of ECDH. (CVE-2011-3210)
798 [Adam Langley (Google)]
800 *) Fix x509_name_ex_d2i memory leak on bad inputs.
803 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
804 signature public key algorithm by using OID xref utilities instead.
805 Before this you could only use some ECC ciphersuites with SHA1 only.
808 *) Add protection against ECDSA timing attacks as mentioned in the paper
809 by Billy Bob Brumley and Nicola Tuveri, see:
811 http://eprint.iacr.org/2011/232.pdf
813 [Billy Bob Brumley and Nicola Tuveri]
815 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
817 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
818 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
820 *) Fix bug in string printing code: if *any* escaping is enabled we must
821 escape the escape character (backslash) or the resulting string is
825 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
827 *) Disable code workaround for ancient and obsolete Netscape browsers
828 and servers: an attacker can use it in a ciphersuite downgrade attack.
829 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
832 *) Fixed J-PAKE implementation error, originally discovered by
833 Sebastien Martini, further info and confirmation from Stefan
834 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
837 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
839 *) Fix extension code to avoid race conditions which can result in a buffer
840 overrun vulnerability: resumed sessions must not be modified as they can
841 be shared by multiple threads. CVE-2010-3864
844 *) Fix WIN32 build system to correctly link an ENGINE directory into
848 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
850 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
852 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
854 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
856 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
857 context. The operation can be customised via the ctrl mechanism in
858 case ENGINEs want to include additional functionality.
861 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
864 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
865 output hashes compatible with older versions of OpenSSL.
866 [Willy Weisz <weisz@vcpc.univie.ac.at>]
868 *) Fix compression algorithm handling: if resuming a session use the
869 compression algorithm of the resumed session instead of determining
870 it from client hello again. Don't allow server to change algorithm.
873 *) Add load_crls() function to apps tidying load_certs() too. Add option
874 to verify utility to allow additional CRLs to be included.
877 *) Update OCSP request code to permit adding custom headers to the request:
878 some responders need this.
881 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
883 [Julia Lawall <julia@diku.dk>]
885 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
886 needlessly dereferenced structures, used obsolete functions and
887 didn't handle all updated verify codes correctly.
890 *) Disable MD2 in the default configuration.
893 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
894 indicate the initial BIO being pushed or popped. This makes it possible
895 to determine whether the BIO is the one explicitly called or as a result
896 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
897 it handles reference counts correctly and doesn't zero out the I/O bio
898 when it is not being explicitly popped. WARNING: applications which
899 included workarounds for the old buggy behaviour will need to be modified
900 or they could free up already freed BIOs.
903 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
904 renaming to all platforms (within the 0.9.8 branch, this was
905 done conditionally on Netware platforms to avoid a name clash).
906 [Guenter <lists@gknw.net>]
908 *) Add ECDHE and PSK support to DTLS.
909 [Michael Tuexen <tuexen@fh-muenster.de>]
911 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
915 *) Add "missing" function EVP_MD_flags() (without this the only way to
916 retrieve a digest flags is by accessing the structure directly. Update
917 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
918 or cipher is registered as in the "from" argument. Print out all
919 registered digests in the dgst usage message instead of manually
920 attempting to work them out.
923 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
924 this allows the use of compression and extensions. Change default cipher
925 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
926 by default unless an application cipher string requests it.
929 *) Alter match criteria in PKCS12_parse(). It used to try to use local
930 key ids to find matching certificates and keys but some PKCS#12 files
931 don't follow the (somewhat unwritten) rules and this strategy fails.
932 Now just gather all certificates together and the first private key
933 then look for the first certificate that matches the key.
936 *) Support use of registered digest and cipher names for dgst and cipher
937 commands instead of having to add each one as a special case. So now
944 openssl dgst -sha256 foo
946 and this works for ENGINE based algorithms too.
950 *) Update Gost ENGINE to support parameter files.
951 [Victor B. Wagner <vitus@cryptocom.ru>]
953 *) Support GeneralizedTime in ca utility.
954 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
956 *) Enhance the hash format used for certificate directory links. The new
957 form uses the canonical encoding (meaning equivalent names will work
958 even if they aren't identical) and uses SHA1 instead of MD5. This form
959 is incompatible with the older format and as a result c_rehash should
960 be used to rebuild symbolic links.
963 *) Make PKCS#8 the default write format for private keys, replacing the
964 traditional format. This form is standardised, more secure and doesn't
965 include an implicit MD5 dependency.
968 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
969 committed to OpenSSL should pass this lot as a minimum.
972 *) Add session ticket override functionality for use by EAP-FAST.
973 [Jouni Malinen <j@w1.fi>]
975 *) Modify HMAC functions to return a value. Since these can be implemented
976 in an ENGINE errors can occur.
979 *) Type-checked OBJ_bsearch_ex.
982 *) Type-checked OBJ_bsearch. Also some constification necessitated
983 by type-checking. Still to come: TXT_DB, bsearch(?),
984 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
988 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
989 seconds to a tm structure directly, instead of going through OS
990 specific date routines. This avoids any issues with OS routines such
991 as the year 2038 bug. New *_adj() functions for ASN1 time structures
992 and X509_time_adj_ex() to cover the extended range. The existing
993 X509_time_adj() is still usable and will no longer have any date issues.
996 *) Delta CRL support. New use deltas option which will attempt to locate
997 and search any appropriate delta CRLs available.
999 This work was sponsored by Google.
1002 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
1003 code and add additional score elements. Validate alternate CRL paths
1004 as part of the CRL checking and indicate a new error "CRL path validation
1005 error" in this case. Applications wanting additional details can use
1006 the verify callback and check the new "parent" field. If this is not
1007 NULL CRL path validation is taking place. Existing applications wont
1008 see this because it requires extended CRL support which is off by
1011 This work was sponsored by Google.
1014 *) Support for freshest CRL extension.
1016 This work was sponsored by Google.
1019 *) Initial indirect CRL support. Currently only supported in the CRLs
1020 passed directly and not via lookup. Process certificate issuer
1021 CRL entry extension and lookup CRL entries by bother issuer name
1022 and serial number. Check and process CRL issuer entry in IDP extension.
1024 This work was sponsored by Google.
1027 *) Add support for distinct certificate and CRL paths. The CRL issuer
1028 certificate is validated separately in this case. Only enabled if
1029 an extended CRL support flag is set: this flag will enable additional
1030 CRL functionality in future.
1032 This work was sponsored by Google.
1035 *) Add support for policy mappings extension.
1037 This work was sponsored by Google.
1040 *) Fixes to pathlength constraint, self issued certificate handling,
1041 policy processing to align with RFC3280 and PKITS tests.
1043 This work was sponsored by Google.
1046 *) Support for name constraints certificate extension. DN, email, DNS
1047 and URI types are currently supported.
1049 This work was sponsored by Google.
1052 *) To cater for systems that provide a pointer-based thread ID rather
1053 than numeric, deprecate the current numeric thread ID mechanism and
1054 replace it with a structure and associated callback type. This
1055 mechanism allows a numeric "hash" to be extracted from a thread ID in
1056 either case, and on platforms where pointers are larger than 'long',
1057 mixing is done to help ensure the numeric 'hash' is usable even if it
1058 can't be guaranteed unique. The default mechanism is to use "&errno"
1059 as a pointer-based thread ID to distinguish between threads.
1061 Applications that want to provide their own thread IDs should now use
1062 CRYPTO_THREADID_set_callback() to register a callback that will call
1063 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
1065 Note that ERR_remove_state() is now deprecated, because it is tied
1066 to the assumption that thread IDs are numeric. ERR_remove_state(0)
1067 to free the current thread's error state should be replaced by
1068 ERR_remove_thread_state(NULL).
1070 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
1071 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
1072 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
1073 application was previously providing a numeric thread callback that
1074 was inappropriate for distinguishing threads, then uniqueness might
1075 have been obtained with &errno that happened immediately in the
1076 intermediate development versions of OpenSSL; this is no longer the
1077 case, the numeric thread callback will now override the automatic use
1079 [Geoff Thorpe, with help from Bodo Moeller]
1081 *) Initial support for different CRL issuing certificates. This covers a
1082 simple case where the self issued certificates in the chain exist and
1083 the real CRL issuer is higher in the existing chain.
1085 This work was sponsored by Google.
1088 *) Removed effectively defunct crypto/store from the build.
1091 *) Revamp of STACK to provide stronger type-checking. Still to come:
1092 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
1093 ASN1_STRING, CONF_VALUE.
1096 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
1097 RAM on SSL connections. This option can save about 34k per idle SSL.
1100 *) Revamp of LHASH to provide stronger type-checking. Still to come:
1101 STACK, TXT_DB, bsearch, qsort.
1104 *) Initial support for Cryptographic Message Syntax (aka CMS) based
1105 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
1106 support for data, signedData, compressedData, digestedData and
1107 encryptedData, envelopedData types included. Scripts to check against
1108 RFC4134 examples draft and interop and consistency checks of many
1109 content types and variants.
1112 *) Add options to enc utility to support use of zlib compression BIO.
1115 *) Extend mk1mf to support importing of options and assembly language
1116 files from Configure script, currently only included in VC-WIN32.
1117 The assembly language rules can now optionally generate the source
1118 files from the associated perl scripts.
1121 *) Implement remaining functionality needed to support GOST ciphersuites.
1122 Interop testing has been performed using CryptoPro implementations.
1123 [Victor B. Wagner <vitus@cryptocom.ru>]
1125 *) s390x assembler pack.
1128 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
1132 *) Implement Opaque PRF Input TLS extension as specified in
1133 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
1134 official specification yet and no extension type assignment by
1135 IANA exists, this extension (for now) will have to be explicitly
1136 enabled when building OpenSSL by providing the extension number
1137 to use. For example, specify an option
1139 -DTLSEXT_TYPE_opaque_prf_input=0x9527
1141 to the "config" or "Configure" script to enable the extension,
1142 assuming extension number 0x9527 (which is a completely arbitrary
1143 and unofficial assignment based on the MD5 hash of the Internet
1144 Draft). Note that by doing so, you potentially lose
1145 interoperability with other TLS implementations since these might
1146 be using the same extension number for other purposes.
1148 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
1149 opaque PRF input value to use in the handshake. This will create
1150 an interal copy of the length-'len' string at 'src', and will
1151 return non-zero for success.
1153 To get more control and flexibility, provide a callback function
1156 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
1157 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
1161 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
1164 Callback function 'cb' will be called in handshakes, and is
1165 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
1166 Argument 'arg' is for application purposes (the value as given to
1167 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
1168 be provided to the callback function). The callback function
1169 has to return non-zero to report success: usually 1 to use opaque
1170 PRF input just if possible, or 2 to enforce use of the opaque PRF
1171 input. In the latter case, the library will abort the handshake
1172 if opaque PRF input is not successfully negotiated.
1174 Arguments 'peerinput' and 'len' given to the callback function
1175 will always be NULL and 0 in the case of a client. A server will
1176 see the client's opaque PRF input through these variables if
1177 available (NULL and 0 otherwise). Note that if the server
1178 provides an opaque PRF input, the length must be the same as the
1179 length of the client's opaque PRF input.
1181 Note that the callback function will only be called when creating
1182 a new session (session resumption can resume whatever was
1183 previously negotiated), and will not be called in SSL 2.0
1184 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
1185 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
1186 for applications that need to enforce opaque PRF input.
1190 *) Update ssl code to support digests other than SHA1+MD5 for handshake
1193 [Victor B. Wagner <vitus@cryptocom.ru>]
1195 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1196 RFC4507bis. The encrypted ticket format is an encrypted encoded
1197 SSL_SESSION structure, that way new session features are automatically
1200 If a client application caches session in an SSL_SESSION structure
1201 support is transparent because tickets are now stored in the encoded
1204 The SSL_CTX structure automatically generates keys for ticket
1205 protection in servers so again support should be possible
1206 with no application modification.
1208 If a client or server wishes to disable RFC4507 support then the option
1209 SSL_OP_NO_TICKET can be set.
1211 Add a TLS extension debugging callback to allow the contents of any client
1212 or server extensions to be examined.
1214 This work was sponsored by Google.
1217 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
1218 OpenSSL should now compile cleanly on gcc 4.2
1219 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
1221 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
1222 support including streaming MAC support: this is required for GOST
1223 ciphersuite support.
1224 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
1226 *) Add option -stream to use PKCS#7 streaming in smime utility. New
1227 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
1228 to output in BER and PEM format.
1231 *) Experimental support for use of HMAC via EVP_PKEY interface. This
1232 allows HMAC to be handled via the EVP_DigestSign*() interface. The
1233 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
1234 ENGINE support for HMAC keys which are unextractable. New -mac and
1235 -macopt options to dgst utility.
1238 *) New option -sigopt to dgst utility. Update dgst to use
1239 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
1240 alternative signing paramaters such as X9.31 or PSS in the dgst
1244 *) Change ssl_cipher_apply_rule(), the internal function that does
1245 the work each time a ciphersuite string requests enabling
1246 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
1247 removing ("!foo+bar") a class of ciphersuites: Now it maintains
1248 the order of disabled ciphersuites such that those ciphersuites
1249 that most recently went from enabled to disabled not only stay
1250 in order with respect to each other, but also have higher priority
1251 than other disabled ciphersuites the next time ciphersuites are
1254 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
1255 the same ciphersuites as with "HIGH" alone, but in a specific
1256 order where the PSK ciphersuites come first (since they are the
1257 most recently disabled ciphersuites when "HIGH" is parsed).
1259 Also, change ssl_create_cipher_list() (using this new
1260 funcionality) such that between otherwise identical
1261 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
1265 *) Change ssl_create_cipher_list() so that it automatically
1266 arranges the ciphersuites in reasonable order before starting
1267 to process the rule string. Thus, the definition for "DEFAULT"
1268 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
1269 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
1270 This makes it much easier to arrive at a reasonable default order
1271 in applications for which anonymous ciphers are OK (meaning
1272 that you can't actually use DEFAULT).
1273 [Bodo Moeller; suggested by Victor Duchovni]
1275 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
1276 processing) into multiple integers instead of setting
1277 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
1278 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
1279 (These masks as well as the individual bit definitions are hidden
1280 away into the non-exported interface ssl/ssl_locl.h, so this
1281 change to the definition of the SSL_CIPHER structure shouldn't
1282 affect applications.) This give us more bits for each of these
1283 categories, so there is no longer a need to coagulate AES128 and
1284 AES256 into a single algorithm bit, and to coagulate Camellia128
1285 and Camellia256 into a single algorithm bit, which has led to all
1288 Thus, among other things, the kludge introduced in 0.9.7m and
1289 0.9.8e for masking out AES256 independently of AES128 or masking
1290 out Camellia256 independently of AES256 is not needed here in 0.9.9.
1292 With the change, we also introduce new ciphersuite aliases that
1293 so far were missing: "AES128", "AES256", "CAMELLIA128", and
1297 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
1298 Use the leftmost N bytes of the signature input if the input is
1299 larger than the prime q (with N being the size in bytes of q).
1302 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
1303 it yet and it is largely untested.
1306 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
1309 *) Initial incomplete changes to avoid need for function casts in OpenSSL
1310 some compilers (gcc 4.2 and later) reject their use. Safestack is
1311 reimplemented. Update ASN1 to avoid use of legacy functions.
1314 *) Win32/64 targets are linked with Winsock2.
1317 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
1318 to external functions. This can be used to increase CRL handling
1319 efficiency especially when CRLs are very large by (for example) storing
1320 the CRL revoked certificates in a database.
1323 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
1324 new CRLs added to a directory can be used. New command line option
1325 -verify_return_error to s_client and s_server. This causes real errors
1326 to be returned by the verify callback instead of carrying on no matter
1327 what. This reflects the way a "real world" verify callback would behave.
1330 *) GOST engine, supporting several GOST algorithms and public key formats.
1331 Kindly donated by Cryptocom.
1334 *) Partial support for Issuing Distribution Point CRL extension. CRLs
1335 partitioned by DP are handled but no indirect CRL or reason partitioning
1336 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
1337 selected via a scoring technique which handles IDP and AKID in CRLs.
1340 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
1341 will ultimately be used for all verify operations: this will remove the
1342 X509_STORE dependency on certificate verification and allow alternative
1343 lookup methods. X509_STORE based implementations of these two callbacks.
1346 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
1347 Modify get_crl() to find a valid (unexpired) CRL if possible.
1350 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
1351 this would be called X509_CRL_cmp() but that name is already used by
1352 a function that just compares CRL issuer names. Cache several CRL
1353 extensions in X509_CRL structure and cache CRLDP in X509.
1356 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
1357 this maps equivalent X509_NAME structures into a consistent structure.
1358 Name comparison can then be performed rapidly using memcmp().
1361 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
1365 *) Allow digests to supply their own micalg string for S/MIME type using
1366 the ctrl EVP_MD_CTRL_MICALG.
1369 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
1370 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
1371 ctrl. It can then customise the structure before and/or after signing
1375 *) New function OBJ_add_sigid() to allow application defined signature OIDs
1376 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
1377 to free up any added signature OIDs.
1380 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
1381 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
1382 digest and cipher tables. New options added to openssl utility:
1383 list-message-digest-algorithms and list-cipher-algorithms.
1386 *) Change the array representation of binary polynomials: the list
1387 of degrees of non-zero coefficients is now terminated with -1.
1388 Previously it was terminated with 0, which was also part of the
1389 value; thus, the array representation was not applicable to
1390 polynomials where t^0 has coefficient zero. This change makes
1391 the array representation useful in a more general context.
1394 *) Various modifications and fixes to SSL/TLS cipher string
1395 handling. For ECC, the code now distinguishes between fixed ECDH
1396 with RSA certificates on the one hand and with ECDSA certificates
1397 on the other hand, since these are separate ciphersuites. The
1398 unused code for Fortezza ciphersuites has been removed.
1400 For consistency with EDH, ephemeral ECDH is now called "EECDH"
1401 (not "ECDHE"). For consistency with the code for DH
1402 certificates, use of ECDH certificates is now considered ECDH
1403 authentication, not RSA or ECDSA authentication (the latter is
1404 merely the CA's signing algorithm and not actively used in the
1407 The temporary ciphersuite alias "ECCdraft" is no longer
1408 available, and ECC ciphersuites are no longer excluded from "ALL"
1409 and "DEFAULT". The following aliases now exist for RFC 4492
1410 ciphersuites, most of these by analogy with the DH case:
1412 kECDHr - ECDH cert, signed with RSA
1413 kECDHe - ECDH cert, signed with ECDSA
1414 kECDH - ECDH cert (signed with either RSA or ECDSA)
1415 kEECDH - ephemeral ECDH
1416 ECDH - ECDH cert or ephemeral ECDH
1422 AECDH - anonymous ECDH
1423 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
1427 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
1428 Use correct micalg parameters depending on digest(s) in signed message.
1431 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
1432 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
1435 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
1436 an engine to register a method. Add ENGINE lookups for methods and
1437 functional reference processing.
1440 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
1441 EVP_{Sign,Verify}* which allow an application to customise the signature
1445 *) New -resign option to smime utility. This adds one or more signers
1446 to an existing PKCS#7 signedData structure. Also -md option to use an
1447 alternative message digest algorithm for signing.
1450 *) Tidy up PKCS#7 routines and add new functions to make it easier to
1451 create PKCS7 structures containing multiple signers. Update smime
1452 application to support multiple signers.
1455 *) New -macalg option to pkcs12 utility to allow setting of an alternative
1459 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
1460 Reorganize PBE internals to lookup from a static table using NIDs,
1461 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
1462 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
1463 PRF which will be automatically used with PBES2.
1466 *) Replace the algorithm specific calls to generate keys in "req" with the
1470 *) Update PKCS#7 enveloped data routines to use new API. This is now
1471 supported by any public key method supporting the encrypt operation. A
1472 ctrl is added to allow the public key algorithm to examine or modify
1473 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
1477 *) Add a ctrl to asn1 method to allow a public key algorithm to express
1478 a default digest type to use. In most cases this will be SHA1 but some
1479 algorithms (such as GOST) need to specify an alternative digest. The
1480 return value indicates how strong the prefernce is 1 means optional and
1481 2 is mandatory (that is it is the only supported type). Modify
1482 ASN1_item_sign() to accept a NULL digest argument to indicate it should
1483 use the default md. Update openssl utilities to use the default digest
1484 type for signing if it is not explicitly indicated.
1487 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
1488 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
1489 signing method from the key type. This effectively removes the link
1490 between digests and public key types.
1493 *) Add an OID cross reference table and utility functions. Its purpose is to
1494 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
1495 rsaEncryption. This will allow some of the algorithm specific hackery
1496 needed to use the correct OID to be removed.
1499 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
1500 structures for PKCS7_sign(). They are now set up by the relevant public
1504 *) Add provisional EC pkey method with support for ECDSA and ECDH.
1507 *) Add support for key derivation (agreement) in the API, DH method and
1511 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
1512 public and private key formats. As a side effect these add additional
1513 command line functionality not previously available: DSA signatures can be
1514 generated and verified using pkeyutl and DH key support and generation in
1519 [Oliver Tappe <zooey@hirschkaefer.de>]
1521 *) New make target "install_html_docs" installs HTML renditions of the
1523 [Oliver Tappe <zooey@hirschkaefer.de>]
1525 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
1526 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
1527 support key and parameter generation and add initial key generation
1528 functionality for RSA.
1531 *) Add functions for main EVP_PKEY_method operations. The undocumented
1532 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
1533 EVP_PKEY_{encrypt,decrypt}_old.
1536 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
1537 key API, doesn't do much yet.
1540 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
1541 public key algorithms. New option to openssl utility:
1542 "list-public-key-algorithms" to print out info.
1545 *) Implement the Supported Elliptic Curves Extension for
1546 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1549 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
1550 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
1553 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
1554 utilities such as rsa, dsa, dsaparam etc except they process any key
1558 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1559 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1560 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
1564 *) Initial support for pluggable public key ASN1.
1565 De-spaghettify the public key ASN1 handling. Move public and private
1566 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
1567 algorithm specific handling to a single module within the relevant
1568 algorithm directory. Add functions to allow (near) opaque processing
1569 of public and private key structures.
1572 *) Implement the Supported Point Formats Extension for
1573 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1576 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
1577 for the psk identity [hint] and the psk callback functions to the
1578 SSL_SESSION, SSL and SSL_CTX structure.
1581 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
1585 SSL_CTX_use_psk_identity_hint
1586 SSL_get_psk_identity_hint
1587 SSL_get_psk_identity
1588 SSL_use_psk_identity_hint
1590 [Mika Kousa and Pasi Eronen of Nokia Corporation]
1592 *) Add RFC 3161 compliant time stamp request creation, response generation
1593 and response verification functionality.
1594 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
1596 *) Add initial support for TLS extensions, specifically for the server_name
1597 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1598 have new members for a host name. The SSL data structure has an
1599 additional member SSL_CTX *initial_ctx so that new sessions can be
1600 stored in that context to allow for session resumption, even after the
1601 SSL has been switched to a new SSL_CTX in reaction to a client's
1602 server_name extension.
1604 New functions (subject to change):
1606 SSL_get_servername()
1607 SSL_get_servername_type()
1610 New CTRL codes and macros (subject to change):
1612 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1613 - SSL_CTX_set_tlsext_servername_callback()
1614 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1615 - SSL_CTX_set_tlsext_servername_arg()
1616 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1618 openssl s_client has a new '-servername ...' option.
1620 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1621 '-key2 ...', '-servername_fatal' (subject to change). This allows
1622 testing the HostName extension for a specific single host name ('-cert'
1623 and '-key' remain fallbacks for handshakes without HostName
1624 negotiation). If the unrecogninzed_name alert has to be sent, this by
1625 default is a warning; it becomes fatal with the '-servername_fatal'
1628 [Peter Sylvester, Remy Allais, Christophe Renou]
1630 *) Whirlpool hash implementation is added.
1633 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1634 bn(64,32). Because of instruction set limitations it doesn't have
1635 any negative impact on performance. This was done mostly in order
1636 to make it possible to share assembler modules, such as bn_mul_mont
1637 implementations, between 32- and 64-bit builds without hassle.
1640 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1641 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1645 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1646 dedicated Montgomery multiplication procedure, is introduced.
1647 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1648 "64-bit" performance on certain 32-bit targets.
1651 *) New option SSL_OP_NO_COMP to disable use of compression selectively
1652 in SSL structures. New SSL ctrl to set maximum send fragment size.
1653 Save memory by seeting the I/O buffer sizes dynamically instead of
1654 using the maximum available value.
1657 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
1658 in addition to the text details.
1661 *) Very, very preliminary EXPERIMENTAL support for printing of general
1662 ASN1 structures. This currently produces rather ugly output and doesn't
1663 handle several customised structures at all.
1666 *) Integrated support for PVK file format and some related formats such
1667 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
1668 these in the 'rsa' and 'dsa' utilities.
1671 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
1674 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
1675 place for the (very old) "NETSCAPE" format certificates which are now
1676 handled using new ASN1 code equivalents.
1679 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
1680 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
1681 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
1684 *) Modify CRL distribution points extension code to print out previously
1685 unsupported fields. Enhance extension setting code to allow setting of
1689 *) Add print and set support for Issuing Distribution Point CRL extension.
1692 *) Change 'Configure' script to enable Camellia by default.
1695 Changes between 0.9.8x and 0.9.8y [5 Feb 2013]
1697 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1699 This addresses the flaw in CBC record processing discovered by
1700 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1701 at: http://www.isg.rhul.ac.uk/tls/
1703 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1704 Security Group at Royal Holloway, University of London
1705 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1706 Emilia Käsper for the initial patch.
1708 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1710 *) Return an error when checking OCSP signatures when key is NULL.
1711 This fixes a DoS attack. (CVE-2013-0166)
1714 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1715 the right response is stapled. Also change SSL_get_certificate()
1716 so it returns the certificate actually sent.
1717 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1718 (This is a backport)
1719 [Rob Stradling <rob.stradling@comodo.com>]
1721 *) Fix possible deadlock when decoding public keys.
1724 Changes between 0.9.8w and 0.9.8x [10 May 2012]
1726 *) Sanity check record length before skipping explicit IV in DTLS
1729 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1730 fuzzing as a service testing platform.
1734 *) Initialise tkeylen properly when encrypting CMS messages.
1735 Thanks to Solar Designer of Openwall for reporting this issue.
1738 Changes between 0.9.8v and 0.9.8w [23 Apr 2012]
1740 *) The fix for CVE-2012-2110 did not take into account that the
1741 'len' argument to BUF_MEM_grow and BUF_MEM_grow_clean is an
1742 int in OpenSSL 0.9.8, making it still vulnerable. Fix by
1743 rejecting negative len parameter. (CVE-2012-2131)
1744 [Tomas Hoger <thoger@redhat.com>]
1746 Changes between 0.9.8u and 0.9.8v [19 Apr 2012]
1748 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1749 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1750 in CRYPTO_realloc_clean.
1752 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1753 issue and to Adam Langley <agl@chromium.org> for fixing it.
1755 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1757 Changes between 0.9.8t and 0.9.8u [12 Mar 2012]
1759 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1760 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1761 content decryption and always return the same error. Note: this attack
1762 needs on average 2^20 messages so it only affects automated senders. The
1763 old behaviour can be reenabled in the CMS code by setting the
1764 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1765 an MMA defence is not necessary.
1766 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1767 this issue. (CVE-2012-0884)
1770 *) Fix CVE-2011-4619: make sure we really are receiving a
1771 client hello before rejecting multiple SGC restarts. Thanks to
1772 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1775 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
1777 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1778 Thanks to Antonio Martin, Enterprise Secure Access Research and
1779 Development, Cisco Systems, Inc. for discovering this bug and
1780 preparing a fix. (CVE-2012-0050)
1783 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
1785 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1786 of the Vaudenay padding oracle attack on CBC mode encryption
1787 which enables an efficient plaintext recovery attack against
1788 the OpenSSL implementation of DTLS. Their attack exploits timing
1789 differences arising during decryption processing. A research
1790 paper describing this attack can be found at:
1791 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1792 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1793 Security Group at Royal Holloway, University of London
1794 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1795 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1796 for preparing the fix. (CVE-2011-4108)
1797 [Robin Seggelmann, Michael Tuexen]
1799 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
1800 [Ben Laurie, Kasper <ekasper@google.com>]
1802 *) Clear bytes used for block padding of SSL 3.0 records.
1804 [Adam Langley (Google)]
1806 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1807 Kadianakis <desnacked@gmail.com> for discovering this issue and
1808 Adam Langley for preparing the fix. (CVE-2011-4619)
1809 [Adam Langley (Google)]
1811 *) Prevent malformed RFC3779 data triggering an assertion failure.
1812 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1813 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1814 [Rob Austein <sra@hactrn.net>]
1816 *) Fix ssl_ciph.c set-up race.
1817 [Adam Langley (Google)]
1819 *) Fix spurious failures in ecdsatest.c.
1820 [Emilia Käsper (Google)]
1822 *) Fix the BIO_f_buffer() implementation (which was mixing different
1823 interpretations of the '..._len' fields).
1824 [Adam Langley (Google)]
1826 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1827 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1828 threads won't reuse the same blinding coefficients.
1830 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1831 lock to call BN_BLINDING_invert_ex, and avoids one use of
1832 BN_BLINDING_update for each BN_BLINDING structure (previously,
1833 the last update always remained unused).
1834 [Emilia Käsper (Google)]
1836 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1837 for multi-threaded use of ECDH.
1838 [Adam Langley (Google)]
1840 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1843 *) Add protection against ECDSA timing attacks as mentioned in the paper
1844 by Billy Bob Brumley and Nicola Tuveri, see:
1846 http://eprint.iacr.org/2011/232.pdf
1848 [Billy Bob Brumley and Nicola Tuveri]
1850 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
1852 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1853 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1855 *) Fix bug in string printing code: if *any* escaping is enabled we must
1856 escape the escape character (backslash) or the resulting string is
1860 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
1862 *) Disable code workaround for ancient and obsolete Netscape browsers
1863 and servers: an attacker can use it in a ciphersuite downgrade attack.
1864 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1867 *) Fixed J-PAKE implementation error, originally discovered by
1868 Sebastien Martini, further info and confirmation from Stefan
1869 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1872 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
1874 *) Fix extension code to avoid race conditions which can result in a buffer
1875 overrun vulnerability: resumed sessions must not be modified as they can
1876 be shared by multiple threads. CVE-2010-3864
1879 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
1882 *) Don't reencode certificate when calculating signature: cache and use
1883 the original encoding instead. This makes signature verification of
1884 some broken encodings work correctly.
1887 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
1888 is also one of the inputs.
1889 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1891 *) Don't repeatedly append PBE algorithms to table if they already exist.
1892 Sort table on each new add. This effectively makes the table read only
1893 after all algorithms are added and subsequent calls to PKCS12_pbe_add
1897 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
1899 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
1902 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
1903 access or freeing data twice (CVE-2010-0742)
1904 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
1906 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
1907 common in certificates and some applications which only call
1908 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
1912 Reduce copying into .apps and .test in makevms.com
1913 Don't try to use blank CA certificate in CA.com
1914 Allow use of C files from original directories in maketests.com
1915 [Steven M. Schweda" <sms@antinode.info>]
1917 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
1919 *) When rejecting SSL/TLS records due to an incorrect version number, never
1920 update s->server with a new major version number. As of
1921 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
1922 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
1923 the previous behavior could result in a read attempt at NULL when
1924 receiving specific incorrect SSL/TLS records once record payload
1925 protection is active. (CVE-2010-0740)
1926 [Bodo Moeller, Adam Langley <agl@chromium.org>]
1928 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
1929 could be crashed if the relevant tables were not present (e.g. chrooted).
1930 [Tomas Hoger <thoger@redhat.com>]
1932 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
1934 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
1935 [Martin Olsson, Neel Mehta]
1937 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
1938 accommodate for stack sorting, always a write lock!).
1941 *) On some versions of WIN32 Heap32Next is very slow. This can cause
1942 excessive delays in the RAND_poll(): over a minute. As a workaround
1943 include a time check in the inner Heap32Next loop too.
1946 *) The code that handled flushing of data in SSL/TLS originally used the
1947 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
1948 the problem outlined in PR#1949. The fix suggested there however can
1949 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
1950 of Apache). So instead simplify the code to flush unconditionally.
1951 This should be fine since flushing with no data to flush is a no op.
1954 *) Handle TLS versions 2.0 and later properly and correctly use the
1955 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
1956 off ancient servers have a habit of sticking around for a while...
1959 *) Modify compression code so it frees up structures without using the
1960 ex_data callbacks. This works around a problem where some applications
1961 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
1962 restarting) then use compression (e.g. SSL with compression) later.
1963 This results in significant per-connection memory leaks and
1964 has caused some security issues including CVE-2008-1678 and
1968 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
1969 change when encrypting or decrypting.
1972 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
1973 connect and renegotiate with servers which do not support RI.
1974 Until RI is more widely deployed this option is enabled by default.
1977 *) Add "missing" ssl ctrls to clear options and mode.
1980 *) If client attempts to renegotiate and doesn't support RI respond with
1981 a no_renegotiation alert as required by RFC5746. Some renegotiating
1982 TLS clients will continue a connection gracefully when they receive
1983 the alert. Unfortunately OpenSSL mishandled this alert and would hang
1984 waiting for a server hello which it will never receive. Now we treat a
1985 received no_renegotiation alert as a fatal error. This is because
1986 applications requesting a renegotiation might well expect it to succeed
1987 and would have no code in place to handle the server denying it so the
1988 only safe thing to do is to terminate the connection.
1991 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
1992 peer supports secure renegotiation and 0 otherwise. Print out peer
1993 renegotiation support in s_client/s_server.
1996 *) Replace the highly broken and deprecated SPKAC certification method with
1997 the updated NID creation version. This should correctly handle UTF8.
2000 *) Implement RFC5746. Re-enable renegotiation but require the extension
2001 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
2002 turns out to be a bad idea. It has been replaced by
2003 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
2004 SSL_CTX_set_options(). This is really not recommended unless you
2005 know what you are doing.
2006 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
2008 *) Fixes to stateless session resumption handling. Use initial_ctx when
2009 issuing and attempting to decrypt tickets in case it has changed during
2010 servername handling. Use a non-zero length session ID when attempting
2011 stateless session resumption: this makes it possible to determine if
2012 a resumption has occurred immediately after receiving server hello
2013 (several places in OpenSSL subtly assume this) instead of later in
2017 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
2018 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
2019 fixes for a few places where the return code is not checked
2021 [Julia Lawall <julia@diku.dk>]
2023 *) Add --strict-warnings option to Configure script to include devteam
2024 warnings in other configurations.
2027 *) Add support for --libdir option and LIBDIR variable in makefiles. This
2028 makes it possible to install openssl libraries in locations which
2029 have names other than "lib", for example "/usr/lib64" which some
2031 [Steve Henson, based on patch from Jeremy Utley]
2033 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
2034 X690 8.9.12 and can produce some misleading textual output of OIDs.
2035 [Steve Henson, reported by Dan Kaminsky]
2037 *) Delete MD2 from algorithm tables. This follows the recommendation in
2038 several standards that it is not used in new applications due to
2039 several cryptographic weaknesses. For binary compatibility reasons
2040 the MD2 API is still compiled in by default.
2043 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
2047 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
2048 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
2050 [Guenter <lists@gknw.net>]
2052 *) Fix the server certificate chain building code to use X509_verify_cert(),
2053 it used to have an ad-hoc builder which was unable to cope with anything
2054 other than a simple chain.
2055 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
2057 *) Don't check self signed certificate signatures in X509_verify_cert()
2058 by default (a flag can override this): it just wastes time without
2059 adding any security. As a useful side effect self signed root CAs
2060 with non-FIPS digests are now usable in FIPS mode.
2063 *) In dtls1_process_out_of_seq_message() the check if the current message
2064 is already buffered was missing. For every new message was memory
2065 allocated, allowing an attacker to perform an denial of service attack
2066 with sending out of seq handshake messages until there is no memory
2067 left. Additionally every future messege was buffered, even if the
2068 sequence number made no sense and would be part of another handshake.
2069 So only messages with sequence numbers less than 10 in advance will be
2070 buffered. (CVE-2009-1378)
2071 [Robin Seggelmann, discovered by Daniel Mentz]
2073 *) Records are buffered if they arrive with a future epoch to be
2074 processed after finishing the corresponding handshake. There is
2075 currently no limitation to this buffer allowing an attacker to perform
2076 a DOS attack with sending records with future epochs until there is no
2077 memory left. This patch adds the pqueue_size() function to detemine
2078 the size of a buffer and limits the record buffer to 100 entries.
2080 [Robin Seggelmann, discovered by Daniel Mentz]
2082 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
2083 parent structure is freed. (CVE-2009-1379)
2086 *) Handle non-blocking I/O properly in SSL_shutdown() call.
2087 [Darryl Miles <darryl-mailinglists@netbauds.net>]
2090 [Ilya O. <vrghost@gmail.com>]
2092 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
2094 *) Disable renegotiation completely - this fixes a severe security
2095 problem (CVE-2009-3555) at the cost of breaking all
2096 renegotiation. Renegotiation can be re-enabled by setting
2097 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
2098 run-time. This is really not recommended unless you know what
2102 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
2104 *) Don't set val to NULL when freeing up structures, it is freed up by
2105 underlying code. If sizeof(void *) > sizeof(long) this can result in
2106 zeroing past the valid field. (CVE-2009-0789)
2107 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
2109 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
2110 checked correctly. This would allow some invalid signed attributes to
2111 appear to verify correctly. (CVE-2009-0591)
2112 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2114 *) Reject UniversalString and BMPString types with invalid lengths. This
2115 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
2116 a legal length. (CVE-2009-0590)
2119 *) Set S/MIME signing as the default purpose rather than setting it
2120 unconditionally. This allows applications to override it at the store
2124 *) Permit restricted recursion of ASN1 strings. This is needed in practice
2125 to handle some structures.
2128 *) Improve efficiency of mem_gets: don't search whole buffer each time
2130 [Jeremy Shapiro <jnshapir@us.ibm.com>]
2132 *) New -hex option for openssl rand.
2135 *) Print out UTF8String and NumericString when parsing ASN1.
2138 *) Support NumericString type for name components.
2141 *) Allow CC in the environment to override the automatically chosen
2142 compiler. Note that nothing is done to ensure flags work with the
2146 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
2148 *) Properly check EVP_VerifyFinal() and similar return values
2150 [Ben Laurie, Bodo Moeller, Google Security Team]
2152 *) Enable TLS extensions by default.
2155 *) Allow the CHIL engine to be loaded, whether the application is
2156 multithreaded or not. (This does not release the developer from the
2157 obligation to set up the dynamic locking callbacks.)
2158 [Sander Temme <sander@temme.net>]
2160 *) Use correct exit code if there is an error in dgst command.
2161 [Steve Henson; problem pointed out by Roland Dirlewanger]
2163 *) Tweak Configure so that you need to say "experimental-jpake" to enable
2164 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
2167 *) Add experimental JPAKE support, including demo authentication in
2168 s_client and s_server.
2171 *) Set the comparison function in v3_addr_canonize().
2172 [Rob Austein <sra@hactrn.net>]
2174 *) Add support for XMPP STARTTLS in s_client.
2175 [Philip Paeps <philip@freebsd.org>]
2177 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
2178 to ensure that even with this option, only ciphersuites in the
2179 server's preference list will be accepted. (Note that the option
2180 applies only when resuming a session, so the earlier behavior was
2181 just about the algorithm choice for symmetric cryptography.)
2184 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
2186 *) Fix NULL pointer dereference if a DTLS server received
2187 ChangeCipherSpec as first record (CVE-2009-1386).
2190 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
2191 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
2194 *) The fix in 0.9.8c that supposedly got rid of unsafe
2195 double-checked locking was incomplete for RSA blinding,
2196 addressing just one layer of what turns out to have been
2197 doubly unsafe triple-checked locking.
2199 So now fix this for real by retiring the MONT_HELPER macro
2200 in crypto/rsa/rsa_eay.c.
2202 [Bodo Moeller; problem pointed out by Marius Schilder]
2204 *) Various precautionary measures:
2206 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
2208 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
2209 (NB: This would require knowledge of the secret session ticket key
2210 to exploit, in which case you'd be SOL either way.)
2212 - Change bn_nist.c so that it will properly handle input BIGNUMs
2213 outside the expected range.
2215 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
2218 [Neel Mehta, Bodo Moeller]
2220 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
2221 the load fails. Useful for distros.
2222 [Ben Laurie and the FreeBSD team]
2224 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
2227 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
2230 *) Expand ENGINE to support engine supplied SSL client certificate functions.
2232 This work was sponsored by Logica.
2235 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
2236 keystores. Support for SSL/TLS client authentication too.
2237 Not compiled unless enable-capieng specified to Configure.
2239 This work was sponsored by Logica.
2242 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
2243 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
2244 attribute creation routines such as certifcate requests and PKCS#12
2248 Changes between 0.9.8g and 0.9.8h [28 May 2008]
2250 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
2251 handshake which could lead to a cilent crash as found using the
2252 Codenomicon TLS test suite (CVE-2008-1672)
2253 [Steve Henson, Mark Cox]
2255 *) Fix double free in TLS server name extensions which could lead to
2256 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
2259 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
2261 Clear the error queue to ensure that error entries left from
2262 older function calls do not interfere with the correct operation.
2263 [Lutz Jaenicke, Erik de Castro Lopo]
2265 *) Remove root CA certificates of commercial CAs:
2267 The OpenSSL project does not recommend any specific CA and does not
2268 have any policy with respect to including or excluding any CA.
2269 Therefore it does not make any sense to ship an arbitrary selection
2270 of root CA certificates with the OpenSSL software.
2273 *) RSA OAEP patches to fix two separate invalid memory reads.
2274 The first one involves inputs when 'lzero' is greater than
2275 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
2276 before the beginning of from). The second one involves inputs where
2277 the 'db' section contains nothing but zeroes (there is a one-byte
2278 invalid read after the end of 'db').
2279 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2281 *) Partial backport from 0.9.9-dev:
2283 Introduce bn_mul_mont (dedicated Montgomery multiplication
2284 procedure) as a candidate for BIGNUM assembler implementation.
2285 While 0.9.9-dev uses assembler for various architectures, only
2286 x86_64 is available by default here in the 0.9.8 branch, and
2287 32-bit x86 is available through a compile-time setting.
2289 To try the 32-bit x86 assembler implementation, use Configure
2290 option "enable-montasm" (which exists only for this backport).
2292 As "enable-montasm" for 32-bit x86 disclaims code stability
2293 anyway, in this constellation we activate additional code
2294 backported from 0.9.9-dev for further performance improvements,
2295 namely BN_from_montgomery_word. (To enable this otherwise,
2296 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
2298 [Andy Polyakov (backport partially by Bodo Moeller)]
2300 *) Add TLS session ticket callback. This allows an application to set
2301 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
2302 values. This is useful for key rollover for example where several key
2303 sets may exist with different names.
2306 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
2307 This was broken until now in 0.9.8 releases, such that the only way
2308 a registered ENGINE could be used (assuming it initialises
2309 successfully on the host) was to explicitly set it as the default
2310 for the relevant algorithms. This is in contradiction with 0.9.7
2311 behaviour and the documentation. With this fix, when an ENGINE is
2312 registered into a given algorithm's table of implementations, the
2313 'uptodate' flag is reset so that auto-discovery will be used next
2314 time a new context for that algorithm attempts to select an
2316 [Ian Lister (tweaked by Geoff Thorpe)]
2318 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
2319 implemention in the following ways:
2321 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
2324 Lack of BER streaming support means one pass streaming processing is
2325 only supported if data is detached: setting the streaming flag is
2326 ignored for embedded content.
2328 CMS support is disabled by default and must be explicitly enabled
2329 with the enable-cms configuration option.
2332 *) Update the GMP engine glue to do direct copies between BIGNUM and
2333 mpz_t when openssl and GMP use the same limb size. Otherwise the
2334 existing "conversion via a text string export" trick is still used.
2335 [Paul Sheer <paulsheer@gmail.com>]
2337 *) Zlib compression BIO. This is a filter BIO which compressed and
2338 uncompresses any data passed through it.
2341 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
2342 RFC3394 compatible AES key wrapping.
2345 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
2346 sets string data without copying. X509_ALGOR_set0() and
2347 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
2348 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
2349 from an X509_ATTRIBUTE structure optionally checking it occurs only
2350 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
2354 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
2355 to get the expected BN_FLG_CONSTTIME behavior.
2356 [Bodo Moeller (Google)]
2360 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
2361 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
2362 - added some more tests to do_tests.pl
2363 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
2364 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
2365 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
2366 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
2367 - various changes to netware.pl to enable gcc-cross builds on Win32
2369 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
2370 - various changes to fix missing prototype warnings
2371 - fixed x86nasm.pl to create correct asm files for NASM COFF output
2372 - added AES, WHIRLPOOL and CPUID assembler code to build files
2373 - added missing AES assembler make rules to mk1mf.pl
2374 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
2375 [Guenter Knauf <eflash@gmx.net>]
2377 *) Implement certificate status request TLS extension defined in RFC3546.
2378 A client can set the appropriate parameters and receive the encoded
2379 OCSP response via a callback. A server can query the supplied parameters
2380 and set the encoded OCSP response in the callback. Add simplified examples
2381 to s_client and s_server.
2384 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
2386 *) Fix various bugs:
2387 + Binary incompatibility of ssl_ctx_st structure
2388 + DTLS interoperation with non-compliant servers
2389 + Don't call get_session_cb() without proposed session
2390 + Fix ia64 assembler code
2391 [Andy Polyakov, Steve Henson]
2393 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
2395 *) DTLS Handshake overhaul. There were longstanding issues with
2396 OpenSSL DTLS implementation, which were making it impossible for
2397 RFC 4347 compliant client to communicate with OpenSSL server.
2398 Unfortunately just fixing these incompatibilities would "cut off"
2399 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
2400 server keeps tolerating non RFC compliant syntax. The opposite is
2401 not true, 0.9.8f client can not communicate with earlier server.
2402 This update even addresses CVE-2007-4995.
2405 *) Changes to avoid need for function casts in OpenSSL: some compilers
2406 (gcc 4.2 and later) reject their use.
2407 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
2410 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2411 RFC4507bis. The encrypted ticket format is an encrypted encoded
2412 SSL_SESSION structure, that way new session features are automatically
2415 If a client application caches session in an SSL_SESSION structure
2416 support is transparent because tickets are now stored in the encoded
2419 The SSL_CTX structure automatically generates keys for ticket
2420 protection in servers so again support should be possible
2421 with no application modification.
2423 If a client or server wishes to disable RFC4507 support then the option
2424 SSL_OP_NO_TICKET can be set.
2426 Add a TLS extension debugging callback to allow the contents of any client
2427 or server extensions to be examined.
2429 This work was sponsored by Google.
2432 *) Add initial support for TLS extensions, specifically for the server_name
2433 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2434 have new members for a host name. The SSL data structure has an
2435 additional member SSL_CTX *initial_ctx so that new sessions can be
2436 stored in that context to allow for session resumption, even after the
2437 SSL has been switched to a new SSL_CTX in reaction to a client's
2438 server_name extension.
2440 New functions (subject to change):
2442 SSL_get_servername()
2443 SSL_get_servername_type()
2446 New CTRL codes and macros (subject to change):
2448 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2449 - SSL_CTX_set_tlsext_servername_callback()
2450 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2451 - SSL_CTX_set_tlsext_servername_arg()
2452 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2454 openssl s_client has a new '-servername ...' option.
2456 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2457 '-key2 ...', '-servername_fatal' (subject to change). This allows
2458 testing the HostName extension for a specific single host name ('-cert'
2459 and '-key' remain fallbacks for handshakes without HostName
2460 negotiation). If the unrecogninzed_name alert has to be sent, this by
2461 default is a warning; it becomes fatal with the '-servername_fatal'
2464 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
2466 *) Add AES and SSE2 assembly language support to VC++ build.
2469 *) Mitigate attack on final subtraction in Montgomery reduction.
2472 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
2473 (which previously caused an internal error).
2476 *) Squeeze another 10% out of IGE mode when in != out.
2479 *) AES IGE mode speedup.
2480 [Dean Gaudet (Google)]
2482 *) Add the Korean symmetric 128-bit cipher SEED (see
2483 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
2484 add SEED ciphersuites from RFC 4162:
2486 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
2487 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
2488 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
2489 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
2491 To minimize changes between patchlevels in the OpenSSL 0.9.8
2492 series, SEED remains excluded from compilation unless OpenSSL
2493 is configured with 'enable-seed'.
2494 [KISA, Bodo Moeller]
2496 *) Mitigate branch prediction attacks, which can be practical if a
2497 single processor is shared, allowing a spy process to extract
2498 information. For detailed background information, see
2499 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
2500 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
2501 and Necessary Software Countermeasures"). The core of the change
2502 are new versions BN_div_no_branch() and
2503 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
2504 respectively, which are slower, but avoid the security-relevant
2505 conditional branches. These are automatically called by BN_div()
2506 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
2507 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
2508 remove a conditional branch.
2510 BN_FLG_CONSTTIME is the new name for the previous
2511 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
2512 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
2513 in the exponent causes BN_mod_exp_mont() to use the alternative
2514 implementation in BN_mod_exp_mont_consttime().) The old name
2515 remains as a deprecated alias.
2517 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
2518 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
2519 constant-time implementations for more than just exponentiation.
2520 Here too the old name is kept as a deprecated alias.
2522 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2523 the BN_BLINDING structure gets an independent copy of the
2524 modulus. This means that the previous "BIGNUM *m" argument to
2525 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2526 essentially becomes "const BIGNUM *m", although we can't actually
2527 change this in the header file before 0.9.9. It allows
2528 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2529 enable BN_FLG_CONSTTIME.
2531 [Matthew D Wood (Intel Corp)]
2533 *) In the SSL/TLS server implementation, be strict about session ID
2534 context matching (which matters if an application uses a single
2535 external cache for different purposes). Previously,
2536 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2537 set. This did ensure strict client verification, but meant that,
2538 with applications using a single external cache for quite
2539 different requirements, clients could circumvent ciphersuite
2540 restrictions for a given session ID context by starting a session
2541 in a different context.
2544 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2545 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2546 authentication-only ciphersuites.
2549 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
2550 not complete and could lead to a possible single byte overflow
2551 (CVE-2007-5135) [Ben Laurie]
2553 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
2555 *) Since AES128 and AES256 (and similarly Camellia128 and
2556 Camellia256) share a single mask bit in the logic of
2557 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2558 kludge to work properly if AES128 is available and AES256 isn't
2559 (or if Camellia128 is available and Camellia256 isn't).
2562 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
2563 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
2564 When a point or a seed is encoded in a BIT STRING, we need to
2565 prevent the removal of trailing zero bits to get the proper DER
2566 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
2567 of a NamedBitList, for which trailing 0 bits need to be removed.)
2570 *) Have SSL/TLS server implementation tolerate "mismatched" record
2571 protocol version while receiving ClientHello even if the
2572 ClientHello is fragmented. (The server can't insist on the
2573 particular protocol version it has chosen before the ServerHello
2574 message has informed the client about his choice.)
2577 *) Add RFC 3779 support.
2578 [Rob Austein for ARIN, Ben Laurie]
2580 *) Load error codes if they are not already present instead of using a
2581 static variable. This allows them to be cleanly unloaded and reloaded.
2582 Improve header file function name parsing.
2585 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
2586 or CAPABILITY handshake as required by RFCs.
2589 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
2591 *) Introduce limits to prevent malicious keys being able to
2592 cause a denial of service. (CVE-2006-2940)
2593 [Steve Henson, Bodo Moeller]
2595 *) Fix ASN.1 parsing of certain invalid structures that can result
2596 in a denial of service. (CVE-2006-2937) [Steve Henson]
2598 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2599 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2601 *) Fix SSL client code which could crash if connecting to a
2602 malicious SSLv2 server. (CVE-2006-4343)
2603 [Tavis Ormandy and Will Drewry, Google Security Team]
2605 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
2606 match only those. Before that, "AES256-SHA" would be interpreted
2607 as a pattern and match "AES128-SHA" too (since AES128-SHA got
2608 the same strength classification in 0.9.7h) as we currently only
2609 have a single AES bit in the ciphersuite description bitmap.
2610 That change, however, also applied to ciphersuite strings such as
2611 "RC4-MD5" that intentionally matched multiple ciphersuites --
2612 namely, SSL 2.0 ciphersuites in addition to the more common ones
2613 from SSL 3.0/TLS 1.0.
2615 So we change the selection algorithm again: Naming an explicit
2616 ciphersuite selects this one ciphersuite, and any other similar
2617 ciphersuite (same bitmap) from *other* protocol versions.
2618 Thus, "RC4-MD5" again will properly select both the SSL 2.0
2619 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
2621 Since SSL 2.0 does not have any ciphersuites for which the
2622 128/256 bit distinction would be relevant, this works for now.
2623 The proper fix will be to use different bits for AES128 and
2624 AES256, which would have avoided the problems from the beginning;
2625 however, bits are scarce, so we can only do this in a new release
2626 (not just a patchlevel) when we can change the SSL_CIPHER
2627 definition to split the single 'unsigned long mask' bitmap into
2628 multiple values to extend the available space.
2632 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2634 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2635 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2637 *) Add AES IGE and biIGE modes.
2640 *) Change the Unix randomness entropy gathering to use poll() when
2641 possible instead of select(), since the latter has some
2642 undesirable limitations.
2643 [Darryl Miles via Richard Levitte and Bodo Moeller]
2645 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2646 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2647 cannot be implicitly activated as part of, e.g., the "AES" alias.
2648 However, please upgrade to OpenSSL 0.9.9[-dev] for
2649 non-experimental use of the ECC ciphersuites to get TLS extension
2650 support, which is required for curve and point format negotiation
2651 to avoid potential handshake problems.
2654 *) Disable rogue ciphersuites:
2656 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2657 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2658 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2660 The latter two were purportedly from
2661 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2664 Also deactivate the remaining ciphersuites from
2665 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2666 unofficial, and the ID has long expired.
2669 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2670 dual-core machines) and other potential thread-safety issues.
2673 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
2674 versions), which is now available for royalty-free use
2675 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
2676 Also, add Camellia TLS ciphersuites from RFC 4132.
2678 To minimize changes between patchlevels in the OpenSSL 0.9.8
2679 series, Camellia remains excluded from compilation unless OpenSSL
2680 is configured with 'enable-camellia'.
2683 *) Disable the padding bug check when compression is in use. The padding
2684 bug check assumes the first packet is of even length, this is not
2685 necessarily true if compresssion is enabled and can result in false
2686 positives causing handshake failure. The actual bug test is ancient
2687 code so it is hoped that implementations will either have fixed it by
2688 now or any which still have the bug do not support compression.
2691 Changes between 0.9.8a and 0.9.8b [04 May 2006]
2693 *) When applying a cipher rule check to see if string match is an explicit
2694 cipher suite and only match that one cipher suite if it is.
2697 *) Link in manifests for VC++ if needed.
2698 [Austin Ziegler <halostatue@gmail.com>]
2700 *) Update support for ECC-based TLS ciphersuites according to
2701 draft-ietf-tls-ecc-12.txt with proposed changes (but without
2702 TLS extensions, which are supported starting with the 0.9.9
2703 branch, not in the OpenSSL 0.9.8 branch).
2706 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
2707 opaque EVP_CIPHER_CTX handling.
2710 *) Fixes and enhancements to zlib compression code. We now only use
2711 "zlib1.dll" and use the default __cdecl calling convention on Win32
2712 to conform with the standards mentioned here:
2713 http://www.zlib.net/DLL_FAQ.txt
2714 Static zlib linking now works on Windows and the new --with-zlib-include
2715 --with-zlib-lib options to Configure can be used to supply the location
2716 of the headers and library. Gracefully handle case where zlib library
2720 *) Several fixes and enhancements to the OID generation code. The old code
2721 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
2722 handle numbers larger than ULONG_MAX, truncated printing and had a
2723 non standard OBJ_obj2txt() behaviour.
2726 *) Add support for building of engines under engine/ as shared libraries
2727 under VC++ build system.
2730 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
2731 Hopefully, we will not see any false combination of paths any more.
2734 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
2736 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2737 (part of SSL_OP_ALL). This option used to disable the
2738 countermeasure against man-in-the-middle protocol-version
2739 rollback in the SSL 2.0 server implementation, which is a bad
2740 idea. (CVE-2005-2969)
2742 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2743 for Information Security, National Institute of Advanced Industrial
2744 Science and Technology [AIST], Japan)]
2746 *) Add two function to clear and return the verify parameter flags.
2749 *) Keep cipherlists sorted in the source instead of sorting them at
2750 runtime, thus removing the need for a lock.
2753 *) Avoid some small subgroup attacks in Diffie-Hellman.
2754 [Nick Mathewson and Ben Laurie]
2756 *) Add functions for well-known primes.
2759 *) Extended Windows CE support.
2760 [Satoshi Nakamura and Andy Polyakov]
2762 *) Initialize SSL_METHOD structures at compile time instead of during
2763 runtime, thus removing the need for a lock.
2766 *) Make PKCS7_decrypt() work even if no certificate is supplied by
2767 attempting to decrypt each encrypted key in turn. Add support to
2771 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
2773 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
2776 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
2779 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
2780 key into the same file any more.
2783 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
2786 *) Add -utf8 command line and config file option to 'ca'.
2787 [Stefan <stf@udoma.org]
2789 *) Removed the macro des_crypt(), as it seems to conflict with some
2790 libraries. Use DES_crypt().
2793 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
2794 involves renaming the source and generated shared-libs for
2795 both. The engines will accept the corrected or legacy ids
2796 ('ncipher' and '4758_cca' respectively) when binding. NB,
2797 this only applies when building 'shared'.
2798 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
2800 *) Add attribute functions to EVP_PKEY structure. Modify
2801 PKCS12_create() to recognize a CSP name attribute and
2802 use it. Make -CSP option work again in pkcs12 utility.
2805 *) Add new functionality to the bn blinding code:
2806 - automatic re-creation of the BN_BLINDING parameters after
2807 a fixed number of uses (currently 32)
2808 - add new function for parameter creation
2809 - introduce flags to control the update behaviour of the
2810 BN_BLINDING parameters
2811 - hide BN_BLINDING structure
2812 Add a second BN_BLINDING slot to the RSA structure to improve
2813 performance when a single RSA object is shared among several
2817 *) Add support for DTLS.
2818 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
2820 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
2821 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
2824 *) Remove buggy and incompletet DH cert support from
2825 ssl/ssl_rsa.c and ssl/s3_both.c
2828 *) Use SHA-1 instead of MD5 as the default digest algorithm for
2829 the apps/openssl applications.
2832 *) Compile clean with "-Wall -Wmissing-prototypes
2833 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
2834 DEBUG_SAFESTACK must also be set.
2837 *) Change ./Configure so that certain algorithms can be disabled by default.
2838 The new counterpiece to "no-xxx" is "enable-xxx".
2840 The patented RC5 and MDC2 algorithms will now be disabled unless
2841 "enable-rc5" and "enable-mdc2", respectively, are specified.
2843 (IDEA remains enabled despite being patented. This is because IDEA
2844 is frequently required for interoperability, and there is no license
2845 fee for non-commercial use. As before, "no-idea" can be used to
2846 avoid this algorithm.)
2850 *) Add processing of proxy certificates (see RFC 3820). This work was
2851 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
2852 EGEE (Enabling Grids for E-science in Europe).
2855 *) RC4 performance overhaul on modern architectures/implementations, such
2856 as Intel P4, IA-64 and AMD64.
2859 *) New utility extract-section.pl. This can be used specify an alternative
2860 section number in a pod file instead of having to treat each file as
2861 a separate case in Makefile. This can be done by adding two lines to the
2864 =for comment openssl_section:XXX
2866 The blank line is mandatory.
2870 *) New arguments -certform, -keyform and -pass for s_client and s_server
2871 to allow alternative format key and certificate files and passphrase
2875 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
2876 update associated structures and add various utility functions.
2878 Add new policy related verify parameters, include policy checking in
2879 standard verify code. Enhance 'smime' application with extra parameters
2880 to support policy checking and print out.
2883 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
2884 Nehemiah processors. These extensions support AES encryption in hardware
2885 as well as RNG (though RNG support is currently disabled).
2886 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
2888 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
2891 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
2892 [Andy Polyakov and a number of other people]
2894 *) Improved PowerPC platform support. Most notably BIGNUM assembler
2895 implementation contributed by IBM.
2896 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
2898 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
2899 exponent rather than 'unsigned long'. There is a corresponding change to
2900 the new 'rsa_keygen' element of the RSA_METHOD structure.
2901 [Jelte Jansen, Geoff Thorpe]
2903 *) Functionality for creating the initial serial number file is now
2904 moved from CA.pl to the 'ca' utility with a new option -create_serial.
2906 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
2907 number file to 1, which is bound to cause problems. To avoid
2908 the problems while respecting compatibility between different 0.9.7
2909 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
2910 CA.pl for serial number initialization. With the new release 0.9.8,
2911 we can fix the problem directly in the 'ca' utility.)
2914 *) Reduced header interdepencies by declaring more opaque objects in
2915 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
2916 give fewer recursive includes, which could break lazy source code - so
2917 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
2918 developers should define this symbol when building and using openssl to
2919 ensure they track the recommended behaviour, interfaces, [etc], but
2920 backwards-compatible behaviour prevails when this isn't defined.
2923 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
2926 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
2927 This will generate a random key of the appropriate length based on the
2928 cipher context. The EVP_CIPHER can provide its own random key generation
2929 routine to support keys of a specific form. This is used in the des and
2930 3des routines to generate a key of the correct parity. Update S/MIME
2931 code to use new functions and hence generate correct parity DES keys.
2932 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
2933 valid (weak or incorrect parity).
2936 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
2937 as looking them up. This is useful when the verified structure may contain
2938 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
2939 present unless the new PKCS7_NO_CRL flag is asserted.
2942 *) Extend ASN1 oid configuration module. It now additionally accepts the
2945 shortName = some long name, 1.2.3.4
2948 *) Reimplemented the BN_CTX implementation. There is now no more static
2949 limitation on the number of variables it can handle nor the depth of the
2950 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
2951 information can now expand as required, and rather than having a single
2952 static array of bignums, BN_CTX now uses a linked-list of such arrays
2953 allowing it to expand on demand whilst maintaining the usefulness of
2954 BN_CTX's "bundling".
2957 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
2958 to allow all RSA operations to function using a single BN_CTX.
2961 *) Preliminary support for certificate policy evaluation and checking. This
2962 is initially intended to pass the tests outlined in "Conformance Testing
2963 of Relying Party Client Certificate Path Processing Logic" v1.07.
2966 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
2967 remained unused and not that useful. A variety of other little bignum
2968 tweaks and fixes have also been made continuing on from the audit (see
2972 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
2973 associated ASN1, EVP and SSL functions and old ASN1 macros.
2976 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
2977 and this should never fail. So the return value from the use of
2978 BN_set_word() (which can fail due to needless expansion) is now deprecated;
2979 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
2982 *) BN_CTX_get() should return zero-valued bignums, providing the same
2983 initialised value as BN_new().
2984 [Geoff Thorpe, suggested by Ulf Möller]
2986 *) Support for inhibitAnyPolicy certificate extension.
2989 *) An audit of the BIGNUM code is underway, for which debugging code is
2990 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
2991 is considered valid when processing BIGNUMs, and causes execution to
2992 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
2993 further steps are taken to deliberately pollute unused data in BIGNUM
2994 structures to try and expose faulty code further on. For now, openssl will
2995 (in its default mode of operation) continue to tolerate the inconsistent
2996 forms that it has tolerated in the past, but authors and packagers should
2997 consider trying openssl and their own applications when compiled with
2998 these debugging symbols defined. It will help highlight potential bugs in
2999 their own code, and will improve the test coverage for OpenSSL itself. At
3000 some point, these tighter rules will become openssl's default to improve
3001 maintainability, though the assert()s and other overheads will remain only
3002 in debugging configurations. See bn.h for more details.
3003 [Geoff Thorpe, Nils Larsch, Ulf Möller]
3005 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
3006 that can only be obtained through BN_CTX_new() (which implicitly
3007 initialises it). The presence of this function only made it possible
3008 to overwrite an existing structure (and cause memory leaks).
3011 *) Because of the callback-based approach for implementing LHASH as a
3012 template type, lh_insert() adds opaque objects to hash-tables and
3013 lh_doall() or lh_doall_arg() are typically used with a destructor callback
3014 to clean up those corresponding objects before destroying the hash table
3015 (and losing the object pointers). So some over-zealous constifications in
3016 LHASH have been relaxed so that lh_insert() does not take (nor store) the
3017 objects as "const" and the lh_doall[_arg] callback wrappers are not
3018 prototyped to have "const" restrictions on the object pointers they are
3019 given (and so aren't required to cast them away any more).
3022 *) The tmdiff.h API was so ugly and minimal that our own timing utility
3023 (speed) prefers to use its own implementation. The two implementations
3024 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
3025 its object type properly exposed (MS_TM) instead of casting to/from "char
3026 *". This may still change yet if someone realises MS_TM and "ms_time_***"
3027 aren't necessarily the greatest nomenclatures - but this is what was used
3028 internally to the implementation so I've used that for now.
3031 *) Ensure that deprecated functions do not get compiled when
3032 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
3033 the self-tests were still using deprecated key-generation functions so
3034 these have been updated also.
3037 *) Reorganise PKCS#7 code to separate the digest location functionality
3038 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
3039 New function PKCS7_set_digest() to set the digest type for PKCS#7
3040 digestedData type. Add additional code to correctly generate the
3041 digestedData type and add support for this type in PKCS7 initialization
3045 *) New function PKCS7_set0_type_other() this initializes a PKCS7
3046 structure of type "other".
3049 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
3050 sure the loop does correctly stop and breaking ("division by zero")
3051 modulus operations are not performed. The (pre-generated) prime
3052 table crypto/bn/bn_prime.h was already correct, but it could not be
3053 re-generated on some platforms because of the "division by zero"
3054 situation in the script.
3055 [Ralf S. Engelschall]
3057 *) Update support for ECC-based TLS ciphersuites according to
3058 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
3059 SHA-1 now is only used for "small" curves (where the
3060 representation of a field element takes up to 24 bytes); for
3061 larger curves, the field element resulting from ECDH is directly
3062 used as premaster secret.
3063 [Douglas Stebila (Sun Microsystems Laboratories)]
3065 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
3066 curve secp160r1 to the tests.
3067 [Douglas Stebila (Sun Microsystems Laboratories)]
3069 *) Add the possibility to load symbols globally with DSO.
3070 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
3072 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
3073 control of the error stack.
3076 *) Add support for STORE in ENGINE.
3079 *) Add the STORE type. The intention is to provide a common interface
3080 to certificate and key stores, be they simple file-based stores, or
3081 HSM-type store, or LDAP stores, or...
3082 NOTE: The code is currently UNTESTED and isn't really used anywhere.
3085 *) Add a generic structure called OPENSSL_ITEM. This can be used to
3086 pass a list of arguments to any function as well as provide a way
3087 for a function to pass data back to the caller.
3090 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
3091 works like BUF_strdup() but can be used to duplicate a portion of
3092 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
3096 *) Add the function sk_find_ex() which works like sk_find(), but will
3097 return an index to an element even if an exact match couldn't be
3098 found. The index is guaranteed to point at the element where the
3099 searched-for key would be inserted to preserve sorting order.
3102 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
3103 takes an extra flags argument for optional functionality. Currently,
3104 the following flags are defined:
3106 OBJ_BSEARCH_VALUE_ON_NOMATCH
3107 This one gets OBJ_bsearch_ex() to return a pointer to the first
3108 element where the comparing function returns a negative or zero
3111 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
3112 This one gets OBJ_bsearch_ex() to return a pointer to the first
3113 element where the comparing function returns zero. This is useful
3114 if there are more than one element where the comparing function
3118 *) Make it possible to create self-signed certificates with 'openssl ca'
3119 in such a way that the self-signed certificate becomes part of the
3120 CA database and uses the same mechanisms for serial number generation
3121 as all other certificate signing. The new flag '-selfsign' enables
3122 this functionality. Adapt CA.sh and CA.pl.in.
3125 *) Add functionality to check the public key of a certificate request
3126 against a given private. This is useful to check that a certificate
3127 request can be signed by that key (self-signing).
3130 *) Make it possible to have multiple active certificates with the same
3131 subject in the CA index file. This is done only if the keyword
3132 'unique_subject' is set to 'no' in the main CA section (default
3133 if 'CA_default') of the configuration file. The value is saved
3134 with the database itself in a separate index attribute file,
3135 named like the index file with '.attr' appended to the name.
3138 *) Generate muti valued AVAs using '+' notation in config files for
3142 *) Support for nameConstraints certificate extension.
3145 *) Support for policyConstraints certificate extension.
3148 *) Support for policyMappings certificate extension.
3151 *) Make sure the default DSA_METHOD implementation only uses its
3152 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
3153 and change its own handlers to be NULL so as to remove unnecessary
3154 indirection. This lets alternative implementations fallback to the
3155 default implementation more easily.
3158 *) Support for directoryName in GeneralName related extensions
3162 *) Make it possible to link applications using Makefile.shared.
3163 Make that possible even when linking against static libraries!
3166 *) Support for single pass processing for S/MIME signing. This now
3167 means that S/MIME signing can be done from a pipe, in addition
3168 cleartext signing (multipart/signed type) is effectively streaming
3169 and the signed data does not need to be all held in memory.
3171 This is done with a new flag PKCS7_STREAM. When this flag is set
3172 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
3173 is done after the data is output (and digests calculated) in
3174 SMIME_write_PKCS7().
3177 *) Add full support for -rpath/-R, both in shared libraries and
3178 applications, at least on the platforms where it's known how
3182 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
3183 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
3184 will now compute a table of multiples of the generator that
3185 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
3186 faster (notably in the case of a single point multiplication,
3187 scalar * generator).
3188 [Nils Larsch, Bodo Moeller]
3190 *) IPv6 support for certificate extensions. The various extensions
3191 which use the IP:a.b.c.d can now take IPv6 addresses using the
3192 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
3196 *) Added an ENGINE that implements RSA by performing private key
3197 exponentiations with the GMP library. The conversions to and from
3198 GMP's mpz_t format aren't optimised nor are any montgomery forms
3199 cached, and on x86 it appears OpenSSL's own performance has caught up.
3200 However there are likely to be other architectures where GMP could
3201 provide a boost. This ENGINE is not built in by default, but it can be
3202 specified at Configure time and should be accompanied by the necessary
3203 linker additions, eg;
3204 ./config -DOPENSSL_USE_GMP -lgmp
3207 *) "openssl engine" will not display ENGINE/DSO load failure errors when
3208 testing availability of engines with "-t" - the old behaviour is
3209 produced by increasing the feature's verbosity with "-tt".
3212 *) ECDSA routines: under certain error conditions uninitialized BN objects
3213 could be freed. Solution: make sure initialization is performed early
3214 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
3218 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
3219 and DH_METHOD (eg. by ENGINE implementations) to override the normal
3220 software implementations. For DSA and DH, parameter generation can
3221 also be overriden by providing the appropriate method callbacks.
3224 *) Change the "progress" mechanism used in key-generation and
3225 primality testing to functions that take a new BN_GENCB pointer in
3226 place of callback/argument pairs. The new API functions have "_ex"
3227 postfixes and the older functions are reimplemented as wrappers for
3228 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
3229 declarations of the old functions to help (graceful) attempts to
3230 migrate to the new functions. Also, the new key-generation API
3231 functions operate on a caller-supplied key-structure and return
3232 success/failure rather than returning a key or NULL - this is to
3233 help make "keygen" another member function of RSA_METHOD etc.
3235 Example for using the new callback interface:
3237 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
3241 BN_GENCB_set(&my_cb, my_callback, my_arg);
3243 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
3244 /* For the meaning of a, b in calls to my_callback(), see the
3245 * documentation of the function that calls the callback.
3246 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
3247 * my_callback should return 1 if it wants BN_is_prime_ex()
3248 * to continue, or 0 to stop.
3253 *) Change the ZLIB compression method to be stateful, and make it
3254 available to TLS with the number defined in
3255 draft-ietf-tls-compression-04.txt.
3258 *) Add the ASN.1 structures and functions for CertificatePair, which
3259 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
3261 CertificatePair ::= SEQUENCE {
3262 forward [0] Certificate OPTIONAL,
3263 reverse [1] Certificate OPTIONAL,
3264 -- at least one of the pair shall be present -- }
3266 Also implement the PEM functions to read and write certificate
3267 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
3269 This needed to be defined, mostly for the sake of the LDAP
3270 attribute crossCertificatePair, but may prove useful elsewhere as
3274 *) Make it possible to inhibit symlinking of shared libraries in
3275 Makefile.shared, for Cygwin's sake.
3278 *) Extend the BIGNUM API by creating a function
3279 void BN_set_negative(BIGNUM *a, int neg);
3280 and a macro that behave like
3281 int BN_is_negative(const BIGNUM *a);
3283 to avoid the need to access 'a->neg' directly in applications.
3286 *) Implement fast modular reduction for pseudo-Mersenne primes
3287 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
3288 EC_GROUP_new_curve_GFp() will now automatically use this
3290 [Nils Larsch <nla@trustcenter.de>]
3292 *) Add new lock type (CRYPTO_LOCK_BN).
3295 *) Change the ENGINE framework to automatically load engines
3296 dynamically from specific directories unless they could be
3297 found to already be built in or loaded. Move all the
3298 current engines except for the cryptodev one to a new
3300 The engines in engines/ are built as shared libraries if
3301 the "shared" options was given to ./Configure or ./config.
3302 Otherwise, they are inserted in libcrypto.a.
3303 /usr/local/ssl/engines is the default directory for dynamic
3304 engines, but that can be overriden at configure time through
3305 the usual use of --prefix and/or --openssldir, and at run
3306 time with the environment variable OPENSSL_ENGINES.
3307 [Geoff Thorpe and Richard Levitte]
3309 *) Add Makefile.shared, a helper makefile to build shared
3310 libraries. Addapt Makefile.org.
3313 *) Add version info to Win32 DLLs.
3314 [Peter 'Luna' Runestig" <peter@runestig.com>]
3316 *) Add new 'medium level' PKCS#12 API. Certificates and keys
3317 can be added using this API to created arbitrary PKCS#12
3318 files while avoiding the low level API.
3320 New options to PKCS12_create(), key or cert can be NULL and
3321 will then be omitted from the output file. The encryption
3322 algorithm NIDs can be set to -1 for no encryption, the mac
3323 iteration count can be set to 0 to omit the mac.
3325 Enhance pkcs12 utility by making the -nokeys and -nocerts
3326 options work when creating a PKCS#12 file. New option -nomac
3327 to omit the mac, NONE can be set for an encryption algorithm.
3328 New code is modified to use the enhanced PKCS12_create()
3329 instead of the low level API.
3332 *) Extend ASN1 encoder to support indefinite length constructed
3333 encoding. This can output sequences tags and octet strings in
3334 this form. Modify pk7_asn1.c to support indefinite length
3335 encoding. This is experimental and needs additional code to
3336 be useful, such as an ASN1 bio and some enhanced streaming
3339 Extend template encode functionality so that tagging is passed
3340 down to the template encoder.
3343 *) Let 'openssl req' fail if an argument to '-newkey' is not
3344 recognized instead of using RSA as a default.
3347 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
3348 As these are not official, they are not included in "ALL";
3349 the "ECCdraft" ciphersuite group alias can be used to select them.
3350 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
3352 *) Add ECDH engine support.
3353 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
3355 *) Add ECDH in new directory crypto/ecdh/.
3356 [Douglas Stebila (Sun Microsystems Laboratories)]
3358 *) Let BN_rand_range() abort with an error after 100 iterations
3359 without success (which indicates a broken PRNG).
3362 *) Change BN_mod_sqrt() so that it verifies that the input value
3363 is really the square of the return value. (Previously,
3364 BN_mod_sqrt would show GIGO behaviour.)
3367 *) Add named elliptic curves over binary fields from X9.62, SECG,
3368 and WAP/WTLS; add OIDs that were still missing.
3370 [Sheueling Chang Shantz and Douglas Stebila
3371 (Sun Microsystems Laboratories)]
3373 *) Extend the EC library for elliptic curves over binary fields
3374 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
3377 EC_GF2m_simple_method
3381 EC_GROUP_new_curve_GF2m
3382 EC_GROUP_set_curve_GF2m
3383 EC_GROUP_get_curve_GF2m
3384 EC_POINT_set_affine_coordinates_GF2m
3385 EC_POINT_get_affine_coordinates_GF2m
3386 EC_POINT_set_compressed_coordinates_GF2m
3388 Point compression for binary fields is disabled by default for
3389 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
3392 As binary polynomials are represented as BIGNUMs, various members
3393 of the EC_GROUP and EC_POINT data structures can be shared
3394 between the implementations for prime fields and binary fields;
3395 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
3396 are essentially identical to their ..._GFp counterparts.
3397 (For simplicity, the '..._GFp' prefix has been dropped from
3398 various internal method names.)
3400 An internal 'field_div' method (similar to 'field_mul' and
3401 'field_sqr') has been added; this is used only for binary fields.
3403 [Sheueling Chang Shantz and Douglas Stebila
3404 (Sun Microsystems Laboratories)]
3406 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
3407 through methods ('mul', 'precompute_mult').
3409 The generic implementations (now internally called 'ec_wNAF_mul'
3410 and 'ec_wNAF_precomputed_mult') remain the default if these
3411 methods are undefined.
3413 [Sheueling Chang Shantz and Douglas Stebila
3414 (Sun Microsystems Laboratories)]
3416 *) New function EC_GROUP_get_degree, which is defined through
3417 EC_METHOD. For curves over prime fields, this returns the bit
3418 length of the modulus.
3420 [Sheueling Chang Shantz and Douglas Stebila
3421 (Sun Microsystems Laboratories)]
3423 *) New functions EC_GROUP_dup, EC_POINT_dup.
3424 (These simply call ..._new and ..._copy).
3426 [Sheueling Chang Shantz and Douglas Stebila
3427 (Sun Microsystems Laboratories)]
3429 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
3430 Polynomials are represented as BIGNUMs (where the sign bit is not
3431 used) in the following functions [macros]:
3434 BN_GF2m_sub [= BN_GF2m_add]
3435 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
3436 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
3437 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
3439 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
3440 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
3441 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
3442 BN_GF2m_cmp [= BN_ucmp]
3444 (Note that only the 'mod' functions are actually for fields GF(2^m).
3445 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
3447 For some functions, an the irreducible polynomial defining a
3448 field can be given as an 'unsigned int[]' with strictly
3449 decreasing elements giving the indices of those bits that are set;
3450 i.e., p[] represents the polynomial
3451 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
3453 p[0] > p[1] > ... > p[k] = 0.
3454 This applies to the following functions:
3459 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
3460 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
3462 BN_GF2m_mod_sqrt_arr
3463 BN_GF2m_mod_solve_quad_arr
3467 Conversion can be performed by the following functions:
3472 bntest.c has additional tests for binary polynomial arithmetic.
3474 Two implementations for BN_GF2m_mod_div() are available.
3475 The default algorithm simply uses BN_GF2m_mod_inv() and
3476 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
3477 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
3478 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
3480 [Sheueling Chang Shantz and Douglas Stebila
3481 (Sun Microsystems Laboratories)]
3483 *) Add new error code 'ERR_R_DISABLED' that can be used when some
3484 functionality is disabled at compile-time.
3485 [Douglas Stebila <douglas.stebila@sun.com>]
3487 *) Change default behaviour of 'openssl asn1parse' so that more
3488 information is visible when viewing, e.g., a certificate:
3490 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
3491 mode the content of non-printable OCTET STRINGs is output in a
3492 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
3493 avoid the appearance of a printable string.
3494 [Nils Larsch <nla@trustcenter.de>]
3496 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
3498 EC_GROUP_set_asn1_flag()
3499 EC_GROUP_get_asn1_flag()
3500 EC_GROUP_set_point_conversion_form()
3501 EC_GROUP_get_point_conversion_form()
3502 These control ASN1 encoding details:
3503 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
3504 has been set to OPENSSL_EC_NAMED_CURVE.
3505 - Points are encoded in uncompressed form by default; options for
3506 asn1_for are as for point2oct, namely
3507 POINT_CONVERSION_COMPRESSED
3508 POINT_CONVERSION_UNCOMPRESSED
3509 POINT_CONVERSION_HYBRID
3511 Also add 'seed' and 'seed_len' members to EC_GROUP with access
3514 EC_GROUP_get0_seed()
3515 EC_GROUP_get_seed_len()
3516 This is used only for ASN1 purposes (so far).
3517 [Nils Larsch <nla@trustcenter.de>]
3519 *) Add 'field_type' member to EC_METHOD, which holds the NID
3520 of the appropriate field type OID. The new function
3521 EC_METHOD_get_field_type() returns this value.
3522 [Nils Larsch <nla@trustcenter.de>]
3527 EC_POINT_point2hex()
3528 EC_POINT_hex2point()
3529 providing useful interfaces to EC_POINT_point2oct() and
3530 EC_POINT_oct2point().
3531 [Nils Larsch <nla@trustcenter.de>]
3533 *) Change internals of the EC library so that the functions
3534 EC_GROUP_set_generator()
3535 EC_GROUP_get_generator()
3536 EC_GROUP_get_order()
3537 EC_GROUP_get_cofactor()
3538 are implemented directly in crypto/ec/ec_lib.c and not dispatched
3539 to methods, which would lead to unnecessary code duplication when
3540 adding different types of curves.
3541 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
3543 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
3544 arithmetic, and such that modified wNAFs are generated
3545 (which avoid length expansion in many cases).
3548 *) Add a function EC_GROUP_check_discriminant() (defined via
3549 EC_METHOD) that verifies that the curve discriminant is non-zero.
3551 Add a function EC_GROUP_check() that makes some sanity tests
3552 on a EC_GROUP, its generator and order. This includes
3553 EC_GROUP_check_discriminant().
3554 [Nils Larsch <nla@trustcenter.de>]
3556 *) Add ECDSA in new directory crypto/ecdsa/.
3558 Add applications 'openssl ecparam' and 'openssl ecdsa'
3559 (these are based on 'openssl dsaparam' and 'openssl dsa').
3561 ECDSA support is also included in various other files across the
3562 library. Most notably,
3563 - 'openssl req' now has a '-newkey ecdsa:file' option;
3564 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
3565 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
3566 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
3567 them suitable for ECDSA where domain parameters must be
3568 extracted before the specific public key;
3569 - ECDSA engine support has been added.
3570 [Nils Larsch <nla@trustcenter.de>]
3572 *) Include some named elliptic curves, and add OIDs from X9.62,
3573 SECG, and WAP/WTLS. Each curve can be obtained from the new
3575 EC_GROUP_new_by_curve_name(),
3576 and the list of available named curves can be obtained with
3577 EC_get_builtin_curves().
3578 Also add a 'curve_name' member to EC_GROUP objects, which can be
3580 EC_GROUP_set_curve_name()
3581 EC_GROUP_get_curve_name()
3582 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
3584 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3585 was actually never needed) and in BN_mul(). The removal in BN_mul()
3586 required a small change in bn_mul_part_recursive() and the addition
3587 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3588 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3589 bn_sub_words() and bn_add_words() except they take arrays with
3593 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
3595 *) Cleanse PEM buffers before freeing them since they may contain
3597 [Benjamin Bennett <ben@psc.edu>]
3599 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3600 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3601 authentication-only ciphersuites.
3604 *) Since AES128 and AES256 share a single mask bit in the logic of
3605 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3606 kludge to work properly if AES128 is available and AES256 isn't.
3609 *) Expand security boundary to match 1.1.1 module.
3612 *) Remove redundant features: hash file source, editing of test vectors
3613 modify fipsld to use external fips_premain.c signature.
3616 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
3617 run algorithm test programs.
3620 *) Make algorithm test programs more tolerant of whitespace.
3623 *) Have SSL/TLS server implementation tolerate "mismatched" record
3624 protocol version while receiving ClientHello even if the
3625 ClientHello is fragmented. (The server can't insist on the
3626 particular protocol version it has chosen before the ServerHello
3627 message has informed the client about his choice.)
3630 *) Load error codes if they are not already present instead of using a
3631 static variable. This allows them to be cleanly unloaded and reloaded.
3634 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3636 *) Introduce limits to prevent malicious keys being able to
3637 cause a denial of service. (CVE-2006-2940)
3638 [Steve Henson, Bodo Moeller]
3640 *) Fix ASN.1 parsing of certain invalid structures that can result
3641 in a denial of service. (CVE-2006-2937) [Steve Henson]
3643 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3644 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3646 *) Fix SSL client code which could crash if connecting to a
3647 malicious SSLv2 server. (CVE-2006-4343)
3648 [Tavis Ormandy and Will Drewry, Google Security Team]
3650 *) Change ciphersuite string processing so that an explicit
3651 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3652 will no longer include "AES128-SHA"), and any other similar
3653 ciphersuite (same bitmap) from *other* protocol versions (so that
3654 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3655 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3656 changes from 0.9.8b and 0.9.8d.
3659 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
3661 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3662 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3664 *) Change the Unix randomness entropy gathering to use poll() when
3665 possible instead of select(), since the latter has some
3666 undesirable limitations.
3667 [Darryl Miles via Richard Levitte and Bodo Moeller]
3669 *) Disable rogue ciphersuites:
3671 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3672 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3673 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3675 The latter two were purportedly from
3676 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3679 Also deactive the remaining ciphersuites from
3680 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3681 unofficial, and the ID has long expired.
3684 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3685 dual-core machines) and other potential thread-safety issues.
3688 Changes between 0.9.7i and 0.9.7j [04 May 2006]
3690 *) Adapt fipsld and the build system to link against the validated FIPS
3691 module in FIPS mode.
3694 *) Fixes for VC++ 2005 build under Windows.
3697 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
3698 from a Windows bash shell such as MSYS. It is autodetected from the
3699 "config" script when run from a VC++ environment. Modify standard VC++
3700 build to use fipscanister.o from the GNU make build.
3703 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
3705 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
3706 The value now differs depending on if you build for FIPS or not.
3707 BEWARE! A program linked with a shared FIPSed libcrypto can't be
3708 safely run with a non-FIPSed libcrypto, as it may crash because of
3709 the difference induced by this change.
3712 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
3714 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3715 (part of SSL_OP_ALL). This option used to disable the
3716 countermeasure against man-in-the-middle protocol-version
3717 rollback in the SSL 2.0 server implementation, which is a bad
3718 idea. (CVE-2005-2969)
3720 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3721 for Information Security, National Institute of Advanced Industrial
3722 Science and Technology [AIST], Japan)]
3724 *) Minimal support for X9.31 signatures and PSS padding modes. This is
3725 mainly for FIPS compliance and not fully integrated at this stage.
3728 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
3729 the exponentiation using a fixed-length exponent. (Otherwise,
3730 the information leaked through timing could expose the secret key
3731 after many signatures; cf. Bleichenbacher's attack on DSA with
3735 *) Make a new fixed-window mod_exp implementation the default for
3736 RSA, DSA, and DH private-key operations so that the sequence of
3737 squares and multiplies and the memory access pattern are
3738 independent of the particular secret key. This will mitigate
3739 cache-timing and potential related attacks.
3741 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
3742 and this is automatically used by BN_mod_exp_mont() if the new flag
3743 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
3744 will use this BN flag for private exponents unless the flag
3745 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
3746 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
3748 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
3750 *) Change the client implementation for SSLv23_method() and
3751 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
3752 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
3753 (Previously, the SSL 2.0 backwards compatible Client Hello
3754 message format would be used even with SSL_OP_NO_SSLv2.)
3757 *) Add support for smime-type MIME parameter in S/MIME messages which some
3761 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
3762 a threadsafe manner. Modify rsa code to use new function and add calls
3763 to dsa and dh code (which had race conditions before).
3766 *) Include the fixed error library code in the C error file definitions
3767 instead of fixing them up at runtime. This keeps the error code
3768 structures constant.
3771 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
3773 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
3776 *) Fixes for newer kerberos headers. NB: the casts are needed because
3777 the 'length' field is signed on one version and unsigned on another
3778 with no (?) obvious way to tell the difference, without these VC++
3779 complains. Also the "definition" of FAR (blank) is no longer included
3780 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
3781 some needed definitions.
3784 *) Undo Cygwin change.
3787 *) Added support for proxy certificates according to RFC 3820.
3788 Because they may be a security thread to unaware applications,
3789 they must be explicitely allowed in run-time. See
3790 docs/HOWTO/proxy_certificates.txt for further information.
3793 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
3795 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
3796 server and client random values. Previously
3797 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
3798 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
3800 This change has negligible security impact because:
3802 1. Server and client random values still have 24 bytes of pseudo random
3805 2. Server and client random values are sent in the clear in the initial
3808 3. The master secret is derived using the premaster secret (48 bytes in
3809 size for static RSA ciphersuites) as well as client server and random
3812 The OpenSSL team would like to thank the UK NISCC for bringing this issue
3815 [Stephen Henson, reported by UK NISCC]
3817 *) Use Windows randomness collection on Cygwin.
3820 *) Fix hang in EGD/PRNGD query when communication socket is closed
3821 prematurely by EGD/PRNGD.
3822 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
3824 *) Prompt for pass phrases when appropriate for PKCS12 input format.
3827 *) Back-port of selected performance improvements from development
3828 branch, as well as improved support for PowerPC platforms.
3831 *) Add lots of checks for memory allocation failure, error codes to indicate
3832 failure and freeing up memory if a failure occurs.
3833 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
3835 *) Add new -passin argument to dgst.
3838 *) Perform some character comparisons of different types in X509_NAME_cmp:
3839 this is needed for some certificates that reencode DNs into UTF8Strings
3840 (in violation of RFC3280) and can't or wont issue name rollover
3844 *) Make an explicit check during certificate validation to see that
3845 the CA setting in each certificate on the chain is correct. As a
3846 side effect always do the following basic checks on extensions,
3847 not just when there's an associated purpose to the check:
3849 - if there is an unhandled critical extension (unless the user
3850 has chosen to ignore this fault)
3851 - if the path length has been exceeded (if one is set at all)
3852 - that certain extensions fit the associated purpose (if one has
3856 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
3858 *) Avoid a race condition when CRLs are checked in a multi threaded
3859 environment. This would happen due to the reordering of the revoked
3860 entries during signature checking and serial number lookup. Now the
3861 encoding is cached and the serial number sort performed under a lock.
3862 Add new STACK function sk_is_sorted().
3865 *) Add Delta CRL to the extension code.
3868 *) Various fixes to s3_pkt.c so alerts are sent properly.
3869 [David Holmes <d.holmes@f5.com>]
3871 *) Reduce the chances of duplicate issuer name and serial numbers (in
3872 violation of RFC3280) using the OpenSSL certificate creation utilities.
3873 This is done by creating a random 64 bit value for the initial serial
3874 number when a serial number file is created or when a self signed
3875 certificate is created using 'openssl req -x509'. The initial serial
3876 number file is created using 'openssl x509 -next_serial' in CA.pl
3877 rather than being initialized to 1.
3880 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
3882 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3883 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3884 [Joe Orton, Steve Henson]
3886 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
3888 [Joe Orton, Steve Henson]
3890 *) Make it possible to have multiple active certificates with the same
3891 subject in the CA index file. This is done only if the keyword
3892 'unique_subject' is set to 'no' in the main CA section (default
3893 if 'CA_default') of the configuration file. The value is saved
3894 with the database itself in a separate index attribute file,
3895 named like the index file with '.attr' appended to the name.
3898 *) X509 verify fixes. Disable broken certificate workarounds when
3899 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
3900 keyUsage extension present. Don't accept CRLs with unhandled critical
3901 extensions: since verify currently doesn't process CRL extensions this
3902 rejects a CRL with *any* critical extensions. Add new verify error codes
3906 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
3907 A clarification of RFC2560 will require the use of OCTET STRINGs and
3908 some implementations cannot handle the current raw format. Since OpenSSL
3909 copies and compares OCSP nonces as opaque blobs without any attempt at
3910 parsing them this should not create any compatibility issues.
3913 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
3914 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
3915 this HMAC (and other) operations are several times slower than OpenSSL
3919 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
3920 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3922 *) Use the correct content when signing type "other".
3925 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
3927 *) Fix various bugs revealed by running the NISCC test suite:
3929 Stop out of bounds reads in the ASN1 code when presented with
3930 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3932 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
3934 If verify callback ignores invalid public key errors don't try to check
3935 certificate signature with the NULL public key.
3939 *) New -ignore_err option in ocsp application to stop the server
3940 exiting on the first error in a request.
3943 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3944 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3948 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3949 extra data after the compression methods not only for TLS 1.0
3950 but also for SSL 3.0 (as required by the specification).
3951 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3953 *) Change X509_certificate_type() to mark the key as exported/exportable
3954 when it's 512 *bits* long, not 512 bytes.
3957 *) Change AES_cbc_encrypt() so it outputs exact multiple of
3958 blocks during encryption.
3961 *) Various fixes to base64 BIO and non blocking I/O. On write
3962 flushes were not handled properly if the BIO retried. On read
3963 data was not being buffered properly and had various logic bugs.
3964 This also affects blocking I/O when the data being decoded is a
3968 *) Various S/MIME bugfixes and compatibility changes:
3969 output correct application/pkcs7 MIME type if
3970 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
3971 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
3972 of files as .eml work). Correctly handle very long lines in MIME
3976 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
3978 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3979 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3980 a protocol version number mismatch like a decryption error
3981 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3984 *) Turn on RSA blinding by default in the default implementation
3985 to avoid a timing attack. Applications that don't want it can call
3986 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3987 They would be ill-advised to do so in most cases.
3988 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3990 *) Change RSA blinding code so that it works when the PRNG is not
3991 seeded (in this case, the secret RSA exponent is abused as
3992 an unpredictable seed -- if it is not unpredictable, there
3993 is no point in blinding anyway). Make RSA blinding thread-safe
3994 by remembering the creator's thread ID in rsa->blinding and
3995 having all other threads use local one-time blinding factors
3996 (this requires more computation than sharing rsa->blinding, but
3997 avoids excessive locking; and if an RSA object is not shared
3998 between threads, blinding will still be very fast).
4001 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
4002 ENGINE as defaults for all supported algorithms irrespective of
4003 the 'flags' parameter. 'flags' is now honoured, so applications
4004 should make sure they are passing it correctly.
4007 *) Target "mingw" now allows native Windows code to be generated in
4008 the Cygwin environment as well as with the MinGW compiler.
4011 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
4013 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4014 via timing by performing a MAC computation even if incorrrect
4015 block cipher padding has been found. This is a countermeasure
4016 against active attacks where the attacker has to distinguish
4017 between bad padding and a MAC verification error. (CVE-2003-0078)
4019 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4020 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4021 Martin Vuagnoux (EPFL, Ilion)]
4023 *) Make the no-err option work as intended. The intention with no-err
4024 is not to have the whole error stack handling routines removed from
4025 libcrypto, it's only intended to remove all the function name and
4026 reason texts, thereby removing some of the footprint that may not
4027 be interesting if those errors aren't displayed anyway.
4029 NOTE: it's still possible for any application or module to have it's
4030 own set of error texts inserted. The routines are there, just not
4031 used by default when no-err is given.
4034 *) Add support for FreeBSD on IA64.
4035 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
4037 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
4038 Kerberos function mit_des_cbc_cksum(). Before this change,
4039 the value returned by DES_cbc_cksum() was like the one from
4040 mit_des_cbc_cksum(), except the bytes were swapped.
4041 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
4043 *) Allow an application to disable the automatic SSL chain building.
4044 Before this a rather primitive chain build was always performed in
4045 ssl3_output_cert_chain(): an application had no way to send the
4046 correct chain if the automatic operation produced an incorrect result.
4048 Now the chain builder is disabled if either:
4050 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
4052 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
4054 The reasoning behind this is that an application would not want the
4055 auto chain building to take place if extra chain certificates are
4056 present and it might also want a means of sending no additional
4057 certificates (for example the chain has two certificates and the
4061 *) Add the possibility to build without the ENGINE framework.
4062 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4064 *) Under Win32 gmtime() can return NULL: check return value in
4065 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
4068 *) DSA routines: under certain error conditions uninitialized BN objects
4069 could be freed. Solution: make sure initialization is performed early
4070 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
4071 Nils Larsch <nla@trustcenter.de> via PR#459)
4074 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
4075 checked on reconnect on the client side, therefore session resumption
4076 could still fail with a "ssl session id is different" error. This
4077 behaviour is masked when SSL_OP_ALL is used due to
4078 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
4079 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4080 followup to PR #377.
4083 *) IA-32 assembler support enhancements: unified ELF targets, support
4084 for SCO/Caldera platforms, fix for Cygwin shared build.
4087 *) Add support for FreeBSD on sparc64. As a consequence, support for
4088 FreeBSD on non-x86 processors is separate from x86 processors on
4089 the config script, much like the NetBSD support.
4090 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
4092 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
4094 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
4097 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
4098 code (06) was taken as the first octet of the session ID and the last
4099 octet was ignored consequently. As a result SSLv2 client side session
4100 caching could not have worked due to the session ID mismatch between
4102 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4106 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
4107 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
4111 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
4112 seems that in spite of existing for more than a year, many application
4113 author have done nothing to provide the necessary callbacks, which
4114 means that this particular engine will not work properly anywhere.
4115 This is a very unfortunate situation which forces us, in the name
4116 of usability, to give the hw_ncipher.c a static lock, which is part
4118 NOTE: This is for the 0.9.7 series ONLY. This hack will never
4119 appear in 0.9.8 or later. We EXPECT application authors to have
4120 dealt properly with this when 0.9.8 is released (unless we actually
4121 make such changes in the libcrypto locking code that changes will
4122 have to be made anyway).
4125 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
4126 octets have been read, EOF or an error occurs. Without this change
4127 some truncated ASN1 structures will not produce an error.
4130 *) Disable Heimdal support, since it hasn't been fully implemented.
4131 Still give the possibility to force the use of Heimdal, but with
4132 warnings and a request that patches get sent to openssl-dev.
4135 *) Add the VC-CE target, introduce the WINCE sysname, and add
4136 INSTALL.WCE and appropriate conditionals to make it build.
4137 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4139 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
4140 cygssl-x.y.z.dll, where x, y and z are the major, minor and
4141 edit numbers of the version.
4142 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4144 *) Introduce safe string copy and catenation functions
4145 (BUF_strlcpy() and BUF_strlcat()).
4146 [Ben Laurie (CHATS) and Richard Levitte]
4148 *) Avoid using fixed-size buffers for one-line DNs.
4149 [Ben Laurie (CHATS)]
4151 *) Add BUF_MEM_grow_clean() to avoid information leakage when
4152 resizing buffers containing secrets, and use where appropriate.
4153 [Ben Laurie (CHATS)]
4155 *) Avoid using fixed size buffers for configuration file location.
4156 [Ben Laurie (CHATS)]
4158 *) Avoid filename truncation for various CA files.
4159 [Ben Laurie (CHATS)]
4161 *) Use sizeof in preference to magic numbers.
4162 [Ben Laurie (CHATS)]
4164 *) Avoid filename truncation in cert requests.
4165 [Ben Laurie (CHATS)]
4167 *) Add assertions to check for (supposedly impossible) buffer
4169 [Ben Laurie (CHATS)]
4171 *) Don't cache truncated DNS entries in the local cache (this could
4172 potentially lead to a spoofing attack).
4173 [Ben Laurie (CHATS)]
4175 *) Fix various buffers to be large enough for hex/decimal
4176 representations in a platform independent manner.
4177 [Ben Laurie (CHATS)]
4179 *) Add CRYPTO_realloc_clean() to avoid information leakage when
4180 resizing buffers containing secrets, and use where appropriate.
4181 [Ben Laurie (CHATS)]
4183 *) Add BIO_indent() to avoid much slightly worrying code to do
4185 [Ben Laurie (CHATS)]
4187 *) Convert sprintf()/BIO_puts() to BIO_printf().
4188 [Ben Laurie (CHATS)]
4190 *) buffer_gets() could terminate with the buffer only half
4192 [Ben Laurie (CHATS)]
4194 *) Add assertions to prevent user-supplied crypto functions from
4195 overflowing internal buffers by having large block sizes, etc.
4196 [Ben Laurie (CHATS)]
4198 *) New OPENSSL_assert() macro (similar to assert(), but enabled
4200 [Ben Laurie (CHATS)]
4202 *) Eliminate unused copy of key in RC4.
4203 [Ben Laurie (CHATS)]
4205 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
4206 [Ben Laurie (CHATS)]
4208 *) Fix off-by-one error in EGD path.
4209 [Ben Laurie (CHATS)]
4211 *) If RANDFILE path is too long, ignore instead of truncating.
4212 [Ben Laurie (CHATS)]
4214 *) Eliminate unused and incorrectly sized X.509 structure
4216 [Ben Laurie (CHATS)]
4218 *) Eliminate unused and dangerous function knumber().
4219 [Ben Laurie (CHATS)]
4221 *) Eliminate unused and dangerous structure, KSSL_ERR.
4222 [Ben Laurie (CHATS)]
4224 *) Protect against overlong session ID context length in an encoded
4225 session object. Since these are local, this does not appear to be
4227 [Ben Laurie (CHATS)]
4229 *) Change from security patch (see 0.9.6e below) that did not affect
4230 the 0.9.6 release series:
4232 Remote buffer overflow in SSL3 protocol - an attacker could
4233 supply an oversized master key in Kerberos-enabled versions.
4235 [Ben Laurie (CHATS)]
4237 *) Change the SSL kerb5 codes to match RFC 2712.
4240 *) Make -nameopt work fully for req and add -reqopt switch.
4241 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
4243 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
4244 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
4246 *) Make sure tests can be performed even if the corresponding algorithms
4247 have been removed entirely. This was also the last step to make
4248 OpenSSL compilable with DJGPP under all reasonable conditions.
4249 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
4251 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
4252 to allow version independent disabling of normally unselected ciphers,
4253 which may be activated as a side-effect of selecting a single cipher.
4255 (E.g., cipher list string "RSA" enables ciphersuites that are left
4256 out of "ALL" because they do not provide symmetric encryption.
4257 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
4258 [Lutz Jaenicke, Bodo Moeller]
4260 *) Add appropriate support for separate platform-dependent build
4261 directories. The recommended way to make a platform-dependent
4262 build directory is the following (tested on Linux), maybe with
4265 # Place yourself outside of the OpenSSL source tree. In
4266 # this example, the environment variable OPENSSL_SOURCE
4267 # is assumed to contain the absolute OpenSSL source directory.
4268 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
4269 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
4270 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
4271 mkdir -p `dirname $F`
4272 ln -s $OPENSSL_SOURCE/$F $F
4275 To be absolutely sure not to disturb the source tree, a "make clean"
4276 is a good thing. If it isn't successfull, don't worry about it,
4277 it probably means the source directory is very clean.
4280 *) Make sure any ENGINE control commands make local copies of string
4281 pointers passed to them whenever necessary. Otherwise it is possible
4282 the caller may have overwritten (or deallocated) the original string
4283 data when a later ENGINE operation tries to use the stored values.
4284 [Götz Babin-Ebell <babinebell@trustcenter.de>]
4286 *) Improve diagnostics in file reading and command-line digests.
4287 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
4289 *) Add AES modes CFB and OFB to the object database. Correct an
4290 error in AES-CFB decryption.
4293 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
4294 allows existing EVP_CIPHER_CTX structures to be reused after
4295 calling EVP_*Final(). This behaviour is used by encryption
4296 BIOs and some applications. This has the side effect that
4297 applications must explicitly clean up cipher contexts with
4298 EVP_CIPHER_CTX_cleanup() or they will leak memory.
4301 *) Check the values of dna and dnb in bn_mul_recursive before calling
4302 bn_mul_comba (a non zero value means the a or b arrays do not contain
4303 n2 elements) and fallback to bn_mul_normal if either is not zero.
4306 *) Fix escaping of non-ASCII characters when using the -subj option
4307 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
4310 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
4311 form for "surname", serialNumber has no short form.
4312 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
4313 therefore remove "mail" short name for "internet 7".
4314 The OID for unique identifiers in X509 certificates is
4315 x500UniqueIdentifier, not uniqueIdentifier.
4316 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
4319 *) Add an "init" command to the ENGINE config module and auto initialize
4320 ENGINEs. Without any "init" command the ENGINE will be initialized
4321 after all ctrl commands have been executed on it. If init=1 the
4322 ENGINE is initailized at that point (ctrls before that point are run
4323 on the uninitialized ENGINE and after on the initialized one). If
4324 init=0 then the ENGINE will not be iniatialized at all.
4327 *) Fix the 'app_verify_callback' interface so that the user-defined
4328 argument is actually passed to the callback: In the
4329 SSL_CTX_set_cert_verify_callback() prototype, the callback
4330 declaration has been changed from
4333 int (*cb)(X509_STORE_CTX *,void *);
4334 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
4335 i=s->ctx->app_verify_callback(&ctx)
4336 has been changed into
4337 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
4339 To update applications using SSL_CTX_set_cert_verify_callback(),
4340 a dummy argument can be added to their callback functions.
4341 [D. K. Smetters <smetters@parc.xerox.com>]
4343 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
4344 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
4346 *) Add and OPENSSL_LOAD_CONF define which will cause
4347 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
4348 This allows older applications to transparently support certain
4349 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
4350 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
4351 load the config file and OPENSSL_add_all_algorithms_conf() which will
4352 always load it have also been added.
4355 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
4356 Adjust NIDs and EVP layer.
4357 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4359 *) Config modules support in openssl utility.
4361 Most commands now load modules from the config file,
4362 though in a few (such as version) this isn't done
4363 because it couldn't be used for anything.
4365 In the case of ca and req the config file used is
4366 the same as the utility itself: that is the -config
4367 command line option can be used to specify an
4371 *) Move default behaviour from OPENSSL_config(). If appname is NULL
4372 use "openssl_conf" if filename is NULL use default openssl config file.
4375 *) Add an argument to OPENSSL_config() to allow the use of an alternative
4376 config section name. Add a new flag to tolerate a missing config file
4377 and move code to CONF_modules_load_file().
4380 *) Support for crypto accelerator cards from Accelerated Encryption
4381 Processing, www.aep.ie. (Use engine 'aep')
4382 The support was copied from 0.9.6c [engine] and adapted/corrected
4383 to work with the new engine framework.
4384 [AEP Inc. and Richard Levitte]
4386 *) Support for SureWare crypto accelerator cards from Baltimore
4387 Technologies. (Use engine 'sureware')
4388 The support was copied from 0.9.6c [engine] and adapted
4389 to work with the new engine framework.
4392 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
4393 make the newer ENGINE framework commands for the CHIL engine work.
4394 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
4396 *) Make it possible to produce shared libraries on ReliantUNIX.
4397 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
4399 *) Add the configuration target debug-linux-ppro.
4400 Make 'openssl rsa' use the general key loading routines
4401 implemented in apps.c, and make those routines able to
4402 handle the key format FORMAT_NETSCAPE and the variant
4404 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4406 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4407 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4409 *) Add -keyform to rsautl, and document -engine.
4410 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
4412 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
4413 BIO_R_NO_SUCH_FILE error code rather than the generic
4414 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
4417 *) Add new functions
4419 ERR_peek_last_error_line
4420 ERR_peek_last_error_line_data.
4421 These are similar to
4424 ERR_peek_error_line_data,
4425 but report on the latest error recorded rather than the first one
4426 still in the error queue.
4427 [Ben Laurie, Bodo Moeller]
4429 *) default_algorithms option in ENGINE config module. This allows things
4431 default_algorithms = ALL
4432 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
4435 *) Prelminary ENGINE config module.
4438 *) New experimental application configuration code.
4441 *) Change the AES code to follow the same name structure as all other
4442 symmetric ciphers, and behave the same way. Move everything to
4443 the directory crypto/aes, thereby obsoleting crypto/rijndael.
4444 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4446 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
4447 [Ben Laurie and Theo de Raadt]
4449 *) Add option to output public keys in req command.
4450 [Massimiliano Pala madwolf@openca.org]
4452 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
4453 (up to about 10% better than before for P-192 and P-224).
4456 *) New functions/macros
4458 SSL_CTX_set_msg_callback(ctx, cb)
4459 SSL_CTX_set_msg_callback_arg(ctx, arg)
4460 SSL_set_msg_callback(ssl, cb)
4461 SSL_set_msg_callback_arg(ssl, arg)
4463 to request calling a callback function
4465 void cb(int write_p, int version, int content_type,
4466 const void *buf, size_t len, SSL *ssl, void *arg)
4468 whenever a protocol message has been completely received
4469 (write_p == 0) or sent (write_p == 1). Here 'version' is the
4470 protocol version according to which the SSL library interprets
4471 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
4472 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
4473 the content type as defined in the SSL 3.0/TLS 1.0 protocol
4474 specification (change_cipher_spec(20), alert(21), handshake(22)).
4475 'buf' and 'len' point to the actual message, 'ssl' to the
4476 SSL object, and 'arg' is the application-defined value set by
4477 SSL[_CTX]_set_msg_callback_arg().
4479 'openssl s_client' and 'openssl s_server' have new '-msg' options
4480 to enable a callback that displays all protocol messages.
4483 *) Change the shared library support so shared libraries are built as
4484 soon as the corresponding static library is finished, and thereby get
4485 openssl and the test programs linked against the shared library.
4486 This still only happens when the keyword "shard" has been given to
4487 the configuration scripts.
4489 NOTE: shared library support is still an experimental thing, and
4490 backward binary compatibility is still not guaranteed.
4491 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
4493 *) Add support for Subject Information Access extension.
4494 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4496 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
4497 additional bytes when new memory had to be allocated, not just
4498 when reusing an existing buffer.
4501 *) New command line and configuration option 'utf8' for the req command.
4502 This allows field values to be specified as UTF8 strings.
4505 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
4506 runs for the former and machine-readable output for the latter.
4509 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
4510 of the e-mail address in the DN (i.e., it will go into a certificate
4511 extension only). The new configuration file option 'email_in_dn = no'
4512 has the same effect.
4513 [Massimiliano Pala madwolf@openca.org]
4515 *) Change all functions with names starting with des_ to be starting
4516 with DES_ instead. Add wrappers that are compatible with libdes,
4517 but are named _ossl_old_des_*. Finally, add macros that map the
4518 des_* symbols to the corresponding _ossl_old_des_* if libdes
4519 compatibility is desired. If OpenSSL 0.9.6c compatibility is
4520 desired, the des_* symbols will be mapped to DES_*, with one
4523 Since we provide two compatibility mappings, the user needs to
4524 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
4525 compatibility is desired. The default (i.e., when that macro
4526 isn't defined) is OpenSSL 0.9.6c compatibility.
4528 There are also macros that enable and disable the support of old
4529 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
4530 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
4531 are defined, the default will apply: to support the old des routines.
4533 In either case, one must include openssl/des.h to get the correct
4534 definitions. Do not try to just include openssl/des_old.h, that
4537 NOTE: This is a major break of an old API into a new one. Software
4538 authors are encouraged to switch to the DES_ style functions. Some
4539 time in the future, des_old.h and the libdes compatibility functions
4540 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
4541 default), and then completely removed.
4544 *) Test for certificates which contain unsupported critical extensions.
4545 If such a certificate is found during a verify operation it is
4546 rejected by default: this behaviour can be overridden by either
4547 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
4548 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
4549 X509_supported_extension() has also been added which returns 1 if a
4550 particular extension is supported.
4553 *) Modify the behaviour of EVP cipher functions in similar way to digests
4554 to retain compatibility with existing code.
4557 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
4558 compatibility with existing code. In particular the 'ctx' parameter does
4559 not have to be to be initialized before the call to EVP_DigestInit() and
4560 it is tidied up after a call to EVP_DigestFinal(). New function
4561 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
4562 EVP_MD_CTX_copy() changed to not require the destination to be
4563 initialized valid and new function EVP_MD_CTX_copy_ex() added which
4564 requires the destination to be valid.
4566 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
4567 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
4570 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
4571 so that complete 'Handshake' protocol structures are kept in memory
4572 instead of overwriting 'msg_type' and 'length' with 'body' data.
4575 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
4576 [Massimo Santin via Richard Levitte]
4578 *) Major restructuring to the underlying ENGINE code. This includes
4579 reduction of linker bloat, separation of pure "ENGINE" manipulation
4580 (initialisation, etc) from functionality dealing with implementations
4581 of specific crypto iterfaces. This change also introduces integrated
4582 support for symmetric ciphers and digest implementations - so ENGINEs
4583 can now accelerate these by providing EVP_CIPHER and EVP_MD
4584 implementations of their own. This is detailed in crypto/engine/README
4585 as it couldn't be adequately described here. However, there are a few
4586 API changes worth noting - some RSA, DSA, DH, and RAND functions that
4587 were changed in the original introduction of ENGINE code have now
4588 reverted back - the hooking from this code to ENGINE is now a good
4589 deal more passive and at run-time, operations deal directly with
4590 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
4591 dereferencing through an ENGINE pointer any more. Also, the ENGINE
4592 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
4593 they were not being used by the framework as there is no concept of a
4594 BIGNUM_METHOD and they could not be generalised to the new
4595 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
4596 ENGINE_cpy() has been removed as it cannot be consistently defined in
4600 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
4603 *) Change mkdef.pl to sort symbols that get the same entry number,
4604 and make sure the automatically generated functions ERR_load_*
4605 become part of libeay.num as well.
4608 *) New function SSL_renegotiate_pending(). This returns true once
4609 renegotiation has been requested (either SSL_renegotiate() call
4610 or HelloRequest/ClientHello receveived from the peer) and becomes
4611 false once a handshake has been completed.
4612 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
4613 sends a HelloRequest, but does not ensure that a handshake takes
4614 place. SSL_renegotiate_pending() is useful for checking if the
4615 client has followed the request.)
4618 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
4619 By default, clients may request session resumption even during
4620 renegotiation (if session ID contexts permit); with this option,
4621 session resumption is possible only in the first handshake.
4623 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4624 more bits available for options that should not be part of
4625 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4628 *) Add some demos for certificate and certificate request creation.
4631 *) Make maximum certificate chain size accepted from the peer application
4632 settable (SSL*_get/set_max_cert_list()), as proposed by
4633 "Douglas E. Engert" <deengert@anl.gov>.
4636 *) Add support for shared libraries for Unixware-7
4637 (Boyd Lynn Gerber <gerberb@zenez.com>).
4640 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4641 be done prior to destruction. Use this to unload error strings from
4642 ENGINEs that load their own error strings. NB: This adds two new API
4643 functions to "get" and "set" this destroy handler in an ENGINE.
4646 *) Alter all existing ENGINE implementations (except "openssl" and
4647 "openbsd") to dynamically instantiate their own error strings. This
4648 makes them more flexible to be built both as statically-linked ENGINEs
4649 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4650 Also, add stub code to each that makes building them as self-contained
4651 shared-libraries easier (see README.ENGINE).
4654 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4655 implementations into applications that are completely implemented in
4656 self-contained shared-libraries. The "dynamic" ENGINE exposes control
4657 commands that can be used to configure what shared-library to load and
4658 to control aspects of the way it is handled. Also, made an update to
4659 the README.ENGINE file that brings its information up-to-date and
4660 provides some information and instructions on the "dynamic" ENGINE
4661 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
4664 *) Make it possible to unload ranges of ERR strings with a new
4665 "ERR_unload_strings" function.
4668 *) Add a copy() function to EVP_MD.
4671 *) Make EVP_MD routines take a context pointer instead of just the
4672 md_data void pointer.
4675 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
4676 that the digest can only process a single chunk of data
4677 (typically because it is provided by a piece of
4678 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
4679 is only going to provide a single chunk of data, and hence the
4680 framework needn't accumulate the data for oneshot drivers.
4683 *) As with "ERR", make it possible to replace the underlying "ex_data"
4684 functions. This change also alters the storage and management of global
4685 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
4686 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
4687 index counters. The API functions that use this state have been changed
4688 to take a "class_index" rather than pointers to the class's local STACK
4689 and counter, and there is now an API function to dynamically create new
4690 classes. This centralisation allows us to (a) plug a lot of the
4691 thread-safety problems that existed, and (b) makes it possible to clean
4692 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
4693 such data would previously have always leaked in application code and
4694 workarounds were in place to make the memory debugging turn a blind eye
4695 to it. Application code that doesn't use this new function will still
4696 leak as before, but their memory debugging output will announce it now
4697 rather than letting it slide.
4699 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
4700 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
4701 has a return value to indicate success or failure.
4704 *) Make it possible to replace the underlying "ERR" functions such that the
4705 global state (2 LHASH tables and 2 locks) is only used by the "default"
4706 implementation. This change also adds two functions to "get" and "set"
4707 the implementation prior to it being automatically set the first time
4708 any other ERR function takes place. Ie. an application can call "get",
4709 pass the return value to a module it has just loaded, and that module
4710 can call its own "set" function using that value. This means the
4711 module's "ERR" operations will use (and modify) the error state in the
4712 application and not in its own statically linked copy of OpenSSL code.
4715 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
4716 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
4717 the operation, and provides a more encapsulated way for external code
4718 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
4719 to use these functions rather than manually incrementing the counts.
4721 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
4724 *) Add EVP test program.
4727 *) Add symmetric cipher support to ENGINE. Expect the API to change!
4730 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
4731 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
4732 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
4733 These allow a CRL to be built without having to access X509_CRL fields
4734 directly. Modify 'ca' application to use new functions.
4737 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
4738 bug workarounds. Rollback attack detection is a security feature.
4739 The problem will only arise on OpenSSL servers when TLSv1 is not
4740 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
4741 Software authors not wanting to support TLSv1 will have special reasons
4742 for their choice and can explicitly enable this option.
4743 [Bodo Moeller, Lutz Jaenicke]
4745 *) Rationalise EVP so it can be extended: don't include a union of
4746 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
4747 (similar to those existing for EVP_CIPHER_CTX).
4752 EVP_MD_CTX_init(&md); /* new function call */
4753 EVP_DigestInit(&md, EVP_sha1());
4754 EVP_DigestUpdate(&md, in, len);
4755 EVP_DigestFinal(&md, out, NULL);
4756 EVP_MD_CTX_cleanup(&md); /* new function call */
4760 *) Make DES key schedule conform to the usual scheme, as well as
4761 correcting its structure. This means that calls to DES functions
4762 now have to pass a pointer to a des_key_schedule instead of a
4763 plain des_key_schedule (which was actually always a pointer
4766 des_key_schedule ks;
4768 des_set_key_checked(..., &ks);
4769 des_ncbc_encrypt(..., &ks, ...);
4771 (Note that a later change renames 'des_...' into 'DES_...'.)
4774 *) Initial reduction of linker bloat: the use of some functions, such as
4775 PEM causes large amounts of unused functions to be linked in due to
4776 poor organisation. For example pem_all.c contains every PEM function
4777 which has a knock on effect of linking in large amounts of (unused)
4778 ASN1 code. Grouping together similar functions and splitting unrelated
4779 functions prevents this.
4782 *) Cleanup of EVP macros.
4785 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
4786 correct _ecb suffix.
4789 *) Add initial OCSP responder support to ocsp application. The
4790 revocation information is handled using the text based index
4791 use by the ca application. The responder can either handle
4792 requests generated internally, supplied in files (for example
4793 via a CGI script) or using an internal minimal server.
4796 *) Add configuration choices to get zlib compression for TLS.
4799 *) Changes to Kerberos SSL for RFC 2712 compliance:
4800 1. Implemented real KerberosWrapper, instead of just using
4801 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
4802 2. Implemented optional authenticator field of KerberosWrapper.
4804 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
4805 and authenticator structs; see crypto/krb5/.
4807 Generalized Kerberos calls to support multiple Kerberos libraries.
4808 [Vern Staats <staatsvr@asc.hpc.mil>,
4809 Jeffrey Altman <jaltman@columbia.edu>
4810 via Richard Levitte]
4812 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
4813 already does with RSA. testdsa.h now has 'priv_key/pub_key'
4814 values for each of the key sizes rather than having just
4815 parameters (and 'speed' generating keys each time).
4818 *) Speed up EVP routines.
4821 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
4822 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
4823 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
4824 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
4826 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
4827 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
4828 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
4831 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
4833 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
4836 *) Added the OS2-EMX target.
4837 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
4839 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
4840 to support NCONF routines in extension code. New function CONF_set_nconf()
4841 to allow functions which take an NCONF to also handle the old LHASH
4842 structure: this means that the old CONF compatible routines can be
4843 retained (in particular wrt extensions) without having to duplicate the
4844 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
4847 *) Enhance the general user interface with mechanisms for inner control
4848 and with possibilities to have yes/no kind of prompts.
4851 *) Change all calls to low level digest routines in the library and
4852 applications to use EVP. Add missing calls to HMAC_cleanup() and
4853 don't assume HMAC_CTX can be copied using memcpy().
4854 [Verdon Walker <VWalker@novell.com>, Steve Henson]
4856 *) Add the possibility to control engines through control names but with
4857 arbitrary arguments instead of just a string.
4858 Change the key loaders to take a UI_METHOD instead of a callback
4859 function pointer. NOTE: this breaks binary compatibility with earlier
4860 versions of OpenSSL [engine].
4861 Adapt the nCipher code for these new conditions and add a card insertion
4865 *) Enhance the general user interface with mechanisms to better support
4866 dialog box interfaces, application-defined prompts, the possibility
4867 to use defaults (for example default passwords from somewhere else)
4868 and interrupts/cancellations.
4871 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
4872 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
4875 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
4876 tidy up some unnecessarily weird code in 'sk_new()').
4877 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
4879 *) Change the key loading routines for ENGINEs to use the same kind
4880 callback (pem_password_cb) as all other routines that need this
4884 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
4885 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
4886 than this minimum value is recommended.
4889 *) New random seeder for OpenVMS, using the system process statistics
4890 that are easily reachable.
4893 *) Windows apparently can't transparently handle global
4894 variables defined in DLLs. Initialisations such as:
4896 const ASN1_ITEM *it = &ASN1_INTEGER_it;
4898 wont compile. This is used by the any applications that need to
4899 declare their own ASN1 modules. This was fixed by adding the option
4900 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
4901 needed for static libraries under Win32.
4904 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
4905 setting of purpose and trust fields. New X509_STORE trust and
4906 purpose functions and tidy up setting in other SSL functions.
4909 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
4910 structure. These are inherited by X509_STORE_CTX when it is
4911 initialised. This allows various defaults to be set in the
4912 X509_STORE structure (such as flags for CRL checking and custom
4913 purpose or trust settings) for functions which only use X509_STORE_CTX
4914 internally such as S/MIME.
4916 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
4917 trust settings if they are not set in X509_STORE. This allows X509_STORE
4918 purposes and trust (in S/MIME for example) to override any set by default.
4920 Add command line options for CRL checking to smime, s_client and s_server
4924 *) Initial CRL based revocation checking. If the CRL checking flag(s)
4925 are set then the CRL is looked up in the X509_STORE structure and
4926 its validity and signature checked, then if the certificate is found
4927 in the CRL the verify fails with a revoked error.
4929 Various new CRL related callbacks added to X509_STORE_CTX structure.
4931 Command line options added to 'verify' application to support this.
4933 This needs some additional work, such as being able to handle multiple
4934 CRLs with different times, extension based lookup (rather than just
4935 by subject name) and ultimately more complete V2 CRL extension
4939 *) Add a general user interface API (crypto/ui/). This is designed
4940 to replace things like des_read_password and friends (backward
4941 compatibility functions using this new API are provided).
4942 The purpose is to remove prompting functions from the DES code
4943 section as well as provide for prompting through dialog boxes in
4944 a window system and the like.
4947 *) Add "ex_data" support to ENGINE so implementations can add state at a
4948 per-structure level rather than having to store it globally.
4951 *) Make it possible for ENGINE structures to be copied when retrieved by
4952 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
4953 This causes the "original" ENGINE structure to act like a template,
4954 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
4955 operational state can be localised to each ENGINE structure, despite the
4956 fact they all share the same "methods". New ENGINE structures returned in
4957 this case have no functional references and the return value is the single
4958 structural reference. This matches the single structural reference returned
4959 by ENGINE_by_id() normally, when it is incremented on the pre-existing
4963 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
4964 needs to match any other type at all we need to manually clear the
4968 *) Changes to the "openssl engine" utility to include;
4969 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
4970 about an ENGINE's available control commands.
4971 - executing control commands from command line arguments using the
4972 '-pre' and '-post' switches. '-post' is only used if '-t' is
4973 specified and the ENGINE is successfully initialised. The syntax for
4974 the individual commands are colon-separated, for example;
4975 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
4978 *) New dynamic control command support for ENGINEs. ENGINEs can now
4979 declare their own commands (numbers), names (strings), descriptions,
4980 and input types for run-time discovery by calling applications. A
4981 subset of these commands are implicitly classed as "executable"
4982 depending on their input type, and only these can be invoked through
4983 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
4984 can be based on user input, config files, etc). The distinction is
4985 that "executable" commands cannot return anything other than a boolean
4986 result and can only support numeric or string input, whereas some
4987 discoverable commands may only be for direct use through
4988 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
4989 pointers, or other custom uses. The "executable" commands are to
4990 support parameterisations of ENGINE behaviour that can be
4991 unambiguously defined by ENGINEs and used consistently across any
4992 OpenSSL-based application. Commands have been added to all the
4993 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
4994 control over shared-library paths without source code alterations.
4997 *) Changed all ENGINE implementations to dynamically allocate their
4998 ENGINEs rather than declaring them statically. Apart from this being
4999 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
5000 this also allows the implementations to compile without using the
5001 internal engine_int.h header.
5004 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
5005 'const' value. Any code that should be able to modify a RAND_METHOD
5006 should already have non-const pointers to it (ie. they should only
5007 modify their own ones).
5010 *) Made a variety of little tweaks to the ENGINE code.
5011 - "atalla" and "ubsec" string definitions were moved from header files
5012 to C code. "nuron" string definitions were placed in variables
5013 rather than hard-coded - allowing parameterisation of these values
5014 later on via ctrl() commands.
5015 - Removed unused "#if 0"'d code.
5016 - Fixed engine list iteration code so it uses ENGINE_free() to release
5017 structural references.
5018 - Constified the RAND_METHOD element of ENGINE structures.
5019 - Constified various get/set functions as appropriate and added
5020 missing functions (including a catch-all ENGINE_cpy that duplicates
5021 all ENGINE values onto a new ENGINE except reference counts/state).
5022 - Removed NULL parameter checks in get/set functions. Setting a method
5023 or function to NULL is a way of cancelling out a previously set
5024 value. Passing a NULL ENGINE parameter is just plain stupid anyway
5025 and doesn't justify the extra error symbols and code.
5026 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
5027 flags from engine_int.h to engine.h.
5028 - Changed prototypes for ENGINE handler functions (init(), finish(),
5029 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
5032 *) Implement binary inversion algorithm for BN_mod_inverse in addition
5033 to the algorithm using long division. The binary algorithm can be
5034 used only if the modulus is odd. On 32-bit systems, it is faster
5035 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
5036 roughly 5-15% for 256-bit moduli), so we use it only for moduli
5037 up to 450 bits. In 64-bit environments, the binary algorithm
5038 appears to be advantageous for much longer moduli; here we use it
5039 for moduli up to 2048 bits.
5042 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
5043 could not support the combine flag in choice fields.
5046 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
5047 extensions from a certificate request to the certificate.
5050 *) Allow multiple 'certopt' and 'nameopt' options to be separated
5051 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
5052 file: this allows the display of the certificate about to be
5053 signed to be customised, to allow certain fields to be included
5054 or excluded and extension details. The old system didn't display
5055 multicharacter strings properly, omitted fields not in the policy
5056 and couldn't display additional details such as extensions.
5059 *) Function EC_POINTs_mul for multiple scalar multiplication
5060 of an arbitrary number of elliptic curve points
5061 \sum scalars[i]*points[i],
5062 optionally including the generator defined for the EC_GROUP:
5063 scalar*generator + \sum scalars[i]*points[i].
5065 EC_POINT_mul is a simple wrapper function for the typical case
5066 that the point list has just one item (besides the optional
5070 *) First EC_METHODs for curves over GF(p):
5072 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
5073 operations and provides various method functions that can also
5074 operate with faster implementations of modular arithmetic.
5076 EC_GFp_mont_method() reuses most functions that are part of
5077 EC_GFp_simple_method, but uses Montgomery arithmetic.
5079 [Bodo Moeller; point addition and point doubling
5080 implementation directly derived from source code provided by
5081 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
5083 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
5084 crypto/ec/ec_lib.c):
5086 Curves are EC_GROUP objects (with an optional group generator)
5087 based on EC_METHODs that are built into the library.
5089 Points are EC_POINT objects based on EC_GROUP objects.
5091 Most of the framework would be able to handle curves over arbitrary
5092 finite fields, but as there are no obvious types for fields other
5093 than GF(p), some functions are limited to that for now.
5096 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
5097 that the file contains a complete HTTP response.
5100 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
5101 change the def and num file printf format specifier from "%-40sXXX"
5102 to "%-39s XXX". The latter will always guarantee a space after the
5103 field while the former will cause them to run together if the field
5104 is 40 of more characters long.
5107 *) Constify the cipher and digest 'method' functions and structures
5108 and modify related functions to take constant EVP_MD and EVP_CIPHER
5112 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
5113 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
5116 *) Modify EVP_Digest*() routines so they now return values. Although the
5117 internal software routines can never fail additional hardware versions
5121 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
5123 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
5124 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
5127 ERR_R_NESTED_ASN1_ERROR
5129 ERR_R_MISSING_ASN1_EOS
5130 were 4 .. 9, conflicting with
5131 ERR_LIB_RSA (= ERR_R_RSA_LIB)
5133 ERR_LIB_PEM (= ERR_R_PEM_LIB).
5134 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
5136 Add new error code 'ERR_R_INTERNAL_ERROR'.
5139 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
5143 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
5144 sets the subject name for a new request or supersedes the
5145 subject name in a given request. Formats that can be parsed are
5146 'CN=Some Name, OU=myOU, C=IT'
5148 'CN=Some Name/OU=myOU/C=IT'.
5150 Add options '-batch' and '-verbose' to 'openssl req'.
5151 [Massimiliano Pala <madwolf@hackmasters.net>]
5153 *) Introduce the possibility to access global variables through
5154 functions on platform were that's the best way to handle exporting
5155 global variables in shared libraries. To enable this functionality,
5156 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
5157 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
5158 is normally done by Configure or something similar).
5160 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
5161 in the source file (foo.c) like this:
5163 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
5164 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
5166 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
5167 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
5169 OPENSSL_DECLARE_GLOBAL(int,foo);
5170 #define foo OPENSSL_GLOBAL_REF(foo)
5171 OPENSSL_DECLARE_GLOBAL(double,bar);
5172 #define bar OPENSSL_GLOBAL_REF(bar)
5174 The #defines are very important, and therefore so is including the
5175 header file everywhere where the defined globals are used.
5177 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
5178 of ASN.1 items, but that structure is a bit different.
5180 The largest change is in util/mkdef.pl which has been enhanced with
5181 better and easier to understand logic to choose which symbols should
5182 go into the Windows .def files as well as a number of fixes and code
5183 cleanup (among others, algorithm keywords are now sorted
5184 lexicographically to avoid constant rewrites).
5187 *) In BN_div() keep a copy of the sign of 'num' before writing the
5188 result to 'rm' because if rm==num the value will be overwritten
5189 and produce the wrong result if 'num' is negative: this caused
5190 problems with BN_mod() and BN_nnmod().
5193 *) Function OCSP_request_verify(). This checks the signature on an
5194 OCSP request and verifies the signer certificate. The signer
5195 certificate is just checked for a generic purpose and OCSP request
5199 *) Add OCSP_check_validity() function to check the validity of OCSP
5200 responses. OCSP responses are prepared in real time and may only
5201 be a few seconds old. Simply checking that the current time lies
5202 between thisUpdate and nextUpdate max reject otherwise valid responses
5203 caused by either OCSP responder or client clock inaccuracy. Instead
5204 we allow thisUpdate and nextUpdate to fall within a certain period of
5205 the current time. The age of the response can also optionally be
5206 checked. Two new options -validity_period and -status_age added to
5210 *) If signature or public key algorithm is unrecognized print out its
5211 OID rather that just UNKNOWN.
5214 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
5215 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
5216 ID to be generated from the issuer certificate alone which can then be
5217 passed to OCSP_id_issuer_cmp().
5220 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
5221 ASN1 modules to export functions returning ASN1_ITEM pointers
5222 instead of the ASN1_ITEM structures themselves. This adds several
5223 new macros which allow the underlying ASN1 function/structure to
5224 be accessed transparently. As a result code should not use ASN1_ITEM
5225 references directly (such as &X509_it) but instead use the relevant
5226 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
5227 use of the new ASN1 code on platforms where exporting structures
5228 is problematical (for example in shared libraries) but exporting
5229 functions returning pointers to structures is not.
5232 *) Add support for overriding the generation of SSL/TLS session IDs.
5233 These callbacks can be registered either in an SSL_CTX or per SSL.
5234 The purpose of this is to allow applications to control, if they wish,
5235 the arbitrary values chosen for use as session IDs, particularly as it
5236 can be useful for session caching in multiple-server environments. A
5237 command-line switch for testing this (and any client code that wishes
5238 to use such a feature) has been added to "s_server".
5239 [Geoff Thorpe, Lutz Jaenicke]
5241 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
5242 of the form '#if defined(...) || defined(...) || ...' and
5243 '#if !defined(...) && !defined(...) && ...'. This also avoids
5244 the growing number of special cases it was previously handling.
5247 *) Make all configuration macros available for application by making
5248 sure they are available in opensslconf.h, by giving them names starting
5249 with "OPENSSL_" to avoid conflicts with other packages and by making
5250 sure e_os2.h will cover all platform-specific cases together with
5252 Additionally, it is now possible to define configuration/platform-
5253 specific names (called "system identities"). In the C code, these
5254 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
5255 macro with the name beginning with "OPENSSL_SYS_", which is determined
5256 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
5260 *) New option -set_serial to 'req' and 'x509' this allows the serial
5261 number to use to be specified on the command line. Previously self
5262 signed certificates were hard coded with serial number 0 and the
5263 CA options of 'x509' had to use a serial number in a file which was
5267 *) New options to 'ca' utility to support V2 CRL entry extensions.
5268 Currently CRL reason, invalidity date and hold instruction are
5269 supported. Add new CRL extensions to V3 code and some new objects.
5272 *) New function EVP_CIPHER_CTX_set_padding() this is used to
5273 disable standard block padding (aka PKCS#5 padding) in the EVP
5274 API, which was previously mandatory. This means that the data is
5275 not padded in any way and so the total length much be a multiple
5276 of the block size, otherwise an error occurs.
5279 *) Initial (incomplete) OCSP SSL support.
5282 *) New function OCSP_parse_url(). This splits up a URL into its host,
5283 port and path components: primarily to parse OCSP URLs. New -url
5284 option to ocsp utility.
5287 *) New nonce behavior. The return value of OCSP_check_nonce() now
5288 reflects the various checks performed. Applications can decide
5289 whether to tolerate certain situations such as an absent nonce
5290 in a response when one was present in a request: the ocsp application
5291 just prints out a warning. New function OCSP_add1_basic_nonce()
5292 this is to allow responders to include a nonce in a response even if
5293 the request is nonce-less.
5296 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
5297 skipped when using openssl x509 multiple times on a single input file,
5298 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
5301 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
5302 set string type: to handle setting ASN1_TIME structures. Fix ca
5303 utility to correctly initialize revocation date of CRLs.
5306 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
5307 the clients preferred ciphersuites and rather use its own preferences.
5308 Should help to work around M$ SGC (Server Gated Cryptography) bug in
5309 Internet Explorer by ensuring unchanged hash method during stepup.
5310 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
5313 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
5314 to aes and add a new 'exist' option to print out symbols that don't
5318 *) Additional options to ocsp utility to allow flags to be set and
5319 additional certificates supplied.
5322 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
5323 OCSP client a number of certificate to only verify the response
5327 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
5328 handle the new API. Currently only ECB, CBC modes supported. Add new
5331 Add TLS AES ciphersuites as described in RFC3268, "Advanced
5332 Encryption Standard (AES) Ciphersuites for Transport Layer
5333 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
5334 not enabled by default and were not part of the "ALL" ciphersuite
5335 alias because they were not yet official; they could be
5336 explicitly requested by specifying the "AESdraft" ciphersuite
5337 group alias. In the final release of OpenSSL 0.9.7, the group
5338 alias is called "AES" and is part of "ALL".)
5339 [Ben Laurie, Steve Henson, Bodo Moeller]
5341 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
5342 request to response.
5345 *) Functions for OCSP responders. OCSP_request_onereq_count(),
5346 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
5347 extract information from a certificate request. OCSP_response_create()
5348 creates a response and optionally adds a basic response structure.
5349 OCSP_basic_add1_status() adds a complete single response to a basic
5350 response and returns the OCSP_SINGLERESP structure just added (to allow
5351 extensions to be included for example). OCSP_basic_add1_cert() adds a
5352 certificate to a basic response and OCSP_basic_sign() signs a basic
5353 response with various flags. New helper functions ASN1_TIME_check()
5354 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
5355 (converts ASN1_TIME to GeneralizedTime).
5358 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
5359 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
5360 structure from a certificate. X509_pubkey_digest() digests the public_key
5361 contents: this is used in various key identifiers.
5364 *) Make sk_sort() tolerate a NULL argument.
5365 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
5367 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
5368 passed by the function are trusted implicitly. If any of them signed the
5369 response then it is assumed to be valid and is not verified.
5372 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
5373 to data. This was previously part of the PKCS7 ASN1 code. This
5374 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
5375 [Steve Henson, reported by Kenneth R. Robinette
5376 <support@securenetterm.com>]
5378 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
5379 routines: without these tracing memory leaks is very painful.
5380 Fix leaks in PKCS12 and PKCS7 routines.
5383 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
5384 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
5385 effectively meant GeneralizedTime would never be used. Now it
5386 is initialised to -1 but X509_time_adj() now has to check the value
5387 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
5388 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
5389 [Steve Henson, reported by Kenneth R. Robinette
5390 <support@securenetterm.com>]
5392 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
5393 result in a zero length in the ASN1_INTEGER structure which was
5394 not consistent with the structure when d2i_ASN1_INTEGER() was used
5395 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
5396 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
5397 where it did not print out a minus for negative ASN1_INTEGER.
5400 *) Add summary printout to ocsp utility. The various functions which
5401 convert status values to strings have been renamed to:
5402 OCSP_response_status_str(), OCSP_cert_status_str() and
5403 OCSP_crl_reason_str() and are no longer static. New options
5404 to verify nonce values and to disable verification. OCSP response
5405 printout format cleaned up.
5408 *) Add additional OCSP certificate checks. These are those specified
5409 in RFC2560. This consists of two separate checks: the CA of the
5410 certificate being checked must either be the OCSP signer certificate
5411 or the issuer of the OCSP signer certificate. In the latter case the
5412 OCSP signer certificate must contain the OCSP signing extended key
5413 usage. This check is performed by attempting to match the OCSP
5414 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
5415 in the OCSP_CERTID structures of the response.
5418 *) Initial OCSP certificate verification added to OCSP_basic_verify()
5419 and related routines. This uses the standard OpenSSL certificate
5420 verify routines to perform initial checks (just CA validity) and
5421 to obtain the certificate chain. Then additional checks will be
5422 performed on the chain. Currently the root CA is checked to see
5423 if it is explicitly trusted for OCSP signing. This is used to set
5424 a root CA as a global signing root: that is any certificate that
5425 chains to that CA is an acceptable OCSP signing certificate.
5428 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
5429 extensions from a separate configuration file.
5430 As when reading extensions from the main configuration file,
5431 the '-extensions ...' option may be used for specifying the
5433 [Massimiliano Pala <madwolf@comune.modena.it>]
5435 *) New OCSP utility. Allows OCSP requests to be generated or
5436 read. The request can be sent to a responder and the output
5437 parsed, outputed or printed in text form. Not complete yet:
5438 still needs to check the OCSP response validity.
5441 *) New subcommands for 'openssl ca':
5442 'openssl ca -status <serial>' prints the status of the cert with
5443 the given serial number (according to the index file).
5444 'openssl ca -updatedb' updates the expiry status of certificates
5446 [Massimiliano Pala <madwolf@comune.modena.it>]
5448 *) New '-newreq-nodes' command option to CA.pl. This is like
5449 '-newreq', but calls 'openssl req' with the '-nodes' option
5450 so that the resulting key is not encrypted.
5451 [Damien Miller <djm@mindrot.org>]
5453 *) New configuration for the GNU Hurd.
5454 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
5456 *) Initial code to implement OCSP basic response verify. This
5457 is currently incomplete. Currently just finds the signer's
5458 certificate and verifies the signature on the response.
5461 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
5462 value of OPENSSLDIR. This is available via the new '-d' option
5463 to 'openssl version', and is also included in 'openssl version -a'.
5466 *) Allowing defining memory allocation callbacks that will be given
5467 file name and line number information in additional arguments
5468 (a const char* and an int). The basic functionality remains, as
5469 well as the original possibility to just replace malloc(),
5470 realloc() and free() by functions that do not know about these
5471 additional arguments. To register and find out the current
5472 settings for extended allocation functions, the following
5473 functions are provided:
5475 CRYPTO_set_mem_ex_functions
5476 CRYPTO_set_locked_mem_ex_functions
5477 CRYPTO_get_mem_ex_functions
5478 CRYPTO_get_locked_mem_ex_functions
5480 These work the same way as CRYPTO_set_mem_functions and friends.
5481 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
5482 extended allocation function is enabled.
5483 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
5484 a conventional allocation function is enabled.
5485 [Richard Levitte, Bodo Moeller]
5487 *) Finish off removing the remaining LHASH function pointer casts.
5488 There should no longer be any prototype-casting required when using
5489 the LHASH abstraction, and any casts that remain are "bugs". See
5490 the callback types and macros at the head of lhash.h for details
5491 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
5494 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
5495 If /dev/[u]random devices are not available or do not return enough
5496 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
5498 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
5499 /etc/entropy will be queried once each in this sequence, quering stops
5500 when enough entropy was collected without querying more sockets.
5503 *) Change the Unix RAND_poll() variant to be able to poll several
5504 random devices, as specified by DEVRANDOM, until a sufficient amount
5505 of data has been collected. We spend at most 10 ms on each file
5506 (select timeout) and read in non-blocking mode. DEVRANDOM now
5507 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
5508 (previously it was just the string "/dev/urandom"), so on typical
5509 platforms the 10 ms delay will never occur.
5510 Also separate out the Unix variant to its own file, rand_unix.c.
5511 For VMS, there's a currently-empty rand_vms.c.
5514 *) Move OCSP client related routines to ocsp_cl.c. These
5515 provide utility functions which an application needing
5516 to issue a request to an OCSP responder and analyse the
5517 response will typically need: as opposed to those which an
5518 OCSP responder itself would need which will be added later.
5520 OCSP_request_sign() signs an OCSP request with an API similar
5521 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
5522 response. OCSP_response_get1_basic() extracts basic response
5523 from response. OCSP_resp_find_status(): finds and extracts status
5524 information from an OCSP_CERTID structure (which will be created
5525 when the request structure is built). These are built from lower
5526 level functions which work on OCSP_SINGLERESP structures but
5527 wont normally be used unless the application wishes to examine
5528 extensions in the OCSP response for example.
5530 Replace nonce routines with a pair of functions.
5531 OCSP_request_add1_nonce() adds a nonce value and optionally
5532 generates a random value. OCSP_check_nonce() checks the
5533 validity of the nonce in an OCSP response.
5536 *) Change function OCSP_request_add() to OCSP_request_add0_id().
5537 This doesn't copy the supplied OCSP_CERTID and avoids the
5538 need to free up the newly created id. Change return type
5539 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
5540 This can then be used to add extensions to the request.
5541 Deleted OCSP_request_new(), since most of its functionality
5542 is now in OCSP_REQUEST_new() (and the case insensitive name
5543 clash) apart from the ability to set the request name which
5544 will be added elsewhere.
5547 *) Update OCSP API. Remove obsolete extensions argument from
5548 various functions. Extensions are now handled using the new
5549 OCSP extension code. New simple OCSP HTTP function which
5550 can be used to send requests and parse the response.
5553 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
5554 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
5555 uses the special reorder version of SET OF to sort the attributes
5556 and reorder them to match the encoded order. This resolves a long
5557 standing problem: a verify on a PKCS7 structure just after signing
5558 it used to fail because the attribute order did not match the
5559 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
5560 it uses the received order. This is necessary to tolerate some broken
5561 software that does not order SET OF. This is handled by encoding
5562 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
5563 to produce the required SET OF.
5566 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
5567 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
5568 files to get correct declarations of the ASN.1 item variables.
5571 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
5572 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
5573 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
5574 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
5575 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
5576 ASN1_ITEM and no wrapper functions.
5579 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
5580 replace the old function pointer based I/O routines. Change most of
5581 the *_d2i_bio() and *_d2i_fp() functions to use these.
5584 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
5585 lines, recognice more "algorithms" that can be deselected, and make
5586 it complain about algorithm deselection that isn't recognised.
5589 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
5590 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
5591 to use new functions. Add NO_ASN1_OLD which can be set to remove
5592 some old style ASN1 functions: this can be used to determine if old
5593 code will still work when these eventually go away.
5596 *) New extension functions for OCSP structures, these follow the
5597 same conventions as certificates and CRLs.
5600 *) New function X509V3_add1_i2d(). This automatically encodes and
5601 adds an extension. Its behaviour can be customised with various
5602 flags to append, replace or delete. Various wrappers added for
5603 certifcates and CRLs.
5606 *) Fix to avoid calling the underlying ASN1 print routine when
5607 an extension cannot be parsed. Correct a typo in the
5608 OCSP_SERVICELOC extension. Tidy up print OCSP format.
5611 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
5612 entries for variables.
5615 *) Add functionality to apps/openssl.c for detecting locking
5616 problems: As the program is single-threaded, all we have
5617 to do is register a locking callback using an array for
5618 storing which locks are currently held by the program.
5621 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
5622 SSL_get_ex_data_X509_STORE_idx(), which is used in
5623 ssl_verify_cert_chain() and thus can be called at any time
5624 during TLS/SSL handshakes so that thread-safety is essential.
5625 Unfortunately, the ex_data design is not at all suited
5626 for multi-threaded use, so it probably should be abolished.
5629 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5630 [Broadcom, tweaked and integrated by Geoff Thorpe]
5632 *) Move common extension printing code to new function
5633 X509V3_print_extensions(). Reorganise OCSP print routines and
5634 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5637 *) New function X509_signature_print() to remove duplication in some
5641 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5642 set (this was treated exactly the same as SET OF previously). This
5643 is used to reorder the STACK representing the structure to match the
5644 encoding. This will be used to get round a problem where a PKCS7
5645 structure which was signed could not be verified because the STACK
5646 order did not reflect the encoded order.
5649 *) Reimplement the OCSP ASN1 module using the new code.
5652 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5653 for its ASN1 operations. The old style function pointers still exist
5654 for now but they will eventually go away.
5657 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
5658 completely replaces the old ASN1 functionality with a table driven
5659 encoder and decoder which interprets an ASN1_ITEM structure describing
5660 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
5661 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
5662 has also been converted to the new form.
5665 *) Change BN_mod_exp_recp so that negative moduli are tolerated
5666 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
5667 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
5668 for negative moduli.
5671 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
5672 of not touching the result's sign bit.
5675 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
5679 *) Changed the LHASH code to use prototypes for callbacks, and created
5680 macros to declare and implement thin (optionally static) functions
5681 that provide type-safety and avoid function pointer casting for the
5682 type-specific callbacks.
5685 *) Added Kerberos Cipher Suites to be used with TLS, as written in
5687 [Veers Staats <staatsvr@asc.hpc.mil>,
5688 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
5690 *) Reformat the FAQ so the different questions and answers can be divided
5691 in sections depending on the subject.
5694 *) Have the zlib compression code load ZLIB.DLL dynamically under
5698 *) New function BN_mod_sqrt for computing square roots modulo a prime
5699 (using the probabilistic Tonelli-Shanks algorithm unless
5700 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
5701 be handled deterministically).
5702 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5704 *) Make BN_mod_inverse faster by explicitly handling small quotients
5705 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
5706 512 bits], about 30% for larger ones [1024 or 2048 bits].)
5709 *) New function BN_kronecker.
5712 *) Fix BN_gcd so that it works on negative inputs; the result is
5713 positive unless both parameters are zero.
5714 Previously something reasonably close to an infinite loop was
5715 possible because numbers could be growing instead of shrinking
5716 in the implementation of Euclid's algorithm.
5719 *) Fix BN_is_word() and BN_is_one() macros to take into account the
5720 sign of the number in question.
5722 Fix BN_is_word(a,w) to work correctly for w == 0.
5724 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
5725 because its test if the absolute value of 'a' equals 'w'.
5726 Note that BN_abs_is_word does *not* handle w == 0 reliably;
5727 it exists mostly for use in the implementations of BN_is_zero(),
5728 BN_is_one(), and BN_is_word().
5731 *) New function BN_swap.
5734 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
5735 the exponentiation functions are more likely to produce reasonable
5736 results on negative inputs.
5739 *) Change BN_mod_mul so that the result is always non-negative.
5740 Previously, it could be negative if one of the factors was negative;
5741 I don't think anyone really wanted that behaviour.
5744 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
5745 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
5746 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
5747 and add new functions:
5756 BN_mod_lshift1_quick
5760 These functions always generate non-negative results.
5762 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
5763 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
5765 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
5766 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
5767 be reduced modulo m.
5768 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5771 The following entry accidentily appeared in the CHANGES file
5772 distributed with OpenSSL 0.9.7. The modifications described in
5773 it do *not* apply to OpenSSL 0.9.7.
5775 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5776 was actually never needed) and in BN_mul(). The removal in BN_mul()
5777 required a small change in bn_mul_part_recursive() and the addition
5778 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5779 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5780 bn_sub_words() and bn_add_words() except they take arrays with
5785 *) In 'openssl passwd', verify passwords read from the terminal
5786 unless the '-salt' option is used (which usually means that
5787 verification would just waste user's time since the resulting
5788 hash is going to be compared with some given password hash)
5789 or the new '-noverify' option is used.
5791 This is an incompatible change, but it does not affect
5792 non-interactive use of 'openssl passwd' (passwords on the command
5793 line, '-stdin' option, '-in ...' option) and thus should not
5797 *) Remove all references to RSAref, since there's no more need for it.
5800 *) Make DSO load along a path given through an environment variable
5801 (SHLIB_PATH) with shl_load().
5804 *) Constify the ENGINE code as a result of BIGNUM constification.
5805 Also constify the RSA code and most things related to it. In a
5806 few places, most notable in the depth of the ASN.1 code, ugly
5807 casts back to non-const were required (to be solved at a later
5811 *) Make it so the openssl application has all engines loaded by default.
5814 *) Constify the BIGNUM routines a little more.
5817 *) Add the following functions:
5819 ENGINE_load_cswift()
5821 ENGINE_load_atalla()
5823 ENGINE_load_builtin_engines()
5825 That way, an application can itself choose if external engines that
5826 are built-in in OpenSSL shall ever be used or not. The benefit is
5827 that applications won't have to be linked with libdl or other dso
5828 libraries unless it's really needed.
5830 Changed 'openssl engine' to load all engines on demand.
5831 Changed the engine header files to avoid the duplication of some
5832 declarations (they differed!).
5835 *) 'openssl engine' can now list capabilities.
5838 *) Better error reporting in 'openssl engine'.
5841 *) Never call load_dh_param(NULL) in s_server.
5844 *) Add engine application. It can currently list engines by name and
5845 identity, and test if they are actually available.
5848 *) Improve RPM specification file by forcing symbolic linking and making
5849 sure the installed documentation is also owned by root.root.
5850 [Damien Miller <djm@mindrot.org>]
5852 *) Give the OpenSSL applications more possibilities to make use of
5853 keys (public as well as private) handled by engines.
5856 *) Add OCSP code that comes from CertCo.
5859 *) Add VMS support for the Rijndael code.
5862 *) Added untested support for Nuron crypto accelerator.
5865 *) Add support for external cryptographic devices. This code was
5866 previously distributed separately as the "engine" branch.
5867 [Geoff Thorpe, Richard Levitte]
5869 *) Rework the filename-translation in the DSO code. It is now possible to
5870 have far greater control over how a "name" is turned into a filename
5871 depending on the operating environment and any oddities about the
5872 different shared library filenames on each system.
5875 *) Support threads on FreeBSD-elf in Configure.
5878 *) Fix for SHA1 assembly problem with MASM: it produces
5879 warnings about corrupt line number information when assembling
5880 with debugging information. This is caused by the overlapping
5882 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
5885 NCONF_get_number() has no error checking at all. As a replacement,
5886 NCONF_get_number_e() is defined (_e for "error checking") and is
5887 promoted strongly. The old NCONF_get_number is kept around for
5888 binary backward compatibility.
5889 Make it possible for methods to load from something other than a BIO,
5890 by providing a function pointer that is given a name instead of a BIO.
5891 For example, this could be used to load configuration data from an
5895 *) Fix for non blocking accept BIOs. Added new I/O special reason
5896 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
5897 with non blocking I/O was not possible because no retry code was
5898 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
5902 *) Added the beginnings of Rijndael support.
5905 *) Fix for bug in DirectoryString mask setting. Add support for
5906 X509_NAME_print_ex() in 'req' and X509_print_ex() function
5907 to allow certificate printing to more controllable, additional
5908 'certopt' option to 'x509' to allow new printing options to be
5912 *) Clean old EAY MD5 hack from e_os.h.
5915 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
5917 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5918 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5919 [Joe Orton, Steve Henson]
5921 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
5923 *) Fix additional bug revealed by the NISCC test suite:
5925 Stop bug triggering large recursion when presented with
5926 certain ASN.1 tags (CVE-2003-0851)
5929 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
5931 *) Fix various bugs revealed by running the NISCC test suite:
5933 Stop out of bounds reads in the ASN1 code when presented with
5934 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5936 If verify callback ignores invalid public key errors don't try to check
5937 certificate signature with the NULL public key.
5941 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5942 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5946 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5947 extra data after the compression methods not only for TLS 1.0
5948 but also for SSL 3.0 (as required by the specification).
5949 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5951 *) Change X509_certificate_type() to mark the key as exported/exportable
5952 when it's 512 *bits* long, not 512 bytes.
5955 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
5957 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5958 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5959 a protocol version number mismatch like a decryption error
5960 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5963 *) Turn on RSA blinding by default in the default implementation
5964 to avoid a timing attack. Applications that don't want it can call
5965 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5966 They would be ill-advised to do so in most cases.
5967 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5969 *) Change RSA blinding code so that it works when the PRNG is not
5970 seeded (in this case, the secret RSA exponent is abused as
5971 an unpredictable seed -- if it is not unpredictable, there
5972 is no point in blinding anyway). Make RSA blinding thread-safe
5973 by remembering the creator's thread ID in rsa->blinding and
5974 having all other threads use local one-time blinding factors
5975 (this requires more computation than sharing rsa->blinding, but
5976 avoids excessive locking; and if an RSA object is not shared
5977 between threads, blinding will still be very fast).
5980 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
5982 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5983 via timing by performing a MAC computation even if incorrrect
5984 block cipher padding has been found. This is a countermeasure
5985 against active attacks where the attacker has to distinguish
5986 between bad padding and a MAC verification error. (CVE-2003-0078)
5988 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5989 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5990 Martin Vuagnoux (EPFL, Ilion)]
5992 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
5994 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
5995 memory from it's contents. This is done with a counter that will
5996 place alternating values in each byte. This can be used to solve
5997 two issues: 1) the removal of calls to memset() by highly optimizing
5998 compilers, and 2) cleansing with other values than 0, since those can
5999 be read through on certain media, for example a swap space on disk.
6002 *) Bugfix: client side session caching did not work with external caching,
6003 because the session->cipher setting was not restored when reloading
6004 from the external cache. This problem was masked, when
6005 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
6006 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
6009 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
6010 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
6011 [Zeev Lieber <zeev-l@yahoo.com>]
6013 *) Undo an undocumented change introduced in 0.9.6e which caused
6014 repeated calls to OpenSSL_add_all_ciphers() and
6015 OpenSSL_add_all_digests() to be ignored, even after calling
6019 *) Change the default configuration reader to deal with last line not
6020 being properly terminated.
6023 *) Change X509_NAME_cmp() so it applies the special rules on handling
6024 DN values that are of type PrintableString, as well as RDNs of type
6025 emailAddress where the value has the type ia5String.
6026 [stefank@valicert.com via Richard Levitte]
6028 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
6029 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
6030 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
6031 the bitwise-OR of the two for use by the majority of applications
6032 wanting this behaviour, and update the docs. The documented
6033 behaviour and actual behaviour were inconsistent and had been
6034 changing anyway, so this is more a bug-fix than a behavioural
6036 [Geoff Thorpe, diagnosed by Nadav Har'El]
6038 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
6039 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
6042 *) Fix initialization code race conditions in
6043 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
6044 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
6045 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
6046 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
6047 ssl2_get_cipher_by_char(),
6048 ssl3_get_cipher_by_char().
6049 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
6051 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
6052 the cached sessions are flushed, as the remove_cb() might use ex_data
6053 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
6054 (see [openssl.org #212]).
6055 [Geoff Thorpe, Lutz Jaenicke]
6057 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
6058 length, instead of the encoding length to d2i_ASN1_OBJECT.
6061 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
6063 *) [In 0.9.6g-engine release:]
6064 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
6065 [Lynn Gazis <lgazis@rainbow.com>]
6067 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
6069 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
6070 and get fix the header length calculation.
6071 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
6072 Alon Kantor <alonk@checkpoint.com> (and others),
6075 *) Use proper error handling instead of 'assertions' in buffer
6076 overflow checks added in 0.9.6e. This prevents DoS (the
6077 assertions could call abort()).
6078 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
6080 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
6082 *) Add various sanity checks to asn1_get_length() to reject
6083 the ASN1 length bytes if they exceed sizeof(long), will appear
6084 negative or the content length exceeds the length of the
6086 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6088 *) Fix cipher selection routines: ciphers without encryption had no flags
6089 for the cipher strength set and where therefore not handled correctly
6090 by the selection routines (PR #130).
6093 *) Fix EVP_dsa_sha macro.
6097 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
6098 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
6099 that was added in OpenSSL 0.9.6d.
6101 As the countermeasure turned out to be incompatible with some
6102 broken SSL implementations, the new option is part of SSL_OP_ALL.
6103 SSL_OP_ALL is usually employed when compatibility with weird SSL
6104 implementations is desired (e.g. '-bugs' option to 's_client' and
6105 's_server'), so the new option is automatically set in many
6109 *) Changes in security patch:
6111 Changes marked "(CHATS)" were sponsored by the Defense Advanced
6112 Research Projects Agency (DARPA) and Air Force Research Laboratory,
6113 Air Force Materiel Command, USAF, under agreement number
6116 *) Add various sanity checks to asn1_get_length() to reject
6117 the ASN1 length bytes if they exceed sizeof(long), will appear
6118 negative or the content length exceeds the length of the
6119 supplied buffer. (CVE-2002-0659)
6120 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6122 *) Assertions for various potential buffer overflows, not known to
6124 [Ben Laurie (CHATS)]
6126 *) Various temporary buffers to hold ASCII versions of integers were
6127 too small for 64 bit platforms. (CVE-2002-0655)
6128 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
6130 *) Remote buffer overflow in SSL3 protocol - an attacker could
6131 supply an oversized session ID to a client. (CVE-2002-0656)
6132 [Ben Laurie (CHATS)]
6134 *) Remote buffer overflow in SSL2 protocol - an attacker could
6135 supply an oversized client master key. (CVE-2002-0656)
6136 [Ben Laurie (CHATS)]
6138 Changes between 0.9.6c and 0.9.6d [9 May 2002]
6140 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
6141 encoded as NULL) with id-dsa-with-sha1.
6142 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
6144 *) Check various X509_...() return values in apps/req.c.
6145 [Nils Larsch <nla@trustcenter.de>]
6147 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
6148 an end-of-file condition would erronously be flagged, when the CRLF
6149 was just at the end of a processed block. The bug was discovered when
6150 processing data through a buffering memory BIO handing the data to a
6151 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
6152 <ptsekov@syntrex.com> and Nedelcho Stanev.
6155 *) Implement a countermeasure against a vulnerability recently found
6156 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
6157 before application data chunks to avoid the use of known IVs
6158 with data potentially chosen by the attacker.
6161 *) Fix length checks in ssl3_get_client_hello().
6164 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
6165 to prevent ssl3_read_internal() from incorrectly assuming that
6166 ssl3_read_bytes() found application data while handshake
6167 processing was enabled when in fact s->s3->in_read_app_data was
6168 merely automatically cleared during the initial handshake.
6169 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
6171 *) Fix object definitions for Private and Enterprise: they were not
6172 recognized in their shortname (=lowercase) representation. Extend
6173 obj_dat.pl to issue an error when using undefined keywords instead
6174 of silently ignoring the problem (Svenning Sorensen
6175 <sss@sss.dnsalias.net>).
6178 *) Fix DH_generate_parameters() so that it works for 'non-standard'
6179 generators, i.e. generators other than 2 and 5. (Previously, the
6180 code did not properly initialise the 'add' and 'rem' values to
6181 BN_generate_prime().)
6183 In the new general case, we do not insist that 'generator' is
6184 actually a primitive root: This requirement is rather pointless;
6185 a generator of the order-q subgroup is just as good, if not
6189 *) Map new X509 verification errors to alerts. Discovered and submitted by
6190 Tom Wu <tom@arcot.com>.
6193 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
6194 returning non-zero before the data has been completely received
6195 when using non-blocking I/O.
6196 [Bodo Moeller; problem pointed out by John Hughes]
6198 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
6199 [Ben Laurie, Lutz Jaenicke]
6201 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
6202 Yoram Zahavi <YoramZ@gilian.com>).
6205 *) Add information about CygWin 1.3 and on, and preserve proper
6206 configuration for the versions before that.
6207 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6209 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
6210 check whether we deal with a copy of a session and do not delete from
6211 the cache in this case. Problem reported by "Izhar Shoshani Levi"
6212 <izhar@checkpoint.com>.
6215 *) Do not store session data into the internal session cache, if it
6216 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
6217 flag is set). Proposed by Aslam <aslam@funk.com>.
6220 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
6224 *) [In 0.9.6d-engine release:]
6225 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6226 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6228 *) Add the configuration target linux-s390x.
6229 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
6231 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
6232 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
6233 variable as an indication that a ClientHello message has been
6234 received. As the flag value will be lost between multiple
6235 invocations of ssl3_accept when using non-blocking I/O, the
6236 function may not be aware that a handshake has actually taken
6237 place, thus preventing a new session from being added to the
6240 To avoid this problem, we now set s->new_session to 2 instead of
6241 using a local variable.
6242 [Lutz Jaenicke, Bodo Moeller]
6244 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
6245 if the SSL_R_LENGTH_MISMATCH error is detected.
6246 [Geoff Thorpe, Bodo Moeller]
6248 *) New 'shared_ldflag' column in Configure platform table.
6251 *) Fix EVP_CIPHER_mode macro.
6252 ["Dan S. Camper" <dan@bti.net>]
6254 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
6255 type, we must throw them away by setting rr->length to 0.
6256 [D P Chang <dpc@qualys.com>]
6258 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
6260 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
6261 <Dominikus.Scherkl@biodata.com>. (The previous implementation
6262 worked incorrectly for those cases where range = 10..._2 and
6263 3*range is two bits longer than range.)
6266 *) Only add signing time to PKCS7 structures if it is not already
6270 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
6271 OBJ_ld_ce should be OBJ_id_ce.
6272 Also some ip-pda OIDs in crypto/objects/objects.txt were
6273 incorrect (cf. RFC 3039).
6274 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
6276 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
6277 returns early because it has nothing to do.
6278 [Andy Schneider <andy.schneider@bjss.co.uk>]
6280 *) [In 0.9.6c-engine release:]
6281 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
6282 [Andy Schneider <andy.schneider@bjss.co.uk>]
6284 *) [In 0.9.6c-engine release:]
6285 Add support for Cryptographic Appliance's keyserver technology.
6286 (Use engine 'keyclient')
6287 [Cryptographic Appliances and Geoff Thorpe]
6289 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
6290 is called via tools/c89.sh because arguments have to be
6291 rearranged (all '-L' options must appear before the first object
6293 [Richard Shapiro <rshapiro@abinitio.com>]
6295 *) [In 0.9.6c-engine release:]
6296 Add support for Broadcom crypto accelerator cards, backported
6298 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
6300 *) [In 0.9.6c-engine release:]
6301 Add support for SureWare crypto accelerator cards from
6302 Baltimore Technologies. (Use engine 'sureware')
6303 [Baltimore Technologies and Mark Cox]
6305 *) [In 0.9.6c-engine release:]
6306 Add support for crypto accelerator cards from Accelerated
6307 Encryption Processing, www.aep.ie. (Use engine 'aep')
6308 [AEP Inc. and Mark Cox]
6310 *) Add a configuration entry for gcc on UnixWare.
6311 [Gary Benson <gbenson@redhat.com>]
6313 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
6314 messages are stored in a single piece (fixed-length part and
6315 variable-length part combined) and fix various bugs found on the way.
6318 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
6319 instead. BIO_gethostbyname() does not know what timeouts are
6320 appropriate, so entries would stay in cache even when they have
6322 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
6324 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
6325 faced with a pathologically small ClientHello fragment that does
6326 not contain client_version: Instead of aborting with an error,
6327 simply choose the highest available protocol version (i.e.,
6328 TLS 1.0 unless it is disabled). In practice, ClientHello
6329 messages are never sent like this, but this change gives us
6330 strictly correct behaviour at least for TLS.
6333 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
6334 never resets s->method to s->ctx->method when called from within
6335 one of the SSL handshake functions.
6336 [Bodo Moeller; problem pointed out by Niko Baric]
6338 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
6339 (sent using the client's version number) if client_version is
6340 smaller than the protocol version in use. Also change
6341 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
6342 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
6343 the client will at least see that alert.
6346 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
6350 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
6351 client receives HelloRequest while in a handshake.
6352 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
6354 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
6355 should end in 'break', not 'goto end' which circuments various
6356 cleanups done in state SSL_ST_OK. But session related stuff
6357 must be disabled for SSL_ST_OK in the case that we just sent a
6360 Also avoid some overhead by not calling ssl_init_wbio_buffer()
6361 before just sending a HelloRequest.
6362 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
6364 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
6365 reveal whether illegal block cipher padding was found or a MAC
6366 verification error occured. (Neither SSLerr() codes nor alerts
6367 are directly visible to potential attackers, but the information
6368 may leak via logfiles.)
6370 Similar changes are not required for the SSL 2.0 implementation
6371 because the number of padding bytes is sent in clear for SSL 2.0,
6372 and the extra bytes are just ignored. However ssl/s2_pkt.c
6373 failed to verify that the purported number of padding bytes is in
6377 *) Add OpenUNIX-8 support including shared libraries
6378 (Boyd Lynn Gerber <gerberb@zenez.com>).
6381 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
6382 'wristwatch attack' using huge encoding parameters (cf.
6383 James H. Manger's CRYPTO 2001 paper). Note that the
6384 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
6385 encoding parameters and hence was not vulnerable.
6388 *) BN_sqr() bug fix.
6389 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
6391 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
6392 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
6393 followed by modular reduction.
6394 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
6396 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
6397 equivalent based on BN_pseudo_rand() instead of BN_rand().
6400 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
6401 This function was broken, as the check for a new client hello message
6402 to handle SGC did not allow these large messages.
6403 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
6406 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
6409 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
6410 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
6413 *) Rework the configuration and shared library support for Tru64 Unix.
6414 The configuration part makes use of modern compiler features and
6415 still retains old compiler behavior for those that run older versions
6416 of the OS. The shared library support part includes a variant that
6417 uses the RPATH feature, and is available through the special
6418 configuration target "alpha-cc-rpath", which will never be selected
6420 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
6422 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
6423 with the same message size as in ssl3_get_certificate_request().
6424 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
6425 messages might inadvertently be reject as too long.
6426 [Petr Lampa <lampa@fee.vutbr.cz>]
6428 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
6431 *) Modified SSL library such that the verify_callback that has been set
6432 specificly for an SSL object with SSL_set_verify() is actually being
6433 used. Before the change, a verify_callback set with this function was
6434 ignored and the verify_callback() set in the SSL_CTX at the time of
6435 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
6436 to allow the necessary settings.
6439 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
6440 explicitly to NULL, as at least on Solaris 8 this seems not always to be
6441 done automatically (in contradiction to the requirements of the C
6442 standard). This made problems when used from OpenSSH.
6445 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
6446 dh->length and always used
6448 BN_rand_range(priv_key, dh->p).
6450 BN_rand_range() is not necessary for Diffie-Hellman, and this
6451 specific range makes Diffie-Hellman unnecessarily inefficient if
6452 dh->length (recommended exponent length) is much smaller than the
6453 length of dh->p. We could use BN_rand_range() if the order of
6454 the subgroup was stored in the DH structure, but we only have
6459 BN_rand(priv_key, l, ...)
6461 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
6467 RSA_eay_public_encrypt
6468 RSA_eay_private_decrypt
6469 RSA_eay_private_encrypt (signing)
6470 RSA_eay_public_decrypt (signature verification)
6472 (default implementations for RSA_public_encrypt,
6473 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
6474 always reject numbers >= n.
6477 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
6478 to synchronize access to 'locking_thread'. This is necessary on
6479 systems where access to 'locking_thread' (an 'unsigned long'
6480 variable) is not atomic.
6483 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
6484 *before* setting the 'crypto_lock_rand' flag. The previous code had
6485 a race condition if 0 is a valid thread ID.
6486 [Travis Vitek <vitek@roguewave.com>]
6488 *) Add support for shared libraries under Irix.
6489 [Albert Chin-A-Young <china@thewrittenword.com>]
6491 *) Add configuration option to build on Linux on both big-endian and
6493 [Ralf Baechle <ralf@uni-koblenz.de>]
6495 *) Add the possibility to create shared libraries on HP-UX.
6498 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
6500 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
6501 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
6502 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
6503 PRNG state recovery was possible based on the output of
6504 one PRNG request appropriately sized to gain knowledge on
6505 'md' followed by enough consecutive 1-byte PRNG requests
6506 to traverse all of 'state'.
6508 1. When updating 'md_local' (the current thread's copy of 'md')
6509 during PRNG output generation, hash all of the previous
6510 'md_local' value, not just the half used for PRNG output.
6512 2. Make the number of bytes from 'state' included into the hash
6513 independent from the number of PRNG bytes requested.
6515 The first measure alone would be sufficient to avoid
6516 Markku-Juhani's attack. (Actually it had never occurred
6517 to me that the half of 'md_local' used for chaining was the
6518 half from which PRNG output bytes were taken -- I had always
6519 assumed that the secret half would be used.) The second
6520 measure makes sure that additional data from 'state' is never
6521 mixed into 'md_local' in small portions; this heuristically
6522 further strengthens the PRNG.
6525 *) Fix crypto/bn/asm/mips3.s.
6528 *) When only the key is given to "enc", the IV is undefined. Print out
6529 an error message in this case.
6532 *) Handle special case when X509_NAME is empty in X509 printing routines.
6535 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
6536 positive and less than q.
6539 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
6540 used: it isn't thread safe and the add_lock_callback should handle
6542 [Paul Rose <Paul.Rose@bridge.com>]
6544 *) Verify that incoming data obeys the block size in
6545 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
6549 [Ulf Möller, Bodo Möller]
6551 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
6552 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
6553 when fixing the server behaviour for backwards-compatible 'client
6554 hello' messages. (Note that the attack is impractical against
6555 SSL 3.0 and TLS 1.0 anyway because length and version checking
6556 means that the probability of guessing a valid ciphertext is
6557 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
6560 Before 0.9.5, the countermeasure (hide the error by generating a
6561 random 'decryption result') did not work properly because
6562 ERR_clear_error() was missing, meaning that SSL_get_error() would
6563 detect the supposedly ignored error.
6565 Both problems are now fixed.
6568 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
6569 (previously it was 1024).
6572 *) Fix for compatibility mode trust settings: ignore trust settings
6573 unless some valid trust or reject settings are present.
6576 *) Fix for blowfish EVP: its a variable length cipher.
6579 *) Fix various bugs related to DSA S/MIME verification. Handle missing
6580 parameters in DSA public key structures and return an error in the
6581 DSA routines if parameters are absent.
6584 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
6585 in the current directory if neither $RANDFILE nor $HOME was set.
6586 RAND_file_name() in 0.9.6a returned NULL in this case. This has
6587 caused some confusion to Windows users who haven't defined $HOME.
6588 Thus RAND_file_name() is changed again: e_os.h can define a
6589 DEFAULT_HOME, which will be used if $HOME is not set.
6590 For Windows, we use "C:"; on other platforms, we still require
6591 environment variables.
6593 *) Move 'if (!initialized) RAND_poll()' into regions protected by
6594 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
6595 having multiple threads call RAND_poll() concurrently.
6598 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
6599 combination of a flag and a thread ID variable.
6600 Otherwise while one thread is in ssleay_rand_bytes (which sets the
6601 flag), *other* threads can enter ssleay_add_bytes without obeying
6602 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
6603 that they do not hold after the first thread unsets add_do_not_lock).
6606 *) Change bctest again: '-x' expressions are not available in all
6610 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
6612 *) Fix a couple of memory leaks in PKCS7_dataDecode()
6613 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
6615 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
6616 the default extension for executables, if any. Also, make the perl
6617 scripts that use symlink() to test if it really exists and use "cp"
6618 if it doesn't. All this made OpenSSL compilable and installable in
6622 *) Fix for asn1_GetSequence() for indefinite length constructed data.
6623 If SEQUENCE is length is indefinite just set c->slen to the total
6624 amount of data available.
6625 [Steve Henson, reported by shige@FreeBSD.org]
6626 [This change does not apply to 0.9.7.]
6628 *) Change bctest to avoid here-documents inside command substitution
6629 (workaround for FreeBSD /bin/sh bug).
6630 For compatibility with Ultrix, avoid shell functions (introduced
6631 in the bctest version that searches along $PATH).
6634 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6635 with des_encrypt() defined on some operating systems, like Solaris
6639 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6640 On the Importance of Eliminating Errors in Cryptographic
6641 Computations, J. Cryptology 14 (2001) 2, 101-119,
6642 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6645 *) MIPS assembler BIGNUM division bug fix.
6648 *) Disabled incorrect Alpha assembler code.
6651 *) Fix PKCS#7 decode routines so they correctly update the length
6652 after reading an EOC for the EXPLICIT tag.
6654 [This change does not apply to 0.9.7.]
6656 *) Fix bug in PKCS#12 key generation routines. This was triggered
6657 if a 3DES key was generated with a 0 initial byte. Include
6658 PKCS12_BROKEN_KEYGEN compilation option to retain the old
6659 (but broken) behaviour.
6662 *) Enhance bctest to search for a working bc along $PATH and print
6664 [Tim Rice <tim@multitalents.net> via Richard Levitte]
6666 *) Fix memory leaks in err.c: free err_data string if necessary;
6667 don't write to the wrong index in ERR_set_error_data.
6670 *) Implement ssl23_peek (analogous to ssl23_read), which previously
6674 *) Replace rdtsc with _emit statements for VC++ version 5.
6675 [Jeremy Cooper <jeremy@baymoo.org>]
6677 *) Make it possible to reuse SSLv2 sessions.
6680 *) In copy_email() check for >= 0 as a return value for
6681 X509_NAME_get_index_by_NID() since 0 is a valid index.
6682 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
6684 *) Avoid coredump with unsupported or invalid public keys by checking if
6685 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
6686 PKCS7_verify() fails with non detached data.
6689 *) Don't use getenv in library functions when run as setuid/setgid.
6690 New function OPENSSL_issetugid().
6693 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
6694 due to incorrect handling of multi-threading:
6696 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
6698 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
6700 3. Count how many times MemCheck_off() has been called so that
6701 nested use can be treated correctly. This also avoids
6702 inband-signalling in the previous code (which relied on the
6703 assumption that thread ID 0 is impossible).
6706 *) Add "-rand" option also to s_client and s_server.
6709 *) Fix CPU detection on Irix 6.x.
6710 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
6711 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6713 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
6716 [This change does not apply to 0.9.7.]
6718 *) Use the cached encoding of an X509_NAME structure rather than
6719 copying it. This is apparently the reason for the libsafe "errors"
6720 but the code is actually correct.
6723 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
6724 Bleichenbacher's DSA attack.
6725 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
6726 to be set and top=0 forces the highest bit to be set; top=-1 is new
6727 and leaves the highest bit random.
6728 [Ulf Moeller, Bodo Moeller]
6730 *) In the NCONF_...-based implementations for CONF_... queries
6731 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
6732 a temporary CONF structure with the data component set to NULL
6733 (which gives segmentation faults in lh_retrieve).
6734 Instead, use NULL for the CONF pointer in CONF_get_string and
6735 CONF_get_number (which may use environment variables) and directly
6736 return NULL from CONF_get_section.
6739 *) Fix potential buffer overrun for EBCDIC.
6742 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
6743 keyUsage if basicConstraints absent for a CA.
6746 *) Make SMIME_write_PKCS7() write mail header values with a format that
6747 is more generally accepted (no spaces before the semicolon), since
6748 some programs can't parse those values properly otherwise. Also make
6749 sure BIO's that break lines after each write do not create invalid
6753 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
6754 macros previously used would not encode an empty SEQUENCE OF
6755 and break the signature.
6757 [This change does not apply to 0.9.7.]
6759 *) Zero the premaster secret after deriving the master secret in
6763 *) Add some EVP_add_digest_alias registrations (as found in
6764 OpenSSL_add_all_digests()) to SSL_library_init()
6765 aka OpenSSL_add_ssl_algorithms(). This provides improved
6766 compatibility with peers using X.509 certificates
6767 with unconventional AlgorithmIdentifier OIDs.
6770 *) Fix for Irix with NO_ASM.
6771 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6773 *) ./config script fixes.
6774 [Ulf Moeller, Richard Levitte]
6776 *) Fix 'openssl passwd -1'.
6779 *) Change PKCS12_key_gen_asc() so it can cope with non null
6780 terminated strings whose length is passed in the passlen
6781 parameter, for example from PEM callbacks. This was done
6782 by adding an extra length parameter to asc2uni().
6783 [Steve Henson, reported by <oddissey@samsung.co.kr>]
6785 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
6786 call failed, free the DSA structure.
6789 *) Fix to uni2asc() to cope with zero length Unicode strings.
6790 These are present in some PKCS#12 files.
6793 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
6794 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
6795 when writing a 32767 byte record.
6796 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
6798 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
6799 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
6801 (RSA objects have a reference count access to which is protected
6802 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
6803 so they are meant to be shared between threads.)
6804 [Bodo Moeller, Geoff Thorpe; original patch submitted by
6805 "Reddie, Steven" <Steven.Reddie@ca.com>]
6807 *) Fix a deadlock in CRYPTO_mem_leaks().
6810 *) Use better test patterns in bntest.
6813 *) rand_win.c fix for Borland C.
6816 *) BN_rshift bugfix for n == 0.
6819 *) Add a 'bctest' script that checks for some known 'bc' bugs
6820 so that 'make test' does not abort just because 'bc' is broken.
6823 *) Store verify_result within SSL_SESSION also for client side to
6824 avoid potential security hole. (Re-used sessions on the client side
6825 always resulted in verify_result==X509_V_OK, not using the original
6826 result of the server certificate verification.)
6829 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
6830 SSL3_RT_APPLICATION_DATA, return 0.
6831 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
6835 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
6836 releases, have been re-implemented by renaming the previous
6837 implementations of ssl2_read and ssl3_read to ssl2_read_internal
6838 and ssl3_read_internal, respectively, and adding 'peek' parameters
6839 to them. The new ssl[23]_{read,peek} functions are calls to
6840 ssl[23]_read_internal with the 'peek' flag set appropriately.
6841 A 'peek' parameter has also been added to ssl3_read_bytes, which
6842 does the actual work for ssl3_read_internal.
6845 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
6846 the method-specific "init()" handler. Also clean up ex_data after
6847 calling the method-specific "finish()" handler. Previously, this was
6848 happening the other way round.
6851 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
6852 The previous value, 12, was not always sufficient for BN_mod_exp().
6855 *) Make sure that shared libraries get the internal name engine with
6856 the full version number and not just 0. This should mark the
6857 shared libraries as not backward compatible. Of course, this should
6858 be changed again when we can guarantee backward binary compatibility.
6861 *) Fix typo in get_cert_by_subject() in by_dir.c
6862 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
6864 *) Rework the system to generate shared libraries:
6866 - Make note of the expected extension for the shared libraries and
6867 if there is a need for symbolic links from for example libcrypto.so.0
6868 to libcrypto.so.0.9.7. There is extended info in Configure for
6871 - Make as few rebuilds of the shared libraries as possible.
6873 - Still avoid linking the OpenSSL programs with the shared libraries.
6875 - When installing, install the shared libraries separately from the
6879 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
6881 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
6882 and not in SSL_clear because the latter is also used by the
6883 accept/connect functions; previously, the settings made by
6884 SSL_set_read_ahead would be lost during the handshake.
6885 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
6887 *) Correct util/mkdef.pl to be selective about disabled algorithms.
6888 Previously, it would create entries for disableed algorithms no
6892 *) Added several new manual pages for SSL_* function.
6895 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
6897 *) In ssl23_get_client_hello, generate an error message when faced
6898 with an initial SSL 3.0/TLS record that is too small to contain the
6899 first two bytes of the ClientHello message, i.e. client_version.
6900 (Note that this is a pathologic case that probably has never happened
6901 in real life.) The previous approach was to use the version number
6902 from the record header as a substitute; but our protocol choice
6903 should not depend on that one because it is not authenticated
6904 by the Finished messages.
6907 *) More robust randomness gathering functions for Windows.
6908 [Jeffrey Altman <jaltman@columbia.edu>]
6910 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
6911 not set then we don't setup the error code for issuer check errors
6912 to avoid possibly overwriting other errors which the callback does
6913 handle. If an application does set the flag then we assume it knows
6914 what it is doing and can handle the new informational codes
6918 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
6919 a general "ANY" type, as such it should be able to decode anything
6920 including tagged types. However it didn't check the class so it would
6921 wrongly interpret tagged types in the same way as their universal
6922 counterpart and unknown types were just rejected. Changed so that the
6923 tagged and unknown types are handled in the same way as a SEQUENCE:
6924 that is the encoding is stored intact. There is also a new type
6925 "V_ASN1_OTHER" which is used when the class is not universal, in this
6926 case we have no idea what the actual type is so we just lump them all
6930 *) On VMS, stdout may very well lead to a file that is written to
6931 in a record-oriented fashion. That means that every write() will
6932 write a separate record, which will be read separately by the
6933 programs trying to read from it. This can be very confusing.
6935 The solution is to put a BIO filter in the way that will buffer
6936 text until a linefeed is reached, and then write everything a
6937 line at a time, so every record written will be an actual line,
6938 not chunks of lines and not (usually doesn't happen, but I've
6939 seen it once) several lines in one record. BIO_f_linebuffer() is
6942 Currently, it's a VMS-only method, because that's where it has
6943 been tested well enough.
6946 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
6947 it can return incorrect results.
6948 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
6949 but it was in 0.9.6-beta[12].)
6952 *) Disable the check for content being present when verifying detached
6953 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
6954 include zero length content when signing messages.
6957 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
6958 BIO_ctrl (for BIO pairs).
6961 *) Add DSO method for VMS.
6964 *) Bug fix: Montgomery multiplication could produce results with the
6968 *) Add RPM specification openssl.spec and modify it to build three
6969 packages. The default package contains applications, application
6970 documentation and run-time libraries. The devel package contains
6971 include files, static libraries and function documentation. The
6972 doc package contains the contents of the doc directory. The original
6973 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
6976 *) Add a large number of documentation files for many SSL routines.
6977 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6979 *) Add a configuration entry for Sony News 4.
6980 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
6982 *) Don't set the two most significant bits to one when generating a
6983 random number < q in the DSA library.
6986 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
6987 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
6988 the underlying transport is blocking) if a handshake took place.
6989 (The default behaviour is needed by applications such as s_client
6990 and s_server that use select() to determine when to use SSL_read;
6991 but for applications that know in advance when to expect data, it
6992 just makes things more complicated.)
6995 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
6999 *) Add a few more EBCDIC conditionals that make `req' and `x509'
7000 work better on such systems.
7001 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7003 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
7004 Update PKCS12_parse() so it copies the friendlyName and the
7005 keyid to the certificates aux info.
7008 *) Fix bug in PKCS7_verify() which caused an infinite loop
7009 if there was more than one signature.
7010 [Sven Uszpelkat <su@celocom.de>]
7012 *) Major change in util/mkdef.pl to include extra information
7013 about each symbol, as well as presentig variables as well
7014 as functions. This change means that there's n more need
7015 to rebuild the .num files when some algorithms are excluded.
7018 *) Allow the verify time to be set by an application,
7019 rather than always using the current time.
7022 *) Phase 2 verify code reorganisation. The certificate
7023 verify code now looks up an issuer certificate by a
7024 number of criteria: subject name, authority key id
7025 and key usage. It also verifies self signed certificates
7026 by the same criteria. The main comparison function is
7027 X509_check_issued() which performs these checks.
7029 Lot of changes were necessary in order to support this
7030 without completely rewriting the lookup code.
7032 Authority and subject key identifier are now cached.
7034 The LHASH 'certs' is X509_STORE has now been replaced
7035 by a STACK_OF(X509_OBJECT). This is mainly because an
7036 LHASH can't store or retrieve multiple objects with
7037 the same hash value.
7039 As a result various functions (which were all internal
7040 use only) have changed to handle the new X509_STORE
7041 structure. This will break anything that messed round
7042 with X509_STORE internally.
7044 The functions X509_STORE_add_cert() now checks for an
7045 exact match, rather than just subject name.
7047 The X509_STORE API doesn't directly support the retrieval
7048 of multiple certificates matching a given criteria, however
7049 this can be worked round by performing a lookup first
7050 (which will fill the cache with candidate certificates)
7051 and then examining the cache for matches. This is probably
7052 the best we can do without throwing out X509_LOOKUP
7053 entirely (maybe later...).
7055 The X509_VERIFY_CTX structure has been enhanced considerably.
7057 All certificate lookup operations now go via a get_issuer()
7058 callback. Although this currently uses an X509_STORE it
7059 can be replaced by custom lookups. This is a simple way
7060 to bypass the X509_STORE hackery necessary to make this
7061 work and makes it possible to use more efficient techniques
7062 in future. A very simple version which uses a simple
7063 STACK for its trusted certificate store is also provided
7064 using X509_STORE_CTX_trusted_stack().
7066 The verify_cb() and verify() callbacks now have equivalents
7067 in the X509_STORE_CTX structure.
7069 X509_STORE_CTX also has a 'flags' field which can be used
7070 to customise the verify behaviour.
7073 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
7074 excludes S/MIME capabilities.
7077 *) When a certificate request is read in keep a copy of the
7078 original encoding of the signed data and use it when outputing
7079 again. Signatures then use the original encoding rather than
7080 a decoded, encoded version which may cause problems if the
7081 request is improperly encoded.
7084 *) For consistency with other BIO_puts implementations, call
7085 buffer_write(b, ...) directly in buffer_puts instead of calling
7088 In BIO_puts, increment b->num_write as in BIO_write.
7089 [Peter.Sylvester@EdelWeb.fr]
7091 *) Fix BN_mul_word for the case where the word is 0. (We have to use
7092 BN_zero, we may not return a BIGNUM with an array consisting of
7096 *) Avoid calling abort() from within the library when problems are
7097 detected, except if preprocessor symbols have been defined
7098 (such as REF_CHECK, BN_DEBUG etc.).
7101 *) New openssl application 'rsautl'. This utility can be
7102 used for low level RSA operations. DER public key
7103 BIO/fp routines also added.
7106 *) New Configure entry and patches for compiling on QNX 4.
7107 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
7109 *) A demo state-machine implementation was sponsored by
7110 Nuron (http://www.nuron.com/) and is now available in
7111 demos/state_machine.
7114 *) New options added to the 'dgst' utility for signature
7115 generation and verification.
7118 *) Unrecognized PKCS#7 content types are now handled via a
7119 catch all ASN1_TYPE structure. This allows unsupported
7120 types to be stored as a "blob" and an application can
7121 encode and decode it manually.
7124 *) Fix various signed/unsigned issues to make a_strex.c
7126 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
7128 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
7129 length if passed a buffer. ASN1_INTEGER_to_BN failed
7130 if passed a NULL BN and its argument was negative.
7131 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
7133 *) Modification to PKCS#7 encoding routines to output definite
7134 length encoding. Since currently the whole structures are in
7135 memory there's not real point in using indefinite length
7136 constructed encoding. However if OpenSSL is compiled with
7137 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
7140 *) Added BIO_vprintf() and BIO_vsnprintf().
7143 *) Added more prefixes to parse for in the the strings written
7144 through a logging bio, to cover all the levels that are available
7145 through syslog. The prefixes are now:
7147 PANIC, EMERG, EMR => LOG_EMERG
7148 ALERT, ALR => LOG_ALERT
7149 CRIT, CRI => LOG_CRIT
7150 ERROR, ERR => LOG_ERR
7151 WARNING, WARN, WAR => LOG_WARNING
7152 NOTICE, NOTE, NOT => LOG_NOTICE
7153 INFO, INF => LOG_INFO
7154 DEBUG, DBG => LOG_DEBUG
7156 and as before, if none of those prefixes are present at the
7157 beginning of the string, LOG_ERR is chosen.
7159 On Win32, the LOG_* levels are mapped according to this:
7161 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
7162 LOG_WARNING => EVENTLOG_WARNING_TYPE
7163 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
7167 *) Made it possible to reconfigure with just the configuration
7168 argument "reconf" or "reconfigure". The command line arguments
7169 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
7170 and are retrieved from there when reconfiguring.
7174 [Assar Westerlund <assar@sics.se>, Richard Levitte]
7176 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
7179 *) The obj_dat.pl script was messing up the sorting of object
7180 names. The reason was that it compared the quoted version
7181 of strings as a result "OCSP" > "OCSP Signing" because
7182 " > SPACE. Changed script to store unquoted versions of
7183 names and add quotes on output. It was also omitting some
7184 names from the lookup table if they were given a default
7185 value (that is if SN is missing it is given the same
7186 value as LN and vice versa), these are now added on the
7187 grounds that if an object has a name we should be able to
7188 look it up. Finally added warning output when duplicate
7189 short or long names are found.
7192 *) Changes needed for Tandem NSK.
7193 [Scott Uroff <scott@xypro.com>]
7195 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
7196 RSA_padding_check_SSLv23(), special padding was never detected
7197 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
7198 version rollback attacks was not effective.
7200 In s23_clnt.c, don't use special rollback-attack detection padding
7201 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
7202 client; similarly, in s23_srvr.c, don't do the rollback check if
7203 SSL 2.0 is the only protocol enabled in the server.
7206 *) Make it possible to get hexdumps of unprintable data with 'openssl
7207 asn1parse'. By implication, the functions ASN1_parse_dump() and
7208 BIO_dump_indent() are added.
7211 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
7212 these print out strings and name structures based on various
7213 flags including RFC2253 support and proper handling of
7214 multibyte characters. Added options to the 'x509' utility
7215 to allow the various flags to be set.
7218 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
7219 Also change the functions X509_cmp_current_time() and
7220 X509_gmtime_adj() work with an ASN1_TIME structure,
7221 this will enable certificates using GeneralizedTime in validity
7222 dates to be checked.
7225 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
7226 negative public key encodings) on by default,
7227 NO_NEG_PUBKEY_BUG can be set to disable it.
7230 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
7231 content octets. An i2c_ASN1_OBJECT is unnecessary because
7232 the encoding can be trivially obtained from the structure.
7235 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
7236 not read locks (CRYPTO_r_[un]lock).
7239 *) A first attempt at creating official support for shared
7240 libraries through configuration. I've kept it so the
7241 default is static libraries only, and the OpenSSL programs
7242 are always statically linked for now, but there are
7243 preparations for dynamic linking in place.
7244 This has been tested on Linux and Tru64.
7247 *) Randomness polling function for Win9x, as described in:
7248 Peter Gutmann, Software Generation of Practically Strong
7252 *) Fix so PRNG is seeded in req if using an already existing
7256 *) New options to smime application. -inform and -outform
7257 allow alternative formats for the S/MIME message including
7258 PEM and DER. The -content option allows the content to be
7259 specified separately. This should allow things like Netscape
7260 form signing output easier to verify.
7263 *) Fix the ASN1 encoding of tags using the 'long form'.
7266 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
7267 STRING types. These convert content octets to and from the
7268 underlying type. The actual tag and length octets are
7269 already assumed to have been read in and checked. These
7270 are needed because all other string types have virtually
7271 identical handling apart from the tag. By having versions
7272 of the ASN1 functions that just operate on content octets
7273 IMPLICIT tagging can be handled properly. It also allows
7274 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
7275 and ASN1_INTEGER are identical apart from the tag.
7278 *) Change the handling of OID objects as follows:
7280 - New object identifiers are inserted in objects.txt, following
7281 the syntax given in objects.README.
7282 - objects.pl is used to process obj_mac.num and create a new
7284 - obj_dat.pl is used to create a new obj_dat.h, using the data in
7287 This is currently kind of a hack, and the perl code in objects.pl
7288 isn't very elegant, but it works as I intended. The simplest way
7289 to check that it worked correctly is to look in obj_dat.h and
7290 check the array nid_objs and make sure the objects haven't moved
7291 around (this is important!). Additions are OK, as well as
7292 consistent name changes.
7295 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
7298 *) Addition of the command line parameter '-rand file' to 'openssl req'.
7299 The given file adds to whatever has already been seeded into the
7300 random pool through the RANDFILE configuration file option or
7301 environment variable, or the default random state file.
7304 *) mkstack.pl now sorts each macro group into lexical order.
7305 Previously the output order depended on the order the files
7306 appeared in the directory, resulting in needless rewriting
7310 *) Patches to make OpenSSL compile under Win32 again. Mostly
7311 work arounds for the VC++ problem that it treats func() as
7312 func(void). Also stripped out the parts of mkdef.pl that
7313 added extra typesafe functions: these no longer exist.
7316 *) Reorganisation of the stack code. The macros are now all
7317 collected in safestack.h . Each macro is defined in terms of
7318 a "stack macro" of the form SKM_<name>(type, a, b). The
7319 DEBUG_SAFESTACK is now handled in terms of function casts,
7320 this has the advantage of retaining type safety without the
7321 use of additional functions. If DEBUG_SAFESTACK is not defined
7322 then the non typesafe macros are used instead. Also modified the
7323 mkstack.pl script to handle the new form. Needs testing to see
7324 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
7325 the default if no major problems. Similar behaviour for ASN1_SET_OF
7326 and PKCS12_STACK_OF.
7329 *) When some versions of IIS use the 'NET' form of private key the
7330 key derivation algorithm is different. Normally MD5(password) is
7331 used as a 128 bit RC4 key. In the modified case
7332 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
7333 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
7334 as the old Netscape_RSA functions except they have an additional
7335 'sgckey' parameter which uses the modified algorithm. Also added
7336 an -sgckey command line option to the rsa utility. Thanks to
7337 Adrian Peck <bertie@ncipher.com> for posting details of the modified
7338 algorithm to openssl-dev.
7341 *) The evp_local.h macros were using 'c.##kname' which resulted in
7342 invalid expansion on some systems (SCO 5.0.5 for example).
7343 Corrected to 'c.kname'.
7344 [Phillip Porch <root@theporch.com>]
7346 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
7347 a STACK of email addresses from a certificate or request, these look
7348 in the subject name and the subject alternative name extensions and
7349 omit any duplicate addresses.
7352 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
7353 This makes DSA verification about 2 % faster.
7356 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
7357 (meaning that now 2^5 values will be precomputed, which is only 4 KB
7358 plus overhead for 1024 bit moduli).
7359 This makes exponentiations about 0.5 % faster for 1024 bit
7360 exponents (as measured by "openssl speed rsa2048").
7363 *) Rename memory handling macros to avoid conflicts with other
7365 Malloc => OPENSSL_malloc
7366 Malloc_locked => OPENSSL_malloc_locked
7367 Realloc => OPENSSL_realloc
7368 Free => OPENSSL_free
7371 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
7372 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
7375 *) CygWin32 support.
7376 [John Jarvie <jjarvie@newsguy.com>]
7378 *) The type-safe stack code has been rejigged. It is now only compiled
7379 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
7380 by default all type-specific stack functions are "#define"d back to
7381 standard stack functions. This results in more streamlined output
7382 but retains the type-safety checking possibilities of the original
7386 *) The STACK code has been cleaned up, and certain type declarations
7387 that didn't make a lot of sense have been brought in line. This has
7388 also involved a cleanup of sorts in safestack.h to more correctly
7389 map type-safe stack functions onto their plain stack counterparts.
7390 This work has also resulted in a variety of "const"ifications of
7391 lots of the code, especially "_cmp" operations which should normally
7392 be prototyped with "const" parameters anyway.
7395 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
7396 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
7397 (The PRNG state consists of two parts, the large pool 'state' and 'md',
7398 where all of 'md' is used each time the PRNG is used, but 'state'
7399 is used only indexed by a cyclic counter. As entropy may not be
7400 well distributed from the beginning, 'md' is important as a
7401 chaining variable. However, the output function chains only half
7402 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
7403 all of 'md', and seeding with STATE_SIZE dummy bytes will result
7404 in all of 'state' being rewritten, with the new values depending
7405 on virtually all of 'md'. This overcomes the 80 bit limitation.)
7408 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
7409 the handshake is continued after ssl_verify_cert_chain();
7410 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
7411 can lead to 'unexplainable' connection aborts later.
7412 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
7414 *) Major EVP API cipher revision.
7415 Add hooks for extra EVP features. This allows various cipher
7416 parameters to be set in the EVP interface. Support added for variable
7417 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
7418 setting of RC2 and RC5 parameters.
7420 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
7423 Remove lots of duplicated code from the EVP library. For example *every*
7424 cipher init() function handles the 'iv' in the same way according to the
7425 cipher mode. They also all do nothing if the 'key' parameter is NULL and
7426 for CFB and OFB modes they zero ctx->num.
7428 New functionality allows removal of S/MIME code RC2 hack.
7430 Most of the routines have the same form and so can be declared in terms
7433 By shifting this to the top level EVP_CipherInit() it can be removed from
7434 all individual ciphers. If the cipher wants to handle IVs or keys
7435 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
7438 Change lots of functions like EVP_EncryptUpdate() to now return a
7439 value: although software versions of the algorithms cannot fail
7440 any installed hardware versions can.
7443 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
7444 this option is set, tolerate broken clients that send the negotiated
7445 protocol version number instead of the requested protocol version
7449 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
7450 i.e. non-zero for export ciphersuites, zero otherwise.
7451 Previous versions had this flag inverted, inconsistent with
7452 rsa_tmp_cb (..._TMP_RSA_CB).
7453 [Bodo Moeller; problem reported by Amit Chopra]
7455 *) Add missing DSA library text string. Work around for some IIS
7456 key files with invalid SEQUENCE encoding.
7459 *) Add a document (doc/standards.txt) that list all kinds of standards
7460 and so on that are implemented in OpenSSL.
7463 *) Enhance c_rehash script. Old version would mishandle certificates
7464 with the same subject name hash and wouldn't handle CRLs at all.
7465 Added -fingerprint option to crl utility, to support new c_rehash
7469 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
7472 *) Fix for SSL server purpose checking. Server checking was
7473 rejecting certificates which had extended key usage present
7474 but no ssl client purpose.
7475 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
7477 *) Make PKCS#12 code work with no password. The PKCS#12 spec
7478 is a little unclear about how a blank password is handled.
7479 Since the password in encoded as a BMPString with terminating
7480 double NULL a zero length password would end up as just the
7481 double NULL. However no password at all is different and is
7482 handled differently in the PKCS#12 key generation code. NS
7483 treats a blank password as zero length. MSIE treats it as no
7484 password on export: but it will try both on import. We now do
7485 the same: PKCS12_parse() tries zero length and no password if
7486 the password is set to "" or NULL (NULL is now a valid password:
7487 it wasn't before) as does the pkcs12 application.
7490 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
7491 perror when PEM_read_bio_X509_REQ fails, the error message must
7492 be obtained from the error queue.
7495 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
7496 it in ERR_remove_state if appropriate, and change ERR_get_state
7497 accordingly to avoid race conditions (this is necessary because
7498 thread_hash is no longer constant once set).
7501 *) Bugfix for linux-elf makefile.one.
7504 *) RSA_get_default_method() will now cause a default
7505 RSA_METHOD to be chosen if one doesn't exist already.
7506 Previously this was only set during a call to RSA_new()
7507 or RSA_new_method(NULL) meaning it was possible for
7508 RSA_get_default_method() to return NULL.
7511 *) Added native name translation to the existing DSO code
7512 that will convert (if the flag to do so is set) filenames
7513 that are sufficiently small and have no path information
7514 into a canonical native form. Eg. "blah" converted to
7515 "libblah.so" or "blah.dll" etc.
7518 *) New function ERR_error_string_n(e, buf, len) which is like
7519 ERR_error_string(e, buf), but writes at most 'len' bytes
7520 including the 0 terminator. For ERR_error_string_n, 'buf'
7522 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
7524 *) CONF library reworked to become more general. A new CONF
7525 configuration file reader "class" is implemented as well as a
7526 new functions (NCONF_*, for "New CONF") to handle it. The now
7527 old CONF_* functions are still there, but are reimplemented to
7528 work in terms of the new functions. Also, a set of functions
7529 to handle the internal storage of the configuration data is
7530 provided to make it easier to write new configuration file
7531 reader "classes" (I can definitely see something reading a
7532 configuration file in XML format, for example), called _CONF_*,
7533 or "the configuration storage API"...
7535 The new configuration file reading functions are:
7537 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
7538 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
7540 NCONF_default, NCONF_WIN32
7542 NCONF_dump_fp, NCONF_dump_bio
7544 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
7545 NCONF_new creates a new CONF object. This works in the same way
7546 as other interfaces in OpenSSL, like the BIO interface.
7547 NCONF_dump_* dump the internal storage of the configuration file,
7548 which is useful for debugging. All other functions take the same
7549 arguments as the old CONF_* functions wth the exception of the
7550 first that must be a `CONF *' instead of a `LHASH *'.
7552 To make it easer to use the new classes with the old CONF_* functions,
7553 the function CONF_set_default_method is provided.
7556 *) Add '-tls1' option to 'openssl ciphers', which was already
7557 mentioned in the documentation but had not been implemented.
7558 (This option is not yet really useful because even the additional
7559 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
7562 *) Initial DSO code added into libcrypto for letting OpenSSL (and
7563 OpenSSL-based applications) load shared libraries and bind to
7564 them in a portable way.
7565 [Geoff Thorpe, with contributions from Richard Levitte]
7567 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
7569 *) Make sure _lrotl and _lrotr are only used with MSVC.
7571 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
7572 (the default implementation of RAND_status).
7574 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
7575 to '-clrext' (= clear extensions), as intended and documented.
7576 [Bodo Moeller; inconsistency pointed out by Michael Attili
7579 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
7580 was larger than the MD block size.
7581 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
7583 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
7584 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
7585 using the passed key: if the passed key was a private key the result
7586 of X509_print(), for example, would be to print out all the private key
7590 *) des_quad_cksum() byte order bug fix.
7591 [Ulf Möller, using the problem description in krb4-0.9.7, where
7592 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
7594 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
7596 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
7598 *) For easily testing in shell scripts whether some command
7599 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
7600 returns with exit code 0 iff no command of the given name is available.
7601 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
7602 the output goes to stdout and nothing is printed to stderr.
7603 Additional arguments are always ignored.
7605 Since for each cipher there is a command of the same name,
7606 the 'no-cipher' compilation switches can be tested this way.
7608 ('openssl no-XXX' is not able to detect pseudo-commands such
7609 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
7612 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
7615 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
7616 is set; it will be thrown away anyway because each handshake creates
7618 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
7619 to parameters -- in previous versions (since OpenSSL 0.9.3) the
7620 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
7621 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7624 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7625 'Q' and 'R' lose their special meanings (quit/renegotiate).
7626 This is part of what -quiet does; unlike -quiet, -ign_eof
7627 does not suppress any output.
7630 *) Add compatibility options to the purpose and trust code. The
7631 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7632 accepts a certificate or CA, this was the previous behaviour,
7633 with all the associated security issues.
7635 X509_TRUST_COMPAT is the old trust behaviour: only and
7636 automatically trust self signed roots in certificate store. A
7637 new trust setting X509_TRUST_DEFAULT is used to specify that
7638 a purpose has no associated trust setting and it should instead
7639 use the value in the default purpose.
7642 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7643 and fix a memory leak.
7646 *) In util/mkerr.pl (which implements 'make errors'), preserve
7647 reason strings from the previous version of the .c file, as
7648 the default to have only downcase letters (and digits) in
7649 automatically generated reasons codes is not always appropriate.
7652 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7653 using strerror. Previously, ERR_reason_error_string() returned
7654 library names as reason strings for SYSerr; but SYSerr is a special
7655 case where small numbers are errno values, not library numbers.
7658 *) Add '-dsaparam' option to 'openssl dhparam' application. This
7659 converts DSA parameters into DH parameters. (When creating parameters,
7660 DSA_generate_parameters is used.)
7663 *) Include 'length' (recommended exponent length) in C code generated
7664 by 'openssl dhparam -C'.
7667 *) The second argument to set_label in perlasm was already being used
7668 so couldn't be used as a "file scope" flag. Moved to third argument
7672 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
7673 instead of RAND_bytes for encryption IVs and salts.
7676 *) Include RAND_status() into RAND_METHOD instead of implementing
7677 it only for md_rand.c Otherwise replacing the PRNG by calling
7678 RAND_set_rand_method would be impossible.
7681 *) Don't let DSA_generate_key() enter an infinite loop if the random
7682 number generation fails.
7685 *) New 'rand' application for creating pseudo-random output.
7688 *) Added configuration support for Linux/IA64
7689 [Rolf Haberrecker <rolf@suse.de>]
7691 *) Assembler module support for Mingw32.
7694 *) Shared library support for HPUX (in shlib/).
7695 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
7697 *) Shared library support for Solaris gcc.
7698 [Lutz Behnke <behnke@trustcenter.de>]
7700 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
7702 *) PKCS7_encrypt() was adding text MIME headers twice because they
7703 were added manually and by SMIME_crlf_copy().
7706 *) In bntest.c don't call BN_rand with zero bits argument.
7707 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
7709 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
7710 case was implemented. This caused BN_div_recp() to fail occasionally.
7713 *) Add an optional second argument to the set_label() in the perl
7714 assembly language builder. If this argument exists and is set
7715 to 1 it signals that the assembler should use a symbol whose
7716 scope is the entire file, not just the current function. This
7717 is needed with MASM which uses the format label:: for this scope.
7718 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
7720 *) Change the ASN1 types so they are typedefs by default. Before
7721 almost all types were #define'd to ASN1_STRING which was causing
7722 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
7726 *) Change names of new functions to the new get1/get0 naming
7727 convention: After 'get1', the caller owns a reference count
7728 and has to call ..._free; 'get0' returns a pointer to some
7729 data structure without incrementing reference counters.
7730 (Some of the existing 'get' functions increment a reference
7731 counter, some don't.)
7732 Similarly, 'set1' and 'add1' functions increase reference
7733 counters or duplicate objects.
7736 *) Allow for the possibility of temp RSA key generation failure:
7737 the code used to assume it always worked and crashed on failure.
7740 *) Fix potential buffer overrun problem in BIO_printf().
7741 [Ulf Möller, using public domain code by Patrick Powell; problem
7742 pointed out by David Sacerdote <das33@cornell.edu>]
7744 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
7745 RAND_egd() and RAND_status(). In the command line application,
7746 the EGD socket can be specified like a seed file using RANDFILE
7750 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
7751 Some CAs (e.g. Verisign) distribute certificates in this form.
7754 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
7755 list to exclude them. This means that no special compilation option
7756 is needed to use anonymous DH: it just needs to be included in the
7760 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
7761 EVP_MD_type. The old functionality is available in a new macro called
7762 EVP_MD_md(). Change code that uses it and update docs.
7765 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
7766 where the 'void *' argument is replaced by a function pointer argument.
7767 Previously 'void *' was abused to point to functions, which works on
7768 many platforms, but is not correct. As these functions are usually
7769 called by macros defined in OpenSSL header files, most source code
7770 should work without changes.
7773 *) <openssl/opensslconf.h> (which is created by Configure) now contains
7774 sections with information on -D... compiler switches used for
7775 compiling the library so that applications can see them. To enable
7776 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
7777 must be defined. E.g.,
7778 #define OPENSSL_ALGORITHM_DEFINES
7779 #include <openssl/opensslconf.h>
7780 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
7781 [Richard Levitte, Ulf and Bodo Möller]
7783 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
7787 *) Change the 'other' type in certificate aux info to a STACK_OF
7788 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
7789 the required ASN1 format: arbitrary types determined by an OID.
7792 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
7793 argument to 'req'. This is not because the function is newer or
7794 better than others it just uses the work 'NEW' in the certificate
7795 request header lines. Some software needs this.
7798 *) Reorganise password command line arguments: now passwords can be
7799 obtained from various sources. Delete the PEM_cb function and make
7800 it the default behaviour: i.e. if the callback is NULL and the
7801 usrdata argument is not NULL interpret it as a null terminated pass
7802 phrase. If usrdata and the callback are NULL then the pass phrase
7803 is prompted for as usual.
7806 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
7807 the support is automatically enabled. The resulting binaries will
7808 autodetect the card and use it if present.
7809 [Ben Laurie and Compaq Inc.]
7811 *) Work around for Netscape hang bug. This sends certificate request
7812 and server done in one record. Since this is perfectly legal in the
7813 SSL/TLS protocol it isn't a "bug" option and is on by default. See
7814 the bugs/SSLv3 entry for more info.
7817 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
7820 *) Add -rand argument to smime and pkcs12 applications and read/write
7824 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
7827 *) Add command line password options to the remaining applications.
7830 *) Bug fix for BN_div_recp() for numerators with an even number of
7834 *) More tests in bntest.c, and changed test_bn output.
7837 *) ./config recognizes MacOS X now.
7840 *) Bug fix for BN_div() when the first words of num and divsor are
7841 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
7844 *) Add support for various broken PKCS#8 formats, and command line
7845 options to produce them.
7848 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
7849 get temporary BIGNUMs from a BN_CTX.
7852 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
7856 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
7857 include a #define from the old name to the new. The original intent
7858 was that statically linked binaries could for example just call
7859 SSLeay_add_all_ciphers() to just add ciphers to the table and not
7860 link with digests. This never worked becayse SSLeay_add_all_digests()
7861 and SSLeay_add_all_ciphers() were in the same source file so calling
7862 one would link with the other. They are now in separate source files.
7865 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
7868 *) Use a less unusual form of the Miller-Rabin primality test (it used
7869 a binary algorithm for exponentiation integrated into the Miller-Rabin
7870 loop, our standard modexp algorithms are faster).
7873 *) Support for the EBCDIC character set completed.
7874 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
7876 *) Source code cleanups: use const where appropriate, eliminate casts,
7877 use void * instead of char * in lhash.
7880 *) Bugfix: ssl3_send_server_key_exchange was not restartable
7881 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
7882 this the server could overwrite ephemeral keys that the client
7886 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
7887 using 50 iterations of the Rabin-Miller test.
7889 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
7890 iterations of the Rabin-Miller test as required by the appendix
7891 to FIPS PUB 186[-1]) instead of DSA_is_prime.
7892 As BN_is_prime_fasttest includes trial division, DSA parameter
7893 generation becomes much faster.
7895 This implies a change for the callback functions in DSA_is_prime
7896 and DSA_generate_parameters: The callback function is called once
7897 for each positive witness in the Rabin-Miller test, not just
7898 occasionally in the inner loop; and the parameters to the
7899 callback function now provide an iteration count for the outer
7900 loop rather than for the current invocation of the inner loop.
7901 DSA_generate_parameters additionally can call the callback
7902 function with an 'iteration count' of -1, meaning that a
7903 candidate has passed the trial division test (when q is generated
7904 from an application-provided seed, trial division is skipped).
7907 *) New function BN_is_prime_fasttest that optionally does trial
7908 division before starting the Rabin-Miller test and has
7909 an additional BN_CTX * argument (whereas BN_is_prime always
7910 has to allocate at least one BN_CTX).
7911 'callback(1, -1, cb_arg)' is called when a number has passed the
7912 trial division stage.
7915 *) Fix for bug in CRL encoding. The validity dates weren't being handled
7919 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
7922 *) New function BN_pseudo_rand().
7925 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
7926 bignum version of BN_from_montgomery() with the working code from
7927 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
7931 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
7932 made it impossible to use the same SSL_SESSION data structure in
7933 SSL2 clients in multiple threads.
7936 *) The return value of RAND_load_file() no longer counts bytes obtained
7937 by stat(). RAND_load_file(..., -1) is new and uses the complete file
7938 to seed the PRNG (previously an explicit byte count was required).
7939 [Ulf Möller, Bodo Möller]
7941 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
7942 used (char *) instead of (void *) and had casts all over the place.
7945 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
7948 *) Retain source code compatibility for BN_prime_checks macro:
7949 BN_is_prime(..., BN_prime_checks, ...) now uses
7950 BN_prime_checks_for_size to determine the appropriate number of
7951 Rabin-Miller iterations.
7954 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
7955 DH_CHECK_P_NOT_SAFE_PRIME.
7956 (Check if this is true? OpenPGP calls them "strong".)
7959 *) Merge the functionality of "dh" and "gendh" programs into a new program
7960 "dhparam". The old programs are retained for now but will handle DH keys
7961 (instead of parameters) in future.
7964 *) Make the ciphers, s_server and s_client programs check the return values
7965 when a new cipher list is set.
7968 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
7969 ciphers. Before when the 56bit ciphers were enabled the sorting was
7972 The syntax for the cipher sorting has been extended to support sorting by
7973 cipher-strength (using the strength_bits hard coded in the tables).
7974 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
7976 Fix a bug in the cipher-command parser: when supplying a cipher command
7977 string with an "undefined" symbol (neither command nor alphanumeric
7978 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
7979 an error is flagged.
7981 Due to the strength-sorting extension, the code of the
7982 ssl_create_cipher_list() function was completely rearranged. I hope that
7983 the readability was also increased :-)
7984 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7986 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
7987 for the first serial number and places 2 in the serial number file. This
7988 avoids problems when the root CA is created with serial number zero and
7989 the first user certificate has the same issuer name and serial number
7993 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
7994 the new code. Add documentation for this stuff.
7997 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
7998 X509_*() to X509at_*() on the grounds that they don't handle X509
7999 structures and behave in an analagous way to the X509v3 functions:
8000 they shouldn't be called directly but wrapper functions should be used
8003 So we also now have some wrapper functions that call the X509at functions
8004 when passed certificate requests. (TO DO: similar things can be done with
8005 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
8006 things. Some of these need some d2i or i2d and print functionality
8007 because they handle more complex structures.)
8010 *) Add missing #ifndefs that caused missing symbols when building libssl
8011 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
8012 NO_RSA in ssl/s2*.c.
8013 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
8015 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
8016 has a return value which indicates the quality of the random data
8017 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
8018 error queue. New function RAND_pseudo_bytes() generates output that is
8019 guaranteed to be unique but not unpredictable. RAND_add is like
8020 RAND_seed, but takes an extra argument for an entropy estimate
8021 (RAND_seed always assumes full entropy).
8024 *) Do more iterations of Rabin-Miller probable prime test (specifically,
8025 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
8026 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
8027 in crypto/bn/bn_prime.c for the complete table). This guarantees a
8028 false-positive rate of at most 2^-80 for random input.
8031 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
8034 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
8035 in the 0.9.5 release), this returns the chain
8036 from an X509_CTX structure with a dup of the stack and all
8037 the X509 reference counts upped: so the stack will exist
8038 after X509_CTX_cleanup() has been called. Modify pkcs12.c
8041 Also make SSL_SESSION_print() print out the verify return
8045 *) Add manpage for the pkcs12 command. Also change the default
8046 behaviour so MAC iteration counts are used unless the new
8047 -nomaciter option is used. This improves file security and
8048 only older versions of MSIE (4.0 for example) need it.
8051 *) Honor the no-xxx Configure options when creating .DEF files.
8054 *) Add PKCS#10 attributes to field table: challengePassword,
8055 unstructuredName and unstructuredAddress. These are taken from
8056 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
8057 international characters are used.
8059 More changes to X509_ATTRIBUTE code: allow the setting of types
8060 based on strings. Remove the 'loc' parameter when adding
8061 attributes because these will be a SET OF encoding which is sorted
8065 *) Initial changes to the 'req' utility to allow request generation
8066 automation. This will allow an application to just generate a template
8067 file containing all the field values and have req construct the
8070 Initial support for X509_ATTRIBUTE handling. Stacks of these are
8071 used all over the place including certificate requests and PKCS#7
8072 structures. They are currently handled manually where necessary with
8073 some primitive wrappers for PKCS#7. The new functions behave in a
8074 manner analogous to the X509 extension functions: they allow
8075 attributes to be looked up by NID and added.
8077 Later something similar to the X509V3 code would be desirable to
8078 automatically handle the encoding, decoding and printing of the
8079 more complex types. The string types like challengePassword can
8080 be handled by the string table functions.
8082 Also modified the multi byte string table handling. Now there is
8083 a 'global mask' which masks out certain types. The table itself
8084 can use the flag STABLE_NO_MASK to ignore the mask setting: this
8085 is useful when for example there is only one permissible type
8086 (as in countryName) and using the mask might result in no valid
8090 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
8091 SSL_get_peer_finished to allow applications to obtain the latest
8092 Finished messages sent to the peer or expected from the peer,
8093 respectively. (SSL_get_peer_finished is usually the Finished message
8094 actually received from the peer, otherwise the protocol will be aborted.)
8096 As the Finished message are message digests of the complete handshake
8097 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
8098 be used for external authentication procedures when the authentication
8099 provided by SSL/TLS is not desired or is not enough.
8102 *) Enhanced support for Alpha Linux is added. Now ./config checks if
8103 the host supports BWX extension and if Compaq C is present on the
8104 $PATH. Just exploiting of the BWX extension results in 20-30%
8105 performance kick for some algorithms, e.g. DES and RC4 to mention
8106 a couple. Compaq C in turn generates ~20% faster code for MD5 and
8110 *) Add support for MS "fast SGC". This is arguably a violation of the
8111 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
8112 weak crypto and after checking the certificate is SGC a second one
8113 with strong crypto. MS SGC stops the first handshake after receiving
8114 the server certificate message and sends a second client hello. Since
8115 a server will typically do all the time consuming operations before
8116 expecting any further messages from the client (server key exchange
8117 is the most expensive) there is little difference between the two.
8119 To get OpenSSL to support MS SGC we have to permit a second client
8120 hello message after we have sent server done. In addition we have to
8121 reset the MAC if we do get this second client hello.
8124 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
8125 if a DER encoded private key is RSA or DSA traditional format. Changed
8126 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
8127 format DER encoded private key. Newer code should use PKCS#8 format which
8128 has the key type encoded in the ASN1 structure. Added DER private key
8129 support to pkcs8 application.
8132 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
8133 ciphersuites has been selected (as required by the SSL 3/TLS 1
8134 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
8135 is set, we interpret this as a request to violate the specification
8136 (the worst that can happen is a handshake failure, and 'correct'
8137 behaviour would result in a handshake failure anyway).
8140 *) In SSL_CTX_add_session, take into account that there might be multiple
8141 SSL_SESSION structures with the same session ID (e.g. when two threads
8142 concurrently obtain them from an external cache).
8143 The internal cache can handle only one SSL_SESSION with a given ID,
8144 so if there's a conflict, we now throw out the old one to achieve
8148 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
8149 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
8150 some routines that use cipher OIDs: some ciphers do not have OIDs
8151 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
8155 *) Simplify the trust setting structure and code. Now we just have
8156 two sequences of OIDs for trusted and rejected settings. These will
8157 typically have values the same as the extended key usage extension
8158 and any application specific purposes.
8160 The trust checking code now has a default behaviour: it will just
8161 check for an object with the same NID as the passed id. Functions can
8162 be provided to override either the default behaviour or the behaviour
8163 for a given id. SSL client, server and email already have functions
8164 in place for compatibility: they check the NID and also return "trusted"
8165 if the certificate is self signed.
8168 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
8169 traditional format into an EVP_PKEY structure.
8172 *) Add a password callback function PEM_cb() which either prompts for
8173 a password if usr_data is NULL or otherwise assumes it is a null
8174 terminated password. Allow passwords to be passed on command line
8175 environment or config files in a few more utilities.
8178 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
8179 keys. Add some short names for PKCS#8 PBE algorithms and allow them
8180 to be specified on the command line for the pkcs8 and pkcs12 utilities.
8181 Update documentation.
8184 *) Support for ASN1 "NULL" type. This could be handled before by using
8185 ASN1_TYPE but there wasn't any function that would try to read a NULL
8186 and produce an error if it couldn't. For compatibility we also have
8187 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
8188 don't allocate anything because they don't need to.
8191 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
8193 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
8195 *) Rebuild of the memory allocation routines used by OpenSSL code and
8196 possibly others as well. The purpose is to make an interface that
8197 provide hooks so anyone can build a separate set of allocation and
8198 deallocation routines to be used by OpenSSL, for example memory
8199 pool implementations, or something else, which was previously hard
8200 since Malloc(), Realloc() and Free() were defined as macros having
8201 the values malloc, realloc and free, respectively (except for Win32
8202 compilations). The same is provided for memory debugging code.
8203 OpenSSL already comes with functionality to find memory leaks, but
8204 this gives people a chance to debug other memory problems.
8206 With these changes, a new set of functions and macros have appeared:
8208 CRYPTO_set_mem_debug_functions() [F]
8209 CRYPTO_get_mem_debug_functions() [F]
8210 CRYPTO_dbg_set_options() [F]
8211 CRYPTO_dbg_get_options() [F]
8212 CRYPTO_malloc_debug_init() [M]
8214 The memory debug functions are NULL by default, unless the library
8215 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
8216 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
8217 gives the standard debugging functions that come with OpenSSL) or
8218 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
8219 provided by the library user) must be used. When the standard
8220 debugging functions are used, CRYPTO_dbg_set_options can be used to
8221 request additional information:
8222 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
8223 the CRYPTO_MDEBUG_xxx macro when compiling the library.
8225 Also, things like CRYPTO_set_mem_functions will always give the
8226 expected result (the new set of functions is used for allocation
8227 and deallocation) at all times, regardless of platform and compiler
8230 To finish it up, some functions that were never use in any other
8231 way than through macros have a new API and new semantic:
8234 CRYPTO_dbg_realloc()
8237 All macros of value have retained their old syntax.
8238 [Richard Levitte and Bodo Moeller]
8240 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
8241 ordering of SMIMECapabilities wasn't in "strength order" and there
8242 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
8246 *) Some ASN1 types with illegal zero length encoding (INTEGER,
8247 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
8248 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
8250 *) Merge in my S/MIME library for OpenSSL. This provides a simple
8251 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
8252 functionality to handle multipart/signed properly) and a utility
8253 called 'smime' to call all this stuff. This is based on code I
8254 originally wrote for Celo who have kindly allowed it to be
8255 included in OpenSSL.
8258 *) Add variants des_set_key_checked and des_set_key_unchecked of
8259 des_set_key (aka des_key_sched). Global variable des_check_key
8260 decides which of these is called by des_set_key; this way
8261 des_check_key behaves as it always did, but applications and
8262 the library itself, which was buggy for des_check_key == 1,
8263 have a cleaner way to pick the version they need.
8266 *) New function PKCS12_newpass() which changes the password of a
8270 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
8271 dynamic mix. In both cases the ids can be used as an index into the
8272 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
8273 functions so they accept a list of the field values and the
8274 application doesn't need to directly manipulate the X509_TRUST
8278 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
8282 *) Modify the way the V3 extension code looks up extensions. This now
8283 works in a similar way to the object code: we have some "standard"
8284 extensions in a static table which is searched with OBJ_bsearch()
8285 and the application can add dynamic ones if needed. The file
8286 crypto/x509v3/ext_dat.h now has the info: this file needs to be
8287 updated whenever a new extension is added to the core code and kept
8288 in ext_nid order. There is a simple program 'tabtest.c' which checks
8289 this. New extensions are not added too often so this file can readily
8290 be maintained manually.
8292 There are two big advantages in doing things this way. The extensions
8293 can be looked up immediately and no longer need to be "added" using
8294 X509V3_add_standard_extensions(): this function now does nothing.
8295 [Side note: I get *lots* of email saying the extension code doesn't
8296 work because people forget to call this function]
8297 Also no dynamic allocation is done unless new extensions are added:
8298 so if we don't add custom extensions there is no need to call
8299 X509V3_EXT_cleanup().
8302 *) Modify enc utility's salting as follows: make salting the default. Add a
8303 magic header, so unsalted files fail gracefully instead of just decrypting
8304 to garbage. This is because not salting is a big security hole, so people
8305 should be discouraged from doing it.
8308 *) Fixes and enhancements to the 'x509' utility. It allowed a message
8309 digest to be passed on the command line but it only used this
8310 parameter when signing a certificate. Modified so all relevant
8311 operations are affected by the digest parameter including the
8312 -fingerprint and -x509toreq options. Also -x509toreq choked if a
8313 DSA key was used because it didn't fix the digest.
8316 *) Initial certificate chain verify code. Currently tests the untrusted
8317 certificates for consistency with the verify purpose (which is set
8318 when the X509_STORE_CTX structure is set up) and checks the pathlength.
8320 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
8321 this is because it will reject chains with invalid extensions whereas
8322 every previous version of OpenSSL and SSLeay made no checks at all.
8324 Trust code: checks the root CA for the relevant trust settings. Trust
8325 settings have an initial value consistent with the verify purpose: e.g.
8326 if the verify purpose is for SSL client use it expects the CA to be
8327 trusted for SSL client use. However the default value can be changed to
8328 permit custom trust settings: one example of this would be to only trust
8329 certificates from a specific "secure" set of CAs.
8331 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
8332 which should be used for version portability: especially since the
8333 verify structure is likely to change more often now.
8335 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
8336 to set them. If not set then assume SSL clients will verify SSL servers
8339 Two new options to the verify program: -untrusted allows a set of
8340 untrusted certificates to be passed in and -purpose which sets the
8341 intended purpose of the certificate. If a purpose is set then the
8342 new chain verify code is used to check extension consistency.
8345 *) Support for the authority information access extension.
8348 *) Modify RSA and DSA PEM read routines to transparently handle
8349 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
8350 public keys in a format compatible with certificate
8351 SubjectPublicKeyInfo structures. Unfortunately there were already
8352 functions called *_PublicKey_* which used various odd formats so
8353 these are retained for compatibility: however the DSA variants were
8354 never in a public release so they have been deleted. Changed dsa/rsa
8355 utilities to handle the new format: note no releases ever handled public
8356 keys so we should be OK.
8358 The primary motivation for this change is to avoid the same fiasco
8359 that dogs private keys: there are several incompatible private key
8360 formats some of which are standard and some OpenSSL specific and
8361 require various evil hacks to allow partial transparent handling and
8362 even then it doesn't work with DER formats. Given the option anything
8363 other than PKCS#8 should be dumped: but the other formats have to
8364 stay in the name of compatibility.
8366 With public keys and the benefit of hindsight one standard format
8367 is used which works with EVP_PKEY, RSA or DSA structures: though
8368 it clearly returns an error if you try to read the wrong kind of key.
8370 Added a -pubkey option to the 'x509' utility to output the public key.
8371 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
8372 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
8373 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
8374 that do the same as the EVP_PKEY_assign_*() except they up the
8375 reference count of the added key (they don't "swallow" the
8379 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
8380 CRLs would fail if the file contained no certificates or no CRLs:
8381 added a new function to read in both types and return the number
8382 read: this means that if none are read it will be an error. The
8383 DER versions of the certificate and CRL reader would always fail
8384 because it isn't possible to mix certificates and CRLs in DER format
8385 without choking one or the other routine. Changed this to just read
8386 a certificate: this is the best we can do. Also modified the code
8387 in apps/verify.c to take notice of return codes: it was previously
8388 attempting to read in certificates from NULL pointers and ignoring
8389 any errors: this is one reason why the cert and CRL reader seemed
8390 to work. It doesn't check return codes from the default certificate
8391 routines: these may well fail if the certificates aren't installed.
8394 *) Code to support otherName option in GeneralName.
8397 *) First update to verify code. Change the verify utility
8398 so it warns if it is passed a self signed certificate:
8399 for consistency with the normal behaviour. X509_verify
8400 has been modified to it will now verify a self signed
8401 certificate if *exactly* the same certificate appears
8402 in the store: it was previously impossible to trust a
8403 single self signed certificate. This means that:
8404 openssl verify ss.pem
8405 now gives a warning about a self signed certificate but
8406 openssl verify -CAfile ss.pem ss.pem
8410 *) For servers, store verify_result in SSL_SESSION data structure
8411 (and add it to external session representation).
8412 This is needed when client certificate verifications fails,
8413 but an application-provided verification callback (set by
8414 SSL_CTX_set_cert_verify_callback) allows accepting the session
8415 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
8416 but returns 1): When the session is reused, we have to set
8417 ssl->verify_result to the appropriate error code to avoid
8419 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
8421 *) Fix a bug in the new PKCS#7 code: it didn't consider the
8422 case in PKCS7_dataInit() where the signed PKCS7 structure
8423 didn't contain any existing data because it was being created.
8424 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
8426 *) Add a salt to the key derivation routines in enc.c. This
8427 forms the first 8 bytes of the encrypted file. Also add a
8428 -S option to allow a salt to be input on the command line.
8431 *) New function X509_cmp(). Oddly enough there wasn't a function
8432 to compare two certificates. We do this by working out the SHA1
8433 hash and comparing that. X509_cmp() will be needed by the trust
8437 *) SSL_get1_session() is like SSL_get_session(), but increments
8438 the reference count in the SSL_SESSION returned.
8439 [Geoff Thorpe <geoff@eu.c2.net>]
8441 *) Fix for 'req': it was adding a null to request attributes.
8442 Also change the X509_LOOKUP and X509_INFO code to handle
8443 certificate auxiliary information.
8446 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
8450 *) Add the possibility to add extra information to the memory leak
8451 detecting output, to form tracebacks, showing from where each
8452 allocation was originated: CRYPTO_push_info("constant string") adds
8453 the string plus current file name and line number to a per-thread
8454 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
8455 is like calling CYRPTO_pop_info() until the stack is empty.
8456 Also updated memory leak detection code to be multi-thread-safe.
8459 *) Add options -text and -noout to pkcs7 utility and delete the
8460 encryption options which never did anything. Update docs.
8463 *) Add options to some of the utilities to allow the pass phrase
8464 to be included on either the command line (not recommended on
8465 OSes like Unix) or read from the environment. Update the
8466 manpages and fix a few bugs.
8469 *) Add a few manpages for some of the openssl commands.
8472 *) Fix the -revoke option in ca. It was freeing up memory twice,
8473 leaking and not finding already revoked certificates.
8476 *) Extensive changes to support certificate auxiliary information.
8477 This involves the use of X509_CERT_AUX structure and X509_AUX
8478 functions. An X509_AUX function such as PEM_read_X509_AUX()
8479 can still read in a certificate file in the usual way but it
8480 will also read in any additional "auxiliary information". By
8481 doing things this way a fair degree of compatibility can be
8482 retained: existing certificates can have this information added
8483 using the new 'x509' options.
8485 Current auxiliary information includes an "alias" and some trust
8486 settings. The trust settings will ultimately be used in enhanced
8487 certificate chain verification routines: currently a certificate
8488 can only be trusted if it is self signed and then it is trusted
8492 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
8493 The problem was that one of the replacement routines had not been working
8494 since SSLeay releases. For now the offending routine has been replaced
8495 with non-optimised assembler. Even so, this now gives around 95%
8496 performance improvement for 1024 bit RSA signs.
8499 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
8500 handling. Most clients have the effective key size in bits equal to
8501 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
8502 A few however don't do this and instead use the size of the decrypted key
8503 to determine the RC2 key length and the AlgorithmIdentifier to determine
8504 the effective key length. In this case the effective key length can still
8505 be 40 bits but the key length can be 168 bits for example. This is fixed
8506 by manually forcing an RC2 key into the EVP_PKEY structure because the
8507 EVP code can't currently handle unusual RC2 key sizes: it always assumes
8508 the key length and effective key length are equal.
8511 *) Add a bunch of functions that should simplify the creation of
8512 X509_NAME structures. Now you should be able to do:
8513 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
8514 and have it automatically work out the correct field type and fill in
8515 the structures. The more adventurous can try:
8516 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
8517 and it will (hopefully) work out the correct multibyte encoding.
8520 *) Change the 'req' utility to use the new field handling and multibyte
8521 copy routines. Before the DN field creation was handled in an ad hoc
8522 way in req, ca, and x509 which was rather broken and didn't support
8523 BMPStrings or UTF8Strings. Since some software doesn't implement
8524 BMPStrings or UTF8Strings yet, they can be enabled using the config file
8525 using the dirstring_type option. See the new comment in the default
8526 openssl.cnf for more info.
8529 *) Make crypto/rand/md_rand.c more robust:
8530 - Assure unique random numbers after fork().
8531 - Make sure that concurrent threads access the global counter and
8532 md serializably so that we never lose entropy in them
8533 or use exactly the same state in multiple threads.
8534 Access to the large state is not always serializable because
8535 the additional locking could be a performance killer, and
8536 md should be large enough anyway.
8539 *) New file apps/app_rand.c with commonly needed functionality
8540 for handling the random seed file.
8542 Use the random seed file in some applications that previously did not:
8544 dsaparam -genkey (which also ignored its '-rand' option),
8547 x509 (when signing).
8548 Except on systems with /dev/urandom, it is crucial to have a random
8549 seed file at least for key creation, DSA signing, and for DH exchanges;
8550 for RSA signatures we could do without one.
8552 gendh and gendsa (unlike genrsa) used to read only the first byte
8553 of each file listed in the '-rand' option. The function as previously
8554 found in genrsa is now in app_rand.c and is used by all programs
8555 that support '-rand'.
8558 *) In RAND_write_file, use mode 0600 for creating files;
8559 don't just chmod when it may be too late.
8562 *) Report an error from X509_STORE_load_locations
8563 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
8566 *) New function ASN1_mbstring_copy() this copies a string in either
8567 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
8568 into an ASN1_STRING type. A mask of permissible types is passed
8569 and it chooses the "minimal" type to use or an error if not type
8573 *) Add function equivalents to the various macros in asn1.h. The old
8574 macros are retained with an M_ prefix. Code inside the library can
8575 use the M_ macros. External code (including the openssl utility)
8576 should *NOT* in order to be "shared library friendly".
8579 *) Add various functions that can check a certificate's extensions
8580 to see if it usable for various purposes such as SSL client,
8581 server or S/MIME and CAs of these types. This is currently
8582 VERY EXPERIMENTAL but will ultimately be used for certificate chain
8583 verification. Also added a -purpose flag to x509 utility to
8584 print out all the purposes.
8587 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
8591 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
8592 for, obtain and decode and extension and obtain its critical flag.
8593 This allows all the necessary extension code to be handled in a
8594 single function call.
8597 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
8598 platforms. See crypto/rc4/rc4_enc.c for further details.
8601 *) New -noout option to asn1parse. This causes no output to be produced
8602 its main use is when combined with -strparse and -out to extract data
8603 from a file (which may not be in ASN.1 format).
8606 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
8607 when producing the local key id.
8608 [Richard Levitte <levitte@stacken.kth.se>]
8610 *) New option -dhparam in s_server. This allows a DH parameter file to be
8611 stated explicitly. If it is not stated then it tries the first server
8612 certificate file. The previous behaviour hard coded the filename
8616 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
8617 a public key to be input or output. For example:
8618 openssl rsa -in key.pem -pubout -out pubkey.pem
8619 Also added necessary DSA public key functions to handle this.
8622 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
8623 in the message. This was handled by allowing
8624 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8625 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8627 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8628 to the end of the strings whereas this didn't. This would cause problems
8629 if strings read with d2i_ASN1_bytes() were later modified.
8630 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8632 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8633 data and it contains EOF it will end up returning an error. This is
8634 caused by input 46 bytes long. The cause is due to the way base64
8635 BIOs find the start of base64 encoded data. They do this by trying a
8636 trial decode on each line until they find one that works. When they
8637 do a flag is set and it starts again knowing it can pass all the
8638 data directly through the decoder. Unfortunately it doesn't reset
8639 the context it uses. This means that if EOF is reached an attempt
8640 is made to pass two EOFs through the context and this causes the
8641 resulting error. This can also cause other problems as well. As is
8642 usual with these problems it takes *ages* to find and the fix is
8643 trivial: move one line.
8644 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8646 *) Ugly workaround to get s_client and s_server working under Windows. The
8647 old code wouldn't work because it needed to select() on sockets and the
8648 tty (for keypresses and to see if data could be written). Win32 only
8649 supports select() on sockets so we select() with a 1s timeout on the
8650 sockets and then see if any characters are waiting to be read, if none
8651 are present then we retry, we also assume we can always write data to
8652 the tty. This isn't nice because the code then blocks until we've
8653 received a complete line of data and it is effectively polling the
8654 keyboard at 1s intervals: however it's quite a bit better than not
8655 working at all :-) A dedicated Windows application might handle this
8656 with an event loop for example.
8659 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
8660 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
8661 will be called when RSA_sign() and RSA_verify() are used. This is useful
8662 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
8663 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
8664 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
8665 This necessitated the support of an extra signature type NID_md5_sha1
8666 for SSL signatures and modifications to the SSL library to use it instead
8667 of calling RSA_public_decrypt() and RSA_private_encrypt().
8670 *) Add new -verify -CAfile and -CApath options to the crl program, these
8671 will lookup a CRL issuers certificate and verify the signature in a
8672 similar way to the verify program. Tidy up the crl program so it
8673 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
8674 less strict. It will now permit CRL extensions even if it is not
8675 a V2 CRL: this will allow it to tolerate some broken CRLs.
8678 *) Initialize all non-automatic variables each time one of the openssl
8679 sub-programs is started (this is necessary as they may be started
8680 multiple times from the "OpenSSL>" prompt).
8681 [Lennart Bang, Bodo Moeller]
8683 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
8684 removing all other RSA functionality (this is what NO_RSA does). This
8685 is so (for example) those in the US can disable those operations covered
8686 by the RSA patent while allowing storage and parsing of RSA keys and RSA
8690 *) Non-copying interface to BIO pairs.
8691 (still largely untested)
8694 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
8695 ASCII string. This was handled independently in various places before.
8698 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
8699 UTF8 strings a character at a time.
8702 *) Use client_version from client hello to select the protocol
8703 (s23_srvr.c) and for RSA client key exchange verification
8704 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
8707 *) Add various utility functions to handle SPKACs, these were previously
8708 handled by poking round in the structure internals. Added new function
8709 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
8710 print, verify and generate SPKACs. Based on an original idea from
8711 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
8714 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
8717 *) Allow the config file extension section to be overwritten on the
8718 command line. Based on an original idea from Massimiliano Pala
8719 <madwolf@comune.modena.it>. The new option is called -extensions
8720 and can be applied to ca, req and x509. Also -reqexts to override
8721 the request extensions in req and -crlexts to override the crl extensions
8725 *) Add new feature to the SPKAC handling in ca. Now you can include
8726 the same field multiple times by preceding it by "XXXX." for example:
8729 this is the same syntax as used in the req config file.
8732 *) Allow certificate extensions to be added to certificate requests. These
8733 are specified in a 'req_extensions' option of the req section of the
8734 config file. They can be printed out with the -text option to req but
8735 are otherwise ignored at present.
8738 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
8739 data read consists of only the final block it would not decrypted because
8740 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
8741 A misplaced 'break' also meant the decrypted final block might not be
8742 copied until the next read.
8745 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
8746 a few extra parameters to the DH structure: these will be useful if
8747 for example we want the value of 'q' or implement X9.42 DH.
8750 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
8751 provides hooks that allow the default DSA functions or functions on a
8752 "per key" basis to be replaced. This allows hardware acceleration and
8753 hardware key storage to be handled without major modification to the
8754 library. Also added low level modexp hooks and CRYPTO_EX structure and
8755 associated functions.
8758 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
8759 as "read only": it can't be written to and the buffer it points to will
8760 not be freed. Reading from a read only BIO is much more efficient than
8761 a normal memory BIO. This was added because there are several times when
8762 an area of memory needs to be read from a BIO. The previous method was
8763 to create a memory BIO and write the data to it, this results in two
8764 copies of the data and an O(n^2) reading algorithm. There is a new
8765 function BIO_new_mem_buf() which creates a read only memory BIO from
8766 an area of memory. Also modified the PKCS#7 routines to use read only
8770 *) Bugfix: ssl23_get_client_hello did not work properly when called in
8771 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
8772 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
8773 but a retry condition occured while trying to read the rest.
8776 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
8777 NID_pkcs7_encrypted by default: this was wrong since this should almost
8778 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
8779 the encrypted data type: this is a more sensible place to put it and it
8780 allows the PKCS#12 code to be tidied up that duplicated this
8784 *) Changed obj_dat.pl script so it takes its input and output files on
8785 the command line. This should avoid shell escape redirection problems
8789 *) Initial support for certificate extension requests, these are included
8790 in things like Xenroll certificate requests. Included functions to allow
8791 extensions to be obtained and added.
8794 *) -crlf option to s_client and s_server for sending newlines as
8795 CRLF (as required by many protocols).
8798 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
8800 *) Install libRSAglue.a when OpenSSL is built with RSAref.
8801 [Ralf S. Engelschall]
8803 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
8804 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
8806 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
8810 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
8811 DH parameters/keys (q is lost during that conversion, but the resulting
8812 DH parameters contain its length).
8814 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
8815 much faster than DH_generate_parameters (which creates parameters
8816 where p = 2*q + 1), and also the smaller q makes DH computations
8817 much more efficient (160-bit exponentiation instead of 1024-bit
8818 exponentiation); so this provides a convenient way to support DHE
8819 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
8820 utter importance to use
8821 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8823 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8824 when such DH parameters are used, because otherwise small subgroup
8825 attacks may become possible!
8828 *) Avoid memory leak in i2d_DHparams.
8831 *) Allow the -k option to be used more than once in the enc program:
8832 this allows the same encrypted message to be read by multiple recipients.
8835 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
8836 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
8837 it will always use the numerical form of the OID, even if it has a short
8841 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
8842 method only got called if p,q,dmp1,dmq1,iqmp components were present,
8843 otherwise bn_mod_exp was called. In the case of hardware keys for example
8844 no private key components need be present and it might store extra data
8845 in the RSA structure, which cannot be accessed from bn_mod_exp.
8846 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
8847 private key operations.
8850 *) Added support for SPARC Linux.
8853 *) pem_password_cb function type incompatibly changed from
8854 typedef int pem_password_cb(char *buf, int size, int rwflag);
8856 ....(char *buf, int size, int rwflag, void *userdata);
8857 so that applications can pass data to their callbacks:
8858 The PEM[_ASN1]_{read,write}... functions and macros now take an
8859 additional void * argument, which is just handed through whenever
8860 the password callback is called.
8861 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
8863 New function SSL_CTX_set_default_passwd_cb_userdata.
8865 Compatibility note: As many C implementations push function arguments
8866 onto the stack in reverse order, the new library version is likely to
8867 interoperate with programs that have been compiled with the old
8868 pem_password_cb definition (PEM_whatever takes some data that
8869 happens to be on the stack as its last argument, and the callback
8870 just ignores this garbage); but there is no guarantee whatsoever that
8873 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
8874 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
8875 problems not only on Windows, but also on some Unix platforms.
8876 To avoid problematic command lines, these definitions are now in an
8877 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
8878 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
8881 *) MIPS III/IV assembler module is reimplemented.
8884 *) More DES library cleanups: remove references to srand/rand and
8885 delete an unused file.
8888 *) Add support for the the free Netwide assembler (NASM) under Win32,
8889 since not many people have MASM (ml) and it can be hard to obtain.
8890 This is currently experimental but it seems to work OK and pass all
8891 the tests. Check out INSTALL.W32 for info.
8894 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
8895 without temporary keys kept an extra copy of the server key,
8896 and connections with temporary keys did not free everything in case
8900 *) New function RSA_check_key and new openssl rsa option -check
8901 for verifying the consistency of RSA keys.
8902 [Ulf Moeller, Bodo Moeller]
8904 *) Various changes to make Win32 compile work:
8905 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
8906 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
8907 comparison" warnings.
8908 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
8911 *) Add a debugging option to PKCS#5 v2 key generation function: when
8912 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
8913 derived keys are printed to stderr.
8916 *) Copy the flags in ASN1_STRING_dup().
8917 [Roman E. Pavlov <pre@mo.msk.ru>]
8919 *) The x509 application mishandled signing requests containing DSA
8920 keys when the signing key was also DSA and the parameters didn't match.
8922 It was supposed to omit the parameters when they matched the signing key:
8923 the verifying software was then supposed to automatically use the CA's
8924 parameters if they were absent from the end user certificate.
8926 Omitting parameters is no longer recommended. The test was also
8927 the wrong way round! This was probably due to unusual behaviour in
8928 EVP_cmp_parameters() which returns 1 if the parameters match.
8929 This meant that parameters were omitted when they *didn't* match and
8930 the certificate was useless. Certificates signed with 'ca' didn't have
8932 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
8934 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
8935 The interface is as follows:
8936 Applications can use
8937 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
8938 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
8939 "off" is now the default.
8940 The library internally uses
8941 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
8942 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
8943 to disable memory-checking temporarily.
8945 Some inconsistent states that previously were possible (and were
8946 even the default) are now avoided.
8948 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
8949 with each memory chunk allocated; this is occasionally more helpful
8950 than just having a counter.
8952 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
8954 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
8958 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
8959 which largely parallels "options", but is for changing API behaviour,
8960 whereas "options" are about protocol behaviour.
8961 Initial "mode" flags are:
8963 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
8964 a single record has been written.
8965 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
8966 retries use the same buffer location.
8967 (But all of the contents must be
8971 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
8974 *) Fix problems with no-hmac etc.
8975 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
8977 *) New functions RSA_get_default_method(), RSA_set_method() and
8978 RSA_get_method(). These allows replacement of RSA_METHODs without having
8979 to mess around with the internals of an RSA structure.
8982 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
8983 Also really enable memory leak checks in openssl.c and in some
8985 [Chad C. Mulligan, Bodo Moeller]
8987 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
8988 up the length of negative integers. This has now been simplified to just
8989 store the length when it is first determined and use it later, rather
8990 than trying to keep track of where data is copied and updating it to
8992 [Steve Henson, reported by Brien Wheeler
8993 <bwheeler@authentica-security.com>]
8995 *) Add a new function PKCS7_signatureVerify. This allows the verification
8996 of a PKCS#7 signature but with the signing certificate passed to the
8997 function itself. This contrasts with PKCS7_dataVerify which assumes the
8998 certificate is present in the PKCS#7 structure. This isn't always the
8999 case: certificates can be omitted from a PKCS#7 structure and be
9000 distributed by "out of band" means (such as a certificate database).
9003 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
9004 function prototypes in pem.h, also change util/mkdef.pl to add the
9005 necessary function names.
9008 *) mk1mf.pl (used by Windows builds) did not properly read the
9009 options set by Configure in the top level Makefile, and Configure
9010 was not even able to write more than one option correctly.
9011 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
9014 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
9015 file to be loaded from a BIO or FILE pointer. The BIO version will
9016 for example allow memory BIOs to contain config info.
9019 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
9020 Whoever hopes to achieve shared-library compatibility across versions
9021 must use this, not the compile-time macro.
9022 (Exercise 0.9.4: Which is the minimum library version required by
9024 Note: All this applies only to multi-threaded programs, others don't
9028 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
9029 through a BIO pair triggered the default case, i.e.
9030 SSLerr(...,SSL_R_UNKNOWN_STATE).
9033 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
9034 can use the SSL library even if none of the specific BIOs is
9038 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
9039 for the encoded length.
9040 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
9042 *) Add initial documentation of the X509V3 functions.
9045 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
9046 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
9047 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
9048 secure PKCS#8 private key format with a high iteration count.
9051 *) Fix determination of Perl interpreter: A perl or perl5
9052 _directory_ in $PATH was also accepted as the interpreter.
9053 [Ralf S. Engelschall]
9055 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
9056 wrong with it but it was very old and did things like calling
9057 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
9061 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
9062 to use the new extension code.
9065 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
9066 with macros. This should make it easier to change their form, add extra
9067 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
9071 *) Add to configuration table a new entry that can specify an alternative
9072 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
9073 according to Mark Crispin <MRC@Panda.COM>.
9077 *) DES CBC did not update the IV. Weird.
9080 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
9081 Changing the behaviour of the former might break existing programs --
9082 where IV updating is needed, des_ncbc_encrypt can be used.
9085 *) When bntest is run from "make test" it drives bc to check its
9086 calculations, as well as internally checking them. If an internal check
9087 fails, it needs to cause bc to give a non-zero result or make test carries
9088 on without noticing the failure. Fixed.
9091 *) DES library cleanups.
9094 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
9095 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
9096 ciphers. NOTE: although the key derivation function has been verified
9097 against some published test vectors it has not been extensively tested
9098 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
9102 *) Instead of "mkdir -p", which is not fully portable, use new
9103 Perl script "util/mkdir-p.pl".
9106 *) Rewrite the way password based encryption (PBE) is handled. It used to
9107 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
9108 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
9109 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
9110 the 'parameter' field of the AlgorithmIdentifier is passed to the
9111 underlying key generation function so it must do its own ASN1 parsing.
9112 This has also changed the EVP_PBE_CipherInit() function which now has a
9113 'parameter' argument instead of literal salt and iteration count values
9114 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
9117 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
9118 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
9119 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
9120 KEY" because this clashed with PKCS#8 unencrypted string. Since this
9121 value was just used as a "magic string" and not used directly its
9122 value doesn't matter.
9125 *) Introduce some semblance of const correctness to BN. Shame C doesn't
9129 *) "linux-sparc64" configuration (ultrapenguin).
9130 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
9131 "linux-sparc" configuration.
9132 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
9134 *) config now generates no-xxx options for missing ciphers.
9137 *) Support the EBCDIC character set (work in progress).
9138 File ebcdic.c not yet included because it has a different license.
9139 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9141 *) Support BS2000/OSD-POSIX.
9142 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9144 *) Make callbacks for key generation use void * instead of char *.
9147 *) Make S/MIME samples compile (not yet tested).
9150 *) Additional typesafe stacks.
9153 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
9157 Changes between 0.9.3 and 0.9.3a [29 May 1999]
9159 *) New configuration variant "sco5-gcc".
9161 *) Updated some demos.
9162 [Sean O Riordain, Wade Scholine]
9164 *) Add missing BIO_free at exit of pkcs12 application.
9167 *) Fix memory leak in conf.c.
9170 *) Updates for Win32 to assembler version of MD5.
9173 *) Set #! path to perl in apps/der_chop to where we found it
9174 instead of using a fixed path.
9177 *) SHA library changes for irix64-mips4-cc.
9180 *) Improvements for VMS support.
9184 Changes between 0.9.2b and 0.9.3 [24 May 1999]
9186 *) Bignum library bug fix. IRIX 6 passes "make test" now!
9187 This also avoids the problems with SC4.2 and unpatched SC5.
9188 [Andy Polyakov <appro@fy.chalmers.se>]
9190 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
9191 These are required because of the typesafe stack would otherwise break
9192 existing code. If old code used a structure member which used to be STACK
9193 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
9194 sk_num or sk_value it would produce an error because the num, data members
9195 are not present in STACK_OF. Now it just produces a warning. sk_set
9196 replaces the old method of assigning a value to sk_value
9197 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
9198 that does this will no longer work (and should use sk_set instead) but
9199 this could be regarded as a "questionable" behaviour anyway.
9202 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
9203 correctly handle encrypted S/MIME data.
9206 *) Change type of various DES function arguments from des_cblock
9207 (which means, in function argument declarations, pointer to char)
9208 to des_cblock * (meaning pointer to array with 8 char elements),
9209 which allows the compiler to do more typechecking; it was like
9210 that back in SSLeay, but with lots of ugly casts.
9212 Introduce new type const_des_cblock.
9215 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
9216 problems: find RecipientInfo structure that matches recipient certificate
9217 and initialise the ASN1 structures properly based on passed cipher.
9220 *) Belatedly make the BN tests actually check the results.
9223 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
9224 to and from BNs: it was completely broken. New compilation option
9225 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
9226 key elements as negative integers.
9229 *) Reorganize and speed up MD5.
9230 [Andy Polyakov <appro@fy.chalmers.se>]
9233 [Richard Levitte <richard@levitte.org>]
9235 *) New option -out to asn1parse to allow the parsed structure to be
9236 output to a file. This is most useful when combined with the -strparse
9237 option to examine the output of things like OCTET STRINGS.
9240 *) Make SSL library a little more fool-proof by not requiring any longer
9241 that SSL_set_{accept,connect}_state be called before
9242 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
9243 in many applications because usually everything *appeared* to work as
9244 intended anyway -- now it really works as intended).
9247 *) Move openssl.cnf out of lib/.
9250 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
9251 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
9252 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
9253 [Ralf S. Engelschall]
9255 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
9256 handle PKCS#7 enveloped data properly.
9257 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
9259 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
9260 copying pointers. The cert_st handling is changed by this in
9261 various ways (and thus what used to be known as ctx->default_cert
9262 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
9263 any longer when s->cert does not give us what we need).
9264 ssl_cert_instantiate becomes obsolete by this change.
9265 As soon as we've got the new code right (possibly it already is?),
9266 we have solved a couple of bugs of the earlier code where s->cert
9267 was used as if it could not have been shared with other SSL structures.
9269 Note that using the SSL API in certain dirty ways now will result
9270 in different behaviour than observed with earlier library versions:
9271 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
9272 does not influence s as it used to.
9274 In order to clean up things more thoroughly, inside SSL_SESSION
9275 we don't use CERT any longer, but a new structure SESS_CERT
9276 that holds per-session data (if available); currently, this is
9277 the peer's certificate chain and, for clients, the server's certificate
9278 and temporary key. CERT holds only those values that can have
9279 meaningful defaults in an SSL_CTX.
9282 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
9283 from the internal representation. Various PKCS#7 fixes: remove some
9284 evil casts and set the enc_dig_alg field properly based on the signing
9288 *) Allow PKCS#12 password to be set from the command line or the
9289 environment. Let 'ca' get its config file name from the environment
9290 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
9294 *) Allow certificate policies extension to use an IA5STRING for the
9295 organization field. This is contrary to the PKIX definition but
9296 VeriSign uses it and IE5 only recognises this form. Document 'x509'
9300 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
9301 without disallowing inline assembler and the like for non-pedantic builds.
9304 *) Support Borland C++ builder.
9305 [Janez Jere <jj@void.si>, modified by Ulf Möller]
9310 *) SHA-1 cleanups and performance enhancements.
9311 [Andy Polyakov <appro@fy.chalmers.se>]
9313 *) Sparc v8plus assembler for the bignum library.
9314 [Andy Polyakov <appro@fy.chalmers.se>]
9316 *) Accept any -xxx and +xxx compiler options in Configure.
9319 *) Update HPUX configuration.
9322 *) Add missing sk_<type>_unshift() function to safestack.h
9323 [Ralf S. Engelschall]
9325 *) New function SSL_CTX_use_certificate_chain_file that sets the
9326 "extra_cert"s in addition to the certificate. (This makes sense
9327 only for "PEM" format files, as chains as a whole are not
9331 *) Support verify_depth from the SSL API.
9332 x509_vfy.c had what can be considered an off-by-one-error:
9333 Its depth (which was not part of the external interface)
9334 was actually counting the number of certificates in a chain;
9335 now it really counts the depth.
9338 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
9339 instead of X509err, which often resulted in confusing error
9340 messages since the error codes are not globally unique
9341 (e.g. an alleged error in ssl3_accept when a certificate
9342 didn't match the private key).
9344 *) New function SSL_CTX_set_session_id_context that allows to set a default
9345 value (so that you don't need SSL_set_session_id_context for each
9346 connection using the SSL_CTX).
9349 *) OAEP decoding bug fix.
9352 *) Support INSTALL_PREFIX for package builders, as proposed by
9356 *) New Configure options "threads" and "no-threads". For systems
9357 where the proper compiler options are known (currently Solaris
9358 and Linux), "threads" is the default.
9361 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
9364 *) Install various scripts to $(OPENSSLDIR)/misc, not to
9365 $(INSTALLTOP)/bin -- they shouldn't clutter directories
9366 such as /usr/local/bin.
9369 *) "make linux-shared" to build shared libraries.
9370 [Niels Poppe <niels@netbox.org>]
9372 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
9375 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
9376 extension adding in x509 utility.
9379 *) Remove NOPROTO sections and error code comments.
9382 *) Partial rewrite of the DEF file generator to now parse the ANSI
9386 *) New Configure options --prefix=DIR and --openssldir=DIR.
9389 *) Complete rewrite of the error code script(s). It is all now handled
9390 by one script at the top level which handles error code gathering,
9391 header rewriting and C source file generation. It should be much better
9392 than the old method: it now uses a modified version of Ulf's parser to
9393 read the ANSI prototypes in all header files (thus the old K&R definitions
9394 aren't needed for error creation any more) and do a better job of
9395 translating function codes into names. The old 'ASN1 error code imbedded
9396 in a comment' is no longer necessary and it doesn't use .err files which
9397 have now been deleted. Also the error code call doesn't have to appear all
9398 on one line (which resulted in some large lines...).
9401 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
9404 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
9405 0 (which usually indicates a closed connection), but continue reading.
9408 *) Fix some race conditions.
9411 *) Add support for CRL distribution points extension. Add Certificate
9412 Policies and CRL distribution points documentation.
9415 *) Move the autogenerated header file parts to crypto/opensslconf.h.
9418 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
9419 8 of keying material. Merlin has also confirmed interop with this fix
9420 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
9421 [Merlin Hughes <merlin@baltimore.ie>]
9423 *) Fix lots of warnings.
9424 [Richard Levitte <levitte@stacken.kth.se>]
9426 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
9427 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
9428 [Richard Levitte <levitte@stacken.kth.se>]
9430 *) Fix problems with sizeof(long) == 8.
9431 [Andy Polyakov <appro@fy.chalmers.se>]
9433 *) Change functions to ANSI C.
9436 *) Fix typos in error codes.
9437 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
9439 *) Remove defunct assembler files from Configure.
9442 *) SPARC v8 assembler BIGNUM implementation.
9443 [Andy Polyakov <appro@fy.chalmers.se>]
9445 *) Support for Certificate Policies extension: both print and set.
9446 Various additions to support the r2i method this uses.
9449 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
9450 return a const string when you are expecting an allocated buffer.
9453 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
9454 types DirectoryString and DisplayText.
9457 *) Add code to allow r2i extensions to access the configuration database,
9458 add an LHASH database driver and add several ctx helper functions.
9461 *) Fix an evil bug in bn_expand2() which caused various BN functions to
9462 fail when they extended the size of a BIGNUM.
9465 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
9466 support typesafe stack.
9469 *) Fix typo in SSL_[gs]et_options().
9470 [Nils Frostberg <nils@medcom.se>]
9472 *) Delete various functions and files that belonged to the (now obsolete)
9473 old X509V3 handling code.
9476 *) New Configure option "rsaref".
9479 *) Don't auto-generate pem.h.
9482 *) Introduce type-safe ASN.1 SETs.
9485 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
9486 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
9488 *) Introduce type-safe STACKs. This will almost certainly break lots of code
9489 that links with OpenSSL (well at least cause lots of warnings), but fear
9490 not: the conversion is trivial, and it eliminates loads of evil casts. A
9491 few STACKed things have been converted already. Feel free to convert more.
9492 In the fullness of time, I'll do away with the STACK type altogether.
9495 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
9496 specified in <certfile> by updating the entry in the index.txt file.
9497 This way one no longer has to edit the index.txt file manually for
9498 revoking a certificate. The -revoke option does the gory details now.
9499 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
9501 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
9502 `-text' option at all and this way the `-noout -text' combination was
9503 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
9504 [Ralf S. Engelschall]
9506 *) Make sure a corresponding plain text error message exists for the
9507 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
9508 verify callback function determined that a certificate was revoked.
9509 [Ralf S. Engelschall]
9511 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
9512 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
9513 all available cipers including rc5, which was forgotten until now.
9514 In order to let the testing shell script know which algorithms
9515 are available, a new (up to now undocumented) command
9516 "openssl list-cipher-commands" is used.
9519 *) Bugfix: s_client occasionally would sleep in select() when
9520 it should have checked SSL_pending() first.
9523 *) New functions DSA_do_sign and DSA_do_verify to provide access to
9524 the raw DSA values prior to ASN.1 encoding.
9527 *) Tweaks to Configure
9528 [Niels Poppe <niels@netbox.org>]
9530 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
9534 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
9537 *) New config option to avoid instructions that are illegal on the 80386.
9538 The default code is faster, but requires at least a 486.
9541 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
9542 SSL2_SERVER_VERSION (not used at all) macros, which are now the
9543 same as SSL2_VERSION anyway.
9546 *) New "-showcerts" option for s_client.
9549 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
9550 application. Various cleanups and fixes.
9553 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
9554 modify error routines to work internally. Add error codes and PBE init
9555 to library startup routines.
9558 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
9559 packing functions to asn1 and evp. Changed function names and error
9560 codes along the way.
9563 *) PKCS12 integration: and so it begins... First of several patches to
9564 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
9565 objects to objects.h
9568 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
9569 and display support for Thawte strong extranet extension.
9572 *) Add LinuxPPC support.
9573 [Jeff Dubrule <igor@pobox.org>]
9575 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
9576 bn_div_words in alpha.s.
9577 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
9579 *) Make sure the RSA OAEP test is skipped under -DRSAref because
9580 OAEP isn't supported when OpenSSL is built with RSAref.
9581 [Ulf Moeller <ulf@fitug.de>]
9583 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
9584 so they no longer are missing under -DNOPROTO.
9585 [Soren S. Jorvang <soren@t.dk>]
9588 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
9590 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
9591 doesn't work when the session is reused. Coming soon!
9594 *) Fix a security hole, that allows sessions to be reused in the wrong
9595 context thus bypassing client cert protection! All software that uses
9596 client certs and session caches in multiple contexts NEEDS PATCHING to
9597 allow session reuse! A fuller solution is in the works.
9598 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
9600 *) Some more source tree cleanups (removed obsolete files
9601 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
9602 permission on "config" script to be executable) and a fix for the INSTALL
9604 [Ulf Moeller <ulf@fitug.de>]
9606 *) Remove some legacy and erroneous uses of malloc, free instead of
9608 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
9610 *) Make rsa_oaep_test return non-zero on error.
9611 [Ulf Moeller <ulf@fitug.de>]
9613 *) Add support for native Solaris shared libraries. Configure
9614 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
9615 if someone would make that last step automatic.
9616 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
9618 *) ctx_size was not built with the right compiler during "make links". Fixed.
9621 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
9622 except NULL ciphers". This means the default cipher list will no longer
9623 enable NULL ciphers. They need to be specifically enabled e.g. with
9624 the string "DEFAULT:eNULL".
9627 *) Fix to RSA private encryption routines: if p < q then it would
9628 occasionally produce an invalid result. This will only happen with
9629 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9632 *) Be less restrictive and allow also `perl util/perlpath.pl
9633 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9634 because this way one can also use an interpreter named `perl5' (which is
9635 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9636 installed as `perl').
9637 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9639 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9640 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9642 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9643 advapi32.lib to Win32 build and change the pem test comparision
9644 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9645 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9646 and crypto/des/ede_cbcm_enc.c.
9649 *) DES quad checksum was broken on big-endian architectures. Fixed.
9652 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9653 Win32 test batch file so it (might) work again. The Win32 test batch file
9654 is horrible: I feel ill....
9657 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
9658 in e_os.h. Audit of header files to check ANSI and non ANSI
9659 sections: 10 functions were absent from non ANSI section and not exported
9660 from Windows DLLs. Fixed up libeay.num for new functions.
9663 *) Make `openssl version' output lines consistent.
9664 [Ralf S. Engelschall]
9666 *) Fix Win32 symbol export lists for BIO functions: Added
9667 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
9668 to ms/libeay{16,32}.def.
9669 [Ralf S. Engelschall]
9671 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
9672 fine under Unix and passes some trivial tests I've now added. But the
9673 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
9674 added to make sure no one expects that this stuff really works in the
9675 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
9676 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
9678 [Ralf S. Engelschall]
9680 *) Fix the generation of two part addresses in perl.
9681 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
9683 *) Add config entry for Linux on MIPS.
9684 [John Tobey <jtobey@channel1.com>]
9686 *) Make links whenever Configure is run, unless we are on Windoze.
9689 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
9690 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
9694 *) Add a useful kludge to allow package maintainers to specify compiler and
9695 other platforms details on the command line without having to patch the
9696 Configure script everytime: One now can use ``perl Configure
9697 <id>:<details>'', i.e. platform ids are allowed to have details appended
9698 to them (seperated by colons). This is treated as there would be a static
9699 pre-configured entry in Configure's %table under key <id> with value
9700 <details> and ``perl Configure <id>'' is called. So, when you want to
9701 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
9702 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
9703 now, which overrides the FreeBSD-elf entry on-the-fly.
9704 [Ralf S. Engelschall]
9706 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
9709 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
9710 on the `perl Configure ...' command line. This way one can compile
9711 OpenSSL libraries with Position Independent Code (PIC) which is needed
9712 for linking it into DSOs.
9713 [Ralf S. Engelschall]
9715 *) Remarkably, export ciphers were totally broken and no-one had noticed!
9719 *) Cleaned up the LICENSE document: The official contact for any license
9720 questions now is the OpenSSL core team under openssl-core@openssl.org.
9721 And add a paragraph about the dual-license situation to make sure people
9722 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
9723 to the OpenSSL toolkit.
9724 [Ralf S. Engelschall]
9726 *) General source tree makefile cleanups: Made `making xxx in yyy...'
9727 display consistent in the source tree and replaced `/bin/rm' by `rm'.
9728 Additonally cleaned up the `make links' target: Remove unnecessary
9729 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
9730 to speed processing and no longer clutter the display with confusing
9731 stuff. Instead only the actually done links are displayed.
9732 [Ralf S. Engelschall]
9734 *) Permit null encryption ciphersuites, used for authentication only. It used
9735 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
9736 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
9740 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
9741 signed attributes when verifying signatures (this would break them),
9742 the detached data encoding was wrong and public keys obtained using
9743 X509_get_pubkey() weren't freed.
9746 *) Add text documentation for the BUFFER functions. Also added a work around
9747 to a Win95 console bug. This was triggered by the password read stuff: the
9748 last character typed gets carried over to the next fread(). If you were
9749 generating a new cert request using 'req' for example then the last
9750 character of the passphrase would be CR which would then enter the first
9754 *) Added the new `Includes OpenSSL Cryptography Software' button as
9755 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
9756 button and can be used by applications based on OpenSSL to show the
9757 relationship to the OpenSSL project.
9758 [Ralf S. Engelschall]
9760 *) Remove confusing variables in function signatures in files
9761 ssl/ssl_lib.c and ssl/ssl.h.
9762 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9764 *) Don't install bss_file.c under PREFIX/include/
9765 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9767 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
9768 functions that return function pointers and has support for NT specific
9769 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
9770 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
9771 unsigned to signed types: this was killing the Win32 compile.
9774 *) Add new certificate file to stack functions,
9775 SSL_add_dir_cert_subjects_to_stack() and
9776 SSL_add_file_cert_subjects_to_stack(). These largely supplant
9777 SSL_load_client_CA_file(), and can be used to add multiple certs easily
9778 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
9779 This means that Apache-SSL and similar packages don't have to mess around
9780 to add as many CAs as they want to the preferred list.
9783 *) Experiment with doxygen documentation. Currently only partially applied to
9785 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
9786 openssl.doxy as the configuration file.
9789 *) Get rid of remaining C++-style comments which strict C compilers hate.
9790 [Ralf S. Engelschall, pointed out by Carlos Amengual]
9792 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
9793 compiled in by default: it has problems with large keys.
9796 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
9797 DH private keys and/or callback functions which directly correspond to
9798 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
9799 is needed for applications which have to configure certificates on a
9800 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
9802 For the RSA certificate situation is makes no difference, but
9803 for the DSA certificate situation this fixes the "no shared cipher"
9804 problem where the OpenSSL cipher selection procedure failed because the
9805 temporary keys were not overtaken from the context and the API provided
9806 no way to reconfigure them.
9807 The new functions now let applications reconfigure the stuff and they
9808 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
9809 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
9810 non-public-API function ssl_cert_instantiate() is used as a helper
9811 function and also to reduce code redundancy inside ssl_rsa.c.
9812 [Ralf S. Engelschall]
9814 *) Move s_server -dcert and -dkey options out of the undocumented feature
9815 area because they are useful for the DSA situation and should be
9816 recognized by the users.
9817 [Ralf S. Engelschall]
9819 *) Fix the cipher decision scheme for export ciphers: the export bits are
9820 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
9821 SSL_EXP_MASK. So, the original variable has to be used instead of the
9822 already masked variable.
9823 [Richard Levitte <levitte@stacken.kth.se>]
9825 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
9826 [Richard Levitte <levitte@stacken.kth.se>]
9828 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
9829 from `int' to `unsigned int' because it's a length and initialized by
9830 EVP_DigestFinal() which expects an `unsigned int *'.
9831 [Richard Levitte <levitte@stacken.kth.se>]
9833 *) Don't hard-code path to Perl interpreter on shebang line of Configure
9834 script. Instead use the usual Shell->Perl transition trick.
9835 [Ralf S. Engelschall]
9837 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
9838 (in addition to RSA certificates) to match the behaviour of `openssl dsa
9839 -noout -modulus' as it's already the case for `openssl rsa -noout
9840 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
9841 currently the public key is printed (a decision which was already done by
9842 `openssl dsa -modulus' in the past) which serves a similar purpose.
9843 Additionally the NO_RSA no longer completely removes the whole -modulus
9844 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
9846 [Ralf S. Engelschall]
9848 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
9849 BIO. See the source (crypto/evp/bio_ok.c) for more info.
9850 [Arne Ansper <arne@ats.cyber.ee>]
9852 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
9853 to be added. Now both 'req' and 'ca' can use new objects defined in the
9857 *) Add cool BIO that does syslog (or event log on NT).
9858 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
9860 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
9861 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
9862 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
9863 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
9866 *) Add preliminary config info for new extension code.
9869 *) Make RSA_NO_PADDING really use no padding.
9870 [Ulf Moeller <ulf@fitug.de>]
9872 *) Generate errors when private/public key check is done.
9875 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
9876 for some CRL extensions and new objects added.
9879 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
9880 key usage extension and fuller support for authority key id.
9883 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
9884 padding method for RSA, which is recommended for new applications in PKCS
9885 #1 v2.0 (RFC 2437, October 1998).
9886 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
9887 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
9888 against Bleichbacher's attack on RSA.
9889 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
9892 *) Updates to the new SSL compression code
9893 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9895 *) Fix so that the version number in the master secret, when passed
9896 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
9897 (because the server will not accept higher), that the version number
9898 is 0x03,0x01, not 0x03,0x00
9899 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9901 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
9902 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
9903 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
9906 *) Support for RAW extensions where an arbitrary extension can be
9907 created by including its DER encoding. See apps/openssl.cnf for
9911 *) Make sure latest Perl versions don't interpret some generated C array
9912 code as Perl array code in the crypto/err/err_genc.pl script.
9913 [Lars Weber <3weber@informatik.uni-hamburg.de>]
9915 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
9916 not many people have the assembler. Various Win32 compilation fixes and
9917 update to the INSTALL.W32 file with (hopefully) more accurate Win32
9921 *) Modify configure script 'Configure' to automatically create crypto/date.h
9922 file under Win32 and also build pem.h from pem.org. New script
9923 util/mkfiles.pl to create the MINFO file on environments that can't do a
9924 'make files': perl util/mkfiles.pl >MINFO should work.
9927 *) Major rework of DES function declarations, in the pursuit of correctness
9928 and purity. As a result, many evil casts evaporated, and some weirdness,
9929 too. You may find this causes warnings in your code. Zapping your evil
9930 casts will probably fix them. Mostly.
9933 *) Fix for a typo in asn1.h. Bug fix to object creation script
9934 obj_dat.pl. It considered a zero in an object definition to mean
9935 "end of object": none of the objects in objects.h have any zeros
9936 so it wasn't spotted.
9937 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
9939 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
9940 Masking (CBCM). In the absence of test vectors, the best I have been able
9941 to do is check that the decrypt undoes the encrypt, so far. Send me test
9942 vectors if you have them.
9945 *) Correct calculation of key length for export ciphers (too much space was
9946 allocated for null ciphers). This has not been tested!
9949 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
9950 message is now correct (it understands "crypto" and "ssl" on its
9951 command line). There is also now an "update" option. This will update
9952 the util/ssleay.num and util/libeay.num files with any new functions.
9954 perl util/mkdef.pl crypto ssl update
9955 it will update them.
9958 *) Overhauled the Perl interface (perl/*):
9959 - ported BN stuff to OpenSSL's different BN library
9960 - made the perl/ source tree CVS-aware
9961 - renamed the package from SSLeay to OpenSSL (the files still contain
9962 their history because I've copied them in the repository)
9963 - removed obsolete files (the test scripts will be replaced
9964 by better Test::Harness variants in the future)
9965 [Ralf S. Engelschall]
9967 *) First cut for a very conservative source tree cleanup:
9968 1. merge various obsolete readme texts into doc/ssleay.txt
9969 where we collect the old documents and readme texts.
9970 2. remove the first part of files where I'm already sure that we no
9971 longer need them because of three reasons: either they are just temporary
9972 files which were left by Eric or they are preserved original files where
9973 I've verified that the diff is also available in the CVS via "cvs diff
9974 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
9975 the crypto/md/ stuff).
9976 [Ralf S. Engelschall]
9978 *) More extension code. Incomplete support for subject and issuer alt
9979 name, issuer and authority key id. Change the i2v function parameters
9980 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
9981 what that's for :-) Fix to ASN1 macro which messed up
9982 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
9985 *) Preliminary support for ENUMERATED type. This is largely copied from the
9989 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
9990 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9992 *) Make sure `make rehash' target really finds the `openssl' program.
9993 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9995 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
9996 like to hear about it if this slows down other processors.
9999 *) Add CygWin32 platform information to Configure script.
10000 [Alan Batie <batie@aahz.jf.intel.com>]
10002 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
10003 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
10005 *) New program nseq to manipulate netscape certificate sequences
10008 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
10012 *) Fixes to BN code. Previously the default was to define BN_RECURSION
10013 but the BN code had some problems that would cause failures when
10014 doing certificate verification and some other functions.
10015 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10017 *) Add ASN1 and PEM code to support netscape certificate sequences.
10020 *) Add ASN1 and PEM code to support netscape certificate sequences.
10023 *) Add several PKIX and private extended key usage OIDs.
10026 *) Modify the 'ca' program to handle the new extension code. Modify
10027 openssl.cnf for new extension format, add comments.
10030 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
10031 and add a sample to openssl.cnf so req -x509 now adds appropriate
10035 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
10036 error code, add initial support to X509_print() and x509 application.
10039 *) Takes a deep breath and start addding X509 V3 extension support code. Add
10040 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
10041 stuff is currently isolated and isn't even compiled yet.
10044 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
10045 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
10046 Removed the versions check from X509 routines when loading extensions:
10047 this allows certain broken certificates that don't set the version
10048 properly to be processed.
10051 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
10052 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
10053 can still be regenerated with "make depend".
10056 *) Spelling mistake in C version of CAST-128.
10057 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
10059 *) Changes to the error generation code. The perl script err-code.pl
10060 now reads in the old error codes and retains the old numbers, only
10061 adding new ones if necessary. It also only changes the .err files if new
10062 codes are added. The makefiles have been modified to only insert errors
10063 when needed (to avoid needlessly modifying header files). This is done
10064 by only inserting errors if the .err file is newer than the auto generated
10065 C file. To rebuild all the error codes from scratch (the old behaviour)
10066 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
10067 or delete all the .err files.
10070 *) CAST-128 was incorrectly implemented for short keys. The C version has
10071 been fixed, but is untested. The assembler versions are also fixed, but
10072 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
10073 to regenerate it if needed.
10074 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
10075 Hagino <itojun@kame.net>]
10077 *) File was opened incorrectly in randfile.c.
10078 [Ulf Möller <ulf@fitug.de>]
10080 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
10081 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
10082 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
10083 al: it's just almost always a UTCTime. Note this patch adds new error
10084 codes so do a "make errors" if there are problems.
10087 *) Correct Linux 1 recognition in config.
10088 [Ulf Möller <ulf@fitug.de>]
10090 *) Remove pointless MD5 hash when using DSA keys in ca.
10091 [Anonymous <nobody@replay.com>]
10093 *) Generate an error if given an empty string as a cert directory. Also
10094 generate an error if handed NULL (previously returned 0 to indicate an
10095 error, but didn't set one).
10096 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
10098 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
10101 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
10102 parameters. This was causing a warning which killed off the Win32 compile.
10105 *) Remove C++ style comments from crypto/bn/bn_local.h.
10106 [Neil Costigan <neil.costigan@celocom.com>]
10108 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
10109 based on a text string, looking up short and long names and finally
10110 "dot" format. The "dot" format stuff didn't work. Added new function
10111 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
10112 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
10113 OID is not part of the table.
10116 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
10117 X509_LOOKUP_by_alias().
10120 *) Sort openssl functions by name.
10123 *) Get the gendsa program working (hopefully) and add it to app list. Remove
10124 encryption from sample DSA keys (in case anyone is interested the password
10128 *) Make _all_ *_free functions accept a NULL pointer.
10129 [Frans Heymans <fheymans@isaserver.be>]
10131 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
10133 [Anonymous <nobody@replay.com>]
10135 *) s_server should send the CAfile as acceptable CAs, not its own cert.
10136 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10138 *) Don't blow it for numeric -newkey arguments to apps/req.
10139 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10141 *) Temp key "for export" tests were wrong in s3_srvr.c.
10142 [Anonymous <nobody@replay.com>]
10144 *) Add prototype for temp key callback functions
10145 SSL_CTX_set_tmp_{rsa,dh}_callback().
10148 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
10149 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
10152 *) X509_name_add_entry() freed the wrong thing after an error.
10153 [Arne Ansper <arne@ats.cyber.ee>]
10155 *) rsa_eay.c would attempt to free a NULL context.
10156 [Arne Ansper <arne@ats.cyber.ee>]
10158 *) BIO_s_socket() had a broken should_retry() on Windoze.
10159 [Arne Ansper <arne@ats.cyber.ee>]
10161 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
10162 [Arne Ansper <arne@ats.cyber.ee>]
10164 *) Make sure the already existing X509_STORE->depth variable is initialized
10165 in X509_STORE_new(), but document the fact that this variable is still
10166 unused in the certificate verification process.
10167 [Ralf S. Engelschall]
10169 *) Fix the various library and apps files to free up pkeys obtained from
10170 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
10173 *) Fix reference counting in X509_PUBKEY_get(). This makes
10174 demos/maurice/example2.c work, amongst others, probably.
10175 [Steve Henson and Ben Laurie]
10177 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
10178 `openssl' and second, the shortcut symlinks for the `openssl <command>'
10179 are no longer created. This way we have a single and consistent command
10180 line interface `openssl <command>', similar to `cvs <command>'.
10181 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
10183 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
10184 BIT STRING wrapper always have zero unused bits.
10187 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
10190 *) Make the top-level INSTALL documentation easier to understand.
10193 *) Makefiles updated to exit if an error occurs in a sub-directory
10194 make (including if user presses ^C) [Paul Sutton]
10196 *) Make Montgomery context stuff explicit in RSA data structure.
10199 *) Fix build order of pem and err to allow for generated pem.h.
10202 *) Fix renumbering bug in X509_NAME_delete_entry().
10205 *) Enhanced the err-ins.pl script so it makes the error library number
10206 global and can add a library name. This is needed for external ASN1 and
10207 other error libraries.
10210 *) Fixed sk_insert which never worked properly.
10213 *) Fix ASN1 macros so they can handle indefinite length construted
10214 EXPLICIT tags. Some non standard certificates use these: they can now
10218 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
10219 into a single doc/ssleay.txt bundle. This way the information is still
10220 preserved but no longer messes up this directory. Now it's new room for
10221 the new set of documenation files.
10222 [Ralf S. Engelschall]
10224 *) SETs were incorrectly DER encoded. This was a major pain, because they
10225 shared code with SEQUENCEs, which aren't coded the same. This means that
10226 almost everything to do with SETs or SEQUENCEs has either changed name or
10227 number of arguments.
10228 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
10230 *) Fix test data to work with the above.
10233 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
10234 was already fixed by Eric for 0.9.1 it seems.
10235 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
10237 *) Autodetect FreeBSD3.
10240 *) Fix various bugs in Configure. This affects the following platforms:
10244 unixware-2.0-pentium
10248 *) Eliminate generated files from CVS. Reorder tests to regenerate files
10249 before they are needed.
10252 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
10256 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
10258 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
10259 changed SSLeay to OpenSSL in version strings.
10260 [Ralf S. Engelschall]
10262 *) Some fixups to the top-level documents.
10265 *) Fixed the nasty bug where rsaref.h was not found under compile-time
10266 because the symlink to include/ was missing.
10267 [Ralf S. Engelschall]
10269 *) Incorporated the popular no-RSA/DSA-only patches
10270 which allow to compile a RSA-free SSLeay.
10271 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
10273 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
10274 when "ssleay" is still not found.
10275 [Ralf S. Engelschall]
10277 *) Added more platforms to Configure: Cray T3E, HPUX 11,
10278 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
10280 *) Updated the README file.
10281 [Ralf S. Engelschall]
10283 *) Added various .cvsignore files in the CVS repository subdirs
10284 to make a "cvs update" really silent.
10285 [Ralf S. Engelschall]
10287 *) Recompiled the error-definition header files and added
10288 missing symbols to the Win32 linker tables.
10289 [Ralf S. Engelschall]
10291 *) Cleaned up the top-level documents;
10292 o new files: CHANGES and LICENSE
10293 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
10294 o merged COPYRIGHT into LICENSE
10295 o removed obsolete TODO file
10296 o renamed MICROSOFT to INSTALL.W32
10297 [Ralf S. Engelschall]
10299 *) Removed dummy files from the 0.9.1b source tree:
10300 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
10301 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
10302 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
10303 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
10304 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
10305 [Ralf S. Engelschall]
10307 *) Added various platform portability fixes.
10310 *) The Genesis of the OpenSSL rpject:
10311 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
10312 Young and Tim J. Hudson created while they were working for C2Net until
10314 [The OpenSSL Project]
10317 Changes between 0.9.0b and 0.9.1b [not released]
10319 *) Updated a few CA certificates under certs/
10322 *) Changed some BIGNUM api stuff.
10325 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
10326 DGUX x86, Linux Alpha, etc.
10329 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
10330 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
10334 *) Add -strparse option to asn1pars program which parses nested
10336 [Dr Stephen Henson <shenson@bigfoot.com>]
10338 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
10341 *) DSA fix for "ca" program.
10344 *) Added "-genkey" option to "dsaparam" program.
10347 *) Added RIPE MD160 (rmd160) message digest.
10350 *) Added -a (all) option to "ssleay version" command.
10353 *) Added PLATFORM define which is the id given to Configure.
10356 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
10359 *) Extended the ASN.1 parser routines.
10362 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
10365 *) Added a BN_CTX to the BN library.
10368 *) Fixed the weak key values in DES library
10371 *) Changed API in EVP library for cipher aliases.
10374 *) Added support for RC2/64bit cipher.
10377 *) Converted the lhash library to the crypto/mem.c functions.
10380 *) Added more recognized ASN.1 object ids.
10383 *) Added more RSA padding checks for SSL/TLS.
10386 *) Added BIO proxy/filter functionality.
10389 *) Added extra_certs to SSL_CTX which can be used
10390 send extra CA certificates to the client in the CA cert chain sending
10391 process. It can be configured with SSL_CTX_add_extra_chain_cert().
10394 *) Now Fortezza is denied in the authentication phase because
10395 this is key exchange mechanism is not supported by SSLeay at all.
10398 *) Additional PKCS1 checks.
10401 *) Support the string "TLSv1" for all TLS v1 ciphers.
10404 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
10405 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
10408 *) Fixed a few memory leaks.
10411 *) Fixed various code and comment typos.
10414 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
10415 bytes sent in the client random.
10416 [Edward Bishop <ebishop@spyglass.com>]