5 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
7 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
8 SRP code can be overrun an internal buffer. Add sanity check that
9 g, A, B < N to SRP code.
11 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
12 Group for discovering this issue.
16 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
17 TLS 1.0 instead of higher protocol versions when the ClientHello message
18 is badly fragmented. This allows a man-in-the-middle attacker to force a
19 downgrade to TLS 1.0 even if both the server and the client support a
20 higher protocol version, by modifying the client's TLS records.
22 Thanks to David Benjamin and Adam Langley (Google) for discovering and
23 researching this issue.
27 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
28 to a denial of service attack. A malicious server can crash the client
29 with a null pointer dereference (read) by specifying an anonymous (EC)DH
30 ciphersuite and sending carefully crafted handshake messages.
32 Thanks to Felix Gröbert (Google) for discovering and researching this
37 *) By sending carefully crafted DTLS packets an attacker could cause openssl
38 to leak memory. This can be exploited through a Denial of Service attack.
39 Thanks to Adam Langley for discovering and researching this issue.
43 *) An attacker can force openssl to consume large amounts of memory whilst
44 processing DTLS handshake messages. This can be exploited through a
45 Denial of Service attack.
46 Thanks to Adam Langley for discovering and researching this issue.
50 *) An attacker can force an error condition which causes openssl to crash
51 whilst processing DTLS packets due to memory being freed twice. This
52 can be exploited through a Denial of Service attack.
53 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
58 *) If a multithreaded client connects to a malicious server using a resumed
59 session and the server sends an ec point format extension it could write
60 up to 255 bytes to freed memory.
62 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
67 *) A malicious server can crash an OpenSSL client with a null pointer
68 dereference (read) by specifying an SRP ciphersuite even though it was not
69 properly negotiated with the client. This can be exploited through a
70 Denial of Service attack.
72 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
73 discovering and researching this issue.
77 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
78 X509_name_oneline, X509_name_print_ex et al. to leak some information
79 from the stack. Applications may be affected if they echo pretty printing
80 output to the attacker.
82 Thanks to Ivan Fratric (Google) for discovering this issue.
84 [Emilia Käsper, and Steve Henson]
86 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
87 for corner cases. (Certain input points at infinity could lead to
88 bogus results, with non-infinity inputs mapped to infinity too.)
91 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
93 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
94 handshake can force the use of weak keying material in OpenSSL
95 SSL/TLS clients and servers.
97 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
98 researching this issue. (CVE-2014-0224)
99 [KIKUCHI Masashi, Steve Henson]
101 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
102 OpenSSL DTLS client the code can be made to recurse eventually crashing
105 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
107 [Imre Rad, Steve Henson]
109 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
110 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
111 client or server. This is potentially exploitable to run arbitrary
112 code on a vulnerable client or server.
114 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
115 [Jüri Aedla, Steve Henson]
117 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
118 are subject to a denial of service attack.
120 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
121 this issue. (CVE-2014-3470)
122 [Felix Gröbert, Ivan Fratric, Steve Henson]
124 *) Harmonize version and its documentation. -f flag is used to display
126 [mancha <mancha1@zoho.com>]
128 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
130 [mancha <mancha1@zoho.com>]
132 *) Fix some double frees. These are not thought to be exploitable.
133 [mancha <mancha1@zoho.com>]
135 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
137 *) A missing bounds check in the handling of the TLS heartbeat extension
138 can be used to reveal up to 64k of memory to a connected client or
141 Thanks for Neel Mehta of Google Security for discovering this bug and to
142 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
143 preparing the fix (CVE-2014-0160)
144 [Adam Langley, Bodo Moeller]
146 *) Fix for the attack described in the paper "Recovering OpenSSL
147 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
148 by Yuval Yarom and Naomi Benger. Details can be obtained from:
149 http://eprint.iacr.org/2014/140
151 Thanks to Yuval Yarom and Naomi Benger for discovering this
152 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
153 [Yuval Yarom and Naomi Benger]
155 *) TLS pad extension: draft-agl-tls-padding-03
157 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
158 TLS client Hello record length value would otherwise be > 255 and
159 less that 512 pad with a dummy extension containing zeroes so it
160 is at least 512 bytes long.
162 [Adam Langley, Steve Henson]
164 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
166 *) Fix for TLS record tampering bug. A carefully crafted invalid
167 handshake could crash OpenSSL with a NULL pointer exception.
168 Thanks to Anton Johansson for reporting this issues.
171 *) Keep original DTLS digest and encryption contexts in retransmission
172 structures so we can use the previous session parameters if they need
173 to be resent. (CVE-2013-6450)
176 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
177 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
178 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
179 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
180 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
181 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
182 [Rob Stradling, Adam Langley]
184 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
186 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
187 supporting platforms or when small records were transferred.
188 [Andy Polyakov, Steve Henson]
190 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
192 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
194 This addresses the flaw in CBC record processing discovered by
195 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
196 at: http://www.isg.rhul.ac.uk/tls/
198 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
199 Security Group at Royal Holloway, University of London
200 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
201 Emilia Käsper for the initial patch.
203 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
205 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
206 ciphersuites which can be exploited in a denial of service attack.
207 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
208 and detecting this bug and to Wolfgang Ettlinger
209 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
213 *) Return an error when checking OCSP signatures when key is NULL.
214 This fixes a DoS attack. (CVE-2013-0166)
217 *) Make openssl verify return errors.
218 [Chris Palmer <palmer@google.com> and Ben Laurie]
220 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
221 the right response is stapled. Also change SSL_get_certificate()
222 so it returns the certificate actually sent.
223 See http://rt.openssl.org/Ticket/Display.html?id=2836.
224 [Rob Stradling <rob.stradling@comodo.com>]
226 *) Fix possible deadlock when decoding public keys.
229 *) Don't use TLS 1.0 record version number in initial client hello
233 Changes between 1.0.1b and 1.0.1c [10 May 2012]
235 *) Sanity check record length before skipping explicit IV in TLS
236 1.2, 1.1 and DTLS to fix DoS attack.
238 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
239 fuzzing as a service testing platform.
243 *) Initialise tkeylen properly when encrypting CMS messages.
244 Thanks to Solar Designer of Openwall for reporting this issue.
247 *) In FIPS mode don't try to use composite ciphers as they are not
251 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
253 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
254 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
255 mean any application compiled against OpenSSL 1.0.0 headers setting
256 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
257 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
258 0x10000000L Any application which was previously compiled against
259 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
260 will need to be recompiled as a result. Letting be results in
261 inability to disable specifically TLS 1.1 and in client context,
262 in unlike event, limit maximum offered version to TLS 1.0 [see below].
265 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
266 disable just protocol X, but all protocols above X *if* there are
267 protocols *below* X still enabled. In more practical terms it means
268 that if application wants to disable TLS1.0 in favor of TLS1.1 and
269 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
270 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
274 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
276 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
277 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
278 in CRYPTO_realloc_clean.
280 Thanks to Tavis Ormandy, Google Security Team, for discovering this
281 issue and to Adam Langley <agl@chromium.org> for fixing it.
283 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
285 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
288 *) Workarounds for some broken servers that "hang" if a client hello
289 record length exceeds 255 bytes.
291 1. Do not use record version number > TLS 1.0 in initial client
292 hello: some (but not all) hanging servers will now work.
293 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
294 the number of ciphers sent in the client hello. This should be
295 set to an even number, such as 50, for example by passing:
296 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
297 Most broken servers should now work.
298 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
299 TLS 1.2 client support entirely.
302 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
305 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
307 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
308 STRING form instead of a DigestInfo.
311 *) The format used for MDC2 RSA signatures is inconsistent between EVP
312 and the RSA_sign/RSA_verify functions. This was made more apparent when
313 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
314 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
315 the correct format in RSA_verify so both forms transparently work.
318 *) Some servers which support TLS 1.0 can choke if we initially indicate
319 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
320 encrypted premaster secret. As a workaround use the maximum pemitted
321 client version in client hello, this should keep such servers happy
322 and still work with previous versions of OpenSSL.
325 *) Add support for TLS/DTLS heartbeats.
326 [Robin Seggelmann <seggelmann@fh-muenster.de>]
328 *) Add support for SCTP.
329 [Robin Seggelmann <seggelmann@fh-muenster.de>]
331 *) Improved PRNG seeding for VOS.
332 [Paul Green <Paul.Green@stratus.com>]
334 *) Extensive assembler packs updates, most notably:
336 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
337 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
338 - x86_64: bit-sliced AES implementation;
339 - ARM: NEON support, contemporary platforms optimizations;
340 - s390x: z196 support;
341 - *: GHASH and GF(2^m) multiplication implementations;
345 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
346 (removal of unnecessary code)
347 [Peter Sylvester <peter.sylvester@edelweb.fr>]
349 *) Add TLS key material exporter from RFC 5705.
352 *) Add DTLS-SRTP negotiation from RFC 5764.
355 *) Add Next Protocol Negotiation,
356 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
357 disabled with a no-npn flag to config or Configure. Code donated
359 [Adam Langley <agl@google.com> and Ben Laurie]
361 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
362 NIST-P256, NIST-P521, with constant-time single point multiplication on
363 typical inputs. Compiler support for the nonstandard type __uint128_t is
364 required to use this (present in gcc 4.4 and later, for 64-bit builds).
365 Code made available under Apache License version 2.0.
367 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
368 line to include this in your build of OpenSSL, and run "make depend" (or
369 "make update"). This enables the following EC_METHODs:
371 EC_GFp_nistp224_method()
372 EC_GFp_nistp256_method()
373 EC_GFp_nistp521_method()
375 EC_GROUP_new_by_curve_name() will automatically use these (while
376 EC_GROUP_new_curve_GFp() currently prefers the more flexible
378 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
380 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
381 all platforms. Move ssize_t definition from e_os.h to the public
382 header file e_os2.h as it now appears in public header file cms.h
385 *) New -sigopt option to the ca, req and x509 utilities. Additional
386 signature parameters can be passed using this option and in
390 *) Add RSA PSS signing function. This will generate and set the
391 appropriate AlgorithmIdentifiers for PSS based on those in the
392 corresponding EVP_MD_CTX structure. No application support yet.
395 *) Support for companion algorithm specific ASN1 signing routines.
396 New function ASN1_item_sign_ctx() signs a pre-initialised
397 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
398 the appropriate parameters.
401 *) Add new algorithm specific ASN1 verification initialisation function
402 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
403 handling will be the same no matter what EVP_PKEY_METHOD is used.
404 Add a PSS handler to support verification of PSS signatures: checked
405 against a number of sample certificates.
408 *) Add signature printing for PSS. Add PSS OIDs.
409 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
411 *) Add algorithm specific signature printing. An individual ASN1 method
412 can now print out signatures instead of the standard hex dump.
414 More complex signatures (e.g. PSS) can print out more meaningful
415 information. Include DSA version that prints out the signature
419 *) Password based recipient info support for CMS library: implementing
423 *) Split password based encryption into PBES2 and PBKDF2 functions. This
424 neatly separates the code into cipher and PBE sections and is required
425 for some algorithms that split PBES2 into separate pieces (such as
429 *) Session-handling fixes:
430 - Fix handling of connections that are resuming with a session ID,
431 but also support Session Tickets.
432 - Fix a bug that suppressed issuing of a new ticket if the client
433 presented a ticket with an expired session.
434 - Try to set the ticket lifetime hint to something reasonable.
435 - Make tickets shorter by excluding irrelevant information.
436 - On the client side, don't ignore renewed tickets.
437 [Adam Langley, Bodo Moeller (Google)]
439 *) Fix PSK session representation.
442 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
444 This work was sponsored by Intel.
447 *) Add GCM support to TLS library. Some custom code is needed to split
448 the IV between the fixed (from PRF) and explicit (from TLS record)
449 portions. This adds all GCM ciphersuites supported by RFC5288 and
450 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
451 add a special AESGCM string for GCM only.
454 *) Expand range of ctrls for AES GCM. Permit setting invocation
455 field on decrypt and retrieval of invocation field only on encrypt.
458 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
459 As required by RFC5289 these ciphersuites cannot be used if for
460 versions of TLS earlier than 1.2.
463 *) For FIPS capable OpenSSL interpret a NULL default public key method
464 as unset and return the appopriate default but do *not* set the default.
465 This means we can return the appopriate method in applications that
466 swicth between FIPS and non-FIPS modes.
469 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
470 ENGINE is used then we cannot handle that in the FIPS module so we
471 keep original code iff non-FIPS operations are allowed.
474 *) Add -attime option to openssl utilities.
475 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
477 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
480 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
481 FIPS EC methods unconditionally for now.
484 *) New build option no-ec2m to disable characteristic 2 code.
487 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
488 all cases can be covered as some introduce binary incompatibilities.
491 *) Redirect RSA operations to FIPS module including keygen,
492 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
495 *) Add similar low level API blocking to ciphers.
498 *) Low level digest APIs are not approved in FIPS mode: any attempt
499 to use these will cause a fatal error. Applications that *really* want
500 to use them can use the private_* version instead.
503 *) Redirect cipher operations to FIPS module for FIPS builds.
506 *) Redirect digest operations to FIPS module for FIPS builds.
509 *) Update build system to add "fips" flag which will link in fipscanister.o
510 for static and shared library builds embedding a signature if needed.
513 *) Output TLS supported curves in preference order instead of numerical
514 order. This is currently hardcoded for the highest order curves first.
515 This should be configurable so applications can judge speed vs strength.
518 *) Add TLS v1.2 server support for client authentication.
521 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
525 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
526 FIPS modules versions.
529 *) Add TLS v1.2 client side support for client authentication. Keep cache
530 of handshake records longer as we don't know the hash algorithm to use
531 until after the certificate request message is received.
534 *) Initial TLS v1.2 client support. Add a default signature algorithms
535 extension including all the algorithms we support. Parse new signature
536 format in client key exchange. Relax some ECC signing restrictions for
537 TLS v1.2 as indicated in RFC5246.
540 *) Add server support for TLS v1.2 signature algorithms extension. Switch
541 to new signature format when needed using client digest preference.
542 All server ciphersuites should now work correctly in TLS v1.2. No client
543 support yet and no support for client certificates.
546 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
547 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
548 ciphersuites. At present only RSA key exchange ciphersuites work with
549 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
550 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
551 and version checking.
554 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
555 with this defined it will not be affected by any changes to ssl internal
556 structures. Add several utility functions to allow openssl application
557 to work with OPENSSL_NO_SSL_INTERN defined.
561 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
563 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
566 *) Permit abbreviated handshakes when renegotiating using the function
567 SSL_renegotiate_abbreviated().
568 [Robin Seggelmann <seggelmann@fh-muenster.de>]
570 *) Add call to ENGINE_register_all_complete() to
571 ENGINE_load_builtin_engines(), so some implementations get used
572 automatically instead of needing explicit application support.
575 *) Add support for TLS key exporter as described in RFC5705.
576 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
578 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
579 a few changes are required:
581 Add SSL_OP_NO_TLSv1_1 flag.
583 Update version checking logic to handle version 1.1.
584 Add explicit IV handling (ported from DTLS code).
585 Add command line options to s_client/s_server.
588 Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
590 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
592 This addresses the flaw in CBC record processing discovered by
593 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
594 at: http://www.isg.rhul.ac.uk/tls/
596 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
597 Security Group at Royal Holloway, University of London
598 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
599 Emilia Käsper for the initial patch.
601 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
603 *) Return an error when checking OCSP signatures when key is NULL.
604 This fixes a DoS attack. (CVE-2013-0166)
607 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
608 the right response is stapled. Also change SSL_get_certificate()
609 so it returns the certificate actually sent.
610 See http://rt.openssl.org/Ticket/Display.html?id=2836.
612 [Rob Stradling <rob.stradling@comodo.com>]
614 *) Fix possible deadlock when decoding public keys.
617 Changes between 1.0.0i and 1.0.0j [10 May 2012]
619 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
622 *) Sanity check record length before skipping explicit IV in DTLS
625 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
626 fuzzing as a service testing platform.
630 *) Initialise tkeylen properly when encrypting CMS messages.
631 Thanks to Solar Designer of Openwall for reporting this issue.
634 Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
636 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
637 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
638 in CRYPTO_realloc_clean.
640 Thanks to Tavis Ormandy, Google Security Team, for discovering this
641 issue and to Adam Langley <agl@chromium.org> for fixing it.
643 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
645 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
647 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
648 in CMS and PKCS7 code. When RSA decryption fails use a random key for
649 content decryption and always return the same error. Note: this attack
650 needs on average 2^20 messages so it only affects automated senders. The
651 old behaviour can be reenabled in the CMS code by setting the
652 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
653 an MMA defence is not necessary.
654 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
655 this issue. (CVE-2012-0884)
658 *) Fix CVE-2011-4619: make sure we really are receiving a
659 client hello before rejecting multiple SGC restarts. Thanks to
660 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
663 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
665 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
666 Thanks to Antonio Martin, Enterprise Secure Access Research and
667 Development, Cisco Systems, Inc. for discovering this bug and
668 preparing a fix. (CVE-2012-0050)
671 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
673 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
674 of the Vaudenay padding oracle attack on CBC mode encryption
675 which enables an efficient plaintext recovery attack against
676 the OpenSSL implementation of DTLS. Their attack exploits timing
677 differences arising during decryption processing. A research
678 paper describing this attack can be found at:
679 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
680 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
681 Security Group at Royal Holloway, University of London
682 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
683 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
684 for preparing the fix. (CVE-2011-4108)
685 [Robin Seggelmann, Michael Tuexen]
687 *) Clear bytes used for block padding of SSL 3.0 records.
689 [Adam Langley (Google)]
691 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
692 Kadianakis <desnacked@gmail.com> for discovering this issue and
693 Adam Langley for preparing the fix. (CVE-2011-4619)
694 [Adam Langley (Google)]
696 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
697 [Andrey Kulikov <amdeich@gmail.com>]
699 *) Prevent malformed RFC3779 data triggering an assertion failure.
700 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
701 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
702 [Rob Austein <sra@hactrn.net>]
704 *) Improved PRNG seeding for VOS.
705 [Paul Green <Paul.Green@stratus.com>]
707 *) Fix ssl_ciph.c set-up race.
708 [Adam Langley (Google)]
710 *) Fix spurious failures in ecdsatest.c.
711 [Emilia Käsper (Google)]
713 *) Fix the BIO_f_buffer() implementation (which was mixing different
714 interpretations of the '..._len' fields).
715 [Adam Langley (Google)]
717 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
718 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
719 threads won't reuse the same blinding coefficients.
721 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
722 lock to call BN_BLINDING_invert_ex, and avoids one use of
723 BN_BLINDING_update for each BN_BLINDING structure (previously,
724 the last update always remained unused).
725 [Emilia Käsper (Google)]
727 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
728 [Bob Buckholz (Google)]
730 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
732 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
733 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
734 [Kaspar Brand <ossl@velox.ch>]
736 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
737 for multi-threaded use of ECDH. (CVE-2011-3210)
738 [Adam Langley (Google)]
740 *) Fix x509_name_ex_d2i memory leak on bad inputs.
743 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
744 signature public key algorithm by using OID xref utilities instead.
745 Before this you could only use some ECC ciphersuites with SHA1 only.
748 *) Add protection against ECDSA timing attacks as mentioned in the paper
749 by Billy Bob Brumley and Nicola Tuveri, see:
751 http://eprint.iacr.org/2011/232.pdf
753 [Billy Bob Brumley and Nicola Tuveri]
755 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
757 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
758 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
760 *) Fix bug in string printing code: if *any* escaping is enabled we must
761 escape the escape character (backslash) or the resulting string is
765 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
767 *) Disable code workaround for ancient and obsolete Netscape browsers
768 and servers: an attacker can use it in a ciphersuite downgrade attack.
769 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
772 *) Fixed J-PAKE implementation error, originally discovered by
773 Sebastien Martini, further info and confirmation from Stefan
774 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
777 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
779 *) Fix extension code to avoid race conditions which can result in a buffer
780 overrun vulnerability: resumed sessions must not be modified as they can
781 be shared by multiple threads. CVE-2010-3864
784 *) Fix WIN32 build system to correctly link an ENGINE directory into
788 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
790 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
792 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
794 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
796 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
797 context. The operation can be customised via the ctrl mechanism in
798 case ENGINEs want to include additional functionality.
801 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
804 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
805 output hashes compatible with older versions of OpenSSL.
806 [Willy Weisz <weisz@vcpc.univie.ac.at>]
808 *) Fix compression algorithm handling: if resuming a session use the
809 compression algorithm of the resumed session instead of determining
810 it from client hello again. Don't allow server to change algorithm.
813 *) Add load_crls() function to apps tidying load_certs() too. Add option
814 to verify utility to allow additional CRLs to be included.
817 *) Update OCSP request code to permit adding custom headers to the request:
818 some responders need this.
821 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
823 [Julia Lawall <julia@diku.dk>]
825 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
826 needlessly dereferenced structures, used obsolete functions and
827 didn't handle all updated verify codes correctly.
830 *) Disable MD2 in the default configuration.
833 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
834 indicate the initial BIO being pushed or popped. This makes it possible
835 to determine whether the BIO is the one explicitly called or as a result
836 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
837 it handles reference counts correctly and doesn't zero out the I/O bio
838 when it is not being explicitly popped. WARNING: applications which
839 included workarounds for the old buggy behaviour will need to be modified
840 or they could free up already freed BIOs.
843 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
844 renaming to all platforms (within the 0.9.8 branch, this was
845 done conditionally on Netware platforms to avoid a name clash).
846 [Guenter <lists@gknw.net>]
848 *) Add ECDHE and PSK support to DTLS.
849 [Michael Tuexen <tuexen@fh-muenster.de>]
851 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
855 *) Add "missing" function EVP_MD_flags() (without this the only way to
856 retrieve a digest flags is by accessing the structure directly. Update
857 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
858 or cipher is registered as in the "from" argument. Print out all
859 registered digests in the dgst usage message instead of manually
860 attempting to work them out.
863 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
864 this allows the use of compression and extensions. Change default cipher
865 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
866 by default unless an application cipher string requests it.
869 *) Alter match criteria in PKCS12_parse(). It used to try to use local
870 key ids to find matching certificates and keys but some PKCS#12 files
871 don't follow the (somewhat unwritten) rules and this strategy fails.
872 Now just gather all certificates together and the first private key
873 then look for the first certificate that matches the key.
876 *) Support use of registered digest and cipher names for dgst and cipher
877 commands instead of having to add each one as a special case. So now
884 openssl dgst -sha256 foo
886 and this works for ENGINE based algorithms too.
890 *) Update Gost ENGINE to support parameter files.
891 [Victor B. Wagner <vitus@cryptocom.ru>]
893 *) Support GeneralizedTime in ca utility.
894 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
896 *) Enhance the hash format used for certificate directory links. The new
897 form uses the canonical encoding (meaning equivalent names will work
898 even if they aren't identical) and uses SHA1 instead of MD5. This form
899 is incompatible with the older format and as a result c_rehash should
900 be used to rebuild symbolic links.
903 *) Make PKCS#8 the default write format for private keys, replacing the
904 traditional format. This form is standardised, more secure and doesn't
905 include an implicit MD5 dependency.
908 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
909 committed to OpenSSL should pass this lot as a minimum.
912 *) Add session ticket override functionality for use by EAP-FAST.
913 [Jouni Malinen <j@w1.fi>]
915 *) Modify HMAC functions to return a value. Since these can be implemented
916 in an ENGINE errors can occur.
919 *) Type-checked OBJ_bsearch_ex.
922 *) Type-checked OBJ_bsearch. Also some constification necessitated
923 by type-checking. Still to come: TXT_DB, bsearch(?),
924 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
928 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
929 seconds to a tm structure directly, instead of going through OS
930 specific date routines. This avoids any issues with OS routines such
931 as the year 2038 bug. New *_adj() functions for ASN1 time structures
932 and X509_time_adj_ex() to cover the extended range. The existing
933 X509_time_adj() is still usable and will no longer have any date issues.
936 *) Delta CRL support. New use deltas option which will attempt to locate
937 and search any appropriate delta CRLs available.
939 This work was sponsored by Google.
942 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
943 code and add additional score elements. Validate alternate CRL paths
944 as part of the CRL checking and indicate a new error "CRL path validation
945 error" in this case. Applications wanting additional details can use
946 the verify callback and check the new "parent" field. If this is not
947 NULL CRL path validation is taking place. Existing applications wont
948 see this because it requires extended CRL support which is off by
951 This work was sponsored by Google.
954 *) Support for freshest CRL extension.
956 This work was sponsored by Google.
959 *) Initial indirect CRL support. Currently only supported in the CRLs
960 passed directly and not via lookup. Process certificate issuer
961 CRL entry extension and lookup CRL entries by bother issuer name
962 and serial number. Check and process CRL issuer entry in IDP extension.
964 This work was sponsored by Google.
967 *) Add support for distinct certificate and CRL paths. The CRL issuer
968 certificate is validated separately in this case. Only enabled if
969 an extended CRL support flag is set: this flag will enable additional
970 CRL functionality in future.
972 This work was sponsored by Google.
975 *) Add support for policy mappings extension.
977 This work was sponsored by Google.
980 *) Fixes to pathlength constraint, self issued certificate handling,
981 policy processing to align with RFC3280 and PKITS tests.
983 This work was sponsored by Google.
986 *) Support for name constraints certificate extension. DN, email, DNS
987 and URI types are currently supported.
989 This work was sponsored by Google.
992 *) To cater for systems that provide a pointer-based thread ID rather
993 than numeric, deprecate the current numeric thread ID mechanism and
994 replace it with a structure and associated callback type. This
995 mechanism allows a numeric "hash" to be extracted from a thread ID in
996 either case, and on platforms where pointers are larger than 'long',
997 mixing is done to help ensure the numeric 'hash' is usable even if it
998 can't be guaranteed unique. The default mechanism is to use "&errno"
999 as a pointer-based thread ID to distinguish between threads.
1001 Applications that want to provide their own thread IDs should now use
1002 CRYPTO_THREADID_set_callback() to register a callback that will call
1003 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
1005 Note that ERR_remove_state() is now deprecated, because it is tied
1006 to the assumption that thread IDs are numeric. ERR_remove_state(0)
1007 to free the current thread's error state should be replaced by
1008 ERR_remove_thread_state(NULL).
1010 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
1011 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
1012 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
1013 application was previously providing a numeric thread callback that
1014 was inappropriate for distinguishing threads, then uniqueness might
1015 have been obtained with &errno that happened immediately in the
1016 intermediate development versions of OpenSSL; this is no longer the
1017 case, the numeric thread callback will now override the automatic use
1019 [Geoff Thorpe, with help from Bodo Moeller]
1021 *) Initial support for different CRL issuing certificates. This covers a
1022 simple case where the self issued certificates in the chain exist and
1023 the real CRL issuer is higher in the existing chain.
1025 This work was sponsored by Google.
1028 *) Removed effectively defunct crypto/store from the build.
1031 *) Revamp of STACK to provide stronger type-checking. Still to come:
1032 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
1033 ASN1_STRING, CONF_VALUE.
1036 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
1037 RAM on SSL connections. This option can save about 34k per idle SSL.
1040 *) Revamp of LHASH to provide stronger type-checking. Still to come:
1041 STACK, TXT_DB, bsearch, qsort.
1044 *) Initial support for Cryptographic Message Syntax (aka CMS) based
1045 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
1046 support for data, signedData, compressedData, digestedData and
1047 encryptedData, envelopedData types included. Scripts to check against
1048 RFC4134 examples draft and interop and consistency checks of many
1049 content types and variants.
1052 *) Add options to enc utility to support use of zlib compression BIO.
1055 *) Extend mk1mf to support importing of options and assembly language
1056 files from Configure script, currently only included in VC-WIN32.
1057 The assembly language rules can now optionally generate the source
1058 files from the associated perl scripts.
1061 *) Implement remaining functionality needed to support GOST ciphersuites.
1062 Interop testing has been performed using CryptoPro implementations.
1063 [Victor B. Wagner <vitus@cryptocom.ru>]
1065 *) s390x assembler pack.
1068 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
1072 *) Implement Opaque PRF Input TLS extension as specified in
1073 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
1074 official specification yet and no extension type assignment by
1075 IANA exists, this extension (for now) will have to be explicitly
1076 enabled when building OpenSSL by providing the extension number
1077 to use. For example, specify an option
1079 -DTLSEXT_TYPE_opaque_prf_input=0x9527
1081 to the "config" or "Configure" script to enable the extension,
1082 assuming extension number 0x9527 (which is a completely arbitrary
1083 and unofficial assignment based on the MD5 hash of the Internet
1084 Draft). Note that by doing so, you potentially lose
1085 interoperability with other TLS implementations since these might
1086 be using the same extension number for other purposes.
1088 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
1089 opaque PRF input value to use in the handshake. This will create
1090 an interal copy of the length-'len' string at 'src', and will
1091 return non-zero for success.
1093 To get more control and flexibility, provide a callback function
1096 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
1097 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
1101 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
1104 Callback function 'cb' will be called in handshakes, and is
1105 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
1106 Argument 'arg' is for application purposes (the value as given to
1107 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
1108 be provided to the callback function). The callback function
1109 has to return non-zero to report success: usually 1 to use opaque
1110 PRF input just if possible, or 2 to enforce use of the opaque PRF
1111 input. In the latter case, the library will abort the handshake
1112 if opaque PRF input is not successfully negotiated.
1114 Arguments 'peerinput' and 'len' given to the callback function
1115 will always be NULL and 0 in the case of a client. A server will
1116 see the client's opaque PRF input through these variables if
1117 available (NULL and 0 otherwise). Note that if the server
1118 provides an opaque PRF input, the length must be the same as the
1119 length of the client's opaque PRF input.
1121 Note that the callback function will only be called when creating
1122 a new session (session resumption can resume whatever was
1123 previously negotiated), and will not be called in SSL 2.0
1124 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
1125 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
1126 for applications that need to enforce opaque PRF input.
1130 *) Update ssl code to support digests other than SHA1+MD5 for handshake
1133 [Victor B. Wagner <vitus@cryptocom.ru>]
1135 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1136 RFC4507bis. The encrypted ticket format is an encrypted encoded
1137 SSL_SESSION structure, that way new session features are automatically
1140 If a client application caches session in an SSL_SESSION structure
1141 support is transparent because tickets are now stored in the encoded
1144 The SSL_CTX structure automatically generates keys for ticket
1145 protection in servers so again support should be possible
1146 with no application modification.
1148 If a client or server wishes to disable RFC4507 support then the option
1149 SSL_OP_NO_TICKET can be set.
1151 Add a TLS extension debugging callback to allow the contents of any client
1152 or server extensions to be examined.
1154 This work was sponsored by Google.
1157 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
1158 OpenSSL should now compile cleanly on gcc 4.2
1159 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
1161 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
1162 support including streaming MAC support: this is required for GOST
1163 ciphersuite support.
1164 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
1166 *) Add option -stream to use PKCS#7 streaming in smime utility. New
1167 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
1168 to output in BER and PEM format.
1171 *) Experimental support for use of HMAC via EVP_PKEY interface. This
1172 allows HMAC to be handled via the EVP_DigestSign*() interface. The
1173 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
1174 ENGINE support for HMAC keys which are unextractable. New -mac and
1175 -macopt options to dgst utility.
1178 *) New option -sigopt to dgst utility. Update dgst to use
1179 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
1180 alternative signing paramaters such as X9.31 or PSS in the dgst
1184 *) Change ssl_cipher_apply_rule(), the internal function that does
1185 the work each time a ciphersuite string requests enabling
1186 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
1187 removing ("!foo+bar") a class of ciphersuites: Now it maintains
1188 the order of disabled ciphersuites such that those ciphersuites
1189 that most recently went from enabled to disabled not only stay
1190 in order with respect to each other, but also have higher priority
1191 than other disabled ciphersuites the next time ciphersuites are
1194 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
1195 the same ciphersuites as with "HIGH" alone, but in a specific
1196 order where the PSK ciphersuites come first (since they are the
1197 most recently disabled ciphersuites when "HIGH" is parsed).
1199 Also, change ssl_create_cipher_list() (using this new
1200 funcionality) such that between otherwise identical
1201 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
1205 *) Change ssl_create_cipher_list() so that it automatically
1206 arranges the ciphersuites in reasonable order before starting
1207 to process the rule string. Thus, the definition for "DEFAULT"
1208 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
1209 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
1210 This makes it much easier to arrive at a reasonable default order
1211 in applications for which anonymous ciphers are OK (meaning
1212 that you can't actually use DEFAULT).
1213 [Bodo Moeller; suggested by Victor Duchovni]
1215 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
1216 processing) into multiple integers instead of setting
1217 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
1218 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
1219 (These masks as well as the individual bit definitions are hidden
1220 away into the non-exported interface ssl/ssl_locl.h, so this
1221 change to the definition of the SSL_CIPHER structure shouldn't
1222 affect applications.) This give us more bits for each of these
1223 categories, so there is no longer a need to coagulate AES128 and
1224 AES256 into a single algorithm bit, and to coagulate Camellia128
1225 and Camellia256 into a single algorithm bit, which has led to all
1228 Thus, among other things, the kludge introduced in 0.9.7m and
1229 0.9.8e for masking out AES256 independently of AES128 or masking
1230 out Camellia256 independently of AES256 is not needed here in 0.9.9.
1232 With the change, we also introduce new ciphersuite aliases that
1233 so far were missing: "AES128", "AES256", "CAMELLIA128", and
1237 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
1238 Use the leftmost N bytes of the signature input if the input is
1239 larger than the prime q (with N being the size in bytes of q).
1242 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
1243 it yet and it is largely untested.
1246 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
1249 *) Initial incomplete changes to avoid need for function casts in OpenSSL
1250 some compilers (gcc 4.2 and later) reject their use. Safestack is
1251 reimplemented. Update ASN1 to avoid use of legacy functions.
1254 *) Win32/64 targets are linked with Winsock2.
1257 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
1258 to external functions. This can be used to increase CRL handling
1259 efficiency especially when CRLs are very large by (for example) storing
1260 the CRL revoked certificates in a database.
1263 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
1264 new CRLs added to a directory can be used. New command line option
1265 -verify_return_error to s_client and s_server. This causes real errors
1266 to be returned by the verify callback instead of carrying on no matter
1267 what. This reflects the way a "real world" verify callback would behave.
1270 *) GOST engine, supporting several GOST algorithms and public key formats.
1271 Kindly donated by Cryptocom.
1274 *) Partial support for Issuing Distribution Point CRL extension. CRLs
1275 partitioned by DP are handled but no indirect CRL or reason partitioning
1276 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
1277 selected via a scoring technique which handles IDP and AKID in CRLs.
1280 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
1281 will ultimately be used for all verify operations: this will remove the
1282 X509_STORE dependency on certificate verification and allow alternative
1283 lookup methods. X509_STORE based implementations of these two callbacks.
1286 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
1287 Modify get_crl() to find a valid (unexpired) CRL if possible.
1290 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
1291 this would be called X509_CRL_cmp() but that name is already used by
1292 a function that just compares CRL issuer names. Cache several CRL
1293 extensions in X509_CRL structure and cache CRLDP in X509.
1296 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
1297 this maps equivalent X509_NAME structures into a consistent structure.
1298 Name comparison can then be performed rapidly using memcmp().
1301 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
1305 *) Allow digests to supply their own micalg string for S/MIME type using
1306 the ctrl EVP_MD_CTRL_MICALG.
1309 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
1310 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
1311 ctrl. It can then customise the structure before and/or after signing
1315 *) New function OBJ_add_sigid() to allow application defined signature OIDs
1316 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
1317 to free up any added signature OIDs.
1320 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
1321 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
1322 digest and cipher tables. New options added to openssl utility:
1323 list-message-digest-algorithms and list-cipher-algorithms.
1326 *) Change the array representation of binary polynomials: the list
1327 of degrees of non-zero coefficients is now terminated with -1.
1328 Previously it was terminated with 0, which was also part of the
1329 value; thus, the array representation was not applicable to
1330 polynomials where t^0 has coefficient zero. This change makes
1331 the array representation useful in a more general context.
1334 *) Various modifications and fixes to SSL/TLS cipher string
1335 handling. For ECC, the code now distinguishes between fixed ECDH
1336 with RSA certificates on the one hand and with ECDSA certificates
1337 on the other hand, since these are separate ciphersuites. The
1338 unused code for Fortezza ciphersuites has been removed.
1340 For consistency with EDH, ephemeral ECDH is now called "EECDH"
1341 (not "ECDHE"). For consistency with the code for DH
1342 certificates, use of ECDH certificates is now considered ECDH
1343 authentication, not RSA or ECDSA authentication (the latter is
1344 merely the CA's signing algorithm and not actively used in the
1347 The temporary ciphersuite alias "ECCdraft" is no longer
1348 available, and ECC ciphersuites are no longer excluded from "ALL"
1349 and "DEFAULT". The following aliases now exist for RFC 4492
1350 ciphersuites, most of these by analogy with the DH case:
1352 kECDHr - ECDH cert, signed with RSA
1353 kECDHe - ECDH cert, signed with ECDSA
1354 kECDH - ECDH cert (signed with either RSA or ECDSA)
1355 kEECDH - ephemeral ECDH
1356 ECDH - ECDH cert or ephemeral ECDH
1362 AECDH - anonymous ECDH
1363 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
1367 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
1368 Use correct micalg parameters depending on digest(s) in signed message.
1371 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
1372 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
1375 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
1376 an engine to register a method. Add ENGINE lookups for methods and
1377 functional reference processing.
1380 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
1381 EVP_{Sign,Verify}* which allow an application to customise the signature
1385 *) New -resign option to smime utility. This adds one or more signers
1386 to an existing PKCS#7 signedData structure. Also -md option to use an
1387 alternative message digest algorithm for signing.
1390 *) Tidy up PKCS#7 routines and add new functions to make it easier to
1391 create PKCS7 structures containing multiple signers. Update smime
1392 application to support multiple signers.
1395 *) New -macalg option to pkcs12 utility to allow setting of an alternative
1399 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
1400 Reorganize PBE internals to lookup from a static table using NIDs,
1401 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
1402 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
1403 PRF which will be automatically used with PBES2.
1406 *) Replace the algorithm specific calls to generate keys in "req" with the
1410 *) Update PKCS#7 enveloped data routines to use new API. This is now
1411 supported by any public key method supporting the encrypt operation. A
1412 ctrl is added to allow the public key algorithm to examine or modify
1413 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
1417 *) Add a ctrl to asn1 method to allow a public key algorithm to express
1418 a default digest type to use. In most cases this will be SHA1 but some
1419 algorithms (such as GOST) need to specify an alternative digest. The
1420 return value indicates how strong the prefernce is 1 means optional and
1421 2 is mandatory (that is it is the only supported type). Modify
1422 ASN1_item_sign() to accept a NULL digest argument to indicate it should
1423 use the default md. Update openssl utilities to use the default digest
1424 type for signing if it is not explicitly indicated.
1427 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
1428 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
1429 signing method from the key type. This effectively removes the link
1430 between digests and public key types.
1433 *) Add an OID cross reference table and utility functions. Its purpose is to
1434 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
1435 rsaEncryption. This will allow some of the algorithm specific hackery
1436 needed to use the correct OID to be removed.
1439 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
1440 structures for PKCS7_sign(). They are now set up by the relevant public
1444 *) Add provisional EC pkey method with support for ECDSA and ECDH.
1447 *) Add support for key derivation (agreement) in the API, DH method and
1451 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
1452 public and private key formats. As a side effect these add additional
1453 command line functionality not previously available: DSA signatures can be
1454 generated and verified using pkeyutl and DH key support and generation in
1459 [Oliver Tappe <zooey@hirschkaefer.de>]
1461 *) New make target "install_html_docs" installs HTML renditions of the
1463 [Oliver Tappe <zooey@hirschkaefer.de>]
1465 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
1466 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
1467 support key and parameter generation and add initial key generation
1468 functionality for RSA.
1471 *) Add functions for main EVP_PKEY_method operations. The undocumented
1472 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
1473 EVP_PKEY_{encrypt,decrypt}_old.
1476 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
1477 key API, doesn't do much yet.
1480 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
1481 public key algorithms. New option to openssl utility:
1482 "list-public-key-algorithms" to print out info.
1485 *) Implement the Supported Elliptic Curves Extension for
1486 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1489 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
1490 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
1493 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
1494 utilities such as rsa, dsa, dsaparam etc except they process any key
1498 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1499 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1500 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
1504 *) Initial support for pluggable public key ASN1.
1505 De-spaghettify the public key ASN1 handling. Move public and private
1506 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
1507 algorithm specific handling to a single module within the relevant
1508 algorithm directory. Add functions to allow (near) opaque processing
1509 of public and private key structures.
1512 *) Implement the Supported Point Formats Extension for
1513 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1516 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
1517 for the psk identity [hint] and the psk callback functions to the
1518 SSL_SESSION, SSL and SSL_CTX structure.
1521 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
1525 SSL_CTX_use_psk_identity_hint
1526 SSL_get_psk_identity_hint
1527 SSL_get_psk_identity
1528 SSL_use_psk_identity_hint
1530 [Mika Kousa and Pasi Eronen of Nokia Corporation]
1532 *) Add RFC 3161 compliant time stamp request creation, response generation
1533 and response verification functionality.
1534 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
1536 *) Add initial support for TLS extensions, specifically for the server_name
1537 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1538 have new members for a host name. The SSL data structure has an
1539 additional member SSL_CTX *initial_ctx so that new sessions can be
1540 stored in that context to allow for session resumption, even after the
1541 SSL has been switched to a new SSL_CTX in reaction to a client's
1542 server_name extension.
1544 New functions (subject to change):
1546 SSL_get_servername()
1547 SSL_get_servername_type()
1550 New CTRL codes and macros (subject to change):
1552 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1553 - SSL_CTX_set_tlsext_servername_callback()
1554 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1555 - SSL_CTX_set_tlsext_servername_arg()
1556 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1558 openssl s_client has a new '-servername ...' option.
1560 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1561 '-key2 ...', '-servername_fatal' (subject to change). This allows
1562 testing the HostName extension for a specific single host name ('-cert'
1563 and '-key' remain fallbacks for handshakes without HostName
1564 negotiation). If the unrecogninzed_name alert has to be sent, this by
1565 default is a warning; it becomes fatal with the '-servername_fatal'
1568 [Peter Sylvester, Remy Allais, Christophe Renou]
1570 *) Whirlpool hash implementation is added.
1573 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1574 bn(64,32). Because of instruction set limitations it doesn't have
1575 any negative impact on performance. This was done mostly in order
1576 to make it possible to share assembler modules, such as bn_mul_mont
1577 implementations, between 32- and 64-bit builds without hassle.
1580 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1581 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1585 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1586 dedicated Montgomery multiplication procedure, is introduced.
1587 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1588 "64-bit" performance on certain 32-bit targets.
1591 *) New option SSL_OP_NO_COMP to disable use of compression selectively
1592 in SSL structures. New SSL ctrl to set maximum send fragment size.
1593 Save memory by seeting the I/O buffer sizes dynamically instead of
1594 using the maximum available value.
1597 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
1598 in addition to the text details.
1601 *) Very, very preliminary EXPERIMENTAL support for printing of general
1602 ASN1 structures. This currently produces rather ugly output and doesn't
1603 handle several customised structures at all.
1606 *) Integrated support for PVK file format and some related formats such
1607 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
1608 these in the 'rsa' and 'dsa' utilities.
1611 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
1614 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
1615 place for the (very old) "NETSCAPE" format certificates which are now
1616 handled using new ASN1 code equivalents.
1619 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
1620 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
1621 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
1624 *) Modify CRL distribution points extension code to print out previously
1625 unsupported fields. Enhance extension setting code to allow setting of
1629 *) Add print and set support for Issuing Distribution Point CRL extension.
1632 *) Change 'Configure' script to enable Camellia by default.
1635 Changes between 0.9.8x and 0.9.8y [5 Feb 2013]
1637 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1639 This addresses the flaw in CBC record processing discovered by
1640 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1641 at: http://www.isg.rhul.ac.uk/tls/
1643 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1644 Security Group at Royal Holloway, University of London
1645 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1646 Emilia Käsper for the initial patch.
1648 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1650 *) Return an error when checking OCSP signatures when key is NULL.
1651 This fixes a DoS attack. (CVE-2013-0166)
1654 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1655 the right response is stapled. Also change SSL_get_certificate()
1656 so it returns the certificate actually sent.
1657 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1658 (This is a backport)
1659 [Rob Stradling <rob.stradling@comodo.com>]
1661 *) Fix possible deadlock when decoding public keys.
1664 Changes between 0.9.8w and 0.9.8x [10 May 2012]
1666 *) Sanity check record length before skipping explicit IV in DTLS
1669 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1670 fuzzing as a service testing platform.
1674 *) Initialise tkeylen properly when encrypting CMS messages.
1675 Thanks to Solar Designer of Openwall for reporting this issue.
1678 Changes between 0.9.8v and 0.9.8w [23 Apr 2012]
1680 *) The fix for CVE-2012-2110 did not take into account that the
1681 'len' argument to BUF_MEM_grow and BUF_MEM_grow_clean is an
1682 int in OpenSSL 0.9.8, making it still vulnerable. Fix by
1683 rejecting negative len parameter. (CVE-2012-2131)
1684 [Tomas Hoger <thoger@redhat.com>]
1686 Changes between 0.9.8u and 0.9.8v [19 Apr 2012]
1688 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1689 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1690 in CRYPTO_realloc_clean.
1692 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1693 issue and to Adam Langley <agl@chromium.org> for fixing it.
1695 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1697 Changes between 0.9.8t and 0.9.8u [12 Mar 2012]
1699 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1700 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1701 content decryption and always return the same error. Note: this attack
1702 needs on average 2^20 messages so it only affects automated senders. The
1703 old behaviour can be reenabled in the CMS code by setting the
1704 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1705 an MMA defence is not necessary.
1706 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1707 this issue. (CVE-2012-0884)
1710 *) Fix CVE-2011-4619: make sure we really are receiving a
1711 client hello before rejecting multiple SGC restarts. Thanks to
1712 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1715 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
1717 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1718 Thanks to Antonio Martin, Enterprise Secure Access Research and
1719 Development, Cisco Systems, Inc. for discovering this bug and
1720 preparing a fix. (CVE-2012-0050)
1723 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
1725 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1726 of the Vaudenay padding oracle attack on CBC mode encryption
1727 which enables an efficient plaintext recovery attack against
1728 the OpenSSL implementation of DTLS. Their attack exploits timing
1729 differences arising during decryption processing. A research
1730 paper describing this attack can be found at:
1731 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1732 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1733 Security Group at Royal Holloway, University of London
1734 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1735 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1736 for preparing the fix. (CVE-2011-4108)
1737 [Robin Seggelmann, Michael Tuexen]
1739 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
1740 [Ben Laurie, Kasper <ekasper@google.com>]
1742 *) Clear bytes used for block padding of SSL 3.0 records.
1744 [Adam Langley (Google)]
1746 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1747 Kadianakis <desnacked@gmail.com> for discovering this issue and
1748 Adam Langley for preparing the fix. (CVE-2011-4619)
1749 [Adam Langley (Google)]
1751 *) Prevent malformed RFC3779 data triggering an assertion failure.
1752 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1753 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1754 [Rob Austein <sra@hactrn.net>]
1756 *) Fix ssl_ciph.c set-up race.
1757 [Adam Langley (Google)]
1759 *) Fix spurious failures in ecdsatest.c.
1760 [Emilia Käsper (Google)]
1762 *) Fix the BIO_f_buffer() implementation (which was mixing different
1763 interpretations of the '..._len' fields).
1764 [Adam Langley (Google)]
1766 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1767 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1768 threads won't reuse the same blinding coefficients.
1770 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1771 lock to call BN_BLINDING_invert_ex, and avoids one use of
1772 BN_BLINDING_update for each BN_BLINDING structure (previously,
1773 the last update always remained unused).
1774 [Emilia Käsper (Google)]
1776 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1777 for multi-threaded use of ECDH.
1778 [Adam Langley (Google)]
1780 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1783 *) Add protection against ECDSA timing attacks as mentioned in the paper
1784 by Billy Bob Brumley and Nicola Tuveri, see:
1786 http://eprint.iacr.org/2011/232.pdf
1788 [Billy Bob Brumley and Nicola Tuveri]
1790 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
1792 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1793 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1795 *) Fix bug in string printing code: if *any* escaping is enabled we must
1796 escape the escape character (backslash) or the resulting string is
1800 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
1802 *) Disable code workaround for ancient and obsolete Netscape browsers
1803 and servers: an attacker can use it in a ciphersuite downgrade attack.
1804 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1807 *) Fixed J-PAKE implementation error, originally discovered by
1808 Sebastien Martini, further info and confirmation from Stefan
1809 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1812 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
1814 *) Fix extension code to avoid race conditions which can result in a buffer
1815 overrun vulnerability: resumed sessions must not be modified as they can
1816 be shared by multiple threads. CVE-2010-3864
1819 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
1822 *) Don't reencode certificate when calculating signature: cache and use
1823 the original encoding instead. This makes signature verification of
1824 some broken encodings work correctly.
1827 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
1828 is also one of the inputs.
1829 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1831 *) Don't repeatedly append PBE algorithms to table if they already exist.
1832 Sort table on each new add. This effectively makes the table read only
1833 after all algorithms are added and subsequent calls to PKCS12_pbe_add
1837 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
1839 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
1842 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
1843 access or freeing data twice (CVE-2010-0742)
1844 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
1846 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
1847 common in certificates and some applications which only call
1848 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
1852 Reduce copying into .apps and .test in makevms.com
1853 Don't try to use blank CA certificate in CA.com
1854 Allow use of C files from original directories in maketests.com
1855 [Steven M. Schweda" <sms@antinode.info>]
1857 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
1859 *) When rejecting SSL/TLS records due to an incorrect version number, never
1860 update s->server with a new major version number. As of
1861 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
1862 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
1863 the previous behavior could result in a read attempt at NULL when
1864 receiving specific incorrect SSL/TLS records once record payload
1865 protection is active. (CVE-2010-0740)
1866 [Bodo Moeller, Adam Langley <agl@chromium.org>]
1868 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
1869 could be crashed if the relevant tables were not present (e.g. chrooted).
1870 [Tomas Hoger <thoger@redhat.com>]
1872 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
1874 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
1875 [Martin Olsson, Neel Mehta]
1877 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
1878 accommodate for stack sorting, always a write lock!).
1881 *) On some versions of WIN32 Heap32Next is very slow. This can cause
1882 excessive delays in the RAND_poll(): over a minute. As a workaround
1883 include a time check in the inner Heap32Next loop too.
1886 *) The code that handled flushing of data in SSL/TLS originally used the
1887 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
1888 the problem outlined in PR#1949. The fix suggested there however can
1889 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
1890 of Apache). So instead simplify the code to flush unconditionally.
1891 This should be fine since flushing with no data to flush is a no op.
1894 *) Handle TLS versions 2.0 and later properly and correctly use the
1895 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
1896 off ancient servers have a habit of sticking around for a while...
1899 *) Modify compression code so it frees up structures without using the
1900 ex_data callbacks. This works around a problem where some applications
1901 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
1902 restarting) then use compression (e.g. SSL with compression) later.
1903 This results in significant per-connection memory leaks and
1904 has caused some security issues including CVE-2008-1678 and
1908 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
1909 change when encrypting or decrypting.
1912 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
1913 connect and renegotiate with servers which do not support RI.
1914 Until RI is more widely deployed this option is enabled by default.
1917 *) Add "missing" ssl ctrls to clear options and mode.
1920 *) If client attempts to renegotiate and doesn't support RI respond with
1921 a no_renegotiation alert as required by RFC5746. Some renegotiating
1922 TLS clients will continue a connection gracefully when they receive
1923 the alert. Unfortunately OpenSSL mishandled this alert and would hang
1924 waiting for a server hello which it will never receive. Now we treat a
1925 received no_renegotiation alert as a fatal error. This is because
1926 applications requesting a renegotiation might well expect it to succeed
1927 and would have no code in place to handle the server denying it so the
1928 only safe thing to do is to terminate the connection.
1931 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
1932 peer supports secure renegotiation and 0 otherwise. Print out peer
1933 renegotiation support in s_client/s_server.
1936 *) Replace the highly broken and deprecated SPKAC certification method with
1937 the updated NID creation version. This should correctly handle UTF8.
1940 *) Implement RFC5746. Re-enable renegotiation but require the extension
1941 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
1942 turns out to be a bad idea. It has been replaced by
1943 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
1944 SSL_CTX_set_options(). This is really not recommended unless you
1945 know what you are doing.
1946 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
1948 *) Fixes to stateless session resumption handling. Use initial_ctx when
1949 issuing and attempting to decrypt tickets in case it has changed during
1950 servername handling. Use a non-zero length session ID when attempting
1951 stateless session resumption: this makes it possible to determine if
1952 a resumption has occurred immediately after receiving server hello
1953 (several places in OpenSSL subtly assume this) instead of later in
1957 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
1958 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
1959 fixes for a few places where the return code is not checked
1961 [Julia Lawall <julia@diku.dk>]
1963 *) Add --strict-warnings option to Configure script to include devteam
1964 warnings in other configurations.
1967 *) Add support for --libdir option and LIBDIR variable in makefiles. This
1968 makes it possible to install openssl libraries in locations which
1969 have names other than "lib", for example "/usr/lib64" which some
1971 [Steve Henson, based on patch from Jeremy Utley]
1973 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
1974 X690 8.9.12 and can produce some misleading textual output of OIDs.
1975 [Steve Henson, reported by Dan Kaminsky]
1977 *) Delete MD2 from algorithm tables. This follows the recommendation in
1978 several standards that it is not used in new applications due to
1979 several cryptographic weaknesses. For binary compatibility reasons
1980 the MD2 API is still compiled in by default.
1983 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
1987 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
1988 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
1990 [Guenter <lists@gknw.net>]
1992 *) Fix the server certificate chain building code to use X509_verify_cert(),
1993 it used to have an ad-hoc builder which was unable to cope with anything
1994 other than a simple chain.
1995 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
1997 *) Don't check self signed certificate signatures in X509_verify_cert()
1998 by default (a flag can override this): it just wastes time without
1999 adding any security. As a useful side effect self signed root CAs
2000 with non-FIPS digests are now usable in FIPS mode.
2003 *) In dtls1_process_out_of_seq_message() the check if the current message
2004 is already buffered was missing. For every new message was memory
2005 allocated, allowing an attacker to perform an denial of service attack
2006 with sending out of seq handshake messages until there is no memory
2007 left. Additionally every future messege was buffered, even if the
2008 sequence number made no sense and would be part of another handshake.
2009 So only messages with sequence numbers less than 10 in advance will be
2010 buffered. (CVE-2009-1378)
2011 [Robin Seggelmann, discovered by Daniel Mentz]
2013 *) Records are buffered if they arrive with a future epoch to be
2014 processed after finishing the corresponding handshake. There is
2015 currently no limitation to this buffer allowing an attacker to perform
2016 a DOS attack with sending records with future epochs until there is no
2017 memory left. This patch adds the pqueue_size() function to detemine
2018 the size of a buffer and limits the record buffer to 100 entries.
2020 [Robin Seggelmann, discovered by Daniel Mentz]
2022 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
2023 parent structure is freed. (CVE-2009-1379)
2026 *) Handle non-blocking I/O properly in SSL_shutdown() call.
2027 [Darryl Miles <darryl-mailinglists@netbauds.net>]
2030 [Ilya O. <vrghost@gmail.com>]
2032 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
2034 *) Disable renegotiation completely - this fixes a severe security
2035 problem (CVE-2009-3555) at the cost of breaking all
2036 renegotiation. Renegotiation can be re-enabled by setting
2037 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
2038 run-time. This is really not recommended unless you know what
2042 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
2044 *) Don't set val to NULL when freeing up structures, it is freed up by
2045 underlying code. If sizeof(void *) > sizeof(long) this can result in
2046 zeroing past the valid field. (CVE-2009-0789)
2047 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
2049 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
2050 checked correctly. This would allow some invalid signed attributes to
2051 appear to verify correctly. (CVE-2009-0591)
2052 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2054 *) Reject UniversalString and BMPString types with invalid lengths. This
2055 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
2056 a legal length. (CVE-2009-0590)
2059 *) Set S/MIME signing as the default purpose rather than setting it
2060 unconditionally. This allows applications to override it at the store
2064 *) Permit restricted recursion of ASN1 strings. This is needed in practice
2065 to handle some structures.
2068 *) Improve efficiency of mem_gets: don't search whole buffer each time
2070 [Jeremy Shapiro <jnshapir@us.ibm.com>]
2072 *) New -hex option for openssl rand.
2075 *) Print out UTF8String and NumericString when parsing ASN1.
2078 *) Support NumericString type for name components.
2081 *) Allow CC in the environment to override the automatically chosen
2082 compiler. Note that nothing is done to ensure flags work with the
2086 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
2088 *) Properly check EVP_VerifyFinal() and similar return values
2090 [Ben Laurie, Bodo Moeller, Google Security Team]
2092 *) Enable TLS extensions by default.
2095 *) Allow the CHIL engine to be loaded, whether the application is
2096 multithreaded or not. (This does not release the developer from the
2097 obligation to set up the dynamic locking callbacks.)
2098 [Sander Temme <sander@temme.net>]
2100 *) Use correct exit code if there is an error in dgst command.
2101 [Steve Henson; problem pointed out by Roland Dirlewanger]
2103 *) Tweak Configure so that you need to say "experimental-jpake" to enable
2104 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
2107 *) Add experimental JPAKE support, including demo authentication in
2108 s_client and s_server.
2111 *) Set the comparison function in v3_addr_canonize().
2112 [Rob Austein <sra@hactrn.net>]
2114 *) Add support for XMPP STARTTLS in s_client.
2115 [Philip Paeps <philip@freebsd.org>]
2117 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
2118 to ensure that even with this option, only ciphersuites in the
2119 server's preference list will be accepted. (Note that the option
2120 applies only when resuming a session, so the earlier behavior was
2121 just about the algorithm choice for symmetric cryptography.)
2124 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
2126 *) Fix NULL pointer dereference if a DTLS server received
2127 ChangeCipherSpec as first record (CVE-2009-1386).
2130 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
2131 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
2134 *) The fix in 0.9.8c that supposedly got rid of unsafe
2135 double-checked locking was incomplete for RSA blinding,
2136 addressing just one layer of what turns out to have been
2137 doubly unsafe triple-checked locking.
2139 So now fix this for real by retiring the MONT_HELPER macro
2140 in crypto/rsa/rsa_eay.c.
2142 [Bodo Moeller; problem pointed out by Marius Schilder]
2144 *) Various precautionary measures:
2146 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
2148 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
2149 (NB: This would require knowledge of the secret session ticket key
2150 to exploit, in which case you'd be SOL either way.)
2152 - Change bn_nist.c so that it will properly handle input BIGNUMs
2153 outside the expected range.
2155 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
2158 [Neel Mehta, Bodo Moeller]
2160 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
2161 the load fails. Useful for distros.
2162 [Ben Laurie and the FreeBSD team]
2164 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
2167 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
2170 *) Expand ENGINE to support engine supplied SSL client certificate functions.
2172 This work was sponsored by Logica.
2175 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
2176 keystores. Support for SSL/TLS client authentication too.
2177 Not compiled unless enable-capieng specified to Configure.
2179 This work was sponsored by Logica.
2182 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
2183 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
2184 attribute creation routines such as certifcate requests and PKCS#12
2188 Changes between 0.9.8g and 0.9.8h [28 May 2008]
2190 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
2191 handshake which could lead to a cilent crash as found using the
2192 Codenomicon TLS test suite (CVE-2008-1672)
2193 [Steve Henson, Mark Cox]
2195 *) Fix double free in TLS server name extensions which could lead to
2196 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
2199 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
2201 Clear the error queue to ensure that error entries left from
2202 older function calls do not interfere with the correct operation.
2203 [Lutz Jaenicke, Erik de Castro Lopo]
2205 *) Remove root CA certificates of commercial CAs:
2207 The OpenSSL project does not recommend any specific CA and does not
2208 have any policy with respect to including or excluding any CA.
2209 Therefore it does not make any sense to ship an arbitrary selection
2210 of root CA certificates with the OpenSSL software.
2213 *) RSA OAEP patches to fix two separate invalid memory reads.
2214 The first one involves inputs when 'lzero' is greater than
2215 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
2216 before the beginning of from). The second one involves inputs where
2217 the 'db' section contains nothing but zeroes (there is a one-byte
2218 invalid read after the end of 'db').
2219 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2221 *) Partial backport from 0.9.9-dev:
2223 Introduce bn_mul_mont (dedicated Montgomery multiplication
2224 procedure) as a candidate for BIGNUM assembler implementation.
2225 While 0.9.9-dev uses assembler for various architectures, only
2226 x86_64 is available by default here in the 0.9.8 branch, and
2227 32-bit x86 is available through a compile-time setting.
2229 To try the 32-bit x86 assembler implementation, use Configure
2230 option "enable-montasm" (which exists only for this backport).
2232 As "enable-montasm" for 32-bit x86 disclaims code stability
2233 anyway, in this constellation we activate additional code
2234 backported from 0.9.9-dev for further performance improvements,
2235 namely BN_from_montgomery_word. (To enable this otherwise,
2236 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
2238 [Andy Polyakov (backport partially by Bodo Moeller)]
2240 *) Add TLS session ticket callback. This allows an application to set
2241 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
2242 values. This is useful for key rollover for example where several key
2243 sets may exist with different names.
2246 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
2247 This was broken until now in 0.9.8 releases, such that the only way
2248 a registered ENGINE could be used (assuming it initialises
2249 successfully on the host) was to explicitly set it as the default
2250 for the relevant algorithms. This is in contradiction with 0.9.7
2251 behaviour and the documentation. With this fix, when an ENGINE is
2252 registered into a given algorithm's table of implementations, the
2253 'uptodate' flag is reset so that auto-discovery will be used next
2254 time a new context for that algorithm attempts to select an
2256 [Ian Lister (tweaked by Geoff Thorpe)]
2258 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
2259 implemention in the following ways:
2261 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
2264 Lack of BER streaming support means one pass streaming processing is
2265 only supported if data is detached: setting the streaming flag is
2266 ignored for embedded content.
2268 CMS support is disabled by default and must be explicitly enabled
2269 with the enable-cms configuration option.
2272 *) Update the GMP engine glue to do direct copies between BIGNUM and
2273 mpz_t when openssl and GMP use the same limb size. Otherwise the
2274 existing "conversion via a text string export" trick is still used.
2275 [Paul Sheer <paulsheer@gmail.com>]
2277 *) Zlib compression BIO. This is a filter BIO which compressed and
2278 uncompresses any data passed through it.
2281 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
2282 RFC3394 compatible AES key wrapping.
2285 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
2286 sets string data without copying. X509_ALGOR_set0() and
2287 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
2288 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
2289 from an X509_ATTRIBUTE structure optionally checking it occurs only
2290 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
2294 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
2295 to get the expected BN_FLG_CONSTTIME behavior.
2296 [Bodo Moeller (Google)]
2300 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
2301 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
2302 - added some more tests to do_tests.pl
2303 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
2304 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
2305 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
2306 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
2307 - various changes to netware.pl to enable gcc-cross builds on Win32
2309 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
2310 - various changes to fix missing prototype warnings
2311 - fixed x86nasm.pl to create correct asm files for NASM COFF output
2312 - added AES, WHIRLPOOL and CPUID assembler code to build files
2313 - added missing AES assembler make rules to mk1mf.pl
2314 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
2315 [Guenter Knauf <eflash@gmx.net>]
2317 *) Implement certificate status request TLS extension defined in RFC3546.
2318 A client can set the appropriate parameters and receive the encoded
2319 OCSP response via a callback. A server can query the supplied parameters
2320 and set the encoded OCSP response in the callback. Add simplified examples
2321 to s_client and s_server.
2324 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
2326 *) Fix various bugs:
2327 + Binary incompatibility of ssl_ctx_st structure
2328 + DTLS interoperation with non-compliant servers
2329 + Don't call get_session_cb() without proposed session
2330 + Fix ia64 assembler code
2331 [Andy Polyakov, Steve Henson]
2333 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
2335 *) DTLS Handshake overhaul. There were longstanding issues with
2336 OpenSSL DTLS implementation, which were making it impossible for
2337 RFC 4347 compliant client to communicate with OpenSSL server.
2338 Unfortunately just fixing these incompatibilities would "cut off"
2339 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
2340 server keeps tolerating non RFC compliant syntax. The opposite is
2341 not true, 0.9.8f client can not communicate with earlier server.
2342 This update even addresses CVE-2007-4995.
2345 *) Changes to avoid need for function casts in OpenSSL: some compilers
2346 (gcc 4.2 and later) reject their use.
2347 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
2350 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2351 RFC4507bis. The encrypted ticket format is an encrypted encoded
2352 SSL_SESSION structure, that way new session features are automatically
2355 If a client application caches session in an SSL_SESSION structure
2356 support is transparent because tickets are now stored in the encoded
2359 The SSL_CTX structure automatically generates keys for ticket
2360 protection in servers so again support should be possible
2361 with no application modification.
2363 If a client or server wishes to disable RFC4507 support then the option
2364 SSL_OP_NO_TICKET can be set.
2366 Add a TLS extension debugging callback to allow the contents of any client
2367 or server extensions to be examined.
2369 This work was sponsored by Google.
2372 *) Add initial support for TLS extensions, specifically for the server_name
2373 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2374 have new members for a host name. The SSL data structure has an
2375 additional member SSL_CTX *initial_ctx so that new sessions can be
2376 stored in that context to allow for session resumption, even after the
2377 SSL has been switched to a new SSL_CTX in reaction to a client's
2378 server_name extension.
2380 New functions (subject to change):
2382 SSL_get_servername()
2383 SSL_get_servername_type()
2386 New CTRL codes and macros (subject to change):
2388 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2389 - SSL_CTX_set_tlsext_servername_callback()
2390 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2391 - SSL_CTX_set_tlsext_servername_arg()
2392 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2394 openssl s_client has a new '-servername ...' option.
2396 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2397 '-key2 ...', '-servername_fatal' (subject to change). This allows
2398 testing the HostName extension for a specific single host name ('-cert'
2399 and '-key' remain fallbacks for handshakes without HostName
2400 negotiation). If the unrecogninzed_name alert has to be sent, this by
2401 default is a warning; it becomes fatal with the '-servername_fatal'
2404 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
2406 *) Add AES and SSE2 assembly language support to VC++ build.
2409 *) Mitigate attack on final subtraction in Montgomery reduction.
2412 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
2413 (which previously caused an internal error).
2416 *) Squeeze another 10% out of IGE mode when in != out.
2419 *) AES IGE mode speedup.
2420 [Dean Gaudet (Google)]
2422 *) Add the Korean symmetric 128-bit cipher SEED (see
2423 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
2424 add SEED ciphersuites from RFC 4162:
2426 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
2427 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
2428 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
2429 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
2431 To minimize changes between patchlevels in the OpenSSL 0.9.8
2432 series, SEED remains excluded from compilation unless OpenSSL
2433 is configured with 'enable-seed'.
2434 [KISA, Bodo Moeller]
2436 *) Mitigate branch prediction attacks, which can be practical if a
2437 single processor is shared, allowing a spy process to extract
2438 information. For detailed background information, see
2439 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
2440 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
2441 and Necessary Software Countermeasures"). The core of the change
2442 are new versions BN_div_no_branch() and
2443 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
2444 respectively, which are slower, but avoid the security-relevant
2445 conditional branches. These are automatically called by BN_div()
2446 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
2447 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
2448 remove a conditional branch.
2450 BN_FLG_CONSTTIME is the new name for the previous
2451 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
2452 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
2453 in the exponent causes BN_mod_exp_mont() to use the alternative
2454 implementation in BN_mod_exp_mont_consttime().) The old name
2455 remains as a deprecated alias.
2457 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
2458 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
2459 constant-time implementations for more than just exponentiation.
2460 Here too the old name is kept as a deprecated alias.
2462 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2463 the BN_BLINDING structure gets an independent copy of the
2464 modulus. This means that the previous "BIGNUM *m" argument to
2465 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2466 essentially becomes "const BIGNUM *m", although we can't actually
2467 change this in the header file before 0.9.9. It allows
2468 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2469 enable BN_FLG_CONSTTIME.
2471 [Matthew D Wood (Intel Corp)]
2473 *) In the SSL/TLS server implementation, be strict about session ID
2474 context matching (which matters if an application uses a single
2475 external cache for different purposes). Previously,
2476 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2477 set. This did ensure strict client verification, but meant that,
2478 with applications using a single external cache for quite
2479 different requirements, clients could circumvent ciphersuite
2480 restrictions for a given session ID context by starting a session
2481 in a different context.
2484 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2485 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2486 authentication-only ciphersuites.
2489 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
2490 not complete and could lead to a possible single byte overflow
2491 (CVE-2007-5135) [Ben Laurie]
2493 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
2495 *) Since AES128 and AES256 (and similarly Camellia128 and
2496 Camellia256) share a single mask bit in the logic of
2497 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2498 kludge to work properly if AES128 is available and AES256 isn't
2499 (or if Camellia128 is available and Camellia256 isn't).
2502 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
2503 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
2504 When a point or a seed is encoded in a BIT STRING, we need to
2505 prevent the removal of trailing zero bits to get the proper DER
2506 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
2507 of a NamedBitList, for which trailing 0 bits need to be removed.)
2510 *) Have SSL/TLS server implementation tolerate "mismatched" record
2511 protocol version while receiving ClientHello even if the
2512 ClientHello is fragmented. (The server can't insist on the
2513 particular protocol version it has chosen before the ServerHello
2514 message has informed the client about his choice.)
2517 *) Add RFC 3779 support.
2518 [Rob Austein for ARIN, Ben Laurie]
2520 *) Load error codes if they are not already present instead of using a
2521 static variable. This allows them to be cleanly unloaded and reloaded.
2522 Improve header file function name parsing.
2525 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
2526 or CAPABILITY handshake as required by RFCs.
2529 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
2531 *) Introduce limits to prevent malicious keys being able to
2532 cause a denial of service. (CVE-2006-2940)
2533 [Steve Henson, Bodo Moeller]
2535 *) Fix ASN.1 parsing of certain invalid structures that can result
2536 in a denial of service. (CVE-2006-2937) [Steve Henson]
2538 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2539 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2541 *) Fix SSL client code which could crash if connecting to a
2542 malicious SSLv2 server. (CVE-2006-4343)
2543 [Tavis Ormandy and Will Drewry, Google Security Team]
2545 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
2546 match only those. Before that, "AES256-SHA" would be interpreted
2547 as a pattern and match "AES128-SHA" too (since AES128-SHA got
2548 the same strength classification in 0.9.7h) as we currently only
2549 have a single AES bit in the ciphersuite description bitmap.
2550 That change, however, also applied to ciphersuite strings such as
2551 "RC4-MD5" that intentionally matched multiple ciphersuites --
2552 namely, SSL 2.0 ciphersuites in addition to the more common ones
2553 from SSL 3.0/TLS 1.0.
2555 So we change the selection algorithm again: Naming an explicit
2556 ciphersuite selects this one ciphersuite, and any other similar
2557 ciphersuite (same bitmap) from *other* protocol versions.
2558 Thus, "RC4-MD5" again will properly select both the SSL 2.0
2559 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
2561 Since SSL 2.0 does not have any ciphersuites for which the
2562 128/256 bit distinction would be relevant, this works for now.
2563 The proper fix will be to use different bits for AES128 and
2564 AES256, which would have avoided the problems from the beginning;
2565 however, bits are scarce, so we can only do this in a new release
2566 (not just a patchlevel) when we can change the SSL_CIPHER
2567 definition to split the single 'unsigned long mask' bitmap into
2568 multiple values to extend the available space.
2572 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2574 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2575 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2577 *) Add AES IGE and biIGE modes.
2580 *) Change the Unix randomness entropy gathering to use poll() when
2581 possible instead of select(), since the latter has some
2582 undesirable limitations.
2583 [Darryl Miles via Richard Levitte and Bodo Moeller]
2585 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2586 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2587 cannot be implicitly activated as part of, e.g., the "AES" alias.
2588 However, please upgrade to OpenSSL 0.9.9[-dev] for
2589 non-experimental use of the ECC ciphersuites to get TLS extension
2590 support, which is required for curve and point format negotiation
2591 to avoid potential handshake problems.
2594 *) Disable rogue ciphersuites:
2596 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2597 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2598 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2600 The latter two were purportedly from
2601 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2604 Also deactivate the remaining ciphersuites from
2605 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2606 unofficial, and the ID has long expired.
2609 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2610 dual-core machines) and other potential thread-safety issues.
2613 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
2614 versions), which is now available for royalty-free use
2615 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
2616 Also, add Camellia TLS ciphersuites from RFC 4132.
2618 To minimize changes between patchlevels in the OpenSSL 0.9.8
2619 series, Camellia remains excluded from compilation unless OpenSSL
2620 is configured with 'enable-camellia'.
2623 *) Disable the padding bug check when compression is in use. The padding
2624 bug check assumes the first packet is of even length, this is not
2625 necessarily true if compresssion is enabled and can result in false
2626 positives causing handshake failure. The actual bug test is ancient
2627 code so it is hoped that implementations will either have fixed it by
2628 now or any which still have the bug do not support compression.
2631 Changes between 0.9.8a and 0.9.8b [04 May 2006]
2633 *) When applying a cipher rule check to see if string match is an explicit
2634 cipher suite and only match that one cipher suite if it is.
2637 *) Link in manifests for VC++ if needed.
2638 [Austin Ziegler <halostatue@gmail.com>]
2640 *) Update support for ECC-based TLS ciphersuites according to
2641 draft-ietf-tls-ecc-12.txt with proposed changes (but without
2642 TLS extensions, which are supported starting with the 0.9.9
2643 branch, not in the OpenSSL 0.9.8 branch).
2646 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
2647 opaque EVP_CIPHER_CTX handling.
2650 *) Fixes and enhancements to zlib compression code. We now only use
2651 "zlib1.dll" and use the default __cdecl calling convention on Win32
2652 to conform with the standards mentioned here:
2653 http://www.zlib.net/DLL_FAQ.txt
2654 Static zlib linking now works on Windows and the new --with-zlib-include
2655 --with-zlib-lib options to Configure can be used to supply the location
2656 of the headers and library. Gracefully handle case where zlib library
2660 *) Several fixes and enhancements to the OID generation code. The old code
2661 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
2662 handle numbers larger than ULONG_MAX, truncated printing and had a
2663 non standard OBJ_obj2txt() behaviour.
2666 *) Add support for building of engines under engine/ as shared libraries
2667 under VC++ build system.
2670 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
2671 Hopefully, we will not see any false combination of paths any more.
2674 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
2676 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2677 (part of SSL_OP_ALL). This option used to disable the
2678 countermeasure against man-in-the-middle protocol-version
2679 rollback in the SSL 2.0 server implementation, which is a bad
2680 idea. (CVE-2005-2969)
2682 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2683 for Information Security, National Institute of Advanced Industrial
2684 Science and Technology [AIST], Japan)]
2686 *) Add two function to clear and return the verify parameter flags.
2689 *) Keep cipherlists sorted in the source instead of sorting them at
2690 runtime, thus removing the need for a lock.
2693 *) Avoid some small subgroup attacks in Diffie-Hellman.
2694 [Nick Mathewson and Ben Laurie]
2696 *) Add functions for well-known primes.
2699 *) Extended Windows CE support.
2700 [Satoshi Nakamura and Andy Polyakov]
2702 *) Initialize SSL_METHOD structures at compile time instead of during
2703 runtime, thus removing the need for a lock.
2706 *) Make PKCS7_decrypt() work even if no certificate is supplied by
2707 attempting to decrypt each encrypted key in turn. Add support to
2711 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
2713 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
2716 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
2719 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
2720 key into the same file any more.
2723 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
2726 *) Add -utf8 command line and config file option to 'ca'.
2727 [Stefan <stf@udoma.org]
2729 *) Removed the macro des_crypt(), as it seems to conflict with some
2730 libraries. Use DES_crypt().
2733 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
2734 involves renaming the source and generated shared-libs for
2735 both. The engines will accept the corrected or legacy ids
2736 ('ncipher' and '4758_cca' respectively) when binding. NB,
2737 this only applies when building 'shared'.
2738 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
2740 *) Add attribute functions to EVP_PKEY structure. Modify
2741 PKCS12_create() to recognize a CSP name attribute and
2742 use it. Make -CSP option work again in pkcs12 utility.
2745 *) Add new functionality to the bn blinding code:
2746 - automatic re-creation of the BN_BLINDING parameters after
2747 a fixed number of uses (currently 32)
2748 - add new function for parameter creation
2749 - introduce flags to control the update behaviour of the
2750 BN_BLINDING parameters
2751 - hide BN_BLINDING structure
2752 Add a second BN_BLINDING slot to the RSA structure to improve
2753 performance when a single RSA object is shared among several
2757 *) Add support for DTLS.
2758 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
2760 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
2761 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
2764 *) Remove buggy and incompletet DH cert support from
2765 ssl/ssl_rsa.c and ssl/s3_both.c
2768 *) Use SHA-1 instead of MD5 as the default digest algorithm for
2769 the apps/openssl applications.
2772 *) Compile clean with "-Wall -Wmissing-prototypes
2773 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
2774 DEBUG_SAFESTACK must also be set.
2777 *) Change ./Configure so that certain algorithms can be disabled by default.
2778 The new counterpiece to "no-xxx" is "enable-xxx".
2780 The patented RC5 and MDC2 algorithms will now be disabled unless
2781 "enable-rc5" and "enable-mdc2", respectively, are specified.
2783 (IDEA remains enabled despite being patented. This is because IDEA
2784 is frequently required for interoperability, and there is no license
2785 fee for non-commercial use. As before, "no-idea" can be used to
2786 avoid this algorithm.)
2790 *) Add processing of proxy certificates (see RFC 3820). This work was
2791 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
2792 EGEE (Enabling Grids for E-science in Europe).
2795 *) RC4 performance overhaul on modern architectures/implementations, such
2796 as Intel P4, IA-64 and AMD64.
2799 *) New utility extract-section.pl. This can be used specify an alternative
2800 section number in a pod file instead of having to treat each file as
2801 a separate case in Makefile. This can be done by adding two lines to the
2804 =for comment openssl_section:XXX
2806 The blank line is mandatory.
2810 *) New arguments -certform, -keyform and -pass for s_client and s_server
2811 to allow alternative format key and certificate files and passphrase
2815 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
2816 update associated structures and add various utility functions.
2818 Add new policy related verify parameters, include policy checking in
2819 standard verify code. Enhance 'smime' application with extra parameters
2820 to support policy checking and print out.
2823 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
2824 Nehemiah processors. These extensions support AES encryption in hardware
2825 as well as RNG (though RNG support is currently disabled).
2826 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
2828 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
2831 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
2832 [Andy Polyakov and a number of other people]
2834 *) Improved PowerPC platform support. Most notably BIGNUM assembler
2835 implementation contributed by IBM.
2836 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
2838 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
2839 exponent rather than 'unsigned long'. There is a corresponding change to
2840 the new 'rsa_keygen' element of the RSA_METHOD structure.
2841 [Jelte Jansen, Geoff Thorpe]
2843 *) Functionality for creating the initial serial number file is now
2844 moved from CA.pl to the 'ca' utility with a new option -create_serial.
2846 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
2847 number file to 1, which is bound to cause problems. To avoid
2848 the problems while respecting compatibility between different 0.9.7
2849 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
2850 CA.pl for serial number initialization. With the new release 0.9.8,
2851 we can fix the problem directly in the 'ca' utility.)
2854 *) Reduced header interdepencies by declaring more opaque objects in
2855 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
2856 give fewer recursive includes, which could break lazy source code - so
2857 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
2858 developers should define this symbol when building and using openssl to
2859 ensure they track the recommended behaviour, interfaces, [etc], but
2860 backwards-compatible behaviour prevails when this isn't defined.
2863 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
2866 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
2867 This will generate a random key of the appropriate length based on the
2868 cipher context. The EVP_CIPHER can provide its own random key generation
2869 routine to support keys of a specific form. This is used in the des and
2870 3des routines to generate a key of the correct parity. Update S/MIME
2871 code to use new functions and hence generate correct parity DES keys.
2872 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
2873 valid (weak or incorrect parity).
2876 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
2877 as looking them up. This is useful when the verified structure may contain
2878 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
2879 present unless the new PKCS7_NO_CRL flag is asserted.
2882 *) Extend ASN1 oid configuration module. It now additionally accepts the
2885 shortName = some long name, 1.2.3.4
2888 *) Reimplemented the BN_CTX implementation. There is now no more static
2889 limitation on the number of variables it can handle nor the depth of the
2890 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
2891 information can now expand as required, and rather than having a single
2892 static array of bignums, BN_CTX now uses a linked-list of such arrays
2893 allowing it to expand on demand whilst maintaining the usefulness of
2894 BN_CTX's "bundling".
2897 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
2898 to allow all RSA operations to function using a single BN_CTX.
2901 *) Preliminary support for certificate policy evaluation and checking. This
2902 is initially intended to pass the tests outlined in "Conformance Testing
2903 of Relying Party Client Certificate Path Processing Logic" v1.07.
2906 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
2907 remained unused and not that useful. A variety of other little bignum
2908 tweaks and fixes have also been made continuing on from the audit (see
2912 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
2913 associated ASN1, EVP and SSL functions and old ASN1 macros.
2916 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
2917 and this should never fail. So the return value from the use of
2918 BN_set_word() (which can fail due to needless expansion) is now deprecated;
2919 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
2922 *) BN_CTX_get() should return zero-valued bignums, providing the same
2923 initialised value as BN_new().
2924 [Geoff Thorpe, suggested by Ulf Möller]
2926 *) Support for inhibitAnyPolicy certificate extension.
2929 *) An audit of the BIGNUM code is underway, for which debugging code is
2930 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
2931 is considered valid when processing BIGNUMs, and causes execution to
2932 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
2933 further steps are taken to deliberately pollute unused data in BIGNUM
2934 structures to try and expose faulty code further on. For now, openssl will
2935 (in its default mode of operation) continue to tolerate the inconsistent
2936 forms that it has tolerated in the past, but authors and packagers should
2937 consider trying openssl and their own applications when compiled with
2938 these debugging symbols defined. It will help highlight potential bugs in
2939 their own code, and will improve the test coverage for OpenSSL itself. At
2940 some point, these tighter rules will become openssl's default to improve
2941 maintainability, though the assert()s and other overheads will remain only
2942 in debugging configurations. See bn.h for more details.
2943 [Geoff Thorpe, Nils Larsch, Ulf Möller]
2945 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
2946 that can only be obtained through BN_CTX_new() (which implicitly
2947 initialises it). The presence of this function only made it possible
2948 to overwrite an existing structure (and cause memory leaks).
2951 *) Because of the callback-based approach for implementing LHASH as a
2952 template type, lh_insert() adds opaque objects to hash-tables and
2953 lh_doall() or lh_doall_arg() are typically used with a destructor callback
2954 to clean up those corresponding objects before destroying the hash table
2955 (and losing the object pointers). So some over-zealous constifications in
2956 LHASH have been relaxed so that lh_insert() does not take (nor store) the
2957 objects as "const" and the lh_doall[_arg] callback wrappers are not
2958 prototyped to have "const" restrictions on the object pointers they are
2959 given (and so aren't required to cast them away any more).
2962 *) The tmdiff.h API was so ugly and minimal that our own timing utility
2963 (speed) prefers to use its own implementation. The two implementations
2964 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
2965 its object type properly exposed (MS_TM) instead of casting to/from "char
2966 *". This may still change yet if someone realises MS_TM and "ms_time_***"
2967 aren't necessarily the greatest nomenclatures - but this is what was used
2968 internally to the implementation so I've used that for now.
2971 *) Ensure that deprecated functions do not get compiled when
2972 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
2973 the self-tests were still using deprecated key-generation functions so
2974 these have been updated also.
2977 *) Reorganise PKCS#7 code to separate the digest location functionality
2978 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
2979 New function PKCS7_set_digest() to set the digest type for PKCS#7
2980 digestedData type. Add additional code to correctly generate the
2981 digestedData type and add support for this type in PKCS7 initialization
2985 *) New function PKCS7_set0_type_other() this initializes a PKCS7
2986 structure of type "other".
2989 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
2990 sure the loop does correctly stop and breaking ("division by zero")
2991 modulus operations are not performed. The (pre-generated) prime
2992 table crypto/bn/bn_prime.h was already correct, but it could not be
2993 re-generated on some platforms because of the "division by zero"
2994 situation in the script.
2995 [Ralf S. Engelschall]
2997 *) Update support for ECC-based TLS ciphersuites according to
2998 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
2999 SHA-1 now is only used for "small" curves (where the
3000 representation of a field element takes up to 24 bytes); for
3001 larger curves, the field element resulting from ECDH is directly
3002 used as premaster secret.
3003 [Douglas Stebila (Sun Microsystems Laboratories)]
3005 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
3006 curve secp160r1 to the tests.
3007 [Douglas Stebila (Sun Microsystems Laboratories)]
3009 *) Add the possibility to load symbols globally with DSO.
3010 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
3012 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
3013 control of the error stack.
3016 *) Add support for STORE in ENGINE.
3019 *) Add the STORE type. The intention is to provide a common interface
3020 to certificate and key stores, be they simple file-based stores, or
3021 HSM-type store, or LDAP stores, or...
3022 NOTE: The code is currently UNTESTED and isn't really used anywhere.
3025 *) Add a generic structure called OPENSSL_ITEM. This can be used to
3026 pass a list of arguments to any function as well as provide a way
3027 for a function to pass data back to the caller.
3030 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
3031 works like BUF_strdup() but can be used to duplicate a portion of
3032 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
3036 *) Add the function sk_find_ex() which works like sk_find(), but will
3037 return an index to an element even if an exact match couldn't be
3038 found. The index is guaranteed to point at the element where the
3039 searched-for key would be inserted to preserve sorting order.
3042 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
3043 takes an extra flags argument for optional functionality. Currently,
3044 the following flags are defined:
3046 OBJ_BSEARCH_VALUE_ON_NOMATCH
3047 This one gets OBJ_bsearch_ex() to return a pointer to the first
3048 element where the comparing function returns a negative or zero
3051 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
3052 This one gets OBJ_bsearch_ex() to return a pointer to the first
3053 element where the comparing function returns zero. This is useful
3054 if there are more than one element where the comparing function
3058 *) Make it possible to create self-signed certificates with 'openssl ca'
3059 in such a way that the self-signed certificate becomes part of the
3060 CA database and uses the same mechanisms for serial number generation
3061 as all other certificate signing. The new flag '-selfsign' enables
3062 this functionality. Adapt CA.sh and CA.pl.in.
3065 *) Add functionality to check the public key of a certificate request
3066 against a given private. This is useful to check that a certificate
3067 request can be signed by that key (self-signing).
3070 *) Make it possible to have multiple active certificates with the same
3071 subject in the CA index file. This is done only if the keyword
3072 'unique_subject' is set to 'no' in the main CA section (default
3073 if 'CA_default') of the configuration file. The value is saved
3074 with the database itself in a separate index attribute file,
3075 named like the index file with '.attr' appended to the name.
3078 *) Generate muti valued AVAs using '+' notation in config files for
3082 *) Support for nameConstraints certificate extension.
3085 *) Support for policyConstraints certificate extension.
3088 *) Support for policyMappings certificate extension.
3091 *) Make sure the default DSA_METHOD implementation only uses its
3092 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
3093 and change its own handlers to be NULL so as to remove unnecessary
3094 indirection. This lets alternative implementations fallback to the
3095 default implementation more easily.
3098 *) Support for directoryName in GeneralName related extensions
3102 *) Make it possible to link applications using Makefile.shared.
3103 Make that possible even when linking against static libraries!
3106 *) Support for single pass processing for S/MIME signing. This now
3107 means that S/MIME signing can be done from a pipe, in addition
3108 cleartext signing (multipart/signed type) is effectively streaming
3109 and the signed data does not need to be all held in memory.
3111 This is done with a new flag PKCS7_STREAM. When this flag is set
3112 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
3113 is done after the data is output (and digests calculated) in
3114 SMIME_write_PKCS7().
3117 *) Add full support for -rpath/-R, both in shared libraries and
3118 applications, at least on the platforms where it's known how
3122 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
3123 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
3124 will now compute a table of multiples of the generator that
3125 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
3126 faster (notably in the case of a single point multiplication,
3127 scalar * generator).
3128 [Nils Larsch, Bodo Moeller]
3130 *) IPv6 support for certificate extensions. The various extensions
3131 which use the IP:a.b.c.d can now take IPv6 addresses using the
3132 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
3136 *) Added an ENGINE that implements RSA by performing private key
3137 exponentiations with the GMP library. The conversions to and from
3138 GMP's mpz_t format aren't optimised nor are any montgomery forms
3139 cached, and on x86 it appears OpenSSL's own performance has caught up.
3140 However there are likely to be other architectures where GMP could
3141 provide a boost. This ENGINE is not built in by default, but it can be
3142 specified at Configure time and should be accompanied by the necessary
3143 linker additions, eg;
3144 ./config -DOPENSSL_USE_GMP -lgmp
3147 *) "openssl engine" will not display ENGINE/DSO load failure errors when
3148 testing availability of engines with "-t" - the old behaviour is
3149 produced by increasing the feature's verbosity with "-tt".
3152 *) ECDSA routines: under certain error conditions uninitialized BN objects
3153 could be freed. Solution: make sure initialization is performed early
3154 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
3158 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
3159 and DH_METHOD (eg. by ENGINE implementations) to override the normal
3160 software implementations. For DSA and DH, parameter generation can
3161 also be overriden by providing the appropriate method callbacks.
3164 *) Change the "progress" mechanism used in key-generation and
3165 primality testing to functions that take a new BN_GENCB pointer in
3166 place of callback/argument pairs. The new API functions have "_ex"
3167 postfixes and the older functions are reimplemented as wrappers for
3168 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
3169 declarations of the old functions to help (graceful) attempts to
3170 migrate to the new functions. Also, the new key-generation API
3171 functions operate on a caller-supplied key-structure and return
3172 success/failure rather than returning a key or NULL - this is to
3173 help make "keygen" another member function of RSA_METHOD etc.
3175 Example for using the new callback interface:
3177 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
3181 BN_GENCB_set(&my_cb, my_callback, my_arg);
3183 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
3184 /* For the meaning of a, b in calls to my_callback(), see the
3185 * documentation of the function that calls the callback.
3186 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
3187 * my_callback should return 1 if it wants BN_is_prime_ex()
3188 * to continue, or 0 to stop.
3193 *) Change the ZLIB compression method to be stateful, and make it
3194 available to TLS with the number defined in
3195 draft-ietf-tls-compression-04.txt.
3198 *) Add the ASN.1 structures and functions for CertificatePair, which
3199 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
3201 CertificatePair ::= SEQUENCE {
3202 forward [0] Certificate OPTIONAL,
3203 reverse [1] Certificate OPTIONAL,
3204 -- at least one of the pair shall be present -- }
3206 Also implement the PEM functions to read and write certificate
3207 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
3209 This needed to be defined, mostly for the sake of the LDAP
3210 attribute crossCertificatePair, but may prove useful elsewhere as
3214 *) Make it possible to inhibit symlinking of shared libraries in
3215 Makefile.shared, for Cygwin's sake.
3218 *) Extend the BIGNUM API by creating a function
3219 void BN_set_negative(BIGNUM *a, int neg);
3220 and a macro that behave like
3221 int BN_is_negative(const BIGNUM *a);
3223 to avoid the need to access 'a->neg' directly in applications.
3226 *) Implement fast modular reduction for pseudo-Mersenne primes
3227 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
3228 EC_GROUP_new_curve_GFp() will now automatically use this
3230 [Nils Larsch <nla@trustcenter.de>]
3232 *) Add new lock type (CRYPTO_LOCK_BN).
3235 *) Change the ENGINE framework to automatically load engines
3236 dynamically from specific directories unless they could be
3237 found to already be built in or loaded. Move all the
3238 current engines except for the cryptodev one to a new
3240 The engines in engines/ are built as shared libraries if
3241 the "shared" options was given to ./Configure or ./config.
3242 Otherwise, they are inserted in libcrypto.a.
3243 /usr/local/ssl/engines is the default directory for dynamic
3244 engines, but that can be overriden at configure time through
3245 the usual use of --prefix and/or --openssldir, and at run
3246 time with the environment variable OPENSSL_ENGINES.
3247 [Geoff Thorpe and Richard Levitte]
3249 *) Add Makefile.shared, a helper makefile to build shared
3250 libraries. Addapt Makefile.org.
3253 *) Add version info to Win32 DLLs.
3254 [Peter 'Luna' Runestig" <peter@runestig.com>]
3256 *) Add new 'medium level' PKCS#12 API. Certificates and keys
3257 can be added using this API to created arbitrary PKCS#12
3258 files while avoiding the low level API.
3260 New options to PKCS12_create(), key or cert can be NULL and
3261 will then be omitted from the output file. The encryption
3262 algorithm NIDs can be set to -1 for no encryption, the mac
3263 iteration count can be set to 0 to omit the mac.
3265 Enhance pkcs12 utility by making the -nokeys and -nocerts
3266 options work when creating a PKCS#12 file. New option -nomac
3267 to omit the mac, NONE can be set for an encryption algorithm.
3268 New code is modified to use the enhanced PKCS12_create()
3269 instead of the low level API.
3272 *) Extend ASN1 encoder to support indefinite length constructed
3273 encoding. This can output sequences tags and octet strings in
3274 this form. Modify pk7_asn1.c to support indefinite length
3275 encoding. This is experimental and needs additional code to
3276 be useful, such as an ASN1 bio and some enhanced streaming
3279 Extend template encode functionality so that tagging is passed
3280 down to the template encoder.
3283 *) Let 'openssl req' fail if an argument to '-newkey' is not
3284 recognized instead of using RSA as a default.
3287 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
3288 As these are not official, they are not included in "ALL";
3289 the "ECCdraft" ciphersuite group alias can be used to select them.
3290 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
3292 *) Add ECDH engine support.
3293 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
3295 *) Add ECDH in new directory crypto/ecdh/.
3296 [Douglas Stebila (Sun Microsystems Laboratories)]
3298 *) Let BN_rand_range() abort with an error after 100 iterations
3299 without success (which indicates a broken PRNG).
3302 *) Change BN_mod_sqrt() so that it verifies that the input value
3303 is really the square of the return value. (Previously,
3304 BN_mod_sqrt would show GIGO behaviour.)
3307 *) Add named elliptic curves over binary fields from X9.62, SECG,
3308 and WAP/WTLS; add OIDs that were still missing.
3310 [Sheueling Chang Shantz and Douglas Stebila
3311 (Sun Microsystems Laboratories)]
3313 *) Extend the EC library for elliptic curves over binary fields
3314 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
3317 EC_GF2m_simple_method
3321 EC_GROUP_new_curve_GF2m
3322 EC_GROUP_set_curve_GF2m
3323 EC_GROUP_get_curve_GF2m
3324 EC_POINT_set_affine_coordinates_GF2m
3325 EC_POINT_get_affine_coordinates_GF2m
3326 EC_POINT_set_compressed_coordinates_GF2m
3328 Point compression for binary fields is disabled by default for
3329 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
3332 As binary polynomials are represented as BIGNUMs, various members
3333 of the EC_GROUP and EC_POINT data structures can be shared
3334 between the implementations for prime fields and binary fields;
3335 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
3336 are essentially identical to their ..._GFp counterparts.
3337 (For simplicity, the '..._GFp' prefix has been dropped from
3338 various internal method names.)
3340 An internal 'field_div' method (similar to 'field_mul' and
3341 'field_sqr') has been added; this is used only for binary fields.
3343 [Sheueling Chang Shantz and Douglas Stebila
3344 (Sun Microsystems Laboratories)]
3346 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
3347 through methods ('mul', 'precompute_mult').
3349 The generic implementations (now internally called 'ec_wNAF_mul'
3350 and 'ec_wNAF_precomputed_mult') remain the default if these
3351 methods are undefined.
3353 [Sheueling Chang Shantz and Douglas Stebila
3354 (Sun Microsystems Laboratories)]
3356 *) New function EC_GROUP_get_degree, which is defined through
3357 EC_METHOD. For curves over prime fields, this returns the bit
3358 length of the modulus.
3360 [Sheueling Chang Shantz and Douglas Stebila
3361 (Sun Microsystems Laboratories)]
3363 *) New functions EC_GROUP_dup, EC_POINT_dup.
3364 (These simply call ..._new and ..._copy).
3366 [Sheueling Chang Shantz and Douglas Stebila
3367 (Sun Microsystems Laboratories)]
3369 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
3370 Polynomials are represented as BIGNUMs (where the sign bit is not
3371 used) in the following functions [macros]:
3374 BN_GF2m_sub [= BN_GF2m_add]
3375 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
3376 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
3377 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
3379 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
3380 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
3381 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
3382 BN_GF2m_cmp [= BN_ucmp]
3384 (Note that only the 'mod' functions are actually for fields GF(2^m).
3385 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
3387 For some functions, an the irreducible polynomial defining a
3388 field can be given as an 'unsigned int[]' with strictly
3389 decreasing elements giving the indices of those bits that are set;
3390 i.e., p[] represents the polynomial
3391 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
3393 p[0] > p[1] > ... > p[k] = 0.
3394 This applies to the following functions:
3399 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
3400 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
3402 BN_GF2m_mod_sqrt_arr
3403 BN_GF2m_mod_solve_quad_arr
3407 Conversion can be performed by the following functions:
3412 bntest.c has additional tests for binary polynomial arithmetic.
3414 Two implementations for BN_GF2m_mod_div() are available.
3415 The default algorithm simply uses BN_GF2m_mod_inv() and
3416 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
3417 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
3418 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
3420 [Sheueling Chang Shantz and Douglas Stebila
3421 (Sun Microsystems Laboratories)]
3423 *) Add new error code 'ERR_R_DISABLED' that can be used when some
3424 functionality is disabled at compile-time.
3425 [Douglas Stebila <douglas.stebila@sun.com>]
3427 *) Change default behaviour of 'openssl asn1parse' so that more
3428 information is visible when viewing, e.g., a certificate:
3430 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
3431 mode the content of non-printable OCTET STRINGs is output in a
3432 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
3433 avoid the appearance of a printable string.
3434 [Nils Larsch <nla@trustcenter.de>]
3436 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
3438 EC_GROUP_set_asn1_flag()
3439 EC_GROUP_get_asn1_flag()
3440 EC_GROUP_set_point_conversion_form()
3441 EC_GROUP_get_point_conversion_form()
3442 These control ASN1 encoding details:
3443 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
3444 has been set to OPENSSL_EC_NAMED_CURVE.
3445 - Points are encoded in uncompressed form by default; options for
3446 asn1_for are as for point2oct, namely
3447 POINT_CONVERSION_COMPRESSED
3448 POINT_CONVERSION_UNCOMPRESSED
3449 POINT_CONVERSION_HYBRID
3451 Also add 'seed' and 'seed_len' members to EC_GROUP with access
3454 EC_GROUP_get0_seed()
3455 EC_GROUP_get_seed_len()
3456 This is used only for ASN1 purposes (so far).
3457 [Nils Larsch <nla@trustcenter.de>]
3459 *) Add 'field_type' member to EC_METHOD, which holds the NID
3460 of the appropriate field type OID. The new function
3461 EC_METHOD_get_field_type() returns this value.
3462 [Nils Larsch <nla@trustcenter.de>]
3467 EC_POINT_point2hex()
3468 EC_POINT_hex2point()
3469 providing useful interfaces to EC_POINT_point2oct() and
3470 EC_POINT_oct2point().
3471 [Nils Larsch <nla@trustcenter.de>]
3473 *) Change internals of the EC library so that the functions
3474 EC_GROUP_set_generator()
3475 EC_GROUP_get_generator()
3476 EC_GROUP_get_order()
3477 EC_GROUP_get_cofactor()
3478 are implemented directly in crypto/ec/ec_lib.c and not dispatched
3479 to methods, which would lead to unnecessary code duplication when
3480 adding different types of curves.
3481 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
3483 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
3484 arithmetic, and such that modified wNAFs are generated
3485 (which avoid length expansion in many cases).
3488 *) Add a function EC_GROUP_check_discriminant() (defined via
3489 EC_METHOD) that verifies that the curve discriminant is non-zero.
3491 Add a function EC_GROUP_check() that makes some sanity tests
3492 on a EC_GROUP, its generator and order. This includes
3493 EC_GROUP_check_discriminant().
3494 [Nils Larsch <nla@trustcenter.de>]
3496 *) Add ECDSA in new directory crypto/ecdsa/.
3498 Add applications 'openssl ecparam' and 'openssl ecdsa'
3499 (these are based on 'openssl dsaparam' and 'openssl dsa').
3501 ECDSA support is also included in various other files across the
3502 library. Most notably,
3503 - 'openssl req' now has a '-newkey ecdsa:file' option;
3504 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
3505 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
3506 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
3507 them suitable for ECDSA where domain parameters must be
3508 extracted before the specific public key;
3509 - ECDSA engine support has been added.
3510 [Nils Larsch <nla@trustcenter.de>]
3512 *) Include some named elliptic curves, and add OIDs from X9.62,
3513 SECG, and WAP/WTLS. Each curve can be obtained from the new
3515 EC_GROUP_new_by_curve_name(),
3516 and the list of available named curves can be obtained with
3517 EC_get_builtin_curves().
3518 Also add a 'curve_name' member to EC_GROUP objects, which can be
3520 EC_GROUP_set_curve_name()
3521 EC_GROUP_get_curve_name()
3522 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
3524 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3525 was actually never needed) and in BN_mul(). The removal in BN_mul()
3526 required a small change in bn_mul_part_recursive() and the addition
3527 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3528 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3529 bn_sub_words() and bn_add_words() except they take arrays with
3533 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
3535 *) Cleanse PEM buffers before freeing them since they may contain
3537 [Benjamin Bennett <ben@psc.edu>]
3539 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3540 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3541 authentication-only ciphersuites.
3544 *) Since AES128 and AES256 share a single mask bit in the logic of
3545 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3546 kludge to work properly if AES128 is available and AES256 isn't.
3549 *) Expand security boundary to match 1.1.1 module.
3552 *) Remove redundant features: hash file source, editing of test vectors
3553 modify fipsld to use external fips_premain.c signature.
3556 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
3557 run algorithm test programs.
3560 *) Make algorithm test programs more tolerant of whitespace.
3563 *) Have SSL/TLS server implementation tolerate "mismatched" record
3564 protocol version while receiving ClientHello even if the
3565 ClientHello is fragmented. (The server can't insist on the
3566 particular protocol version it has chosen before the ServerHello
3567 message has informed the client about his choice.)
3570 *) Load error codes if they are not already present instead of using a
3571 static variable. This allows them to be cleanly unloaded and reloaded.
3574 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3576 *) Introduce limits to prevent malicious keys being able to
3577 cause a denial of service. (CVE-2006-2940)
3578 [Steve Henson, Bodo Moeller]
3580 *) Fix ASN.1 parsing of certain invalid structures that can result
3581 in a denial of service. (CVE-2006-2937) [Steve Henson]
3583 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3584 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3586 *) Fix SSL client code which could crash if connecting to a
3587 malicious SSLv2 server. (CVE-2006-4343)
3588 [Tavis Ormandy and Will Drewry, Google Security Team]
3590 *) Change ciphersuite string processing so that an explicit
3591 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3592 will no longer include "AES128-SHA"), and any other similar
3593 ciphersuite (same bitmap) from *other* protocol versions (so that
3594 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3595 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3596 changes from 0.9.8b and 0.9.8d.
3599 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
3601 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3602 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3604 *) Change the Unix randomness entropy gathering to use poll() when
3605 possible instead of select(), since the latter has some
3606 undesirable limitations.
3607 [Darryl Miles via Richard Levitte and Bodo Moeller]
3609 *) Disable rogue ciphersuites:
3611 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3612 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3613 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3615 The latter two were purportedly from
3616 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3619 Also deactive the remaining ciphersuites from
3620 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3621 unofficial, and the ID has long expired.
3624 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3625 dual-core machines) and other potential thread-safety issues.
3628 Changes between 0.9.7i and 0.9.7j [04 May 2006]
3630 *) Adapt fipsld and the build system to link against the validated FIPS
3631 module in FIPS mode.
3634 *) Fixes for VC++ 2005 build under Windows.
3637 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
3638 from a Windows bash shell such as MSYS. It is autodetected from the
3639 "config" script when run from a VC++ environment. Modify standard VC++
3640 build to use fipscanister.o from the GNU make build.
3643 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
3645 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
3646 The value now differs depending on if you build for FIPS or not.
3647 BEWARE! A program linked with a shared FIPSed libcrypto can't be
3648 safely run with a non-FIPSed libcrypto, as it may crash because of
3649 the difference induced by this change.
3652 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
3654 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3655 (part of SSL_OP_ALL). This option used to disable the
3656 countermeasure against man-in-the-middle protocol-version
3657 rollback in the SSL 2.0 server implementation, which is a bad
3658 idea. (CVE-2005-2969)
3660 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3661 for Information Security, National Institute of Advanced Industrial
3662 Science and Technology [AIST], Japan)]
3664 *) Minimal support for X9.31 signatures and PSS padding modes. This is
3665 mainly for FIPS compliance and not fully integrated at this stage.
3668 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
3669 the exponentiation using a fixed-length exponent. (Otherwise,
3670 the information leaked through timing could expose the secret key
3671 after many signatures; cf. Bleichenbacher's attack on DSA with
3675 *) Make a new fixed-window mod_exp implementation the default for
3676 RSA, DSA, and DH private-key operations so that the sequence of
3677 squares and multiplies and the memory access pattern are
3678 independent of the particular secret key. This will mitigate
3679 cache-timing and potential related attacks.
3681 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
3682 and this is automatically used by BN_mod_exp_mont() if the new flag
3683 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
3684 will use this BN flag for private exponents unless the flag
3685 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
3686 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
3688 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
3690 *) Change the client implementation for SSLv23_method() and
3691 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
3692 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
3693 (Previously, the SSL 2.0 backwards compatible Client Hello
3694 message format would be used even with SSL_OP_NO_SSLv2.)
3697 *) Add support for smime-type MIME parameter in S/MIME messages which some
3701 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
3702 a threadsafe manner. Modify rsa code to use new function and add calls
3703 to dsa and dh code (which had race conditions before).
3706 *) Include the fixed error library code in the C error file definitions
3707 instead of fixing them up at runtime. This keeps the error code
3708 structures constant.
3711 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
3713 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
3716 *) Fixes for newer kerberos headers. NB: the casts are needed because
3717 the 'length' field is signed on one version and unsigned on another
3718 with no (?) obvious way to tell the difference, without these VC++
3719 complains. Also the "definition" of FAR (blank) is no longer included
3720 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
3721 some needed definitions.
3724 *) Undo Cygwin change.
3727 *) Added support for proxy certificates according to RFC 3820.
3728 Because they may be a security thread to unaware applications,
3729 they must be explicitely allowed in run-time. See
3730 docs/HOWTO/proxy_certificates.txt for further information.
3733 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
3735 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
3736 server and client random values. Previously
3737 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
3738 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
3740 This change has negligible security impact because:
3742 1. Server and client random values still have 24 bytes of pseudo random
3745 2. Server and client random values are sent in the clear in the initial
3748 3. The master secret is derived using the premaster secret (48 bytes in
3749 size for static RSA ciphersuites) as well as client server and random
3752 The OpenSSL team would like to thank the UK NISCC for bringing this issue
3755 [Stephen Henson, reported by UK NISCC]
3757 *) Use Windows randomness collection on Cygwin.
3760 *) Fix hang in EGD/PRNGD query when communication socket is closed
3761 prematurely by EGD/PRNGD.
3762 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
3764 *) Prompt for pass phrases when appropriate for PKCS12 input format.
3767 *) Back-port of selected performance improvements from development
3768 branch, as well as improved support for PowerPC platforms.
3771 *) Add lots of checks for memory allocation failure, error codes to indicate
3772 failure and freeing up memory if a failure occurs.
3773 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
3775 *) Add new -passin argument to dgst.
3778 *) Perform some character comparisons of different types in X509_NAME_cmp:
3779 this is needed for some certificates that reencode DNs into UTF8Strings
3780 (in violation of RFC3280) and can't or wont issue name rollover
3784 *) Make an explicit check during certificate validation to see that
3785 the CA setting in each certificate on the chain is correct. As a
3786 side effect always do the following basic checks on extensions,
3787 not just when there's an associated purpose to the check:
3789 - if there is an unhandled critical extension (unless the user
3790 has chosen to ignore this fault)
3791 - if the path length has been exceeded (if one is set at all)
3792 - that certain extensions fit the associated purpose (if one has
3796 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
3798 *) Avoid a race condition when CRLs are checked in a multi threaded
3799 environment. This would happen due to the reordering of the revoked
3800 entries during signature checking and serial number lookup. Now the
3801 encoding is cached and the serial number sort performed under a lock.
3802 Add new STACK function sk_is_sorted().
3805 *) Add Delta CRL to the extension code.
3808 *) Various fixes to s3_pkt.c so alerts are sent properly.
3809 [David Holmes <d.holmes@f5.com>]
3811 *) Reduce the chances of duplicate issuer name and serial numbers (in
3812 violation of RFC3280) using the OpenSSL certificate creation utilities.
3813 This is done by creating a random 64 bit value for the initial serial
3814 number when a serial number file is created or when a self signed
3815 certificate is created using 'openssl req -x509'. The initial serial
3816 number file is created using 'openssl x509 -next_serial' in CA.pl
3817 rather than being initialized to 1.
3820 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
3822 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3823 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3824 [Joe Orton, Steve Henson]
3826 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
3828 [Joe Orton, Steve Henson]
3830 *) Make it possible to have multiple active certificates with the same
3831 subject in the CA index file. This is done only if the keyword
3832 'unique_subject' is set to 'no' in the main CA section (default
3833 if 'CA_default') of the configuration file. The value is saved
3834 with the database itself in a separate index attribute file,
3835 named like the index file with '.attr' appended to the name.
3838 *) X509 verify fixes. Disable broken certificate workarounds when
3839 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
3840 keyUsage extension present. Don't accept CRLs with unhandled critical
3841 extensions: since verify currently doesn't process CRL extensions this
3842 rejects a CRL with *any* critical extensions. Add new verify error codes
3846 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
3847 A clarification of RFC2560 will require the use of OCTET STRINGs and
3848 some implementations cannot handle the current raw format. Since OpenSSL
3849 copies and compares OCSP nonces as opaque blobs without any attempt at
3850 parsing them this should not create any compatibility issues.
3853 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
3854 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
3855 this HMAC (and other) operations are several times slower than OpenSSL
3859 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
3860 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3862 *) Use the correct content when signing type "other".
3865 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
3867 *) Fix various bugs revealed by running the NISCC test suite:
3869 Stop out of bounds reads in the ASN1 code when presented with
3870 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3872 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
3874 If verify callback ignores invalid public key errors don't try to check
3875 certificate signature with the NULL public key.
3879 *) New -ignore_err option in ocsp application to stop the server
3880 exiting on the first error in a request.
3883 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3884 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3888 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3889 extra data after the compression methods not only for TLS 1.0
3890 but also for SSL 3.0 (as required by the specification).
3891 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3893 *) Change X509_certificate_type() to mark the key as exported/exportable
3894 when it's 512 *bits* long, not 512 bytes.
3897 *) Change AES_cbc_encrypt() so it outputs exact multiple of
3898 blocks during encryption.
3901 *) Various fixes to base64 BIO and non blocking I/O. On write
3902 flushes were not handled properly if the BIO retried. On read
3903 data was not being buffered properly and had various logic bugs.
3904 This also affects blocking I/O when the data being decoded is a
3908 *) Various S/MIME bugfixes and compatibility changes:
3909 output correct application/pkcs7 MIME type if
3910 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
3911 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
3912 of files as .eml work). Correctly handle very long lines in MIME
3916 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
3918 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3919 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3920 a protocol version number mismatch like a decryption error
3921 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3924 *) Turn on RSA blinding by default in the default implementation
3925 to avoid a timing attack. Applications that don't want it can call
3926 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3927 They would be ill-advised to do so in most cases.
3928 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3930 *) Change RSA blinding code so that it works when the PRNG is not
3931 seeded (in this case, the secret RSA exponent is abused as
3932 an unpredictable seed -- if it is not unpredictable, there
3933 is no point in blinding anyway). Make RSA blinding thread-safe
3934 by remembering the creator's thread ID in rsa->blinding and
3935 having all other threads use local one-time blinding factors
3936 (this requires more computation than sharing rsa->blinding, but
3937 avoids excessive locking; and if an RSA object is not shared
3938 between threads, blinding will still be very fast).
3941 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
3942 ENGINE as defaults for all supported algorithms irrespective of
3943 the 'flags' parameter. 'flags' is now honoured, so applications
3944 should make sure they are passing it correctly.
3947 *) Target "mingw" now allows native Windows code to be generated in
3948 the Cygwin environment as well as with the MinGW compiler.
3951 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
3953 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3954 via timing by performing a MAC computation even if incorrrect
3955 block cipher padding has been found. This is a countermeasure
3956 against active attacks where the attacker has to distinguish
3957 between bad padding and a MAC verification error. (CVE-2003-0078)
3959 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3960 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3961 Martin Vuagnoux (EPFL, Ilion)]
3963 *) Make the no-err option work as intended. The intention with no-err
3964 is not to have the whole error stack handling routines removed from
3965 libcrypto, it's only intended to remove all the function name and
3966 reason texts, thereby removing some of the footprint that may not
3967 be interesting if those errors aren't displayed anyway.
3969 NOTE: it's still possible for any application or module to have it's
3970 own set of error texts inserted. The routines are there, just not
3971 used by default when no-err is given.
3974 *) Add support for FreeBSD on IA64.
3975 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
3977 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
3978 Kerberos function mit_des_cbc_cksum(). Before this change,
3979 the value returned by DES_cbc_cksum() was like the one from
3980 mit_des_cbc_cksum(), except the bytes were swapped.
3981 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
3983 *) Allow an application to disable the automatic SSL chain building.
3984 Before this a rather primitive chain build was always performed in
3985 ssl3_output_cert_chain(): an application had no way to send the
3986 correct chain if the automatic operation produced an incorrect result.
3988 Now the chain builder is disabled if either:
3990 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
3992 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
3994 The reasoning behind this is that an application would not want the
3995 auto chain building to take place if extra chain certificates are
3996 present and it might also want a means of sending no additional
3997 certificates (for example the chain has two certificates and the
4001 *) Add the possibility to build without the ENGINE framework.
4002 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4004 *) Under Win32 gmtime() can return NULL: check return value in
4005 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
4008 *) DSA routines: under certain error conditions uninitialized BN objects
4009 could be freed. Solution: make sure initialization is performed early
4010 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
4011 Nils Larsch <nla@trustcenter.de> via PR#459)
4014 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
4015 checked on reconnect on the client side, therefore session resumption
4016 could still fail with a "ssl session id is different" error. This
4017 behaviour is masked when SSL_OP_ALL is used due to
4018 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
4019 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4020 followup to PR #377.
4023 *) IA-32 assembler support enhancements: unified ELF targets, support
4024 for SCO/Caldera platforms, fix for Cygwin shared build.
4027 *) Add support for FreeBSD on sparc64. As a consequence, support for
4028 FreeBSD on non-x86 processors is separate from x86 processors on
4029 the config script, much like the NetBSD support.
4030 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
4032 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
4034 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
4037 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
4038 code (06) was taken as the first octet of the session ID and the last
4039 octet was ignored consequently. As a result SSLv2 client side session
4040 caching could not have worked due to the session ID mismatch between
4042 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4046 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
4047 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
4051 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
4052 seems that in spite of existing for more than a year, many application
4053 author have done nothing to provide the necessary callbacks, which
4054 means that this particular engine will not work properly anywhere.
4055 This is a very unfortunate situation which forces us, in the name
4056 of usability, to give the hw_ncipher.c a static lock, which is part
4058 NOTE: This is for the 0.9.7 series ONLY. This hack will never
4059 appear in 0.9.8 or later. We EXPECT application authors to have
4060 dealt properly with this when 0.9.8 is released (unless we actually
4061 make such changes in the libcrypto locking code that changes will
4062 have to be made anyway).
4065 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
4066 octets have been read, EOF or an error occurs. Without this change
4067 some truncated ASN1 structures will not produce an error.
4070 *) Disable Heimdal support, since it hasn't been fully implemented.
4071 Still give the possibility to force the use of Heimdal, but with
4072 warnings and a request that patches get sent to openssl-dev.
4075 *) Add the VC-CE target, introduce the WINCE sysname, and add
4076 INSTALL.WCE and appropriate conditionals to make it build.
4077 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4079 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
4080 cygssl-x.y.z.dll, where x, y and z are the major, minor and
4081 edit numbers of the version.
4082 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4084 *) Introduce safe string copy and catenation functions
4085 (BUF_strlcpy() and BUF_strlcat()).
4086 [Ben Laurie (CHATS) and Richard Levitte]
4088 *) Avoid using fixed-size buffers for one-line DNs.
4089 [Ben Laurie (CHATS)]
4091 *) Add BUF_MEM_grow_clean() to avoid information leakage when
4092 resizing buffers containing secrets, and use where appropriate.
4093 [Ben Laurie (CHATS)]
4095 *) Avoid using fixed size buffers for configuration file location.
4096 [Ben Laurie (CHATS)]
4098 *) Avoid filename truncation for various CA files.
4099 [Ben Laurie (CHATS)]
4101 *) Use sizeof in preference to magic numbers.
4102 [Ben Laurie (CHATS)]
4104 *) Avoid filename truncation in cert requests.
4105 [Ben Laurie (CHATS)]
4107 *) Add assertions to check for (supposedly impossible) buffer
4109 [Ben Laurie (CHATS)]
4111 *) Don't cache truncated DNS entries in the local cache (this could
4112 potentially lead to a spoofing attack).
4113 [Ben Laurie (CHATS)]
4115 *) Fix various buffers to be large enough for hex/decimal
4116 representations in a platform independent manner.
4117 [Ben Laurie (CHATS)]
4119 *) Add CRYPTO_realloc_clean() to avoid information leakage when
4120 resizing buffers containing secrets, and use where appropriate.
4121 [Ben Laurie (CHATS)]
4123 *) Add BIO_indent() to avoid much slightly worrying code to do
4125 [Ben Laurie (CHATS)]
4127 *) Convert sprintf()/BIO_puts() to BIO_printf().
4128 [Ben Laurie (CHATS)]
4130 *) buffer_gets() could terminate with the buffer only half
4132 [Ben Laurie (CHATS)]
4134 *) Add assertions to prevent user-supplied crypto functions from
4135 overflowing internal buffers by having large block sizes, etc.
4136 [Ben Laurie (CHATS)]
4138 *) New OPENSSL_assert() macro (similar to assert(), but enabled
4140 [Ben Laurie (CHATS)]
4142 *) Eliminate unused copy of key in RC4.
4143 [Ben Laurie (CHATS)]
4145 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
4146 [Ben Laurie (CHATS)]
4148 *) Fix off-by-one error in EGD path.
4149 [Ben Laurie (CHATS)]
4151 *) If RANDFILE path is too long, ignore instead of truncating.
4152 [Ben Laurie (CHATS)]
4154 *) Eliminate unused and incorrectly sized X.509 structure
4156 [Ben Laurie (CHATS)]
4158 *) Eliminate unused and dangerous function knumber().
4159 [Ben Laurie (CHATS)]
4161 *) Eliminate unused and dangerous structure, KSSL_ERR.
4162 [Ben Laurie (CHATS)]
4164 *) Protect against overlong session ID context length in an encoded
4165 session object. Since these are local, this does not appear to be
4167 [Ben Laurie (CHATS)]
4169 *) Change from security patch (see 0.9.6e below) that did not affect
4170 the 0.9.6 release series:
4172 Remote buffer overflow in SSL3 protocol - an attacker could
4173 supply an oversized master key in Kerberos-enabled versions.
4175 [Ben Laurie (CHATS)]
4177 *) Change the SSL kerb5 codes to match RFC 2712.
4180 *) Make -nameopt work fully for req and add -reqopt switch.
4181 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
4183 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
4184 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
4186 *) Make sure tests can be performed even if the corresponding algorithms
4187 have been removed entirely. This was also the last step to make
4188 OpenSSL compilable with DJGPP under all reasonable conditions.
4189 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
4191 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
4192 to allow version independent disabling of normally unselected ciphers,
4193 which may be activated as a side-effect of selecting a single cipher.
4195 (E.g., cipher list string "RSA" enables ciphersuites that are left
4196 out of "ALL" because they do not provide symmetric encryption.
4197 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
4198 [Lutz Jaenicke, Bodo Moeller]
4200 *) Add appropriate support for separate platform-dependent build
4201 directories. The recommended way to make a platform-dependent
4202 build directory is the following (tested on Linux), maybe with
4205 # Place yourself outside of the OpenSSL source tree. In
4206 # this example, the environment variable OPENSSL_SOURCE
4207 # is assumed to contain the absolute OpenSSL source directory.
4208 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
4209 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
4210 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
4211 mkdir -p `dirname $F`
4212 ln -s $OPENSSL_SOURCE/$F $F
4215 To be absolutely sure not to disturb the source tree, a "make clean"
4216 is a good thing. If it isn't successfull, don't worry about it,
4217 it probably means the source directory is very clean.
4220 *) Make sure any ENGINE control commands make local copies of string
4221 pointers passed to them whenever necessary. Otherwise it is possible
4222 the caller may have overwritten (or deallocated) the original string
4223 data when a later ENGINE operation tries to use the stored values.
4224 [Götz Babin-Ebell <babinebell@trustcenter.de>]
4226 *) Improve diagnostics in file reading and command-line digests.
4227 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
4229 *) Add AES modes CFB and OFB to the object database. Correct an
4230 error in AES-CFB decryption.
4233 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
4234 allows existing EVP_CIPHER_CTX structures to be reused after
4235 calling EVP_*Final(). This behaviour is used by encryption
4236 BIOs and some applications. This has the side effect that
4237 applications must explicitly clean up cipher contexts with
4238 EVP_CIPHER_CTX_cleanup() or they will leak memory.
4241 *) Check the values of dna and dnb in bn_mul_recursive before calling
4242 bn_mul_comba (a non zero value means the a or b arrays do not contain
4243 n2 elements) and fallback to bn_mul_normal if either is not zero.
4246 *) Fix escaping of non-ASCII characters when using the -subj option
4247 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
4250 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
4251 form for "surname", serialNumber has no short form.
4252 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
4253 therefore remove "mail" short name for "internet 7".
4254 The OID for unique identifiers in X509 certificates is
4255 x500UniqueIdentifier, not uniqueIdentifier.
4256 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
4259 *) Add an "init" command to the ENGINE config module and auto initialize
4260 ENGINEs. Without any "init" command the ENGINE will be initialized
4261 after all ctrl commands have been executed on it. If init=1 the
4262 ENGINE is initailized at that point (ctrls before that point are run
4263 on the uninitialized ENGINE and after on the initialized one). If
4264 init=0 then the ENGINE will not be iniatialized at all.
4267 *) Fix the 'app_verify_callback' interface so that the user-defined
4268 argument is actually passed to the callback: In the
4269 SSL_CTX_set_cert_verify_callback() prototype, the callback
4270 declaration has been changed from
4273 int (*cb)(X509_STORE_CTX *,void *);
4274 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
4275 i=s->ctx->app_verify_callback(&ctx)
4276 has been changed into
4277 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
4279 To update applications using SSL_CTX_set_cert_verify_callback(),
4280 a dummy argument can be added to their callback functions.
4281 [D. K. Smetters <smetters@parc.xerox.com>]
4283 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
4284 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
4286 *) Add and OPENSSL_LOAD_CONF define which will cause
4287 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
4288 This allows older applications to transparently support certain
4289 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
4290 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
4291 load the config file and OPENSSL_add_all_algorithms_conf() which will
4292 always load it have also been added.
4295 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
4296 Adjust NIDs and EVP layer.
4297 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4299 *) Config modules support in openssl utility.
4301 Most commands now load modules from the config file,
4302 though in a few (such as version) this isn't done
4303 because it couldn't be used for anything.
4305 In the case of ca and req the config file used is
4306 the same as the utility itself: that is the -config
4307 command line option can be used to specify an
4311 *) Move default behaviour from OPENSSL_config(). If appname is NULL
4312 use "openssl_conf" if filename is NULL use default openssl config file.
4315 *) Add an argument to OPENSSL_config() to allow the use of an alternative
4316 config section name. Add a new flag to tolerate a missing config file
4317 and move code to CONF_modules_load_file().
4320 *) Support for crypto accelerator cards from Accelerated Encryption
4321 Processing, www.aep.ie. (Use engine 'aep')
4322 The support was copied from 0.9.6c [engine] and adapted/corrected
4323 to work with the new engine framework.
4324 [AEP Inc. and Richard Levitte]
4326 *) Support for SureWare crypto accelerator cards from Baltimore
4327 Technologies. (Use engine 'sureware')
4328 The support was copied from 0.9.6c [engine] and adapted
4329 to work with the new engine framework.
4332 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
4333 make the newer ENGINE framework commands for the CHIL engine work.
4334 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
4336 *) Make it possible to produce shared libraries on ReliantUNIX.
4337 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
4339 *) Add the configuration target debug-linux-ppro.
4340 Make 'openssl rsa' use the general key loading routines
4341 implemented in apps.c, and make those routines able to
4342 handle the key format FORMAT_NETSCAPE and the variant
4344 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4346 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4347 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4349 *) Add -keyform to rsautl, and document -engine.
4350 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
4352 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
4353 BIO_R_NO_SUCH_FILE error code rather than the generic
4354 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
4357 *) Add new functions
4359 ERR_peek_last_error_line
4360 ERR_peek_last_error_line_data.
4361 These are similar to
4364 ERR_peek_error_line_data,
4365 but report on the latest error recorded rather than the first one
4366 still in the error queue.
4367 [Ben Laurie, Bodo Moeller]
4369 *) default_algorithms option in ENGINE config module. This allows things
4371 default_algorithms = ALL
4372 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
4375 *) Prelminary ENGINE config module.
4378 *) New experimental application configuration code.
4381 *) Change the AES code to follow the same name structure as all other
4382 symmetric ciphers, and behave the same way. Move everything to
4383 the directory crypto/aes, thereby obsoleting crypto/rijndael.
4384 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4386 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
4387 [Ben Laurie and Theo de Raadt]
4389 *) Add option to output public keys in req command.
4390 [Massimiliano Pala madwolf@openca.org]
4392 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
4393 (up to about 10% better than before for P-192 and P-224).
4396 *) New functions/macros
4398 SSL_CTX_set_msg_callback(ctx, cb)
4399 SSL_CTX_set_msg_callback_arg(ctx, arg)
4400 SSL_set_msg_callback(ssl, cb)
4401 SSL_set_msg_callback_arg(ssl, arg)
4403 to request calling a callback function
4405 void cb(int write_p, int version, int content_type,
4406 const void *buf, size_t len, SSL *ssl, void *arg)
4408 whenever a protocol message has been completely received
4409 (write_p == 0) or sent (write_p == 1). Here 'version' is the
4410 protocol version according to which the SSL library interprets
4411 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
4412 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
4413 the content type as defined in the SSL 3.0/TLS 1.0 protocol
4414 specification (change_cipher_spec(20), alert(21), handshake(22)).
4415 'buf' and 'len' point to the actual message, 'ssl' to the
4416 SSL object, and 'arg' is the application-defined value set by
4417 SSL[_CTX]_set_msg_callback_arg().
4419 'openssl s_client' and 'openssl s_server' have new '-msg' options
4420 to enable a callback that displays all protocol messages.
4423 *) Change the shared library support so shared libraries are built as
4424 soon as the corresponding static library is finished, and thereby get
4425 openssl and the test programs linked against the shared library.
4426 This still only happens when the keyword "shard" has been given to
4427 the configuration scripts.
4429 NOTE: shared library support is still an experimental thing, and
4430 backward binary compatibility is still not guaranteed.
4431 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
4433 *) Add support for Subject Information Access extension.
4434 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4436 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
4437 additional bytes when new memory had to be allocated, not just
4438 when reusing an existing buffer.
4441 *) New command line and configuration option 'utf8' for the req command.
4442 This allows field values to be specified as UTF8 strings.
4445 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
4446 runs for the former and machine-readable output for the latter.
4449 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
4450 of the e-mail address in the DN (i.e., it will go into a certificate
4451 extension only). The new configuration file option 'email_in_dn = no'
4452 has the same effect.
4453 [Massimiliano Pala madwolf@openca.org]
4455 *) Change all functions with names starting with des_ to be starting
4456 with DES_ instead. Add wrappers that are compatible with libdes,
4457 but are named _ossl_old_des_*. Finally, add macros that map the
4458 des_* symbols to the corresponding _ossl_old_des_* if libdes
4459 compatibility is desired. If OpenSSL 0.9.6c compatibility is
4460 desired, the des_* symbols will be mapped to DES_*, with one
4463 Since we provide two compatibility mappings, the user needs to
4464 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
4465 compatibility is desired. The default (i.e., when that macro
4466 isn't defined) is OpenSSL 0.9.6c compatibility.
4468 There are also macros that enable and disable the support of old
4469 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
4470 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
4471 are defined, the default will apply: to support the old des routines.
4473 In either case, one must include openssl/des.h to get the correct
4474 definitions. Do not try to just include openssl/des_old.h, that
4477 NOTE: This is a major break of an old API into a new one. Software
4478 authors are encouraged to switch to the DES_ style functions. Some
4479 time in the future, des_old.h and the libdes compatibility functions
4480 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
4481 default), and then completely removed.
4484 *) Test for certificates which contain unsupported critical extensions.
4485 If such a certificate is found during a verify operation it is
4486 rejected by default: this behaviour can be overridden by either
4487 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
4488 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
4489 X509_supported_extension() has also been added which returns 1 if a
4490 particular extension is supported.
4493 *) Modify the behaviour of EVP cipher functions in similar way to digests
4494 to retain compatibility with existing code.
4497 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
4498 compatibility with existing code. In particular the 'ctx' parameter does
4499 not have to be to be initialized before the call to EVP_DigestInit() and
4500 it is tidied up after a call to EVP_DigestFinal(). New function
4501 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
4502 EVP_MD_CTX_copy() changed to not require the destination to be
4503 initialized valid and new function EVP_MD_CTX_copy_ex() added which
4504 requires the destination to be valid.
4506 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
4507 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
4510 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
4511 so that complete 'Handshake' protocol structures are kept in memory
4512 instead of overwriting 'msg_type' and 'length' with 'body' data.
4515 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
4516 [Massimo Santin via Richard Levitte]
4518 *) Major restructuring to the underlying ENGINE code. This includes
4519 reduction of linker bloat, separation of pure "ENGINE" manipulation
4520 (initialisation, etc) from functionality dealing with implementations
4521 of specific crypto iterfaces. This change also introduces integrated
4522 support for symmetric ciphers and digest implementations - so ENGINEs
4523 can now accelerate these by providing EVP_CIPHER and EVP_MD
4524 implementations of their own. This is detailed in crypto/engine/README
4525 as it couldn't be adequately described here. However, there are a few
4526 API changes worth noting - some RSA, DSA, DH, and RAND functions that
4527 were changed in the original introduction of ENGINE code have now
4528 reverted back - the hooking from this code to ENGINE is now a good
4529 deal more passive and at run-time, operations deal directly with
4530 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
4531 dereferencing through an ENGINE pointer any more. Also, the ENGINE
4532 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
4533 they were not being used by the framework as there is no concept of a
4534 BIGNUM_METHOD and they could not be generalised to the new
4535 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
4536 ENGINE_cpy() has been removed as it cannot be consistently defined in
4540 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
4543 *) Change mkdef.pl to sort symbols that get the same entry number,
4544 and make sure the automatically generated functions ERR_load_*
4545 become part of libeay.num as well.
4548 *) New function SSL_renegotiate_pending(). This returns true once
4549 renegotiation has been requested (either SSL_renegotiate() call
4550 or HelloRequest/ClientHello receveived from the peer) and becomes
4551 false once a handshake has been completed.
4552 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
4553 sends a HelloRequest, but does not ensure that a handshake takes
4554 place. SSL_renegotiate_pending() is useful for checking if the
4555 client has followed the request.)
4558 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
4559 By default, clients may request session resumption even during
4560 renegotiation (if session ID contexts permit); with this option,
4561 session resumption is possible only in the first handshake.
4563 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4564 more bits available for options that should not be part of
4565 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4568 *) Add some demos for certificate and certificate request creation.
4571 *) Make maximum certificate chain size accepted from the peer application
4572 settable (SSL*_get/set_max_cert_list()), as proposed by
4573 "Douglas E. Engert" <deengert@anl.gov>.
4576 *) Add support for shared libraries for Unixware-7
4577 (Boyd Lynn Gerber <gerberb@zenez.com>).
4580 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4581 be done prior to destruction. Use this to unload error strings from
4582 ENGINEs that load their own error strings. NB: This adds two new API
4583 functions to "get" and "set" this destroy handler in an ENGINE.
4586 *) Alter all existing ENGINE implementations (except "openssl" and
4587 "openbsd") to dynamically instantiate their own error strings. This
4588 makes them more flexible to be built both as statically-linked ENGINEs
4589 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4590 Also, add stub code to each that makes building them as self-contained
4591 shared-libraries easier (see README.ENGINE).
4594 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4595 implementations into applications that are completely implemented in
4596 self-contained shared-libraries. The "dynamic" ENGINE exposes control
4597 commands that can be used to configure what shared-library to load and
4598 to control aspects of the way it is handled. Also, made an update to
4599 the README.ENGINE file that brings its information up-to-date and
4600 provides some information and instructions on the "dynamic" ENGINE
4601 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
4604 *) Make it possible to unload ranges of ERR strings with a new
4605 "ERR_unload_strings" function.
4608 *) Add a copy() function to EVP_MD.
4611 *) Make EVP_MD routines take a context pointer instead of just the
4612 md_data void pointer.
4615 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
4616 that the digest can only process a single chunk of data
4617 (typically because it is provided by a piece of
4618 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
4619 is only going to provide a single chunk of data, and hence the
4620 framework needn't accumulate the data for oneshot drivers.
4623 *) As with "ERR", make it possible to replace the underlying "ex_data"
4624 functions. This change also alters the storage and management of global
4625 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
4626 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
4627 index counters. The API functions that use this state have been changed
4628 to take a "class_index" rather than pointers to the class's local STACK
4629 and counter, and there is now an API function to dynamically create new
4630 classes. This centralisation allows us to (a) plug a lot of the
4631 thread-safety problems that existed, and (b) makes it possible to clean
4632 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
4633 such data would previously have always leaked in application code and
4634 workarounds were in place to make the memory debugging turn a blind eye
4635 to it. Application code that doesn't use this new function will still
4636 leak as before, but their memory debugging output will announce it now
4637 rather than letting it slide.
4639 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
4640 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
4641 has a return value to indicate success or failure.
4644 *) Make it possible to replace the underlying "ERR" functions such that the
4645 global state (2 LHASH tables and 2 locks) is only used by the "default"
4646 implementation. This change also adds two functions to "get" and "set"
4647 the implementation prior to it being automatically set the first time
4648 any other ERR function takes place. Ie. an application can call "get",
4649 pass the return value to a module it has just loaded, and that module
4650 can call its own "set" function using that value. This means the
4651 module's "ERR" operations will use (and modify) the error state in the
4652 application and not in its own statically linked copy of OpenSSL code.
4655 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
4656 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
4657 the operation, and provides a more encapsulated way for external code
4658 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
4659 to use these functions rather than manually incrementing the counts.
4661 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
4664 *) Add EVP test program.
4667 *) Add symmetric cipher support to ENGINE. Expect the API to change!
4670 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
4671 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
4672 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
4673 These allow a CRL to be built without having to access X509_CRL fields
4674 directly. Modify 'ca' application to use new functions.
4677 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
4678 bug workarounds. Rollback attack detection is a security feature.
4679 The problem will only arise on OpenSSL servers when TLSv1 is not
4680 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
4681 Software authors not wanting to support TLSv1 will have special reasons
4682 for their choice and can explicitly enable this option.
4683 [Bodo Moeller, Lutz Jaenicke]
4685 *) Rationalise EVP so it can be extended: don't include a union of
4686 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
4687 (similar to those existing for EVP_CIPHER_CTX).
4692 EVP_MD_CTX_init(&md); /* new function call */
4693 EVP_DigestInit(&md, EVP_sha1());
4694 EVP_DigestUpdate(&md, in, len);
4695 EVP_DigestFinal(&md, out, NULL);
4696 EVP_MD_CTX_cleanup(&md); /* new function call */
4700 *) Make DES key schedule conform to the usual scheme, as well as
4701 correcting its structure. This means that calls to DES functions
4702 now have to pass a pointer to a des_key_schedule instead of a
4703 plain des_key_schedule (which was actually always a pointer
4706 des_key_schedule ks;
4708 des_set_key_checked(..., &ks);
4709 des_ncbc_encrypt(..., &ks, ...);
4711 (Note that a later change renames 'des_...' into 'DES_...'.)
4714 *) Initial reduction of linker bloat: the use of some functions, such as
4715 PEM causes large amounts of unused functions to be linked in due to
4716 poor organisation. For example pem_all.c contains every PEM function
4717 which has a knock on effect of linking in large amounts of (unused)
4718 ASN1 code. Grouping together similar functions and splitting unrelated
4719 functions prevents this.
4722 *) Cleanup of EVP macros.
4725 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
4726 correct _ecb suffix.
4729 *) Add initial OCSP responder support to ocsp application. The
4730 revocation information is handled using the text based index
4731 use by the ca application. The responder can either handle
4732 requests generated internally, supplied in files (for example
4733 via a CGI script) or using an internal minimal server.
4736 *) Add configuration choices to get zlib compression for TLS.
4739 *) Changes to Kerberos SSL for RFC 2712 compliance:
4740 1. Implemented real KerberosWrapper, instead of just using
4741 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
4742 2. Implemented optional authenticator field of KerberosWrapper.
4744 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
4745 and authenticator structs; see crypto/krb5/.
4747 Generalized Kerberos calls to support multiple Kerberos libraries.
4748 [Vern Staats <staatsvr@asc.hpc.mil>,
4749 Jeffrey Altman <jaltman@columbia.edu>
4750 via Richard Levitte]
4752 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
4753 already does with RSA. testdsa.h now has 'priv_key/pub_key'
4754 values for each of the key sizes rather than having just
4755 parameters (and 'speed' generating keys each time).
4758 *) Speed up EVP routines.
4761 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
4762 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
4763 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
4764 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
4766 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
4767 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
4768 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
4771 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
4773 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
4776 *) Added the OS2-EMX target.
4777 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
4779 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
4780 to support NCONF routines in extension code. New function CONF_set_nconf()
4781 to allow functions which take an NCONF to also handle the old LHASH
4782 structure: this means that the old CONF compatible routines can be
4783 retained (in particular wrt extensions) without having to duplicate the
4784 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
4787 *) Enhance the general user interface with mechanisms for inner control
4788 and with possibilities to have yes/no kind of prompts.
4791 *) Change all calls to low level digest routines in the library and
4792 applications to use EVP. Add missing calls to HMAC_cleanup() and
4793 don't assume HMAC_CTX can be copied using memcpy().
4794 [Verdon Walker <VWalker@novell.com>, Steve Henson]
4796 *) Add the possibility to control engines through control names but with
4797 arbitrary arguments instead of just a string.
4798 Change the key loaders to take a UI_METHOD instead of a callback
4799 function pointer. NOTE: this breaks binary compatibility with earlier
4800 versions of OpenSSL [engine].
4801 Adapt the nCipher code for these new conditions and add a card insertion
4805 *) Enhance the general user interface with mechanisms to better support
4806 dialog box interfaces, application-defined prompts, the possibility
4807 to use defaults (for example default passwords from somewhere else)
4808 and interrupts/cancellations.
4811 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
4812 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
4815 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
4816 tidy up some unnecessarily weird code in 'sk_new()').
4817 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
4819 *) Change the key loading routines for ENGINEs to use the same kind
4820 callback (pem_password_cb) as all other routines that need this
4824 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
4825 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
4826 than this minimum value is recommended.
4829 *) New random seeder for OpenVMS, using the system process statistics
4830 that are easily reachable.
4833 *) Windows apparently can't transparently handle global
4834 variables defined in DLLs. Initialisations such as:
4836 const ASN1_ITEM *it = &ASN1_INTEGER_it;
4838 wont compile. This is used by the any applications that need to
4839 declare their own ASN1 modules. This was fixed by adding the option
4840 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
4841 needed for static libraries under Win32.
4844 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
4845 setting of purpose and trust fields. New X509_STORE trust and
4846 purpose functions and tidy up setting in other SSL functions.
4849 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
4850 structure. These are inherited by X509_STORE_CTX when it is
4851 initialised. This allows various defaults to be set in the
4852 X509_STORE structure (such as flags for CRL checking and custom
4853 purpose or trust settings) for functions which only use X509_STORE_CTX
4854 internally such as S/MIME.
4856 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
4857 trust settings if they are not set in X509_STORE. This allows X509_STORE
4858 purposes and trust (in S/MIME for example) to override any set by default.
4860 Add command line options for CRL checking to smime, s_client and s_server
4864 *) Initial CRL based revocation checking. If the CRL checking flag(s)
4865 are set then the CRL is looked up in the X509_STORE structure and
4866 its validity and signature checked, then if the certificate is found
4867 in the CRL the verify fails with a revoked error.
4869 Various new CRL related callbacks added to X509_STORE_CTX structure.
4871 Command line options added to 'verify' application to support this.
4873 This needs some additional work, such as being able to handle multiple
4874 CRLs with different times, extension based lookup (rather than just
4875 by subject name) and ultimately more complete V2 CRL extension
4879 *) Add a general user interface API (crypto/ui/). This is designed
4880 to replace things like des_read_password and friends (backward
4881 compatibility functions using this new API are provided).
4882 The purpose is to remove prompting functions from the DES code
4883 section as well as provide for prompting through dialog boxes in
4884 a window system and the like.
4887 *) Add "ex_data" support to ENGINE so implementations can add state at a
4888 per-structure level rather than having to store it globally.
4891 *) Make it possible for ENGINE structures to be copied when retrieved by
4892 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
4893 This causes the "original" ENGINE structure to act like a template,
4894 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
4895 operational state can be localised to each ENGINE structure, despite the
4896 fact they all share the same "methods". New ENGINE structures returned in
4897 this case have no functional references and the return value is the single
4898 structural reference. This matches the single structural reference returned
4899 by ENGINE_by_id() normally, when it is incremented on the pre-existing
4903 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
4904 needs to match any other type at all we need to manually clear the
4908 *) Changes to the "openssl engine" utility to include;
4909 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
4910 about an ENGINE's available control commands.
4911 - executing control commands from command line arguments using the
4912 '-pre' and '-post' switches. '-post' is only used if '-t' is
4913 specified and the ENGINE is successfully initialised. The syntax for
4914 the individual commands are colon-separated, for example;
4915 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
4918 *) New dynamic control command support for ENGINEs. ENGINEs can now
4919 declare their own commands (numbers), names (strings), descriptions,
4920 and input types for run-time discovery by calling applications. A
4921 subset of these commands are implicitly classed as "executable"
4922 depending on their input type, and only these can be invoked through
4923 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
4924 can be based on user input, config files, etc). The distinction is
4925 that "executable" commands cannot return anything other than a boolean
4926 result and can only support numeric or string input, whereas some
4927 discoverable commands may only be for direct use through
4928 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
4929 pointers, or other custom uses. The "executable" commands are to
4930 support parameterisations of ENGINE behaviour that can be
4931 unambiguously defined by ENGINEs and used consistently across any
4932 OpenSSL-based application. Commands have been added to all the
4933 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
4934 control over shared-library paths without source code alterations.
4937 *) Changed all ENGINE implementations to dynamically allocate their
4938 ENGINEs rather than declaring them statically. Apart from this being
4939 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
4940 this also allows the implementations to compile without using the
4941 internal engine_int.h header.
4944 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
4945 'const' value. Any code that should be able to modify a RAND_METHOD
4946 should already have non-const pointers to it (ie. they should only
4947 modify their own ones).
4950 *) Made a variety of little tweaks to the ENGINE code.
4951 - "atalla" and "ubsec" string definitions were moved from header files
4952 to C code. "nuron" string definitions were placed in variables
4953 rather than hard-coded - allowing parameterisation of these values
4954 later on via ctrl() commands.
4955 - Removed unused "#if 0"'d code.
4956 - Fixed engine list iteration code so it uses ENGINE_free() to release
4957 structural references.
4958 - Constified the RAND_METHOD element of ENGINE structures.
4959 - Constified various get/set functions as appropriate and added
4960 missing functions (including a catch-all ENGINE_cpy that duplicates
4961 all ENGINE values onto a new ENGINE except reference counts/state).
4962 - Removed NULL parameter checks in get/set functions. Setting a method
4963 or function to NULL is a way of cancelling out a previously set
4964 value. Passing a NULL ENGINE parameter is just plain stupid anyway
4965 and doesn't justify the extra error symbols and code.
4966 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
4967 flags from engine_int.h to engine.h.
4968 - Changed prototypes for ENGINE handler functions (init(), finish(),
4969 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
4972 *) Implement binary inversion algorithm for BN_mod_inverse in addition
4973 to the algorithm using long division. The binary algorithm can be
4974 used only if the modulus is odd. On 32-bit systems, it is faster
4975 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
4976 roughly 5-15% for 256-bit moduli), so we use it only for moduli
4977 up to 450 bits. In 64-bit environments, the binary algorithm
4978 appears to be advantageous for much longer moduli; here we use it
4979 for moduli up to 2048 bits.
4982 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
4983 could not support the combine flag in choice fields.
4986 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
4987 extensions from a certificate request to the certificate.
4990 *) Allow multiple 'certopt' and 'nameopt' options to be separated
4991 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
4992 file: this allows the display of the certificate about to be
4993 signed to be customised, to allow certain fields to be included
4994 or excluded and extension details. The old system didn't display
4995 multicharacter strings properly, omitted fields not in the policy
4996 and couldn't display additional details such as extensions.
4999 *) Function EC_POINTs_mul for multiple scalar multiplication
5000 of an arbitrary number of elliptic curve points
5001 \sum scalars[i]*points[i],
5002 optionally including the generator defined for the EC_GROUP:
5003 scalar*generator + \sum scalars[i]*points[i].
5005 EC_POINT_mul is a simple wrapper function for the typical case
5006 that the point list has just one item (besides the optional
5010 *) First EC_METHODs for curves over GF(p):
5012 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
5013 operations and provides various method functions that can also
5014 operate with faster implementations of modular arithmetic.
5016 EC_GFp_mont_method() reuses most functions that are part of
5017 EC_GFp_simple_method, but uses Montgomery arithmetic.
5019 [Bodo Moeller; point addition and point doubling
5020 implementation directly derived from source code provided by
5021 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
5023 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
5024 crypto/ec/ec_lib.c):
5026 Curves are EC_GROUP objects (with an optional group generator)
5027 based on EC_METHODs that are built into the library.
5029 Points are EC_POINT objects based on EC_GROUP objects.
5031 Most of the framework would be able to handle curves over arbitrary
5032 finite fields, but as there are no obvious types for fields other
5033 than GF(p), some functions are limited to that for now.
5036 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
5037 that the file contains a complete HTTP response.
5040 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
5041 change the def and num file printf format specifier from "%-40sXXX"
5042 to "%-39s XXX". The latter will always guarantee a space after the
5043 field while the former will cause them to run together if the field
5044 is 40 of more characters long.
5047 *) Constify the cipher and digest 'method' functions and structures
5048 and modify related functions to take constant EVP_MD and EVP_CIPHER
5052 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
5053 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
5056 *) Modify EVP_Digest*() routines so they now return values. Although the
5057 internal software routines can never fail additional hardware versions
5061 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
5063 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
5064 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
5067 ERR_R_NESTED_ASN1_ERROR
5069 ERR_R_MISSING_ASN1_EOS
5070 were 4 .. 9, conflicting with
5071 ERR_LIB_RSA (= ERR_R_RSA_LIB)
5073 ERR_LIB_PEM (= ERR_R_PEM_LIB).
5074 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
5076 Add new error code 'ERR_R_INTERNAL_ERROR'.
5079 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
5083 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
5084 sets the subject name for a new request or supersedes the
5085 subject name in a given request. Formats that can be parsed are
5086 'CN=Some Name, OU=myOU, C=IT'
5088 'CN=Some Name/OU=myOU/C=IT'.
5090 Add options '-batch' and '-verbose' to 'openssl req'.
5091 [Massimiliano Pala <madwolf@hackmasters.net>]
5093 *) Introduce the possibility to access global variables through
5094 functions on platform were that's the best way to handle exporting
5095 global variables in shared libraries. To enable this functionality,
5096 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
5097 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
5098 is normally done by Configure or something similar).
5100 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
5101 in the source file (foo.c) like this:
5103 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
5104 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
5106 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
5107 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
5109 OPENSSL_DECLARE_GLOBAL(int,foo);
5110 #define foo OPENSSL_GLOBAL_REF(foo)
5111 OPENSSL_DECLARE_GLOBAL(double,bar);
5112 #define bar OPENSSL_GLOBAL_REF(bar)
5114 The #defines are very important, and therefore so is including the
5115 header file everywhere where the defined globals are used.
5117 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
5118 of ASN.1 items, but that structure is a bit different.
5120 The largest change is in util/mkdef.pl which has been enhanced with
5121 better and easier to understand logic to choose which symbols should
5122 go into the Windows .def files as well as a number of fixes and code
5123 cleanup (among others, algorithm keywords are now sorted
5124 lexicographically to avoid constant rewrites).
5127 *) In BN_div() keep a copy of the sign of 'num' before writing the
5128 result to 'rm' because if rm==num the value will be overwritten
5129 and produce the wrong result if 'num' is negative: this caused
5130 problems with BN_mod() and BN_nnmod().
5133 *) Function OCSP_request_verify(). This checks the signature on an
5134 OCSP request and verifies the signer certificate. The signer
5135 certificate is just checked for a generic purpose and OCSP request
5139 *) Add OCSP_check_validity() function to check the validity of OCSP
5140 responses. OCSP responses are prepared in real time and may only
5141 be a few seconds old. Simply checking that the current time lies
5142 between thisUpdate and nextUpdate max reject otherwise valid responses
5143 caused by either OCSP responder or client clock inaccuracy. Instead
5144 we allow thisUpdate and nextUpdate to fall within a certain period of
5145 the current time. The age of the response can also optionally be
5146 checked. Two new options -validity_period and -status_age added to
5150 *) If signature or public key algorithm is unrecognized print out its
5151 OID rather that just UNKNOWN.
5154 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
5155 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
5156 ID to be generated from the issuer certificate alone which can then be
5157 passed to OCSP_id_issuer_cmp().
5160 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
5161 ASN1 modules to export functions returning ASN1_ITEM pointers
5162 instead of the ASN1_ITEM structures themselves. This adds several
5163 new macros which allow the underlying ASN1 function/structure to
5164 be accessed transparently. As a result code should not use ASN1_ITEM
5165 references directly (such as &X509_it) but instead use the relevant
5166 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
5167 use of the new ASN1 code on platforms where exporting structures
5168 is problematical (for example in shared libraries) but exporting
5169 functions returning pointers to structures is not.
5172 *) Add support for overriding the generation of SSL/TLS session IDs.
5173 These callbacks can be registered either in an SSL_CTX or per SSL.
5174 The purpose of this is to allow applications to control, if they wish,
5175 the arbitrary values chosen for use as session IDs, particularly as it
5176 can be useful for session caching in multiple-server environments. A
5177 command-line switch for testing this (and any client code that wishes
5178 to use such a feature) has been added to "s_server".
5179 [Geoff Thorpe, Lutz Jaenicke]
5181 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
5182 of the form '#if defined(...) || defined(...) || ...' and
5183 '#if !defined(...) && !defined(...) && ...'. This also avoids
5184 the growing number of special cases it was previously handling.
5187 *) Make all configuration macros available for application by making
5188 sure they are available in opensslconf.h, by giving them names starting
5189 with "OPENSSL_" to avoid conflicts with other packages and by making
5190 sure e_os2.h will cover all platform-specific cases together with
5192 Additionally, it is now possible to define configuration/platform-
5193 specific names (called "system identities"). In the C code, these
5194 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
5195 macro with the name beginning with "OPENSSL_SYS_", which is determined
5196 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
5200 *) New option -set_serial to 'req' and 'x509' this allows the serial
5201 number to use to be specified on the command line. Previously self
5202 signed certificates were hard coded with serial number 0 and the
5203 CA options of 'x509' had to use a serial number in a file which was
5207 *) New options to 'ca' utility to support V2 CRL entry extensions.
5208 Currently CRL reason, invalidity date and hold instruction are
5209 supported. Add new CRL extensions to V3 code and some new objects.
5212 *) New function EVP_CIPHER_CTX_set_padding() this is used to
5213 disable standard block padding (aka PKCS#5 padding) in the EVP
5214 API, which was previously mandatory. This means that the data is
5215 not padded in any way and so the total length much be a multiple
5216 of the block size, otherwise an error occurs.
5219 *) Initial (incomplete) OCSP SSL support.
5222 *) New function OCSP_parse_url(). This splits up a URL into its host,
5223 port and path components: primarily to parse OCSP URLs. New -url
5224 option to ocsp utility.
5227 *) New nonce behavior. The return value of OCSP_check_nonce() now
5228 reflects the various checks performed. Applications can decide
5229 whether to tolerate certain situations such as an absent nonce
5230 in a response when one was present in a request: the ocsp application
5231 just prints out a warning. New function OCSP_add1_basic_nonce()
5232 this is to allow responders to include a nonce in a response even if
5233 the request is nonce-less.
5236 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
5237 skipped when using openssl x509 multiple times on a single input file,
5238 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
5241 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
5242 set string type: to handle setting ASN1_TIME structures. Fix ca
5243 utility to correctly initialize revocation date of CRLs.
5246 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
5247 the clients preferred ciphersuites and rather use its own preferences.
5248 Should help to work around M$ SGC (Server Gated Cryptography) bug in
5249 Internet Explorer by ensuring unchanged hash method during stepup.
5250 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
5253 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
5254 to aes and add a new 'exist' option to print out symbols that don't
5258 *) Additional options to ocsp utility to allow flags to be set and
5259 additional certificates supplied.
5262 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
5263 OCSP client a number of certificate to only verify the response
5267 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
5268 handle the new API. Currently only ECB, CBC modes supported. Add new
5271 Add TLS AES ciphersuites as described in RFC3268, "Advanced
5272 Encryption Standard (AES) Ciphersuites for Transport Layer
5273 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
5274 not enabled by default and were not part of the "ALL" ciphersuite
5275 alias because they were not yet official; they could be
5276 explicitly requested by specifying the "AESdraft" ciphersuite
5277 group alias. In the final release of OpenSSL 0.9.7, the group
5278 alias is called "AES" and is part of "ALL".)
5279 [Ben Laurie, Steve Henson, Bodo Moeller]
5281 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
5282 request to response.
5285 *) Functions for OCSP responders. OCSP_request_onereq_count(),
5286 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
5287 extract information from a certificate request. OCSP_response_create()
5288 creates a response and optionally adds a basic response structure.
5289 OCSP_basic_add1_status() adds a complete single response to a basic
5290 response and returns the OCSP_SINGLERESP structure just added (to allow
5291 extensions to be included for example). OCSP_basic_add1_cert() adds a
5292 certificate to a basic response and OCSP_basic_sign() signs a basic
5293 response with various flags. New helper functions ASN1_TIME_check()
5294 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
5295 (converts ASN1_TIME to GeneralizedTime).
5298 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
5299 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
5300 structure from a certificate. X509_pubkey_digest() digests the public_key
5301 contents: this is used in various key identifiers.
5304 *) Make sk_sort() tolerate a NULL argument.
5305 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
5307 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
5308 passed by the function are trusted implicitly. If any of them signed the
5309 response then it is assumed to be valid and is not verified.
5312 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
5313 to data. This was previously part of the PKCS7 ASN1 code. This
5314 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
5315 [Steve Henson, reported by Kenneth R. Robinette
5316 <support@securenetterm.com>]
5318 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
5319 routines: without these tracing memory leaks is very painful.
5320 Fix leaks in PKCS12 and PKCS7 routines.
5323 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
5324 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
5325 effectively meant GeneralizedTime would never be used. Now it
5326 is initialised to -1 but X509_time_adj() now has to check the value
5327 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
5328 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
5329 [Steve Henson, reported by Kenneth R. Robinette
5330 <support@securenetterm.com>]
5332 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
5333 result in a zero length in the ASN1_INTEGER structure which was
5334 not consistent with the structure when d2i_ASN1_INTEGER() was used
5335 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
5336 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
5337 where it did not print out a minus for negative ASN1_INTEGER.
5340 *) Add summary printout to ocsp utility. The various functions which
5341 convert status values to strings have been renamed to:
5342 OCSP_response_status_str(), OCSP_cert_status_str() and
5343 OCSP_crl_reason_str() and are no longer static. New options
5344 to verify nonce values and to disable verification. OCSP response
5345 printout format cleaned up.
5348 *) Add additional OCSP certificate checks. These are those specified
5349 in RFC2560. This consists of two separate checks: the CA of the
5350 certificate being checked must either be the OCSP signer certificate
5351 or the issuer of the OCSP signer certificate. In the latter case the
5352 OCSP signer certificate must contain the OCSP signing extended key
5353 usage. This check is performed by attempting to match the OCSP
5354 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
5355 in the OCSP_CERTID structures of the response.
5358 *) Initial OCSP certificate verification added to OCSP_basic_verify()
5359 and related routines. This uses the standard OpenSSL certificate
5360 verify routines to perform initial checks (just CA validity) and
5361 to obtain the certificate chain. Then additional checks will be
5362 performed on the chain. Currently the root CA is checked to see
5363 if it is explicitly trusted for OCSP signing. This is used to set
5364 a root CA as a global signing root: that is any certificate that
5365 chains to that CA is an acceptable OCSP signing certificate.
5368 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
5369 extensions from a separate configuration file.
5370 As when reading extensions from the main configuration file,
5371 the '-extensions ...' option may be used for specifying the
5373 [Massimiliano Pala <madwolf@comune.modena.it>]
5375 *) New OCSP utility. Allows OCSP requests to be generated or
5376 read. The request can be sent to a responder and the output
5377 parsed, outputed or printed in text form. Not complete yet:
5378 still needs to check the OCSP response validity.
5381 *) New subcommands for 'openssl ca':
5382 'openssl ca -status <serial>' prints the status of the cert with
5383 the given serial number (according to the index file).
5384 'openssl ca -updatedb' updates the expiry status of certificates
5386 [Massimiliano Pala <madwolf@comune.modena.it>]
5388 *) New '-newreq-nodes' command option to CA.pl. This is like
5389 '-newreq', but calls 'openssl req' with the '-nodes' option
5390 so that the resulting key is not encrypted.
5391 [Damien Miller <djm@mindrot.org>]
5393 *) New configuration for the GNU Hurd.
5394 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
5396 *) Initial code to implement OCSP basic response verify. This
5397 is currently incomplete. Currently just finds the signer's
5398 certificate and verifies the signature on the response.
5401 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
5402 value of OPENSSLDIR. This is available via the new '-d' option
5403 to 'openssl version', and is also included in 'openssl version -a'.
5406 *) Allowing defining memory allocation callbacks that will be given
5407 file name and line number information in additional arguments
5408 (a const char* and an int). The basic functionality remains, as
5409 well as the original possibility to just replace malloc(),
5410 realloc() and free() by functions that do not know about these
5411 additional arguments. To register and find out the current
5412 settings for extended allocation functions, the following
5413 functions are provided:
5415 CRYPTO_set_mem_ex_functions
5416 CRYPTO_set_locked_mem_ex_functions
5417 CRYPTO_get_mem_ex_functions
5418 CRYPTO_get_locked_mem_ex_functions
5420 These work the same way as CRYPTO_set_mem_functions and friends.
5421 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
5422 extended allocation function is enabled.
5423 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
5424 a conventional allocation function is enabled.
5425 [Richard Levitte, Bodo Moeller]
5427 *) Finish off removing the remaining LHASH function pointer casts.
5428 There should no longer be any prototype-casting required when using
5429 the LHASH abstraction, and any casts that remain are "bugs". See
5430 the callback types and macros at the head of lhash.h for details
5431 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
5434 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
5435 If /dev/[u]random devices are not available or do not return enough
5436 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
5438 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
5439 /etc/entropy will be queried once each in this sequence, quering stops
5440 when enough entropy was collected without querying more sockets.
5443 *) Change the Unix RAND_poll() variant to be able to poll several
5444 random devices, as specified by DEVRANDOM, until a sufficient amount
5445 of data has been collected. We spend at most 10 ms on each file
5446 (select timeout) and read in non-blocking mode. DEVRANDOM now
5447 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
5448 (previously it was just the string "/dev/urandom"), so on typical
5449 platforms the 10 ms delay will never occur.
5450 Also separate out the Unix variant to its own file, rand_unix.c.
5451 For VMS, there's a currently-empty rand_vms.c.
5454 *) Move OCSP client related routines to ocsp_cl.c. These
5455 provide utility functions which an application needing
5456 to issue a request to an OCSP responder and analyse the
5457 response will typically need: as opposed to those which an
5458 OCSP responder itself would need which will be added later.
5460 OCSP_request_sign() signs an OCSP request with an API similar
5461 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
5462 response. OCSP_response_get1_basic() extracts basic response
5463 from response. OCSP_resp_find_status(): finds and extracts status
5464 information from an OCSP_CERTID structure (which will be created
5465 when the request structure is built). These are built from lower
5466 level functions which work on OCSP_SINGLERESP structures but
5467 wont normally be used unless the application wishes to examine
5468 extensions in the OCSP response for example.
5470 Replace nonce routines with a pair of functions.
5471 OCSP_request_add1_nonce() adds a nonce value and optionally
5472 generates a random value. OCSP_check_nonce() checks the
5473 validity of the nonce in an OCSP response.
5476 *) Change function OCSP_request_add() to OCSP_request_add0_id().
5477 This doesn't copy the supplied OCSP_CERTID and avoids the
5478 need to free up the newly created id. Change return type
5479 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
5480 This can then be used to add extensions to the request.
5481 Deleted OCSP_request_new(), since most of its functionality
5482 is now in OCSP_REQUEST_new() (and the case insensitive name
5483 clash) apart from the ability to set the request name which
5484 will be added elsewhere.
5487 *) Update OCSP API. Remove obsolete extensions argument from
5488 various functions. Extensions are now handled using the new
5489 OCSP extension code. New simple OCSP HTTP function which
5490 can be used to send requests and parse the response.
5493 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
5494 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
5495 uses the special reorder version of SET OF to sort the attributes
5496 and reorder them to match the encoded order. This resolves a long
5497 standing problem: a verify on a PKCS7 structure just after signing
5498 it used to fail because the attribute order did not match the
5499 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
5500 it uses the received order. This is necessary to tolerate some broken
5501 software that does not order SET OF. This is handled by encoding
5502 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
5503 to produce the required SET OF.
5506 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
5507 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
5508 files to get correct declarations of the ASN.1 item variables.
5511 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
5512 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
5513 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
5514 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
5515 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
5516 ASN1_ITEM and no wrapper functions.
5519 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
5520 replace the old function pointer based I/O routines. Change most of
5521 the *_d2i_bio() and *_d2i_fp() functions to use these.
5524 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
5525 lines, recognice more "algorithms" that can be deselected, and make
5526 it complain about algorithm deselection that isn't recognised.
5529 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
5530 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
5531 to use new functions. Add NO_ASN1_OLD which can be set to remove
5532 some old style ASN1 functions: this can be used to determine if old
5533 code will still work when these eventually go away.
5536 *) New extension functions for OCSP structures, these follow the
5537 same conventions as certificates and CRLs.
5540 *) New function X509V3_add1_i2d(). This automatically encodes and
5541 adds an extension. Its behaviour can be customised with various
5542 flags to append, replace or delete. Various wrappers added for
5543 certifcates and CRLs.
5546 *) Fix to avoid calling the underlying ASN1 print routine when
5547 an extension cannot be parsed. Correct a typo in the
5548 OCSP_SERVICELOC extension. Tidy up print OCSP format.
5551 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
5552 entries for variables.
5555 *) Add functionality to apps/openssl.c for detecting locking
5556 problems: As the program is single-threaded, all we have
5557 to do is register a locking callback using an array for
5558 storing which locks are currently held by the program.
5561 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
5562 SSL_get_ex_data_X509_STORE_idx(), which is used in
5563 ssl_verify_cert_chain() and thus can be called at any time
5564 during TLS/SSL handshakes so that thread-safety is essential.
5565 Unfortunately, the ex_data design is not at all suited
5566 for multi-threaded use, so it probably should be abolished.
5569 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5570 [Broadcom, tweaked and integrated by Geoff Thorpe]
5572 *) Move common extension printing code to new function
5573 X509V3_print_extensions(). Reorganise OCSP print routines and
5574 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5577 *) New function X509_signature_print() to remove duplication in some
5581 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5582 set (this was treated exactly the same as SET OF previously). This
5583 is used to reorder the STACK representing the structure to match the
5584 encoding. This will be used to get round a problem where a PKCS7
5585 structure which was signed could not be verified because the STACK
5586 order did not reflect the encoded order.
5589 *) Reimplement the OCSP ASN1 module using the new code.
5592 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5593 for its ASN1 operations. The old style function pointers still exist
5594 for now but they will eventually go away.
5597 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
5598 completely replaces the old ASN1 functionality with a table driven
5599 encoder and decoder which interprets an ASN1_ITEM structure describing
5600 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
5601 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
5602 has also been converted to the new form.
5605 *) Change BN_mod_exp_recp so that negative moduli are tolerated
5606 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
5607 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
5608 for negative moduli.
5611 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
5612 of not touching the result's sign bit.
5615 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
5619 *) Changed the LHASH code to use prototypes for callbacks, and created
5620 macros to declare and implement thin (optionally static) functions
5621 that provide type-safety and avoid function pointer casting for the
5622 type-specific callbacks.
5625 *) Added Kerberos Cipher Suites to be used with TLS, as written in
5627 [Veers Staats <staatsvr@asc.hpc.mil>,
5628 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
5630 *) Reformat the FAQ so the different questions and answers can be divided
5631 in sections depending on the subject.
5634 *) Have the zlib compression code load ZLIB.DLL dynamically under
5638 *) New function BN_mod_sqrt for computing square roots modulo a prime
5639 (using the probabilistic Tonelli-Shanks algorithm unless
5640 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
5641 be handled deterministically).
5642 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5644 *) Make BN_mod_inverse faster by explicitly handling small quotients
5645 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
5646 512 bits], about 30% for larger ones [1024 or 2048 bits].)
5649 *) New function BN_kronecker.
5652 *) Fix BN_gcd so that it works on negative inputs; the result is
5653 positive unless both parameters are zero.
5654 Previously something reasonably close to an infinite loop was
5655 possible because numbers could be growing instead of shrinking
5656 in the implementation of Euclid's algorithm.
5659 *) Fix BN_is_word() and BN_is_one() macros to take into account the
5660 sign of the number in question.
5662 Fix BN_is_word(a,w) to work correctly for w == 0.
5664 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
5665 because its test if the absolute value of 'a' equals 'w'.
5666 Note that BN_abs_is_word does *not* handle w == 0 reliably;
5667 it exists mostly for use in the implementations of BN_is_zero(),
5668 BN_is_one(), and BN_is_word().
5671 *) New function BN_swap.
5674 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
5675 the exponentiation functions are more likely to produce reasonable
5676 results on negative inputs.
5679 *) Change BN_mod_mul so that the result is always non-negative.
5680 Previously, it could be negative if one of the factors was negative;
5681 I don't think anyone really wanted that behaviour.
5684 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
5685 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
5686 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
5687 and add new functions:
5696 BN_mod_lshift1_quick
5700 These functions always generate non-negative results.
5702 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
5703 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
5705 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
5706 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
5707 be reduced modulo m.
5708 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5711 The following entry accidentily appeared in the CHANGES file
5712 distributed with OpenSSL 0.9.7. The modifications described in
5713 it do *not* apply to OpenSSL 0.9.7.
5715 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5716 was actually never needed) and in BN_mul(). The removal in BN_mul()
5717 required a small change in bn_mul_part_recursive() and the addition
5718 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5719 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5720 bn_sub_words() and bn_add_words() except they take arrays with
5725 *) In 'openssl passwd', verify passwords read from the terminal
5726 unless the '-salt' option is used (which usually means that
5727 verification would just waste user's time since the resulting
5728 hash is going to be compared with some given password hash)
5729 or the new '-noverify' option is used.
5731 This is an incompatible change, but it does not affect
5732 non-interactive use of 'openssl passwd' (passwords on the command
5733 line, '-stdin' option, '-in ...' option) and thus should not
5737 *) Remove all references to RSAref, since there's no more need for it.
5740 *) Make DSO load along a path given through an environment variable
5741 (SHLIB_PATH) with shl_load().
5744 *) Constify the ENGINE code as a result of BIGNUM constification.
5745 Also constify the RSA code and most things related to it. In a
5746 few places, most notable in the depth of the ASN.1 code, ugly
5747 casts back to non-const were required (to be solved at a later
5751 *) Make it so the openssl application has all engines loaded by default.
5754 *) Constify the BIGNUM routines a little more.
5757 *) Add the following functions:
5759 ENGINE_load_cswift()
5761 ENGINE_load_atalla()
5763 ENGINE_load_builtin_engines()
5765 That way, an application can itself choose if external engines that
5766 are built-in in OpenSSL shall ever be used or not. The benefit is
5767 that applications won't have to be linked with libdl or other dso
5768 libraries unless it's really needed.
5770 Changed 'openssl engine' to load all engines on demand.
5771 Changed the engine header files to avoid the duplication of some
5772 declarations (they differed!).
5775 *) 'openssl engine' can now list capabilities.
5778 *) Better error reporting in 'openssl engine'.
5781 *) Never call load_dh_param(NULL) in s_server.
5784 *) Add engine application. It can currently list engines by name and
5785 identity, and test if they are actually available.
5788 *) Improve RPM specification file by forcing symbolic linking and making
5789 sure the installed documentation is also owned by root.root.
5790 [Damien Miller <djm@mindrot.org>]
5792 *) Give the OpenSSL applications more possibilities to make use of
5793 keys (public as well as private) handled by engines.
5796 *) Add OCSP code that comes from CertCo.
5799 *) Add VMS support for the Rijndael code.
5802 *) Added untested support for Nuron crypto accelerator.
5805 *) Add support for external cryptographic devices. This code was
5806 previously distributed separately as the "engine" branch.
5807 [Geoff Thorpe, Richard Levitte]
5809 *) Rework the filename-translation in the DSO code. It is now possible to
5810 have far greater control over how a "name" is turned into a filename
5811 depending on the operating environment and any oddities about the
5812 different shared library filenames on each system.
5815 *) Support threads on FreeBSD-elf in Configure.
5818 *) Fix for SHA1 assembly problem with MASM: it produces
5819 warnings about corrupt line number information when assembling
5820 with debugging information. This is caused by the overlapping
5822 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
5825 NCONF_get_number() has no error checking at all. As a replacement,
5826 NCONF_get_number_e() is defined (_e for "error checking") and is
5827 promoted strongly. The old NCONF_get_number is kept around for
5828 binary backward compatibility.
5829 Make it possible for methods to load from something other than a BIO,
5830 by providing a function pointer that is given a name instead of a BIO.
5831 For example, this could be used to load configuration data from an
5835 *) Fix for non blocking accept BIOs. Added new I/O special reason
5836 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
5837 with non blocking I/O was not possible because no retry code was
5838 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
5842 *) Added the beginnings of Rijndael support.
5845 *) Fix for bug in DirectoryString mask setting. Add support for
5846 X509_NAME_print_ex() in 'req' and X509_print_ex() function
5847 to allow certificate printing to more controllable, additional
5848 'certopt' option to 'x509' to allow new printing options to be
5852 *) Clean old EAY MD5 hack from e_os.h.
5855 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
5857 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5858 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5859 [Joe Orton, Steve Henson]
5861 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
5863 *) Fix additional bug revealed by the NISCC test suite:
5865 Stop bug triggering large recursion when presented with
5866 certain ASN.1 tags (CVE-2003-0851)
5869 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
5871 *) Fix various bugs revealed by running the NISCC test suite:
5873 Stop out of bounds reads in the ASN1 code when presented with
5874 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5876 If verify callback ignores invalid public key errors don't try to check
5877 certificate signature with the NULL public key.
5881 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5882 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5886 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5887 extra data after the compression methods not only for TLS 1.0
5888 but also for SSL 3.0 (as required by the specification).
5889 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5891 *) Change X509_certificate_type() to mark the key as exported/exportable
5892 when it's 512 *bits* long, not 512 bytes.
5895 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
5897 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5898 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5899 a protocol version number mismatch like a decryption error
5900 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5903 *) Turn on RSA blinding by default in the default implementation
5904 to avoid a timing attack. Applications that don't want it can call
5905 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5906 They would be ill-advised to do so in most cases.
5907 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5909 *) Change RSA blinding code so that it works when the PRNG is not
5910 seeded (in this case, the secret RSA exponent is abused as
5911 an unpredictable seed -- if it is not unpredictable, there
5912 is no point in blinding anyway). Make RSA blinding thread-safe
5913 by remembering the creator's thread ID in rsa->blinding and
5914 having all other threads use local one-time blinding factors
5915 (this requires more computation than sharing rsa->blinding, but
5916 avoids excessive locking; and if an RSA object is not shared
5917 between threads, blinding will still be very fast).
5920 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
5922 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5923 via timing by performing a MAC computation even if incorrrect
5924 block cipher padding has been found. This is a countermeasure
5925 against active attacks where the attacker has to distinguish
5926 between bad padding and a MAC verification error. (CVE-2003-0078)
5928 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5929 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5930 Martin Vuagnoux (EPFL, Ilion)]
5932 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
5934 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
5935 memory from it's contents. This is done with a counter that will
5936 place alternating values in each byte. This can be used to solve
5937 two issues: 1) the removal of calls to memset() by highly optimizing
5938 compilers, and 2) cleansing with other values than 0, since those can
5939 be read through on certain media, for example a swap space on disk.
5942 *) Bugfix: client side session caching did not work with external caching,
5943 because the session->cipher setting was not restored when reloading
5944 from the external cache. This problem was masked, when
5945 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
5946 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
5949 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
5950 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
5951 [Zeev Lieber <zeev-l@yahoo.com>]
5953 *) Undo an undocumented change introduced in 0.9.6e which caused
5954 repeated calls to OpenSSL_add_all_ciphers() and
5955 OpenSSL_add_all_digests() to be ignored, even after calling
5959 *) Change the default configuration reader to deal with last line not
5960 being properly terminated.
5963 *) Change X509_NAME_cmp() so it applies the special rules on handling
5964 DN values that are of type PrintableString, as well as RDNs of type
5965 emailAddress where the value has the type ia5String.
5966 [stefank@valicert.com via Richard Levitte]
5968 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
5969 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
5970 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
5971 the bitwise-OR of the two for use by the majority of applications
5972 wanting this behaviour, and update the docs. The documented
5973 behaviour and actual behaviour were inconsistent and had been
5974 changing anyway, so this is more a bug-fix than a behavioural
5976 [Geoff Thorpe, diagnosed by Nadav Har'El]
5978 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
5979 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
5982 *) Fix initialization code race conditions in
5983 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
5984 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
5985 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
5986 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
5987 ssl2_get_cipher_by_char(),
5988 ssl3_get_cipher_by_char().
5989 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
5991 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
5992 the cached sessions are flushed, as the remove_cb() might use ex_data
5993 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
5994 (see [openssl.org #212]).
5995 [Geoff Thorpe, Lutz Jaenicke]
5997 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
5998 length, instead of the encoding length to d2i_ASN1_OBJECT.
6001 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
6003 *) [In 0.9.6g-engine release:]
6004 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
6005 [Lynn Gazis <lgazis@rainbow.com>]
6007 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
6009 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
6010 and get fix the header length calculation.
6011 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
6012 Alon Kantor <alonk@checkpoint.com> (and others),
6015 *) Use proper error handling instead of 'assertions' in buffer
6016 overflow checks added in 0.9.6e. This prevents DoS (the
6017 assertions could call abort()).
6018 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
6020 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
6022 *) Add various sanity checks to asn1_get_length() to reject
6023 the ASN1 length bytes if they exceed sizeof(long), will appear
6024 negative or the content length exceeds the length of the
6026 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6028 *) Fix cipher selection routines: ciphers without encryption had no flags
6029 for the cipher strength set and where therefore not handled correctly
6030 by the selection routines (PR #130).
6033 *) Fix EVP_dsa_sha macro.
6037 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
6038 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
6039 that was added in OpenSSL 0.9.6d.
6041 As the countermeasure turned out to be incompatible with some
6042 broken SSL implementations, the new option is part of SSL_OP_ALL.
6043 SSL_OP_ALL is usually employed when compatibility with weird SSL
6044 implementations is desired (e.g. '-bugs' option to 's_client' and
6045 's_server'), so the new option is automatically set in many
6049 *) Changes in security patch:
6051 Changes marked "(CHATS)" were sponsored by the Defense Advanced
6052 Research Projects Agency (DARPA) and Air Force Research Laboratory,
6053 Air Force Materiel Command, USAF, under agreement number
6056 *) Add various sanity checks to asn1_get_length() to reject
6057 the ASN1 length bytes if they exceed sizeof(long), will appear
6058 negative or the content length exceeds the length of the
6059 supplied buffer. (CVE-2002-0659)
6060 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6062 *) Assertions for various potential buffer overflows, not known to
6064 [Ben Laurie (CHATS)]
6066 *) Various temporary buffers to hold ASCII versions of integers were
6067 too small for 64 bit platforms. (CVE-2002-0655)
6068 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
6070 *) Remote buffer overflow in SSL3 protocol - an attacker could
6071 supply an oversized session ID to a client. (CVE-2002-0656)
6072 [Ben Laurie (CHATS)]
6074 *) Remote buffer overflow in SSL2 protocol - an attacker could
6075 supply an oversized client master key. (CVE-2002-0656)
6076 [Ben Laurie (CHATS)]
6078 Changes between 0.9.6c and 0.9.6d [9 May 2002]
6080 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
6081 encoded as NULL) with id-dsa-with-sha1.
6082 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
6084 *) Check various X509_...() return values in apps/req.c.
6085 [Nils Larsch <nla@trustcenter.de>]
6087 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
6088 an end-of-file condition would erronously be flagged, when the CRLF
6089 was just at the end of a processed block. The bug was discovered when
6090 processing data through a buffering memory BIO handing the data to a
6091 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
6092 <ptsekov@syntrex.com> and Nedelcho Stanev.
6095 *) Implement a countermeasure against a vulnerability recently found
6096 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
6097 before application data chunks to avoid the use of known IVs
6098 with data potentially chosen by the attacker.
6101 *) Fix length checks in ssl3_get_client_hello().
6104 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
6105 to prevent ssl3_read_internal() from incorrectly assuming that
6106 ssl3_read_bytes() found application data while handshake
6107 processing was enabled when in fact s->s3->in_read_app_data was
6108 merely automatically cleared during the initial handshake.
6109 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
6111 *) Fix object definitions for Private and Enterprise: they were not
6112 recognized in their shortname (=lowercase) representation. Extend
6113 obj_dat.pl to issue an error when using undefined keywords instead
6114 of silently ignoring the problem (Svenning Sorensen
6115 <sss@sss.dnsalias.net>).
6118 *) Fix DH_generate_parameters() so that it works for 'non-standard'
6119 generators, i.e. generators other than 2 and 5. (Previously, the
6120 code did not properly initialise the 'add' and 'rem' values to
6121 BN_generate_prime().)
6123 In the new general case, we do not insist that 'generator' is
6124 actually a primitive root: This requirement is rather pointless;
6125 a generator of the order-q subgroup is just as good, if not
6129 *) Map new X509 verification errors to alerts. Discovered and submitted by
6130 Tom Wu <tom@arcot.com>.
6133 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
6134 returning non-zero before the data has been completely received
6135 when using non-blocking I/O.
6136 [Bodo Moeller; problem pointed out by John Hughes]
6138 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
6139 [Ben Laurie, Lutz Jaenicke]
6141 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
6142 Yoram Zahavi <YoramZ@gilian.com>).
6145 *) Add information about CygWin 1.3 and on, and preserve proper
6146 configuration for the versions before that.
6147 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6149 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
6150 check whether we deal with a copy of a session and do not delete from
6151 the cache in this case. Problem reported by "Izhar Shoshani Levi"
6152 <izhar@checkpoint.com>.
6155 *) Do not store session data into the internal session cache, if it
6156 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
6157 flag is set). Proposed by Aslam <aslam@funk.com>.
6160 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
6164 *) [In 0.9.6d-engine release:]
6165 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6166 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6168 *) Add the configuration target linux-s390x.
6169 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
6171 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
6172 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
6173 variable as an indication that a ClientHello message has been
6174 received. As the flag value will be lost between multiple
6175 invocations of ssl3_accept when using non-blocking I/O, the
6176 function may not be aware that a handshake has actually taken
6177 place, thus preventing a new session from being added to the
6180 To avoid this problem, we now set s->new_session to 2 instead of
6181 using a local variable.
6182 [Lutz Jaenicke, Bodo Moeller]
6184 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
6185 if the SSL_R_LENGTH_MISMATCH error is detected.
6186 [Geoff Thorpe, Bodo Moeller]
6188 *) New 'shared_ldflag' column in Configure platform table.
6191 *) Fix EVP_CIPHER_mode macro.
6192 ["Dan S. Camper" <dan@bti.net>]
6194 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
6195 type, we must throw them away by setting rr->length to 0.
6196 [D P Chang <dpc@qualys.com>]
6198 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
6200 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
6201 <Dominikus.Scherkl@biodata.com>. (The previous implementation
6202 worked incorrectly for those cases where range = 10..._2 and
6203 3*range is two bits longer than range.)
6206 *) Only add signing time to PKCS7 structures if it is not already
6210 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
6211 OBJ_ld_ce should be OBJ_id_ce.
6212 Also some ip-pda OIDs in crypto/objects/objects.txt were
6213 incorrect (cf. RFC 3039).
6214 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
6216 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
6217 returns early because it has nothing to do.
6218 [Andy Schneider <andy.schneider@bjss.co.uk>]
6220 *) [In 0.9.6c-engine release:]
6221 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
6222 [Andy Schneider <andy.schneider@bjss.co.uk>]
6224 *) [In 0.9.6c-engine release:]
6225 Add support for Cryptographic Appliance's keyserver technology.
6226 (Use engine 'keyclient')
6227 [Cryptographic Appliances and Geoff Thorpe]
6229 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
6230 is called via tools/c89.sh because arguments have to be
6231 rearranged (all '-L' options must appear before the first object
6233 [Richard Shapiro <rshapiro@abinitio.com>]
6235 *) [In 0.9.6c-engine release:]
6236 Add support for Broadcom crypto accelerator cards, backported
6238 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
6240 *) [In 0.9.6c-engine release:]
6241 Add support for SureWare crypto accelerator cards from
6242 Baltimore Technologies. (Use engine 'sureware')
6243 [Baltimore Technologies and Mark Cox]
6245 *) [In 0.9.6c-engine release:]
6246 Add support for crypto accelerator cards from Accelerated
6247 Encryption Processing, www.aep.ie. (Use engine 'aep')
6248 [AEP Inc. and Mark Cox]
6250 *) Add a configuration entry for gcc on UnixWare.
6251 [Gary Benson <gbenson@redhat.com>]
6253 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
6254 messages are stored in a single piece (fixed-length part and
6255 variable-length part combined) and fix various bugs found on the way.
6258 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
6259 instead. BIO_gethostbyname() does not know what timeouts are
6260 appropriate, so entries would stay in cache even when they have
6262 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
6264 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
6265 faced with a pathologically small ClientHello fragment that does
6266 not contain client_version: Instead of aborting with an error,
6267 simply choose the highest available protocol version (i.e.,
6268 TLS 1.0 unless it is disabled). In practice, ClientHello
6269 messages are never sent like this, but this change gives us
6270 strictly correct behaviour at least for TLS.
6273 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
6274 never resets s->method to s->ctx->method when called from within
6275 one of the SSL handshake functions.
6276 [Bodo Moeller; problem pointed out by Niko Baric]
6278 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
6279 (sent using the client's version number) if client_version is
6280 smaller than the protocol version in use. Also change
6281 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
6282 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
6283 the client will at least see that alert.
6286 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
6290 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
6291 client receives HelloRequest while in a handshake.
6292 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
6294 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
6295 should end in 'break', not 'goto end' which circuments various
6296 cleanups done in state SSL_ST_OK. But session related stuff
6297 must be disabled for SSL_ST_OK in the case that we just sent a
6300 Also avoid some overhead by not calling ssl_init_wbio_buffer()
6301 before just sending a HelloRequest.
6302 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
6304 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
6305 reveal whether illegal block cipher padding was found or a MAC
6306 verification error occured. (Neither SSLerr() codes nor alerts
6307 are directly visible to potential attackers, but the information
6308 may leak via logfiles.)
6310 Similar changes are not required for the SSL 2.0 implementation
6311 because the number of padding bytes is sent in clear for SSL 2.0,
6312 and the extra bytes are just ignored. However ssl/s2_pkt.c
6313 failed to verify that the purported number of padding bytes is in
6317 *) Add OpenUNIX-8 support including shared libraries
6318 (Boyd Lynn Gerber <gerberb@zenez.com>).
6321 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
6322 'wristwatch attack' using huge encoding parameters (cf.
6323 James H. Manger's CRYPTO 2001 paper). Note that the
6324 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
6325 encoding parameters and hence was not vulnerable.
6328 *) BN_sqr() bug fix.
6329 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
6331 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
6332 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
6333 followed by modular reduction.
6334 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
6336 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
6337 equivalent based on BN_pseudo_rand() instead of BN_rand().
6340 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
6341 This function was broken, as the check for a new client hello message
6342 to handle SGC did not allow these large messages.
6343 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
6346 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
6349 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
6350 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
6353 *) Rework the configuration and shared library support for Tru64 Unix.
6354 The configuration part makes use of modern compiler features and
6355 still retains old compiler behavior for those that run older versions
6356 of the OS. The shared library support part includes a variant that
6357 uses the RPATH feature, and is available through the special
6358 configuration target "alpha-cc-rpath", which will never be selected
6360 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
6362 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
6363 with the same message size as in ssl3_get_certificate_request().
6364 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
6365 messages might inadvertently be reject as too long.
6366 [Petr Lampa <lampa@fee.vutbr.cz>]
6368 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
6371 *) Modified SSL library such that the verify_callback that has been set
6372 specificly for an SSL object with SSL_set_verify() is actually being
6373 used. Before the change, a verify_callback set with this function was
6374 ignored and the verify_callback() set in the SSL_CTX at the time of
6375 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
6376 to allow the necessary settings.
6379 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
6380 explicitly to NULL, as at least on Solaris 8 this seems not always to be
6381 done automatically (in contradiction to the requirements of the C
6382 standard). This made problems when used from OpenSSH.
6385 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
6386 dh->length and always used
6388 BN_rand_range(priv_key, dh->p).
6390 BN_rand_range() is not necessary for Diffie-Hellman, and this
6391 specific range makes Diffie-Hellman unnecessarily inefficient if
6392 dh->length (recommended exponent length) is much smaller than the
6393 length of dh->p. We could use BN_rand_range() if the order of
6394 the subgroup was stored in the DH structure, but we only have
6399 BN_rand(priv_key, l, ...)
6401 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
6407 RSA_eay_public_encrypt
6408 RSA_eay_private_decrypt
6409 RSA_eay_private_encrypt (signing)
6410 RSA_eay_public_decrypt (signature verification)
6412 (default implementations for RSA_public_encrypt,
6413 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
6414 always reject numbers >= n.
6417 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
6418 to synchronize access to 'locking_thread'. This is necessary on
6419 systems where access to 'locking_thread' (an 'unsigned long'
6420 variable) is not atomic.
6423 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
6424 *before* setting the 'crypto_lock_rand' flag. The previous code had
6425 a race condition if 0 is a valid thread ID.
6426 [Travis Vitek <vitek@roguewave.com>]
6428 *) Add support for shared libraries under Irix.
6429 [Albert Chin-A-Young <china@thewrittenword.com>]
6431 *) Add configuration option to build on Linux on both big-endian and
6433 [Ralf Baechle <ralf@uni-koblenz.de>]
6435 *) Add the possibility to create shared libraries on HP-UX.
6438 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
6440 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
6441 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
6442 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
6443 PRNG state recovery was possible based on the output of
6444 one PRNG request appropriately sized to gain knowledge on
6445 'md' followed by enough consecutive 1-byte PRNG requests
6446 to traverse all of 'state'.
6448 1. When updating 'md_local' (the current thread's copy of 'md')
6449 during PRNG output generation, hash all of the previous
6450 'md_local' value, not just the half used for PRNG output.
6452 2. Make the number of bytes from 'state' included into the hash
6453 independent from the number of PRNG bytes requested.
6455 The first measure alone would be sufficient to avoid
6456 Markku-Juhani's attack. (Actually it had never occurred
6457 to me that the half of 'md_local' used for chaining was the
6458 half from which PRNG output bytes were taken -- I had always
6459 assumed that the secret half would be used.) The second
6460 measure makes sure that additional data from 'state' is never
6461 mixed into 'md_local' in small portions; this heuristically
6462 further strengthens the PRNG.
6465 *) Fix crypto/bn/asm/mips3.s.
6468 *) When only the key is given to "enc", the IV is undefined. Print out
6469 an error message in this case.
6472 *) Handle special case when X509_NAME is empty in X509 printing routines.
6475 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
6476 positive and less than q.
6479 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
6480 used: it isn't thread safe and the add_lock_callback should handle
6482 [Paul Rose <Paul.Rose@bridge.com>]
6484 *) Verify that incoming data obeys the block size in
6485 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
6489 [Ulf Möller, Bodo Möller]
6491 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
6492 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
6493 when fixing the server behaviour for backwards-compatible 'client
6494 hello' messages. (Note that the attack is impractical against
6495 SSL 3.0 and TLS 1.0 anyway because length and version checking
6496 means that the probability of guessing a valid ciphertext is
6497 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
6500 Before 0.9.5, the countermeasure (hide the error by generating a
6501 random 'decryption result') did not work properly because
6502 ERR_clear_error() was missing, meaning that SSL_get_error() would
6503 detect the supposedly ignored error.
6505 Both problems are now fixed.
6508 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
6509 (previously it was 1024).
6512 *) Fix for compatibility mode trust settings: ignore trust settings
6513 unless some valid trust or reject settings are present.
6516 *) Fix for blowfish EVP: its a variable length cipher.
6519 *) Fix various bugs related to DSA S/MIME verification. Handle missing
6520 parameters in DSA public key structures and return an error in the
6521 DSA routines if parameters are absent.
6524 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
6525 in the current directory if neither $RANDFILE nor $HOME was set.
6526 RAND_file_name() in 0.9.6a returned NULL in this case. This has
6527 caused some confusion to Windows users who haven't defined $HOME.
6528 Thus RAND_file_name() is changed again: e_os.h can define a
6529 DEFAULT_HOME, which will be used if $HOME is not set.
6530 For Windows, we use "C:"; on other platforms, we still require
6531 environment variables.
6533 *) Move 'if (!initialized) RAND_poll()' into regions protected by
6534 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
6535 having multiple threads call RAND_poll() concurrently.
6538 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
6539 combination of a flag and a thread ID variable.
6540 Otherwise while one thread is in ssleay_rand_bytes (which sets the
6541 flag), *other* threads can enter ssleay_add_bytes without obeying
6542 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
6543 that they do not hold after the first thread unsets add_do_not_lock).
6546 *) Change bctest again: '-x' expressions are not available in all
6550 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
6552 *) Fix a couple of memory leaks in PKCS7_dataDecode()
6553 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
6555 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
6556 the default extension for executables, if any. Also, make the perl
6557 scripts that use symlink() to test if it really exists and use "cp"
6558 if it doesn't. All this made OpenSSL compilable and installable in
6562 *) Fix for asn1_GetSequence() for indefinite length constructed data.
6563 If SEQUENCE is length is indefinite just set c->slen to the total
6564 amount of data available.
6565 [Steve Henson, reported by shige@FreeBSD.org]
6566 [This change does not apply to 0.9.7.]
6568 *) Change bctest to avoid here-documents inside command substitution
6569 (workaround for FreeBSD /bin/sh bug).
6570 For compatibility with Ultrix, avoid shell functions (introduced
6571 in the bctest version that searches along $PATH).
6574 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6575 with des_encrypt() defined on some operating systems, like Solaris
6579 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6580 On the Importance of Eliminating Errors in Cryptographic
6581 Computations, J. Cryptology 14 (2001) 2, 101-119,
6582 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6585 *) MIPS assembler BIGNUM division bug fix.
6588 *) Disabled incorrect Alpha assembler code.
6591 *) Fix PKCS#7 decode routines so they correctly update the length
6592 after reading an EOC for the EXPLICIT tag.
6594 [This change does not apply to 0.9.7.]
6596 *) Fix bug in PKCS#12 key generation routines. This was triggered
6597 if a 3DES key was generated with a 0 initial byte. Include
6598 PKCS12_BROKEN_KEYGEN compilation option to retain the old
6599 (but broken) behaviour.
6602 *) Enhance bctest to search for a working bc along $PATH and print
6604 [Tim Rice <tim@multitalents.net> via Richard Levitte]
6606 *) Fix memory leaks in err.c: free err_data string if necessary;
6607 don't write to the wrong index in ERR_set_error_data.
6610 *) Implement ssl23_peek (analogous to ssl23_read), which previously
6614 *) Replace rdtsc with _emit statements for VC++ version 5.
6615 [Jeremy Cooper <jeremy@baymoo.org>]
6617 *) Make it possible to reuse SSLv2 sessions.
6620 *) In copy_email() check for >= 0 as a return value for
6621 X509_NAME_get_index_by_NID() since 0 is a valid index.
6622 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
6624 *) Avoid coredump with unsupported or invalid public keys by checking if
6625 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
6626 PKCS7_verify() fails with non detached data.
6629 *) Don't use getenv in library functions when run as setuid/setgid.
6630 New function OPENSSL_issetugid().
6633 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
6634 due to incorrect handling of multi-threading:
6636 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
6638 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
6640 3. Count how many times MemCheck_off() has been called so that
6641 nested use can be treated correctly. This also avoids
6642 inband-signalling in the previous code (which relied on the
6643 assumption that thread ID 0 is impossible).
6646 *) Add "-rand" option also to s_client and s_server.
6649 *) Fix CPU detection on Irix 6.x.
6650 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
6651 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6653 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
6656 [This change does not apply to 0.9.7.]
6658 *) Use the cached encoding of an X509_NAME structure rather than
6659 copying it. This is apparently the reason for the libsafe "errors"
6660 but the code is actually correct.
6663 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
6664 Bleichenbacher's DSA attack.
6665 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
6666 to be set and top=0 forces the highest bit to be set; top=-1 is new
6667 and leaves the highest bit random.
6668 [Ulf Moeller, Bodo Moeller]
6670 *) In the NCONF_...-based implementations for CONF_... queries
6671 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
6672 a temporary CONF structure with the data component set to NULL
6673 (which gives segmentation faults in lh_retrieve).
6674 Instead, use NULL for the CONF pointer in CONF_get_string and
6675 CONF_get_number (which may use environment variables) and directly
6676 return NULL from CONF_get_section.
6679 *) Fix potential buffer overrun for EBCDIC.
6682 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
6683 keyUsage if basicConstraints absent for a CA.
6686 *) Make SMIME_write_PKCS7() write mail header values with a format that
6687 is more generally accepted (no spaces before the semicolon), since
6688 some programs can't parse those values properly otherwise. Also make
6689 sure BIO's that break lines after each write do not create invalid
6693 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
6694 macros previously used would not encode an empty SEQUENCE OF
6695 and break the signature.
6697 [This change does not apply to 0.9.7.]
6699 *) Zero the premaster secret after deriving the master secret in
6703 *) Add some EVP_add_digest_alias registrations (as found in
6704 OpenSSL_add_all_digests()) to SSL_library_init()
6705 aka OpenSSL_add_ssl_algorithms(). This provides improved
6706 compatibility with peers using X.509 certificates
6707 with unconventional AlgorithmIdentifier OIDs.
6710 *) Fix for Irix with NO_ASM.
6711 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6713 *) ./config script fixes.
6714 [Ulf Moeller, Richard Levitte]
6716 *) Fix 'openssl passwd -1'.
6719 *) Change PKCS12_key_gen_asc() so it can cope with non null
6720 terminated strings whose length is passed in the passlen
6721 parameter, for example from PEM callbacks. This was done
6722 by adding an extra length parameter to asc2uni().
6723 [Steve Henson, reported by <oddissey@samsung.co.kr>]
6725 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
6726 call failed, free the DSA structure.
6729 *) Fix to uni2asc() to cope with zero length Unicode strings.
6730 These are present in some PKCS#12 files.
6733 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
6734 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
6735 when writing a 32767 byte record.
6736 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
6738 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
6739 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
6741 (RSA objects have a reference count access to which is protected
6742 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
6743 so they are meant to be shared between threads.)
6744 [Bodo Moeller, Geoff Thorpe; original patch submitted by
6745 "Reddie, Steven" <Steven.Reddie@ca.com>]
6747 *) Fix a deadlock in CRYPTO_mem_leaks().
6750 *) Use better test patterns in bntest.
6753 *) rand_win.c fix for Borland C.
6756 *) BN_rshift bugfix for n == 0.
6759 *) Add a 'bctest' script that checks for some known 'bc' bugs
6760 so that 'make test' does not abort just because 'bc' is broken.
6763 *) Store verify_result within SSL_SESSION also for client side to
6764 avoid potential security hole. (Re-used sessions on the client side
6765 always resulted in verify_result==X509_V_OK, not using the original
6766 result of the server certificate verification.)
6769 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
6770 SSL3_RT_APPLICATION_DATA, return 0.
6771 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
6775 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
6776 releases, have been re-implemented by renaming the previous
6777 implementations of ssl2_read and ssl3_read to ssl2_read_internal
6778 and ssl3_read_internal, respectively, and adding 'peek' parameters
6779 to them. The new ssl[23]_{read,peek} functions are calls to
6780 ssl[23]_read_internal with the 'peek' flag set appropriately.
6781 A 'peek' parameter has also been added to ssl3_read_bytes, which
6782 does the actual work for ssl3_read_internal.
6785 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
6786 the method-specific "init()" handler. Also clean up ex_data after
6787 calling the method-specific "finish()" handler. Previously, this was
6788 happening the other way round.
6791 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
6792 The previous value, 12, was not always sufficient for BN_mod_exp().
6795 *) Make sure that shared libraries get the internal name engine with
6796 the full version number and not just 0. This should mark the
6797 shared libraries as not backward compatible. Of course, this should
6798 be changed again when we can guarantee backward binary compatibility.
6801 *) Fix typo in get_cert_by_subject() in by_dir.c
6802 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
6804 *) Rework the system to generate shared libraries:
6806 - Make note of the expected extension for the shared libraries and
6807 if there is a need for symbolic links from for example libcrypto.so.0
6808 to libcrypto.so.0.9.7. There is extended info in Configure for
6811 - Make as few rebuilds of the shared libraries as possible.
6813 - Still avoid linking the OpenSSL programs with the shared libraries.
6815 - When installing, install the shared libraries separately from the
6819 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
6821 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
6822 and not in SSL_clear because the latter is also used by the
6823 accept/connect functions; previously, the settings made by
6824 SSL_set_read_ahead would be lost during the handshake.
6825 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
6827 *) Correct util/mkdef.pl to be selective about disabled algorithms.
6828 Previously, it would create entries for disableed algorithms no
6832 *) Added several new manual pages for SSL_* function.
6835 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
6837 *) In ssl23_get_client_hello, generate an error message when faced
6838 with an initial SSL 3.0/TLS record that is too small to contain the
6839 first two bytes of the ClientHello message, i.e. client_version.
6840 (Note that this is a pathologic case that probably has never happened
6841 in real life.) The previous approach was to use the version number
6842 from the record header as a substitute; but our protocol choice
6843 should not depend on that one because it is not authenticated
6844 by the Finished messages.
6847 *) More robust randomness gathering functions for Windows.
6848 [Jeffrey Altman <jaltman@columbia.edu>]
6850 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
6851 not set then we don't setup the error code for issuer check errors
6852 to avoid possibly overwriting other errors which the callback does
6853 handle. If an application does set the flag then we assume it knows
6854 what it is doing and can handle the new informational codes
6858 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
6859 a general "ANY" type, as such it should be able to decode anything
6860 including tagged types. However it didn't check the class so it would
6861 wrongly interpret tagged types in the same way as their universal
6862 counterpart and unknown types were just rejected. Changed so that the
6863 tagged and unknown types are handled in the same way as a SEQUENCE:
6864 that is the encoding is stored intact. There is also a new type
6865 "V_ASN1_OTHER" which is used when the class is not universal, in this
6866 case we have no idea what the actual type is so we just lump them all
6870 *) On VMS, stdout may very well lead to a file that is written to
6871 in a record-oriented fashion. That means that every write() will
6872 write a separate record, which will be read separately by the
6873 programs trying to read from it. This can be very confusing.
6875 The solution is to put a BIO filter in the way that will buffer
6876 text until a linefeed is reached, and then write everything a
6877 line at a time, so every record written will be an actual line,
6878 not chunks of lines and not (usually doesn't happen, but I've
6879 seen it once) several lines in one record. BIO_f_linebuffer() is
6882 Currently, it's a VMS-only method, because that's where it has
6883 been tested well enough.
6886 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
6887 it can return incorrect results.
6888 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
6889 but it was in 0.9.6-beta[12].)
6892 *) Disable the check for content being present when verifying detached
6893 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
6894 include zero length content when signing messages.
6897 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
6898 BIO_ctrl (for BIO pairs).
6901 *) Add DSO method for VMS.
6904 *) Bug fix: Montgomery multiplication could produce results with the
6908 *) Add RPM specification openssl.spec and modify it to build three
6909 packages. The default package contains applications, application
6910 documentation and run-time libraries. The devel package contains
6911 include files, static libraries and function documentation. The
6912 doc package contains the contents of the doc directory. The original
6913 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
6916 *) Add a large number of documentation files for many SSL routines.
6917 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6919 *) Add a configuration entry for Sony News 4.
6920 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
6922 *) Don't set the two most significant bits to one when generating a
6923 random number < q in the DSA library.
6926 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
6927 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
6928 the underlying transport is blocking) if a handshake took place.
6929 (The default behaviour is needed by applications such as s_client
6930 and s_server that use select() to determine when to use SSL_read;
6931 but for applications that know in advance when to expect data, it
6932 just makes things more complicated.)
6935 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
6939 *) Add a few more EBCDIC conditionals that make `req' and `x509'
6940 work better on such systems.
6941 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6943 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
6944 Update PKCS12_parse() so it copies the friendlyName and the
6945 keyid to the certificates aux info.
6948 *) Fix bug in PKCS7_verify() which caused an infinite loop
6949 if there was more than one signature.
6950 [Sven Uszpelkat <su@celocom.de>]
6952 *) Major change in util/mkdef.pl to include extra information
6953 about each symbol, as well as presentig variables as well
6954 as functions. This change means that there's n more need
6955 to rebuild the .num files when some algorithms are excluded.
6958 *) Allow the verify time to be set by an application,
6959 rather than always using the current time.
6962 *) Phase 2 verify code reorganisation. The certificate
6963 verify code now looks up an issuer certificate by a
6964 number of criteria: subject name, authority key id
6965 and key usage. It also verifies self signed certificates
6966 by the same criteria. The main comparison function is
6967 X509_check_issued() which performs these checks.
6969 Lot of changes were necessary in order to support this
6970 without completely rewriting the lookup code.
6972 Authority and subject key identifier are now cached.
6974 The LHASH 'certs' is X509_STORE has now been replaced
6975 by a STACK_OF(X509_OBJECT). This is mainly because an
6976 LHASH can't store or retrieve multiple objects with
6977 the same hash value.
6979 As a result various functions (which were all internal
6980 use only) have changed to handle the new X509_STORE
6981 structure. This will break anything that messed round
6982 with X509_STORE internally.
6984 The functions X509_STORE_add_cert() now checks for an
6985 exact match, rather than just subject name.
6987 The X509_STORE API doesn't directly support the retrieval
6988 of multiple certificates matching a given criteria, however
6989 this can be worked round by performing a lookup first
6990 (which will fill the cache with candidate certificates)
6991 and then examining the cache for matches. This is probably
6992 the best we can do without throwing out X509_LOOKUP
6993 entirely (maybe later...).
6995 The X509_VERIFY_CTX structure has been enhanced considerably.
6997 All certificate lookup operations now go via a get_issuer()
6998 callback. Although this currently uses an X509_STORE it
6999 can be replaced by custom lookups. This is a simple way
7000 to bypass the X509_STORE hackery necessary to make this
7001 work and makes it possible to use more efficient techniques
7002 in future. A very simple version which uses a simple
7003 STACK for its trusted certificate store is also provided
7004 using X509_STORE_CTX_trusted_stack().
7006 The verify_cb() and verify() callbacks now have equivalents
7007 in the X509_STORE_CTX structure.
7009 X509_STORE_CTX also has a 'flags' field which can be used
7010 to customise the verify behaviour.
7013 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
7014 excludes S/MIME capabilities.
7017 *) When a certificate request is read in keep a copy of the
7018 original encoding of the signed data and use it when outputing
7019 again. Signatures then use the original encoding rather than
7020 a decoded, encoded version which may cause problems if the
7021 request is improperly encoded.
7024 *) For consistency with other BIO_puts implementations, call
7025 buffer_write(b, ...) directly in buffer_puts instead of calling
7028 In BIO_puts, increment b->num_write as in BIO_write.
7029 [Peter.Sylvester@EdelWeb.fr]
7031 *) Fix BN_mul_word for the case where the word is 0. (We have to use
7032 BN_zero, we may not return a BIGNUM with an array consisting of
7036 *) Avoid calling abort() from within the library when problems are
7037 detected, except if preprocessor symbols have been defined
7038 (such as REF_CHECK, BN_DEBUG etc.).
7041 *) New openssl application 'rsautl'. This utility can be
7042 used for low level RSA operations. DER public key
7043 BIO/fp routines also added.
7046 *) New Configure entry and patches for compiling on QNX 4.
7047 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
7049 *) A demo state-machine implementation was sponsored by
7050 Nuron (http://www.nuron.com/) and is now available in
7051 demos/state_machine.
7054 *) New options added to the 'dgst' utility for signature
7055 generation and verification.
7058 *) Unrecognized PKCS#7 content types are now handled via a
7059 catch all ASN1_TYPE structure. This allows unsupported
7060 types to be stored as a "blob" and an application can
7061 encode and decode it manually.
7064 *) Fix various signed/unsigned issues to make a_strex.c
7066 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
7068 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
7069 length if passed a buffer. ASN1_INTEGER_to_BN failed
7070 if passed a NULL BN and its argument was negative.
7071 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
7073 *) Modification to PKCS#7 encoding routines to output definite
7074 length encoding. Since currently the whole structures are in
7075 memory there's not real point in using indefinite length
7076 constructed encoding. However if OpenSSL is compiled with
7077 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
7080 *) Added BIO_vprintf() and BIO_vsnprintf().
7083 *) Added more prefixes to parse for in the the strings written
7084 through a logging bio, to cover all the levels that are available
7085 through syslog. The prefixes are now:
7087 PANIC, EMERG, EMR => LOG_EMERG
7088 ALERT, ALR => LOG_ALERT
7089 CRIT, CRI => LOG_CRIT
7090 ERROR, ERR => LOG_ERR
7091 WARNING, WARN, WAR => LOG_WARNING
7092 NOTICE, NOTE, NOT => LOG_NOTICE
7093 INFO, INF => LOG_INFO
7094 DEBUG, DBG => LOG_DEBUG
7096 and as before, if none of those prefixes are present at the
7097 beginning of the string, LOG_ERR is chosen.
7099 On Win32, the LOG_* levels are mapped according to this:
7101 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
7102 LOG_WARNING => EVENTLOG_WARNING_TYPE
7103 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
7107 *) Made it possible to reconfigure with just the configuration
7108 argument "reconf" or "reconfigure". The command line arguments
7109 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
7110 and are retrieved from there when reconfiguring.
7114 [Assar Westerlund <assar@sics.se>, Richard Levitte]
7116 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
7119 *) The obj_dat.pl script was messing up the sorting of object
7120 names. The reason was that it compared the quoted version
7121 of strings as a result "OCSP" > "OCSP Signing" because
7122 " > SPACE. Changed script to store unquoted versions of
7123 names and add quotes on output. It was also omitting some
7124 names from the lookup table if they were given a default
7125 value (that is if SN is missing it is given the same
7126 value as LN and vice versa), these are now added on the
7127 grounds that if an object has a name we should be able to
7128 look it up. Finally added warning output when duplicate
7129 short or long names are found.
7132 *) Changes needed for Tandem NSK.
7133 [Scott Uroff <scott@xypro.com>]
7135 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
7136 RSA_padding_check_SSLv23(), special padding was never detected
7137 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
7138 version rollback attacks was not effective.
7140 In s23_clnt.c, don't use special rollback-attack detection padding
7141 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
7142 client; similarly, in s23_srvr.c, don't do the rollback check if
7143 SSL 2.0 is the only protocol enabled in the server.
7146 *) Make it possible to get hexdumps of unprintable data with 'openssl
7147 asn1parse'. By implication, the functions ASN1_parse_dump() and
7148 BIO_dump_indent() are added.
7151 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
7152 these print out strings and name structures based on various
7153 flags including RFC2253 support and proper handling of
7154 multibyte characters. Added options to the 'x509' utility
7155 to allow the various flags to be set.
7158 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
7159 Also change the functions X509_cmp_current_time() and
7160 X509_gmtime_adj() work with an ASN1_TIME structure,
7161 this will enable certificates using GeneralizedTime in validity
7162 dates to be checked.
7165 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
7166 negative public key encodings) on by default,
7167 NO_NEG_PUBKEY_BUG can be set to disable it.
7170 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
7171 content octets. An i2c_ASN1_OBJECT is unnecessary because
7172 the encoding can be trivially obtained from the structure.
7175 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
7176 not read locks (CRYPTO_r_[un]lock).
7179 *) A first attempt at creating official support for shared
7180 libraries through configuration. I've kept it so the
7181 default is static libraries only, and the OpenSSL programs
7182 are always statically linked for now, but there are
7183 preparations for dynamic linking in place.
7184 This has been tested on Linux and Tru64.
7187 *) Randomness polling function for Win9x, as described in:
7188 Peter Gutmann, Software Generation of Practically Strong
7192 *) Fix so PRNG is seeded in req if using an already existing
7196 *) New options to smime application. -inform and -outform
7197 allow alternative formats for the S/MIME message including
7198 PEM and DER. The -content option allows the content to be
7199 specified separately. This should allow things like Netscape
7200 form signing output easier to verify.
7203 *) Fix the ASN1 encoding of tags using the 'long form'.
7206 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
7207 STRING types. These convert content octets to and from the
7208 underlying type. The actual tag and length octets are
7209 already assumed to have been read in and checked. These
7210 are needed because all other string types have virtually
7211 identical handling apart from the tag. By having versions
7212 of the ASN1 functions that just operate on content octets
7213 IMPLICIT tagging can be handled properly. It also allows
7214 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
7215 and ASN1_INTEGER are identical apart from the tag.
7218 *) Change the handling of OID objects as follows:
7220 - New object identifiers are inserted in objects.txt, following
7221 the syntax given in objects.README.
7222 - objects.pl is used to process obj_mac.num and create a new
7224 - obj_dat.pl is used to create a new obj_dat.h, using the data in
7227 This is currently kind of a hack, and the perl code in objects.pl
7228 isn't very elegant, but it works as I intended. The simplest way
7229 to check that it worked correctly is to look in obj_dat.h and
7230 check the array nid_objs and make sure the objects haven't moved
7231 around (this is important!). Additions are OK, as well as
7232 consistent name changes.
7235 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
7238 *) Addition of the command line parameter '-rand file' to 'openssl req'.
7239 The given file adds to whatever has already been seeded into the
7240 random pool through the RANDFILE configuration file option or
7241 environment variable, or the default random state file.
7244 *) mkstack.pl now sorts each macro group into lexical order.
7245 Previously the output order depended on the order the files
7246 appeared in the directory, resulting in needless rewriting
7250 *) Patches to make OpenSSL compile under Win32 again. Mostly
7251 work arounds for the VC++ problem that it treats func() as
7252 func(void). Also stripped out the parts of mkdef.pl that
7253 added extra typesafe functions: these no longer exist.
7256 *) Reorganisation of the stack code. The macros are now all
7257 collected in safestack.h . Each macro is defined in terms of
7258 a "stack macro" of the form SKM_<name>(type, a, b). The
7259 DEBUG_SAFESTACK is now handled in terms of function casts,
7260 this has the advantage of retaining type safety without the
7261 use of additional functions. If DEBUG_SAFESTACK is not defined
7262 then the non typesafe macros are used instead. Also modified the
7263 mkstack.pl script to handle the new form. Needs testing to see
7264 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
7265 the default if no major problems. Similar behaviour for ASN1_SET_OF
7266 and PKCS12_STACK_OF.
7269 *) When some versions of IIS use the 'NET' form of private key the
7270 key derivation algorithm is different. Normally MD5(password) is
7271 used as a 128 bit RC4 key. In the modified case
7272 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
7273 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
7274 as the old Netscape_RSA functions except they have an additional
7275 'sgckey' parameter which uses the modified algorithm. Also added
7276 an -sgckey command line option to the rsa utility. Thanks to
7277 Adrian Peck <bertie@ncipher.com> for posting details of the modified
7278 algorithm to openssl-dev.
7281 *) The evp_local.h macros were using 'c.##kname' which resulted in
7282 invalid expansion on some systems (SCO 5.0.5 for example).
7283 Corrected to 'c.kname'.
7284 [Phillip Porch <root@theporch.com>]
7286 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
7287 a STACK of email addresses from a certificate or request, these look
7288 in the subject name and the subject alternative name extensions and
7289 omit any duplicate addresses.
7292 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
7293 This makes DSA verification about 2 % faster.
7296 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
7297 (meaning that now 2^5 values will be precomputed, which is only 4 KB
7298 plus overhead for 1024 bit moduli).
7299 This makes exponentiations about 0.5 % faster for 1024 bit
7300 exponents (as measured by "openssl speed rsa2048").
7303 *) Rename memory handling macros to avoid conflicts with other
7305 Malloc => OPENSSL_malloc
7306 Malloc_locked => OPENSSL_malloc_locked
7307 Realloc => OPENSSL_realloc
7308 Free => OPENSSL_free
7311 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
7312 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
7315 *) CygWin32 support.
7316 [John Jarvie <jjarvie@newsguy.com>]
7318 *) The type-safe stack code has been rejigged. It is now only compiled
7319 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
7320 by default all type-specific stack functions are "#define"d back to
7321 standard stack functions. This results in more streamlined output
7322 but retains the type-safety checking possibilities of the original
7326 *) The STACK code has been cleaned up, and certain type declarations
7327 that didn't make a lot of sense have been brought in line. This has
7328 also involved a cleanup of sorts in safestack.h to more correctly
7329 map type-safe stack functions onto their plain stack counterparts.
7330 This work has also resulted in a variety of "const"ifications of
7331 lots of the code, especially "_cmp" operations which should normally
7332 be prototyped with "const" parameters anyway.
7335 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
7336 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
7337 (The PRNG state consists of two parts, the large pool 'state' and 'md',
7338 where all of 'md' is used each time the PRNG is used, but 'state'
7339 is used only indexed by a cyclic counter. As entropy may not be
7340 well distributed from the beginning, 'md' is important as a
7341 chaining variable. However, the output function chains only half
7342 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
7343 all of 'md', and seeding with STATE_SIZE dummy bytes will result
7344 in all of 'state' being rewritten, with the new values depending
7345 on virtually all of 'md'. This overcomes the 80 bit limitation.)
7348 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
7349 the handshake is continued after ssl_verify_cert_chain();
7350 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
7351 can lead to 'unexplainable' connection aborts later.
7352 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
7354 *) Major EVP API cipher revision.
7355 Add hooks for extra EVP features. This allows various cipher
7356 parameters to be set in the EVP interface. Support added for variable
7357 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
7358 setting of RC2 and RC5 parameters.
7360 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
7363 Remove lots of duplicated code from the EVP library. For example *every*
7364 cipher init() function handles the 'iv' in the same way according to the
7365 cipher mode. They also all do nothing if the 'key' parameter is NULL and
7366 for CFB and OFB modes they zero ctx->num.
7368 New functionality allows removal of S/MIME code RC2 hack.
7370 Most of the routines have the same form and so can be declared in terms
7373 By shifting this to the top level EVP_CipherInit() it can be removed from
7374 all individual ciphers. If the cipher wants to handle IVs or keys
7375 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
7378 Change lots of functions like EVP_EncryptUpdate() to now return a
7379 value: although software versions of the algorithms cannot fail
7380 any installed hardware versions can.
7383 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
7384 this option is set, tolerate broken clients that send the negotiated
7385 protocol version number instead of the requested protocol version
7389 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
7390 i.e. non-zero for export ciphersuites, zero otherwise.
7391 Previous versions had this flag inverted, inconsistent with
7392 rsa_tmp_cb (..._TMP_RSA_CB).
7393 [Bodo Moeller; problem reported by Amit Chopra]
7395 *) Add missing DSA library text string. Work around for some IIS
7396 key files with invalid SEQUENCE encoding.
7399 *) Add a document (doc/standards.txt) that list all kinds of standards
7400 and so on that are implemented in OpenSSL.
7403 *) Enhance c_rehash script. Old version would mishandle certificates
7404 with the same subject name hash and wouldn't handle CRLs at all.
7405 Added -fingerprint option to crl utility, to support new c_rehash
7409 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
7412 *) Fix for SSL server purpose checking. Server checking was
7413 rejecting certificates which had extended key usage present
7414 but no ssl client purpose.
7415 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
7417 *) Make PKCS#12 code work with no password. The PKCS#12 spec
7418 is a little unclear about how a blank password is handled.
7419 Since the password in encoded as a BMPString with terminating
7420 double NULL a zero length password would end up as just the
7421 double NULL. However no password at all is different and is
7422 handled differently in the PKCS#12 key generation code. NS
7423 treats a blank password as zero length. MSIE treats it as no
7424 password on export: but it will try both on import. We now do
7425 the same: PKCS12_parse() tries zero length and no password if
7426 the password is set to "" or NULL (NULL is now a valid password:
7427 it wasn't before) as does the pkcs12 application.
7430 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
7431 perror when PEM_read_bio_X509_REQ fails, the error message must
7432 be obtained from the error queue.
7435 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
7436 it in ERR_remove_state if appropriate, and change ERR_get_state
7437 accordingly to avoid race conditions (this is necessary because
7438 thread_hash is no longer constant once set).
7441 *) Bugfix for linux-elf makefile.one.
7444 *) RSA_get_default_method() will now cause a default
7445 RSA_METHOD to be chosen if one doesn't exist already.
7446 Previously this was only set during a call to RSA_new()
7447 or RSA_new_method(NULL) meaning it was possible for
7448 RSA_get_default_method() to return NULL.
7451 *) Added native name translation to the existing DSO code
7452 that will convert (if the flag to do so is set) filenames
7453 that are sufficiently small and have no path information
7454 into a canonical native form. Eg. "blah" converted to
7455 "libblah.so" or "blah.dll" etc.
7458 *) New function ERR_error_string_n(e, buf, len) which is like
7459 ERR_error_string(e, buf), but writes at most 'len' bytes
7460 including the 0 terminator. For ERR_error_string_n, 'buf'
7462 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
7464 *) CONF library reworked to become more general. A new CONF
7465 configuration file reader "class" is implemented as well as a
7466 new functions (NCONF_*, for "New CONF") to handle it. The now
7467 old CONF_* functions are still there, but are reimplemented to
7468 work in terms of the new functions. Also, a set of functions
7469 to handle the internal storage of the configuration data is
7470 provided to make it easier to write new configuration file
7471 reader "classes" (I can definitely see something reading a
7472 configuration file in XML format, for example), called _CONF_*,
7473 or "the configuration storage API"...
7475 The new configuration file reading functions are:
7477 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
7478 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
7480 NCONF_default, NCONF_WIN32
7482 NCONF_dump_fp, NCONF_dump_bio
7484 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
7485 NCONF_new creates a new CONF object. This works in the same way
7486 as other interfaces in OpenSSL, like the BIO interface.
7487 NCONF_dump_* dump the internal storage of the configuration file,
7488 which is useful for debugging. All other functions take the same
7489 arguments as the old CONF_* functions wth the exception of the
7490 first that must be a `CONF *' instead of a `LHASH *'.
7492 To make it easer to use the new classes with the old CONF_* functions,
7493 the function CONF_set_default_method is provided.
7496 *) Add '-tls1' option to 'openssl ciphers', which was already
7497 mentioned in the documentation but had not been implemented.
7498 (This option is not yet really useful because even the additional
7499 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
7502 *) Initial DSO code added into libcrypto for letting OpenSSL (and
7503 OpenSSL-based applications) load shared libraries and bind to
7504 them in a portable way.
7505 [Geoff Thorpe, with contributions from Richard Levitte]
7507 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
7509 *) Make sure _lrotl and _lrotr are only used with MSVC.
7511 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
7512 (the default implementation of RAND_status).
7514 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
7515 to '-clrext' (= clear extensions), as intended and documented.
7516 [Bodo Moeller; inconsistency pointed out by Michael Attili
7519 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
7520 was larger than the MD block size.
7521 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
7523 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
7524 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
7525 using the passed key: if the passed key was a private key the result
7526 of X509_print(), for example, would be to print out all the private key
7530 *) des_quad_cksum() byte order bug fix.
7531 [Ulf Möller, using the problem description in krb4-0.9.7, where
7532 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
7534 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
7536 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
7538 *) For easily testing in shell scripts whether some command
7539 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
7540 returns with exit code 0 iff no command of the given name is available.
7541 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
7542 the output goes to stdout and nothing is printed to stderr.
7543 Additional arguments are always ignored.
7545 Since for each cipher there is a command of the same name,
7546 the 'no-cipher' compilation switches can be tested this way.
7548 ('openssl no-XXX' is not able to detect pseudo-commands such
7549 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
7552 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
7555 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
7556 is set; it will be thrown away anyway because each handshake creates
7558 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
7559 to parameters -- in previous versions (since OpenSSL 0.9.3) the
7560 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
7561 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7564 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7565 'Q' and 'R' lose their special meanings (quit/renegotiate).
7566 This is part of what -quiet does; unlike -quiet, -ign_eof
7567 does not suppress any output.
7570 *) Add compatibility options to the purpose and trust code. The
7571 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7572 accepts a certificate or CA, this was the previous behaviour,
7573 with all the associated security issues.
7575 X509_TRUST_COMPAT is the old trust behaviour: only and
7576 automatically trust self signed roots in certificate store. A
7577 new trust setting X509_TRUST_DEFAULT is used to specify that
7578 a purpose has no associated trust setting and it should instead
7579 use the value in the default purpose.
7582 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7583 and fix a memory leak.
7586 *) In util/mkerr.pl (which implements 'make errors'), preserve
7587 reason strings from the previous version of the .c file, as
7588 the default to have only downcase letters (and digits) in
7589 automatically generated reasons codes is not always appropriate.
7592 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7593 using strerror. Previously, ERR_reason_error_string() returned
7594 library names as reason strings for SYSerr; but SYSerr is a special
7595 case where small numbers are errno values, not library numbers.
7598 *) Add '-dsaparam' option to 'openssl dhparam' application. This
7599 converts DSA parameters into DH parameters. (When creating parameters,
7600 DSA_generate_parameters is used.)
7603 *) Include 'length' (recommended exponent length) in C code generated
7604 by 'openssl dhparam -C'.
7607 *) The second argument to set_label in perlasm was already being used
7608 so couldn't be used as a "file scope" flag. Moved to third argument
7612 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
7613 instead of RAND_bytes for encryption IVs and salts.
7616 *) Include RAND_status() into RAND_METHOD instead of implementing
7617 it only for md_rand.c Otherwise replacing the PRNG by calling
7618 RAND_set_rand_method would be impossible.
7621 *) Don't let DSA_generate_key() enter an infinite loop if the random
7622 number generation fails.
7625 *) New 'rand' application for creating pseudo-random output.
7628 *) Added configuration support for Linux/IA64
7629 [Rolf Haberrecker <rolf@suse.de>]
7631 *) Assembler module support for Mingw32.
7634 *) Shared library support for HPUX (in shlib/).
7635 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
7637 *) Shared library support for Solaris gcc.
7638 [Lutz Behnke <behnke@trustcenter.de>]
7640 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
7642 *) PKCS7_encrypt() was adding text MIME headers twice because they
7643 were added manually and by SMIME_crlf_copy().
7646 *) In bntest.c don't call BN_rand with zero bits argument.
7647 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
7649 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
7650 case was implemented. This caused BN_div_recp() to fail occasionally.
7653 *) Add an optional second argument to the set_label() in the perl
7654 assembly language builder. If this argument exists and is set
7655 to 1 it signals that the assembler should use a symbol whose
7656 scope is the entire file, not just the current function. This
7657 is needed with MASM which uses the format label:: for this scope.
7658 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
7660 *) Change the ASN1 types so they are typedefs by default. Before
7661 almost all types were #define'd to ASN1_STRING which was causing
7662 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
7666 *) Change names of new functions to the new get1/get0 naming
7667 convention: After 'get1', the caller owns a reference count
7668 and has to call ..._free; 'get0' returns a pointer to some
7669 data structure without incrementing reference counters.
7670 (Some of the existing 'get' functions increment a reference
7671 counter, some don't.)
7672 Similarly, 'set1' and 'add1' functions increase reference
7673 counters or duplicate objects.
7676 *) Allow for the possibility of temp RSA key generation failure:
7677 the code used to assume it always worked and crashed on failure.
7680 *) Fix potential buffer overrun problem in BIO_printf().
7681 [Ulf Möller, using public domain code by Patrick Powell; problem
7682 pointed out by David Sacerdote <das33@cornell.edu>]
7684 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
7685 RAND_egd() and RAND_status(). In the command line application,
7686 the EGD socket can be specified like a seed file using RANDFILE
7690 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
7691 Some CAs (e.g. Verisign) distribute certificates in this form.
7694 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
7695 list to exclude them. This means that no special compilation option
7696 is needed to use anonymous DH: it just needs to be included in the
7700 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
7701 EVP_MD_type. The old functionality is available in a new macro called
7702 EVP_MD_md(). Change code that uses it and update docs.
7705 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
7706 where the 'void *' argument is replaced by a function pointer argument.
7707 Previously 'void *' was abused to point to functions, which works on
7708 many platforms, but is not correct. As these functions are usually
7709 called by macros defined in OpenSSL header files, most source code
7710 should work without changes.
7713 *) <openssl/opensslconf.h> (which is created by Configure) now contains
7714 sections with information on -D... compiler switches used for
7715 compiling the library so that applications can see them. To enable
7716 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
7717 must be defined. E.g.,
7718 #define OPENSSL_ALGORITHM_DEFINES
7719 #include <openssl/opensslconf.h>
7720 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
7721 [Richard Levitte, Ulf and Bodo Möller]
7723 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
7727 *) Change the 'other' type in certificate aux info to a STACK_OF
7728 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
7729 the required ASN1 format: arbitrary types determined by an OID.
7732 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
7733 argument to 'req'. This is not because the function is newer or
7734 better than others it just uses the work 'NEW' in the certificate
7735 request header lines. Some software needs this.
7738 *) Reorganise password command line arguments: now passwords can be
7739 obtained from various sources. Delete the PEM_cb function and make
7740 it the default behaviour: i.e. if the callback is NULL and the
7741 usrdata argument is not NULL interpret it as a null terminated pass
7742 phrase. If usrdata and the callback are NULL then the pass phrase
7743 is prompted for as usual.
7746 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
7747 the support is automatically enabled. The resulting binaries will
7748 autodetect the card and use it if present.
7749 [Ben Laurie and Compaq Inc.]
7751 *) Work around for Netscape hang bug. This sends certificate request
7752 and server done in one record. Since this is perfectly legal in the
7753 SSL/TLS protocol it isn't a "bug" option and is on by default. See
7754 the bugs/SSLv3 entry for more info.
7757 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
7760 *) Add -rand argument to smime and pkcs12 applications and read/write
7764 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
7767 *) Add command line password options to the remaining applications.
7770 *) Bug fix for BN_div_recp() for numerators with an even number of
7774 *) More tests in bntest.c, and changed test_bn output.
7777 *) ./config recognizes MacOS X now.
7780 *) Bug fix for BN_div() when the first words of num and divsor are
7781 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
7784 *) Add support for various broken PKCS#8 formats, and command line
7785 options to produce them.
7788 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
7789 get temporary BIGNUMs from a BN_CTX.
7792 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
7796 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
7797 include a #define from the old name to the new. The original intent
7798 was that statically linked binaries could for example just call
7799 SSLeay_add_all_ciphers() to just add ciphers to the table and not
7800 link with digests. This never worked becayse SSLeay_add_all_digests()
7801 and SSLeay_add_all_ciphers() were in the same source file so calling
7802 one would link with the other. They are now in separate source files.
7805 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
7808 *) Use a less unusual form of the Miller-Rabin primality test (it used
7809 a binary algorithm for exponentiation integrated into the Miller-Rabin
7810 loop, our standard modexp algorithms are faster).
7813 *) Support for the EBCDIC character set completed.
7814 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
7816 *) Source code cleanups: use const where appropriate, eliminate casts,
7817 use void * instead of char * in lhash.
7820 *) Bugfix: ssl3_send_server_key_exchange was not restartable
7821 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
7822 this the server could overwrite ephemeral keys that the client
7826 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
7827 using 50 iterations of the Rabin-Miller test.
7829 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
7830 iterations of the Rabin-Miller test as required by the appendix
7831 to FIPS PUB 186[-1]) instead of DSA_is_prime.
7832 As BN_is_prime_fasttest includes trial division, DSA parameter
7833 generation becomes much faster.
7835 This implies a change for the callback functions in DSA_is_prime
7836 and DSA_generate_parameters: The callback function is called once
7837 for each positive witness in the Rabin-Miller test, not just
7838 occasionally in the inner loop; and the parameters to the
7839 callback function now provide an iteration count for the outer
7840 loop rather than for the current invocation of the inner loop.
7841 DSA_generate_parameters additionally can call the callback
7842 function with an 'iteration count' of -1, meaning that a
7843 candidate has passed the trial division test (when q is generated
7844 from an application-provided seed, trial division is skipped).
7847 *) New function BN_is_prime_fasttest that optionally does trial
7848 division before starting the Rabin-Miller test and has
7849 an additional BN_CTX * argument (whereas BN_is_prime always
7850 has to allocate at least one BN_CTX).
7851 'callback(1, -1, cb_arg)' is called when a number has passed the
7852 trial division stage.
7855 *) Fix for bug in CRL encoding. The validity dates weren't being handled
7859 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
7862 *) New function BN_pseudo_rand().
7865 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
7866 bignum version of BN_from_montgomery() with the working code from
7867 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
7871 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
7872 made it impossible to use the same SSL_SESSION data structure in
7873 SSL2 clients in multiple threads.
7876 *) The return value of RAND_load_file() no longer counts bytes obtained
7877 by stat(). RAND_load_file(..., -1) is new and uses the complete file
7878 to seed the PRNG (previously an explicit byte count was required).
7879 [Ulf Möller, Bodo Möller]
7881 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
7882 used (char *) instead of (void *) and had casts all over the place.
7885 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
7888 *) Retain source code compatibility for BN_prime_checks macro:
7889 BN_is_prime(..., BN_prime_checks, ...) now uses
7890 BN_prime_checks_for_size to determine the appropriate number of
7891 Rabin-Miller iterations.
7894 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
7895 DH_CHECK_P_NOT_SAFE_PRIME.
7896 (Check if this is true? OpenPGP calls them "strong".)
7899 *) Merge the functionality of "dh" and "gendh" programs into a new program
7900 "dhparam". The old programs are retained for now but will handle DH keys
7901 (instead of parameters) in future.
7904 *) Make the ciphers, s_server and s_client programs check the return values
7905 when a new cipher list is set.
7908 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
7909 ciphers. Before when the 56bit ciphers were enabled the sorting was
7912 The syntax for the cipher sorting has been extended to support sorting by
7913 cipher-strength (using the strength_bits hard coded in the tables).
7914 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
7916 Fix a bug in the cipher-command parser: when supplying a cipher command
7917 string with an "undefined" symbol (neither command nor alphanumeric
7918 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
7919 an error is flagged.
7921 Due to the strength-sorting extension, the code of the
7922 ssl_create_cipher_list() function was completely rearranged. I hope that
7923 the readability was also increased :-)
7924 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7926 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
7927 for the first serial number and places 2 in the serial number file. This
7928 avoids problems when the root CA is created with serial number zero and
7929 the first user certificate has the same issuer name and serial number
7933 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
7934 the new code. Add documentation for this stuff.
7937 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
7938 X509_*() to X509at_*() on the grounds that they don't handle X509
7939 structures and behave in an analagous way to the X509v3 functions:
7940 they shouldn't be called directly but wrapper functions should be used
7943 So we also now have some wrapper functions that call the X509at functions
7944 when passed certificate requests. (TO DO: similar things can be done with
7945 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
7946 things. Some of these need some d2i or i2d and print functionality
7947 because they handle more complex structures.)
7950 *) Add missing #ifndefs that caused missing symbols when building libssl
7951 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
7952 NO_RSA in ssl/s2*.c.
7953 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
7955 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
7956 has a return value which indicates the quality of the random data
7957 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
7958 error queue. New function RAND_pseudo_bytes() generates output that is
7959 guaranteed to be unique but not unpredictable. RAND_add is like
7960 RAND_seed, but takes an extra argument for an entropy estimate
7961 (RAND_seed always assumes full entropy).
7964 *) Do more iterations of Rabin-Miller probable prime test (specifically,
7965 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
7966 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
7967 in crypto/bn/bn_prime.c for the complete table). This guarantees a
7968 false-positive rate of at most 2^-80 for random input.
7971 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
7974 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
7975 in the 0.9.5 release), this returns the chain
7976 from an X509_CTX structure with a dup of the stack and all
7977 the X509 reference counts upped: so the stack will exist
7978 after X509_CTX_cleanup() has been called. Modify pkcs12.c
7981 Also make SSL_SESSION_print() print out the verify return
7985 *) Add manpage for the pkcs12 command. Also change the default
7986 behaviour so MAC iteration counts are used unless the new
7987 -nomaciter option is used. This improves file security and
7988 only older versions of MSIE (4.0 for example) need it.
7991 *) Honor the no-xxx Configure options when creating .DEF files.
7994 *) Add PKCS#10 attributes to field table: challengePassword,
7995 unstructuredName and unstructuredAddress. These are taken from
7996 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
7997 international characters are used.
7999 More changes to X509_ATTRIBUTE code: allow the setting of types
8000 based on strings. Remove the 'loc' parameter when adding
8001 attributes because these will be a SET OF encoding which is sorted
8005 *) Initial changes to the 'req' utility to allow request generation
8006 automation. This will allow an application to just generate a template
8007 file containing all the field values and have req construct the
8010 Initial support for X509_ATTRIBUTE handling. Stacks of these are
8011 used all over the place including certificate requests and PKCS#7
8012 structures. They are currently handled manually where necessary with
8013 some primitive wrappers for PKCS#7. The new functions behave in a
8014 manner analogous to the X509 extension functions: they allow
8015 attributes to be looked up by NID and added.
8017 Later something similar to the X509V3 code would be desirable to
8018 automatically handle the encoding, decoding and printing of the
8019 more complex types. The string types like challengePassword can
8020 be handled by the string table functions.
8022 Also modified the multi byte string table handling. Now there is
8023 a 'global mask' which masks out certain types. The table itself
8024 can use the flag STABLE_NO_MASK to ignore the mask setting: this
8025 is useful when for example there is only one permissible type
8026 (as in countryName) and using the mask might result in no valid
8030 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
8031 SSL_get_peer_finished to allow applications to obtain the latest
8032 Finished messages sent to the peer or expected from the peer,
8033 respectively. (SSL_get_peer_finished is usually the Finished message
8034 actually received from the peer, otherwise the protocol will be aborted.)
8036 As the Finished message are message digests of the complete handshake
8037 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
8038 be used for external authentication procedures when the authentication
8039 provided by SSL/TLS is not desired or is not enough.
8042 *) Enhanced support for Alpha Linux is added. Now ./config checks if
8043 the host supports BWX extension and if Compaq C is present on the
8044 $PATH. Just exploiting of the BWX extension results in 20-30%
8045 performance kick for some algorithms, e.g. DES and RC4 to mention
8046 a couple. Compaq C in turn generates ~20% faster code for MD5 and
8050 *) Add support for MS "fast SGC". This is arguably a violation of the
8051 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
8052 weak crypto and after checking the certificate is SGC a second one
8053 with strong crypto. MS SGC stops the first handshake after receiving
8054 the server certificate message and sends a second client hello. Since
8055 a server will typically do all the time consuming operations before
8056 expecting any further messages from the client (server key exchange
8057 is the most expensive) there is little difference between the two.
8059 To get OpenSSL to support MS SGC we have to permit a second client
8060 hello message after we have sent server done. In addition we have to
8061 reset the MAC if we do get this second client hello.
8064 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
8065 if a DER encoded private key is RSA or DSA traditional format. Changed
8066 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
8067 format DER encoded private key. Newer code should use PKCS#8 format which
8068 has the key type encoded in the ASN1 structure. Added DER private key
8069 support to pkcs8 application.
8072 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
8073 ciphersuites has been selected (as required by the SSL 3/TLS 1
8074 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
8075 is set, we interpret this as a request to violate the specification
8076 (the worst that can happen is a handshake failure, and 'correct'
8077 behaviour would result in a handshake failure anyway).
8080 *) In SSL_CTX_add_session, take into account that there might be multiple
8081 SSL_SESSION structures with the same session ID (e.g. when two threads
8082 concurrently obtain them from an external cache).
8083 The internal cache can handle only one SSL_SESSION with a given ID,
8084 so if there's a conflict, we now throw out the old one to achieve
8088 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
8089 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
8090 some routines that use cipher OIDs: some ciphers do not have OIDs
8091 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
8095 *) Simplify the trust setting structure and code. Now we just have
8096 two sequences of OIDs for trusted and rejected settings. These will
8097 typically have values the same as the extended key usage extension
8098 and any application specific purposes.
8100 The trust checking code now has a default behaviour: it will just
8101 check for an object with the same NID as the passed id. Functions can
8102 be provided to override either the default behaviour or the behaviour
8103 for a given id. SSL client, server and email already have functions
8104 in place for compatibility: they check the NID and also return "trusted"
8105 if the certificate is self signed.
8108 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
8109 traditional format into an EVP_PKEY structure.
8112 *) Add a password callback function PEM_cb() which either prompts for
8113 a password if usr_data is NULL or otherwise assumes it is a null
8114 terminated password. Allow passwords to be passed on command line
8115 environment or config files in a few more utilities.
8118 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
8119 keys. Add some short names for PKCS#8 PBE algorithms and allow them
8120 to be specified on the command line for the pkcs8 and pkcs12 utilities.
8121 Update documentation.
8124 *) Support for ASN1 "NULL" type. This could be handled before by using
8125 ASN1_TYPE but there wasn't any function that would try to read a NULL
8126 and produce an error if it couldn't. For compatibility we also have
8127 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
8128 don't allocate anything because they don't need to.
8131 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
8133 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
8135 *) Rebuild of the memory allocation routines used by OpenSSL code and
8136 possibly others as well. The purpose is to make an interface that
8137 provide hooks so anyone can build a separate set of allocation and
8138 deallocation routines to be used by OpenSSL, for example memory
8139 pool implementations, or something else, which was previously hard
8140 since Malloc(), Realloc() and Free() were defined as macros having
8141 the values malloc, realloc and free, respectively (except for Win32
8142 compilations). The same is provided for memory debugging code.
8143 OpenSSL already comes with functionality to find memory leaks, but
8144 this gives people a chance to debug other memory problems.
8146 With these changes, a new set of functions and macros have appeared:
8148 CRYPTO_set_mem_debug_functions() [F]
8149 CRYPTO_get_mem_debug_functions() [F]
8150 CRYPTO_dbg_set_options() [F]
8151 CRYPTO_dbg_get_options() [F]
8152 CRYPTO_malloc_debug_init() [M]
8154 The memory debug functions are NULL by default, unless the library
8155 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
8156 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
8157 gives the standard debugging functions that come with OpenSSL) or
8158 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
8159 provided by the library user) must be used. When the standard
8160 debugging functions are used, CRYPTO_dbg_set_options can be used to
8161 request additional information:
8162 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
8163 the CRYPTO_MDEBUG_xxx macro when compiling the library.
8165 Also, things like CRYPTO_set_mem_functions will always give the
8166 expected result (the new set of functions is used for allocation
8167 and deallocation) at all times, regardless of platform and compiler
8170 To finish it up, some functions that were never use in any other
8171 way than through macros have a new API and new semantic:
8174 CRYPTO_dbg_realloc()
8177 All macros of value have retained their old syntax.
8178 [Richard Levitte and Bodo Moeller]
8180 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
8181 ordering of SMIMECapabilities wasn't in "strength order" and there
8182 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
8186 *) Some ASN1 types with illegal zero length encoding (INTEGER,
8187 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
8188 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
8190 *) Merge in my S/MIME library for OpenSSL. This provides a simple
8191 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
8192 functionality to handle multipart/signed properly) and a utility
8193 called 'smime' to call all this stuff. This is based on code I
8194 originally wrote for Celo who have kindly allowed it to be
8195 included in OpenSSL.
8198 *) Add variants des_set_key_checked and des_set_key_unchecked of
8199 des_set_key (aka des_key_sched). Global variable des_check_key
8200 decides which of these is called by des_set_key; this way
8201 des_check_key behaves as it always did, but applications and
8202 the library itself, which was buggy for des_check_key == 1,
8203 have a cleaner way to pick the version they need.
8206 *) New function PKCS12_newpass() which changes the password of a
8210 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
8211 dynamic mix. In both cases the ids can be used as an index into the
8212 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
8213 functions so they accept a list of the field values and the
8214 application doesn't need to directly manipulate the X509_TRUST
8218 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
8222 *) Modify the way the V3 extension code looks up extensions. This now
8223 works in a similar way to the object code: we have some "standard"
8224 extensions in a static table which is searched with OBJ_bsearch()
8225 and the application can add dynamic ones if needed. The file
8226 crypto/x509v3/ext_dat.h now has the info: this file needs to be
8227 updated whenever a new extension is added to the core code and kept
8228 in ext_nid order. There is a simple program 'tabtest.c' which checks
8229 this. New extensions are not added too often so this file can readily
8230 be maintained manually.
8232 There are two big advantages in doing things this way. The extensions
8233 can be looked up immediately and no longer need to be "added" using
8234 X509V3_add_standard_extensions(): this function now does nothing.
8235 [Side note: I get *lots* of email saying the extension code doesn't
8236 work because people forget to call this function]
8237 Also no dynamic allocation is done unless new extensions are added:
8238 so if we don't add custom extensions there is no need to call
8239 X509V3_EXT_cleanup().
8242 *) Modify enc utility's salting as follows: make salting the default. Add a
8243 magic header, so unsalted files fail gracefully instead of just decrypting
8244 to garbage. This is because not salting is a big security hole, so people
8245 should be discouraged from doing it.
8248 *) Fixes and enhancements to the 'x509' utility. It allowed a message
8249 digest to be passed on the command line but it only used this
8250 parameter when signing a certificate. Modified so all relevant
8251 operations are affected by the digest parameter including the
8252 -fingerprint and -x509toreq options. Also -x509toreq choked if a
8253 DSA key was used because it didn't fix the digest.
8256 *) Initial certificate chain verify code. Currently tests the untrusted
8257 certificates for consistency with the verify purpose (which is set
8258 when the X509_STORE_CTX structure is set up) and checks the pathlength.
8260 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
8261 this is because it will reject chains with invalid extensions whereas
8262 every previous version of OpenSSL and SSLeay made no checks at all.
8264 Trust code: checks the root CA for the relevant trust settings. Trust
8265 settings have an initial value consistent with the verify purpose: e.g.
8266 if the verify purpose is for SSL client use it expects the CA to be
8267 trusted for SSL client use. However the default value can be changed to
8268 permit custom trust settings: one example of this would be to only trust
8269 certificates from a specific "secure" set of CAs.
8271 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
8272 which should be used for version portability: especially since the
8273 verify structure is likely to change more often now.
8275 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
8276 to set them. If not set then assume SSL clients will verify SSL servers
8279 Two new options to the verify program: -untrusted allows a set of
8280 untrusted certificates to be passed in and -purpose which sets the
8281 intended purpose of the certificate. If a purpose is set then the
8282 new chain verify code is used to check extension consistency.
8285 *) Support for the authority information access extension.
8288 *) Modify RSA and DSA PEM read routines to transparently handle
8289 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
8290 public keys in a format compatible with certificate
8291 SubjectPublicKeyInfo structures. Unfortunately there were already
8292 functions called *_PublicKey_* which used various odd formats so
8293 these are retained for compatibility: however the DSA variants were
8294 never in a public release so they have been deleted. Changed dsa/rsa
8295 utilities to handle the new format: note no releases ever handled public
8296 keys so we should be OK.
8298 The primary motivation for this change is to avoid the same fiasco
8299 that dogs private keys: there are several incompatible private key
8300 formats some of which are standard and some OpenSSL specific and
8301 require various evil hacks to allow partial transparent handling and
8302 even then it doesn't work with DER formats. Given the option anything
8303 other than PKCS#8 should be dumped: but the other formats have to
8304 stay in the name of compatibility.
8306 With public keys and the benefit of hindsight one standard format
8307 is used which works with EVP_PKEY, RSA or DSA structures: though
8308 it clearly returns an error if you try to read the wrong kind of key.
8310 Added a -pubkey option to the 'x509' utility to output the public key.
8311 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
8312 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
8313 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
8314 that do the same as the EVP_PKEY_assign_*() except they up the
8315 reference count of the added key (they don't "swallow" the
8319 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
8320 CRLs would fail if the file contained no certificates or no CRLs:
8321 added a new function to read in both types and return the number
8322 read: this means that if none are read it will be an error. The
8323 DER versions of the certificate and CRL reader would always fail
8324 because it isn't possible to mix certificates and CRLs in DER format
8325 without choking one or the other routine. Changed this to just read
8326 a certificate: this is the best we can do. Also modified the code
8327 in apps/verify.c to take notice of return codes: it was previously
8328 attempting to read in certificates from NULL pointers and ignoring
8329 any errors: this is one reason why the cert and CRL reader seemed
8330 to work. It doesn't check return codes from the default certificate
8331 routines: these may well fail if the certificates aren't installed.
8334 *) Code to support otherName option in GeneralName.
8337 *) First update to verify code. Change the verify utility
8338 so it warns if it is passed a self signed certificate:
8339 for consistency with the normal behaviour. X509_verify
8340 has been modified to it will now verify a self signed
8341 certificate if *exactly* the same certificate appears
8342 in the store: it was previously impossible to trust a
8343 single self signed certificate. This means that:
8344 openssl verify ss.pem
8345 now gives a warning about a self signed certificate but
8346 openssl verify -CAfile ss.pem ss.pem
8350 *) For servers, store verify_result in SSL_SESSION data structure
8351 (and add it to external session representation).
8352 This is needed when client certificate verifications fails,
8353 but an application-provided verification callback (set by
8354 SSL_CTX_set_cert_verify_callback) allows accepting the session
8355 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
8356 but returns 1): When the session is reused, we have to set
8357 ssl->verify_result to the appropriate error code to avoid
8359 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
8361 *) Fix a bug in the new PKCS#7 code: it didn't consider the
8362 case in PKCS7_dataInit() where the signed PKCS7 structure
8363 didn't contain any existing data because it was being created.
8364 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
8366 *) Add a salt to the key derivation routines in enc.c. This
8367 forms the first 8 bytes of the encrypted file. Also add a
8368 -S option to allow a salt to be input on the command line.
8371 *) New function X509_cmp(). Oddly enough there wasn't a function
8372 to compare two certificates. We do this by working out the SHA1
8373 hash and comparing that. X509_cmp() will be needed by the trust
8377 *) SSL_get1_session() is like SSL_get_session(), but increments
8378 the reference count in the SSL_SESSION returned.
8379 [Geoff Thorpe <geoff@eu.c2.net>]
8381 *) Fix for 'req': it was adding a null to request attributes.
8382 Also change the X509_LOOKUP and X509_INFO code to handle
8383 certificate auxiliary information.
8386 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
8390 *) Add the possibility to add extra information to the memory leak
8391 detecting output, to form tracebacks, showing from where each
8392 allocation was originated: CRYPTO_push_info("constant string") adds
8393 the string plus current file name and line number to a per-thread
8394 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
8395 is like calling CYRPTO_pop_info() until the stack is empty.
8396 Also updated memory leak detection code to be multi-thread-safe.
8399 *) Add options -text and -noout to pkcs7 utility and delete the
8400 encryption options which never did anything. Update docs.
8403 *) Add options to some of the utilities to allow the pass phrase
8404 to be included on either the command line (not recommended on
8405 OSes like Unix) or read from the environment. Update the
8406 manpages and fix a few bugs.
8409 *) Add a few manpages for some of the openssl commands.
8412 *) Fix the -revoke option in ca. It was freeing up memory twice,
8413 leaking and not finding already revoked certificates.
8416 *) Extensive changes to support certificate auxiliary information.
8417 This involves the use of X509_CERT_AUX structure and X509_AUX
8418 functions. An X509_AUX function such as PEM_read_X509_AUX()
8419 can still read in a certificate file in the usual way but it
8420 will also read in any additional "auxiliary information". By
8421 doing things this way a fair degree of compatibility can be
8422 retained: existing certificates can have this information added
8423 using the new 'x509' options.
8425 Current auxiliary information includes an "alias" and some trust
8426 settings. The trust settings will ultimately be used in enhanced
8427 certificate chain verification routines: currently a certificate
8428 can only be trusted if it is self signed and then it is trusted
8432 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
8433 The problem was that one of the replacement routines had not been working
8434 since SSLeay releases. For now the offending routine has been replaced
8435 with non-optimised assembler. Even so, this now gives around 95%
8436 performance improvement for 1024 bit RSA signs.
8439 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
8440 handling. Most clients have the effective key size in bits equal to
8441 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
8442 A few however don't do this and instead use the size of the decrypted key
8443 to determine the RC2 key length and the AlgorithmIdentifier to determine
8444 the effective key length. In this case the effective key length can still
8445 be 40 bits but the key length can be 168 bits for example. This is fixed
8446 by manually forcing an RC2 key into the EVP_PKEY structure because the
8447 EVP code can't currently handle unusual RC2 key sizes: it always assumes
8448 the key length and effective key length are equal.
8451 *) Add a bunch of functions that should simplify the creation of
8452 X509_NAME structures. Now you should be able to do:
8453 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
8454 and have it automatically work out the correct field type and fill in
8455 the structures. The more adventurous can try:
8456 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
8457 and it will (hopefully) work out the correct multibyte encoding.
8460 *) Change the 'req' utility to use the new field handling and multibyte
8461 copy routines. Before the DN field creation was handled in an ad hoc
8462 way in req, ca, and x509 which was rather broken and didn't support
8463 BMPStrings or UTF8Strings. Since some software doesn't implement
8464 BMPStrings or UTF8Strings yet, they can be enabled using the config file
8465 using the dirstring_type option. See the new comment in the default
8466 openssl.cnf for more info.
8469 *) Make crypto/rand/md_rand.c more robust:
8470 - Assure unique random numbers after fork().
8471 - Make sure that concurrent threads access the global counter and
8472 md serializably so that we never lose entropy in them
8473 or use exactly the same state in multiple threads.
8474 Access to the large state is not always serializable because
8475 the additional locking could be a performance killer, and
8476 md should be large enough anyway.
8479 *) New file apps/app_rand.c with commonly needed functionality
8480 for handling the random seed file.
8482 Use the random seed file in some applications that previously did not:
8484 dsaparam -genkey (which also ignored its '-rand' option),
8487 x509 (when signing).
8488 Except on systems with /dev/urandom, it is crucial to have a random
8489 seed file at least for key creation, DSA signing, and for DH exchanges;
8490 for RSA signatures we could do without one.
8492 gendh and gendsa (unlike genrsa) used to read only the first byte
8493 of each file listed in the '-rand' option. The function as previously
8494 found in genrsa is now in app_rand.c and is used by all programs
8495 that support '-rand'.
8498 *) In RAND_write_file, use mode 0600 for creating files;
8499 don't just chmod when it may be too late.
8502 *) Report an error from X509_STORE_load_locations
8503 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
8506 *) New function ASN1_mbstring_copy() this copies a string in either
8507 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
8508 into an ASN1_STRING type. A mask of permissible types is passed
8509 and it chooses the "minimal" type to use or an error if not type
8513 *) Add function equivalents to the various macros in asn1.h. The old
8514 macros are retained with an M_ prefix. Code inside the library can
8515 use the M_ macros. External code (including the openssl utility)
8516 should *NOT* in order to be "shared library friendly".
8519 *) Add various functions that can check a certificate's extensions
8520 to see if it usable for various purposes such as SSL client,
8521 server or S/MIME and CAs of these types. This is currently
8522 VERY EXPERIMENTAL but will ultimately be used for certificate chain
8523 verification. Also added a -purpose flag to x509 utility to
8524 print out all the purposes.
8527 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
8531 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
8532 for, obtain and decode and extension and obtain its critical flag.
8533 This allows all the necessary extension code to be handled in a
8534 single function call.
8537 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
8538 platforms. See crypto/rc4/rc4_enc.c for further details.
8541 *) New -noout option to asn1parse. This causes no output to be produced
8542 its main use is when combined with -strparse and -out to extract data
8543 from a file (which may not be in ASN.1 format).
8546 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
8547 when producing the local key id.
8548 [Richard Levitte <levitte@stacken.kth.se>]
8550 *) New option -dhparam in s_server. This allows a DH parameter file to be
8551 stated explicitly. If it is not stated then it tries the first server
8552 certificate file. The previous behaviour hard coded the filename
8556 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
8557 a public key to be input or output. For example:
8558 openssl rsa -in key.pem -pubout -out pubkey.pem
8559 Also added necessary DSA public key functions to handle this.
8562 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
8563 in the message. This was handled by allowing
8564 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8565 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8567 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8568 to the end of the strings whereas this didn't. This would cause problems
8569 if strings read with d2i_ASN1_bytes() were later modified.
8570 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8572 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8573 data and it contains EOF it will end up returning an error. This is
8574 caused by input 46 bytes long. The cause is due to the way base64
8575 BIOs find the start of base64 encoded data. They do this by trying a
8576 trial decode on each line until they find one that works. When they
8577 do a flag is set and it starts again knowing it can pass all the
8578 data directly through the decoder. Unfortunately it doesn't reset
8579 the context it uses. This means that if EOF is reached an attempt
8580 is made to pass two EOFs through the context and this causes the
8581 resulting error. This can also cause other problems as well. As is
8582 usual with these problems it takes *ages* to find and the fix is
8583 trivial: move one line.
8584 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8586 *) Ugly workaround to get s_client and s_server working under Windows. The
8587 old code wouldn't work because it needed to select() on sockets and the
8588 tty (for keypresses and to see if data could be written). Win32 only
8589 supports select() on sockets so we select() with a 1s timeout on the
8590 sockets and then see if any characters are waiting to be read, if none
8591 are present then we retry, we also assume we can always write data to
8592 the tty. This isn't nice because the code then blocks until we've
8593 received a complete line of data and it is effectively polling the
8594 keyboard at 1s intervals: however it's quite a bit better than not
8595 working at all :-) A dedicated Windows application might handle this
8596 with an event loop for example.
8599 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
8600 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
8601 will be called when RSA_sign() and RSA_verify() are used. This is useful
8602 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
8603 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
8604 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
8605 This necessitated the support of an extra signature type NID_md5_sha1
8606 for SSL signatures and modifications to the SSL library to use it instead
8607 of calling RSA_public_decrypt() and RSA_private_encrypt().
8610 *) Add new -verify -CAfile and -CApath options to the crl program, these
8611 will lookup a CRL issuers certificate and verify the signature in a
8612 similar way to the verify program. Tidy up the crl program so it
8613 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
8614 less strict. It will now permit CRL extensions even if it is not
8615 a V2 CRL: this will allow it to tolerate some broken CRLs.
8618 *) Initialize all non-automatic variables each time one of the openssl
8619 sub-programs is started (this is necessary as they may be started
8620 multiple times from the "OpenSSL>" prompt).
8621 [Lennart Bang, Bodo Moeller]
8623 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
8624 removing all other RSA functionality (this is what NO_RSA does). This
8625 is so (for example) those in the US can disable those operations covered
8626 by the RSA patent while allowing storage and parsing of RSA keys and RSA
8630 *) Non-copying interface to BIO pairs.
8631 (still largely untested)
8634 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
8635 ASCII string. This was handled independently in various places before.
8638 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
8639 UTF8 strings a character at a time.
8642 *) Use client_version from client hello to select the protocol
8643 (s23_srvr.c) and for RSA client key exchange verification
8644 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
8647 *) Add various utility functions to handle SPKACs, these were previously
8648 handled by poking round in the structure internals. Added new function
8649 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
8650 print, verify and generate SPKACs. Based on an original idea from
8651 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
8654 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
8657 *) Allow the config file extension section to be overwritten on the
8658 command line. Based on an original idea from Massimiliano Pala
8659 <madwolf@comune.modena.it>. The new option is called -extensions
8660 and can be applied to ca, req and x509. Also -reqexts to override
8661 the request extensions in req and -crlexts to override the crl extensions
8665 *) Add new feature to the SPKAC handling in ca. Now you can include
8666 the same field multiple times by preceding it by "XXXX." for example:
8669 this is the same syntax as used in the req config file.
8672 *) Allow certificate extensions to be added to certificate requests. These
8673 are specified in a 'req_extensions' option of the req section of the
8674 config file. They can be printed out with the -text option to req but
8675 are otherwise ignored at present.
8678 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
8679 data read consists of only the final block it would not decrypted because
8680 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
8681 A misplaced 'break' also meant the decrypted final block might not be
8682 copied until the next read.
8685 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
8686 a few extra parameters to the DH structure: these will be useful if
8687 for example we want the value of 'q' or implement X9.42 DH.
8690 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
8691 provides hooks that allow the default DSA functions or functions on a
8692 "per key" basis to be replaced. This allows hardware acceleration and
8693 hardware key storage to be handled without major modification to the
8694 library. Also added low level modexp hooks and CRYPTO_EX structure and
8695 associated functions.
8698 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
8699 as "read only": it can't be written to and the buffer it points to will
8700 not be freed. Reading from a read only BIO is much more efficient than
8701 a normal memory BIO. This was added because there are several times when
8702 an area of memory needs to be read from a BIO. The previous method was
8703 to create a memory BIO and write the data to it, this results in two
8704 copies of the data and an O(n^2) reading algorithm. There is a new
8705 function BIO_new_mem_buf() which creates a read only memory BIO from
8706 an area of memory. Also modified the PKCS#7 routines to use read only
8710 *) Bugfix: ssl23_get_client_hello did not work properly when called in
8711 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
8712 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
8713 but a retry condition occured while trying to read the rest.
8716 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
8717 NID_pkcs7_encrypted by default: this was wrong since this should almost
8718 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
8719 the encrypted data type: this is a more sensible place to put it and it
8720 allows the PKCS#12 code to be tidied up that duplicated this
8724 *) Changed obj_dat.pl script so it takes its input and output files on
8725 the command line. This should avoid shell escape redirection problems
8729 *) Initial support for certificate extension requests, these are included
8730 in things like Xenroll certificate requests. Included functions to allow
8731 extensions to be obtained and added.
8734 *) -crlf option to s_client and s_server for sending newlines as
8735 CRLF (as required by many protocols).
8738 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
8740 *) Install libRSAglue.a when OpenSSL is built with RSAref.
8741 [Ralf S. Engelschall]
8743 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
8744 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
8746 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
8750 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
8751 DH parameters/keys (q is lost during that conversion, but the resulting
8752 DH parameters contain its length).
8754 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
8755 much faster than DH_generate_parameters (which creates parameters
8756 where p = 2*q + 1), and also the smaller q makes DH computations
8757 much more efficient (160-bit exponentiation instead of 1024-bit
8758 exponentiation); so this provides a convenient way to support DHE
8759 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
8760 utter importance to use
8761 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8763 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8764 when such DH parameters are used, because otherwise small subgroup
8765 attacks may become possible!
8768 *) Avoid memory leak in i2d_DHparams.
8771 *) Allow the -k option to be used more than once in the enc program:
8772 this allows the same encrypted message to be read by multiple recipients.
8775 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
8776 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
8777 it will always use the numerical form of the OID, even if it has a short
8781 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
8782 method only got called if p,q,dmp1,dmq1,iqmp components were present,
8783 otherwise bn_mod_exp was called. In the case of hardware keys for example
8784 no private key components need be present and it might store extra data
8785 in the RSA structure, which cannot be accessed from bn_mod_exp.
8786 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
8787 private key operations.
8790 *) Added support for SPARC Linux.
8793 *) pem_password_cb function type incompatibly changed from
8794 typedef int pem_password_cb(char *buf, int size, int rwflag);
8796 ....(char *buf, int size, int rwflag, void *userdata);
8797 so that applications can pass data to their callbacks:
8798 The PEM[_ASN1]_{read,write}... functions and macros now take an
8799 additional void * argument, which is just handed through whenever
8800 the password callback is called.
8801 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
8803 New function SSL_CTX_set_default_passwd_cb_userdata.
8805 Compatibility note: As many C implementations push function arguments
8806 onto the stack in reverse order, the new library version is likely to
8807 interoperate with programs that have been compiled with the old
8808 pem_password_cb definition (PEM_whatever takes some data that
8809 happens to be on the stack as its last argument, and the callback
8810 just ignores this garbage); but there is no guarantee whatsoever that
8813 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
8814 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
8815 problems not only on Windows, but also on some Unix platforms.
8816 To avoid problematic command lines, these definitions are now in an
8817 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
8818 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
8821 *) MIPS III/IV assembler module is reimplemented.
8824 *) More DES library cleanups: remove references to srand/rand and
8825 delete an unused file.
8828 *) Add support for the the free Netwide assembler (NASM) under Win32,
8829 since not many people have MASM (ml) and it can be hard to obtain.
8830 This is currently experimental but it seems to work OK and pass all
8831 the tests. Check out INSTALL.W32 for info.
8834 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
8835 without temporary keys kept an extra copy of the server key,
8836 and connections with temporary keys did not free everything in case
8840 *) New function RSA_check_key and new openssl rsa option -check
8841 for verifying the consistency of RSA keys.
8842 [Ulf Moeller, Bodo Moeller]
8844 *) Various changes to make Win32 compile work:
8845 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
8846 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
8847 comparison" warnings.
8848 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
8851 *) Add a debugging option to PKCS#5 v2 key generation function: when
8852 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
8853 derived keys are printed to stderr.
8856 *) Copy the flags in ASN1_STRING_dup().
8857 [Roman E. Pavlov <pre@mo.msk.ru>]
8859 *) The x509 application mishandled signing requests containing DSA
8860 keys when the signing key was also DSA and the parameters didn't match.
8862 It was supposed to omit the parameters when they matched the signing key:
8863 the verifying software was then supposed to automatically use the CA's
8864 parameters if they were absent from the end user certificate.
8866 Omitting parameters is no longer recommended. The test was also
8867 the wrong way round! This was probably due to unusual behaviour in
8868 EVP_cmp_parameters() which returns 1 if the parameters match.
8869 This meant that parameters were omitted when they *didn't* match and
8870 the certificate was useless. Certificates signed with 'ca' didn't have
8872 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
8874 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
8875 The interface is as follows:
8876 Applications can use
8877 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
8878 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
8879 "off" is now the default.
8880 The library internally uses
8881 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
8882 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
8883 to disable memory-checking temporarily.
8885 Some inconsistent states that previously were possible (and were
8886 even the default) are now avoided.
8888 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
8889 with each memory chunk allocated; this is occasionally more helpful
8890 than just having a counter.
8892 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
8894 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
8898 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
8899 which largely parallels "options", but is for changing API behaviour,
8900 whereas "options" are about protocol behaviour.
8901 Initial "mode" flags are:
8903 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
8904 a single record has been written.
8905 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
8906 retries use the same buffer location.
8907 (But all of the contents must be
8911 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
8914 *) Fix problems with no-hmac etc.
8915 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
8917 *) New functions RSA_get_default_method(), RSA_set_method() and
8918 RSA_get_method(). These allows replacement of RSA_METHODs without having
8919 to mess around with the internals of an RSA structure.
8922 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
8923 Also really enable memory leak checks in openssl.c and in some
8925 [Chad C. Mulligan, Bodo Moeller]
8927 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
8928 up the length of negative integers. This has now been simplified to just
8929 store the length when it is first determined and use it later, rather
8930 than trying to keep track of where data is copied and updating it to
8932 [Steve Henson, reported by Brien Wheeler
8933 <bwheeler@authentica-security.com>]
8935 *) Add a new function PKCS7_signatureVerify. This allows the verification
8936 of a PKCS#7 signature but with the signing certificate passed to the
8937 function itself. This contrasts with PKCS7_dataVerify which assumes the
8938 certificate is present in the PKCS#7 structure. This isn't always the
8939 case: certificates can be omitted from a PKCS#7 structure and be
8940 distributed by "out of band" means (such as a certificate database).
8943 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
8944 function prototypes in pem.h, also change util/mkdef.pl to add the
8945 necessary function names.
8948 *) mk1mf.pl (used by Windows builds) did not properly read the
8949 options set by Configure in the top level Makefile, and Configure
8950 was not even able to write more than one option correctly.
8951 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
8954 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
8955 file to be loaded from a BIO or FILE pointer. The BIO version will
8956 for example allow memory BIOs to contain config info.
8959 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
8960 Whoever hopes to achieve shared-library compatibility across versions
8961 must use this, not the compile-time macro.
8962 (Exercise 0.9.4: Which is the minimum library version required by
8964 Note: All this applies only to multi-threaded programs, others don't
8968 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
8969 through a BIO pair triggered the default case, i.e.
8970 SSLerr(...,SSL_R_UNKNOWN_STATE).
8973 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
8974 can use the SSL library even if none of the specific BIOs is
8978 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
8979 for the encoded length.
8980 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
8982 *) Add initial documentation of the X509V3 functions.
8985 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
8986 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
8987 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
8988 secure PKCS#8 private key format with a high iteration count.
8991 *) Fix determination of Perl interpreter: A perl or perl5
8992 _directory_ in $PATH was also accepted as the interpreter.
8993 [Ralf S. Engelschall]
8995 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
8996 wrong with it but it was very old and did things like calling
8997 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
9001 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
9002 to use the new extension code.
9005 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
9006 with macros. This should make it easier to change their form, add extra
9007 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
9011 *) Add to configuration table a new entry that can specify an alternative
9012 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
9013 according to Mark Crispin <MRC@Panda.COM>.
9017 *) DES CBC did not update the IV. Weird.
9020 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
9021 Changing the behaviour of the former might break existing programs --
9022 where IV updating is needed, des_ncbc_encrypt can be used.
9025 *) When bntest is run from "make test" it drives bc to check its
9026 calculations, as well as internally checking them. If an internal check
9027 fails, it needs to cause bc to give a non-zero result or make test carries
9028 on without noticing the failure. Fixed.
9031 *) DES library cleanups.
9034 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
9035 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
9036 ciphers. NOTE: although the key derivation function has been verified
9037 against some published test vectors it has not been extensively tested
9038 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
9042 *) Instead of "mkdir -p", which is not fully portable, use new
9043 Perl script "util/mkdir-p.pl".
9046 *) Rewrite the way password based encryption (PBE) is handled. It used to
9047 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
9048 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
9049 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
9050 the 'parameter' field of the AlgorithmIdentifier is passed to the
9051 underlying key generation function so it must do its own ASN1 parsing.
9052 This has also changed the EVP_PBE_CipherInit() function which now has a
9053 'parameter' argument instead of literal salt and iteration count values
9054 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
9057 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
9058 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
9059 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
9060 KEY" because this clashed with PKCS#8 unencrypted string. Since this
9061 value was just used as a "magic string" and not used directly its
9062 value doesn't matter.
9065 *) Introduce some semblance of const correctness to BN. Shame C doesn't
9069 *) "linux-sparc64" configuration (ultrapenguin).
9070 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
9071 "linux-sparc" configuration.
9072 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
9074 *) config now generates no-xxx options for missing ciphers.
9077 *) Support the EBCDIC character set (work in progress).
9078 File ebcdic.c not yet included because it has a different license.
9079 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9081 *) Support BS2000/OSD-POSIX.
9082 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9084 *) Make callbacks for key generation use void * instead of char *.
9087 *) Make S/MIME samples compile (not yet tested).
9090 *) Additional typesafe stacks.
9093 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
9097 Changes between 0.9.3 and 0.9.3a [29 May 1999]
9099 *) New configuration variant "sco5-gcc".
9101 *) Updated some demos.
9102 [Sean O Riordain, Wade Scholine]
9104 *) Add missing BIO_free at exit of pkcs12 application.
9107 *) Fix memory leak in conf.c.
9110 *) Updates for Win32 to assembler version of MD5.
9113 *) Set #! path to perl in apps/der_chop to where we found it
9114 instead of using a fixed path.
9117 *) SHA library changes for irix64-mips4-cc.
9120 *) Improvements for VMS support.
9124 Changes between 0.9.2b and 0.9.3 [24 May 1999]
9126 *) Bignum library bug fix. IRIX 6 passes "make test" now!
9127 This also avoids the problems with SC4.2 and unpatched SC5.
9128 [Andy Polyakov <appro@fy.chalmers.se>]
9130 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
9131 These are required because of the typesafe stack would otherwise break
9132 existing code. If old code used a structure member which used to be STACK
9133 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
9134 sk_num or sk_value it would produce an error because the num, data members
9135 are not present in STACK_OF. Now it just produces a warning. sk_set
9136 replaces the old method of assigning a value to sk_value
9137 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
9138 that does this will no longer work (and should use sk_set instead) but
9139 this could be regarded as a "questionable" behaviour anyway.
9142 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
9143 correctly handle encrypted S/MIME data.
9146 *) Change type of various DES function arguments from des_cblock
9147 (which means, in function argument declarations, pointer to char)
9148 to des_cblock * (meaning pointer to array with 8 char elements),
9149 which allows the compiler to do more typechecking; it was like
9150 that back in SSLeay, but with lots of ugly casts.
9152 Introduce new type const_des_cblock.
9155 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
9156 problems: find RecipientInfo structure that matches recipient certificate
9157 and initialise the ASN1 structures properly based on passed cipher.
9160 *) Belatedly make the BN tests actually check the results.
9163 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
9164 to and from BNs: it was completely broken. New compilation option
9165 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
9166 key elements as negative integers.
9169 *) Reorganize and speed up MD5.
9170 [Andy Polyakov <appro@fy.chalmers.se>]
9173 [Richard Levitte <richard@levitte.org>]
9175 *) New option -out to asn1parse to allow the parsed structure to be
9176 output to a file. This is most useful when combined with the -strparse
9177 option to examine the output of things like OCTET STRINGS.
9180 *) Make SSL library a little more fool-proof by not requiring any longer
9181 that SSL_set_{accept,connect}_state be called before
9182 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
9183 in many applications because usually everything *appeared* to work as
9184 intended anyway -- now it really works as intended).
9187 *) Move openssl.cnf out of lib/.
9190 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
9191 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
9192 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
9193 [Ralf S. Engelschall]
9195 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
9196 handle PKCS#7 enveloped data properly.
9197 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
9199 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
9200 copying pointers. The cert_st handling is changed by this in
9201 various ways (and thus what used to be known as ctx->default_cert
9202 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
9203 any longer when s->cert does not give us what we need).
9204 ssl_cert_instantiate becomes obsolete by this change.
9205 As soon as we've got the new code right (possibly it already is?),
9206 we have solved a couple of bugs of the earlier code where s->cert
9207 was used as if it could not have been shared with other SSL structures.
9209 Note that using the SSL API in certain dirty ways now will result
9210 in different behaviour than observed with earlier library versions:
9211 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
9212 does not influence s as it used to.
9214 In order to clean up things more thoroughly, inside SSL_SESSION
9215 we don't use CERT any longer, but a new structure SESS_CERT
9216 that holds per-session data (if available); currently, this is
9217 the peer's certificate chain and, for clients, the server's certificate
9218 and temporary key. CERT holds only those values that can have
9219 meaningful defaults in an SSL_CTX.
9222 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
9223 from the internal representation. Various PKCS#7 fixes: remove some
9224 evil casts and set the enc_dig_alg field properly based on the signing
9228 *) Allow PKCS#12 password to be set from the command line or the
9229 environment. Let 'ca' get its config file name from the environment
9230 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
9234 *) Allow certificate policies extension to use an IA5STRING for the
9235 organization field. This is contrary to the PKIX definition but
9236 VeriSign uses it and IE5 only recognises this form. Document 'x509'
9240 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
9241 without disallowing inline assembler and the like for non-pedantic builds.
9244 *) Support Borland C++ builder.
9245 [Janez Jere <jj@void.si>, modified by Ulf Möller]
9250 *) SHA-1 cleanups and performance enhancements.
9251 [Andy Polyakov <appro@fy.chalmers.se>]
9253 *) Sparc v8plus assembler for the bignum library.
9254 [Andy Polyakov <appro@fy.chalmers.se>]
9256 *) Accept any -xxx and +xxx compiler options in Configure.
9259 *) Update HPUX configuration.
9262 *) Add missing sk_<type>_unshift() function to safestack.h
9263 [Ralf S. Engelschall]
9265 *) New function SSL_CTX_use_certificate_chain_file that sets the
9266 "extra_cert"s in addition to the certificate. (This makes sense
9267 only for "PEM" format files, as chains as a whole are not
9271 *) Support verify_depth from the SSL API.
9272 x509_vfy.c had what can be considered an off-by-one-error:
9273 Its depth (which was not part of the external interface)
9274 was actually counting the number of certificates in a chain;
9275 now it really counts the depth.
9278 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
9279 instead of X509err, which often resulted in confusing error
9280 messages since the error codes are not globally unique
9281 (e.g. an alleged error in ssl3_accept when a certificate
9282 didn't match the private key).
9284 *) New function SSL_CTX_set_session_id_context that allows to set a default
9285 value (so that you don't need SSL_set_session_id_context for each
9286 connection using the SSL_CTX).
9289 *) OAEP decoding bug fix.
9292 *) Support INSTALL_PREFIX for package builders, as proposed by
9296 *) New Configure options "threads" and "no-threads". For systems
9297 where the proper compiler options are known (currently Solaris
9298 and Linux), "threads" is the default.
9301 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
9304 *) Install various scripts to $(OPENSSLDIR)/misc, not to
9305 $(INSTALLTOP)/bin -- they shouldn't clutter directories
9306 such as /usr/local/bin.
9309 *) "make linux-shared" to build shared libraries.
9310 [Niels Poppe <niels@netbox.org>]
9312 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
9315 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
9316 extension adding in x509 utility.
9319 *) Remove NOPROTO sections and error code comments.
9322 *) Partial rewrite of the DEF file generator to now parse the ANSI
9326 *) New Configure options --prefix=DIR and --openssldir=DIR.
9329 *) Complete rewrite of the error code script(s). It is all now handled
9330 by one script at the top level which handles error code gathering,
9331 header rewriting and C source file generation. It should be much better
9332 than the old method: it now uses a modified version of Ulf's parser to
9333 read the ANSI prototypes in all header files (thus the old K&R definitions
9334 aren't needed for error creation any more) and do a better job of
9335 translating function codes into names. The old 'ASN1 error code imbedded
9336 in a comment' is no longer necessary and it doesn't use .err files which
9337 have now been deleted. Also the error code call doesn't have to appear all
9338 on one line (which resulted in some large lines...).
9341 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
9344 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
9345 0 (which usually indicates a closed connection), but continue reading.
9348 *) Fix some race conditions.
9351 *) Add support for CRL distribution points extension. Add Certificate
9352 Policies and CRL distribution points documentation.
9355 *) Move the autogenerated header file parts to crypto/opensslconf.h.
9358 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
9359 8 of keying material. Merlin has also confirmed interop with this fix
9360 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
9361 [Merlin Hughes <merlin@baltimore.ie>]
9363 *) Fix lots of warnings.
9364 [Richard Levitte <levitte@stacken.kth.se>]
9366 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
9367 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
9368 [Richard Levitte <levitte@stacken.kth.se>]
9370 *) Fix problems with sizeof(long) == 8.
9371 [Andy Polyakov <appro@fy.chalmers.se>]
9373 *) Change functions to ANSI C.
9376 *) Fix typos in error codes.
9377 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
9379 *) Remove defunct assembler files from Configure.
9382 *) SPARC v8 assembler BIGNUM implementation.
9383 [Andy Polyakov <appro@fy.chalmers.se>]
9385 *) Support for Certificate Policies extension: both print and set.
9386 Various additions to support the r2i method this uses.
9389 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
9390 return a const string when you are expecting an allocated buffer.
9393 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
9394 types DirectoryString and DisplayText.
9397 *) Add code to allow r2i extensions to access the configuration database,
9398 add an LHASH database driver and add several ctx helper functions.
9401 *) Fix an evil bug in bn_expand2() which caused various BN functions to
9402 fail when they extended the size of a BIGNUM.
9405 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
9406 support typesafe stack.
9409 *) Fix typo in SSL_[gs]et_options().
9410 [Nils Frostberg <nils@medcom.se>]
9412 *) Delete various functions and files that belonged to the (now obsolete)
9413 old X509V3 handling code.
9416 *) New Configure option "rsaref".
9419 *) Don't auto-generate pem.h.
9422 *) Introduce type-safe ASN.1 SETs.
9425 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
9426 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
9428 *) Introduce type-safe STACKs. This will almost certainly break lots of code
9429 that links with OpenSSL (well at least cause lots of warnings), but fear
9430 not: the conversion is trivial, and it eliminates loads of evil casts. A
9431 few STACKed things have been converted already. Feel free to convert more.
9432 In the fullness of time, I'll do away with the STACK type altogether.
9435 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
9436 specified in <certfile> by updating the entry in the index.txt file.
9437 This way one no longer has to edit the index.txt file manually for
9438 revoking a certificate. The -revoke option does the gory details now.
9439 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
9441 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
9442 `-text' option at all and this way the `-noout -text' combination was
9443 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
9444 [Ralf S. Engelschall]
9446 *) Make sure a corresponding plain text error message exists for the
9447 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
9448 verify callback function determined that a certificate was revoked.
9449 [Ralf S. Engelschall]
9451 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
9452 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
9453 all available cipers including rc5, which was forgotten until now.
9454 In order to let the testing shell script know which algorithms
9455 are available, a new (up to now undocumented) command
9456 "openssl list-cipher-commands" is used.
9459 *) Bugfix: s_client occasionally would sleep in select() when
9460 it should have checked SSL_pending() first.
9463 *) New functions DSA_do_sign and DSA_do_verify to provide access to
9464 the raw DSA values prior to ASN.1 encoding.
9467 *) Tweaks to Configure
9468 [Niels Poppe <niels@netbox.org>]
9470 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
9474 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
9477 *) New config option to avoid instructions that are illegal on the 80386.
9478 The default code is faster, but requires at least a 486.
9481 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
9482 SSL2_SERVER_VERSION (not used at all) macros, which are now the
9483 same as SSL2_VERSION anyway.
9486 *) New "-showcerts" option for s_client.
9489 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
9490 application. Various cleanups and fixes.
9493 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
9494 modify error routines to work internally. Add error codes and PBE init
9495 to library startup routines.
9498 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
9499 packing functions to asn1 and evp. Changed function names and error
9500 codes along the way.
9503 *) PKCS12 integration: and so it begins... First of several patches to
9504 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
9505 objects to objects.h
9508 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
9509 and display support for Thawte strong extranet extension.
9512 *) Add LinuxPPC support.
9513 [Jeff Dubrule <igor@pobox.org>]
9515 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
9516 bn_div_words in alpha.s.
9517 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
9519 *) Make sure the RSA OAEP test is skipped under -DRSAref because
9520 OAEP isn't supported when OpenSSL is built with RSAref.
9521 [Ulf Moeller <ulf@fitug.de>]
9523 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
9524 so they no longer are missing under -DNOPROTO.
9525 [Soren S. Jorvang <soren@t.dk>]
9528 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
9530 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
9531 doesn't work when the session is reused. Coming soon!
9534 *) Fix a security hole, that allows sessions to be reused in the wrong
9535 context thus bypassing client cert protection! All software that uses
9536 client certs and session caches in multiple contexts NEEDS PATCHING to
9537 allow session reuse! A fuller solution is in the works.
9538 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
9540 *) Some more source tree cleanups (removed obsolete files
9541 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
9542 permission on "config" script to be executable) and a fix for the INSTALL
9544 [Ulf Moeller <ulf@fitug.de>]
9546 *) Remove some legacy and erroneous uses of malloc, free instead of
9548 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
9550 *) Make rsa_oaep_test return non-zero on error.
9551 [Ulf Moeller <ulf@fitug.de>]
9553 *) Add support for native Solaris shared libraries. Configure
9554 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
9555 if someone would make that last step automatic.
9556 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
9558 *) ctx_size was not built with the right compiler during "make links". Fixed.
9561 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
9562 except NULL ciphers". This means the default cipher list will no longer
9563 enable NULL ciphers. They need to be specifically enabled e.g. with
9564 the string "DEFAULT:eNULL".
9567 *) Fix to RSA private encryption routines: if p < q then it would
9568 occasionally produce an invalid result. This will only happen with
9569 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9572 *) Be less restrictive and allow also `perl util/perlpath.pl
9573 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9574 because this way one can also use an interpreter named `perl5' (which is
9575 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9576 installed as `perl').
9577 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9579 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9580 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9582 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9583 advapi32.lib to Win32 build and change the pem test comparision
9584 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9585 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9586 and crypto/des/ede_cbcm_enc.c.
9589 *) DES quad checksum was broken on big-endian architectures. Fixed.
9592 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9593 Win32 test batch file so it (might) work again. The Win32 test batch file
9594 is horrible: I feel ill....
9597 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
9598 in e_os.h. Audit of header files to check ANSI and non ANSI
9599 sections: 10 functions were absent from non ANSI section and not exported
9600 from Windows DLLs. Fixed up libeay.num for new functions.
9603 *) Make `openssl version' output lines consistent.
9604 [Ralf S. Engelschall]
9606 *) Fix Win32 symbol export lists for BIO functions: Added
9607 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
9608 to ms/libeay{16,32}.def.
9609 [Ralf S. Engelschall]
9611 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
9612 fine under Unix and passes some trivial tests I've now added. But the
9613 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
9614 added to make sure no one expects that this stuff really works in the
9615 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
9616 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
9618 [Ralf S. Engelschall]
9620 *) Fix the generation of two part addresses in perl.
9621 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
9623 *) Add config entry for Linux on MIPS.
9624 [John Tobey <jtobey@channel1.com>]
9626 *) Make links whenever Configure is run, unless we are on Windoze.
9629 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
9630 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
9634 *) Add a useful kludge to allow package maintainers to specify compiler and
9635 other platforms details on the command line without having to patch the
9636 Configure script everytime: One now can use ``perl Configure
9637 <id>:<details>'', i.e. platform ids are allowed to have details appended
9638 to them (seperated by colons). This is treated as there would be a static
9639 pre-configured entry in Configure's %table under key <id> with value
9640 <details> and ``perl Configure <id>'' is called. So, when you want to
9641 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
9642 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
9643 now, which overrides the FreeBSD-elf entry on-the-fly.
9644 [Ralf S. Engelschall]
9646 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
9649 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
9650 on the `perl Configure ...' command line. This way one can compile
9651 OpenSSL libraries with Position Independent Code (PIC) which is needed
9652 for linking it into DSOs.
9653 [Ralf S. Engelschall]
9655 *) Remarkably, export ciphers were totally broken and no-one had noticed!
9659 *) Cleaned up the LICENSE document: The official contact for any license
9660 questions now is the OpenSSL core team under openssl-core@openssl.org.
9661 And add a paragraph about the dual-license situation to make sure people
9662 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
9663 to the OpenSSL toolkit.
9664 [Ralf S. Engelschall]
9666 *) General source tree makefile cleanups: Made `making xxx in yyy...'
9667 display consistent in the source tree and replaced `/bin/rm' by `rm'.
9668 Additonally cleaned up the `make links' target: Remove unnecessary
9669 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
9670 to speed processing and no longer clutter the display with confusing
9671 stuff. Instead only the actually done links are displayed.
9672 [Ralf S. Engelschall]
9674 *) Permit null encryption ciphersuites, used for authentication only. It used
9675 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
9676 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
9680 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
9681 signed attributes when verifying signatures (this would break them),
9682 the detached data encoding was wrong and public keys obtained using
9683 X509_get_pubkey() weren't freed.
9686 *) Add text documentation for the BUFFER functions. Also added a work around
9687 to a Win95 console bug. This was triggered by the password read stuff: the
9688 last character typed gets carried over to the next fread(). If you were
9689 generating a new cert request using 'req' for example then the last
9690 character of the passphrase would be CR which would then enter the first
9694 *) Added the new `Includes OpenSSL Cryptography Software' button as
9695 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
9696 button and can be used by applications based on OpenSSL to show the
9697 relationship to the OpenSSL project.
9698 [Ralf S. Engelschall]
9700 *) Remove confusing variables in function signatures in files
9701 ssl/ssl_lib.c and ssl/ssl.h.
9702 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9704 *) Don't install bss_file.c under PREFIX/include/
9705 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9707 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
9708 functions that return function pointers and has support for NT specific
9709 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
9710 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
9711 unsigned to signed types: this was killing the Win32 compile.
9714 *) Add new certificate file to stack functions,
9715 SSL_add_dir_cert_subjects_to_stack() and
9716 SSL_add_file_cert_subjects_to_stack(). These largely supplant
9717 SSL_load_client_CA_file(), and can be used to add multiple certs easily
9718 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
9719 This means that Apache-SSL and similar packages don't have to mess around
9720 to add as many CAs as they want to the preferred list.
9723 *) Experiment with doxygen documentation. Currently only partially applied to
9725 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
9726 openssl.doxy as the configuration file.
9729 *) Get rid of remaining C++-style comments which strict C compilers hate.
9730 [Ralf S. Engelschall, pointed out by Carlos Amengual]
9732 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
9733 compiled in by default: it has problems with large keys.
9736 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
9737 DH private keys and/or callback functions which directly correspond to
9738 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
9739 is needed for applications which have to configure certificates on a
9740 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
9742 For the RSA certificate situation is makes no difference, but
9743 for the DSA certificate situation this fixes the "no shared cipher"
9744 problem where the OpenSSL cipher selection procedure failed because the
9745 temporary keys were not overtaken from the context and the API provided
9746 no way to reconfigure them.
9747 The new functions now let applications reconfigure the stuff and they
9748 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
9749 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
9750 non-public-API function ssl_cert_instantiate() is used as a helper
9751 function and also to reduce code redundancy inside ssl_rsa.c.
9752 [Ralf S. Engelschall]
9754 *) Move s_server -dcert and -dkey options out of the undocumented feature
9755 area because they are useful for the DSA situation and should be
9756 recognized by the users.
9757 [Ralf S. Engelschall]
9759 *) Fix the cipher decision scheme for export ciphers: the export bits are
9760 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
9761 SSL_EXP_MASK. So, the original variable has to be used instead of the
9762 already masked variable.
9763 [Richard Levitte <levitte@stacken.kth.se>]
9765 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
9766 [Richard Levitte <levitte@stacken.kth.se>]
9768 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
9769 from `int' to `unsigned int' because it's a length and initialized by
9770 EVP_DigestFinal() which expects an `unsigned int *'.
9771 [Richard Levitte <levitte@stacken.kth.se>]
9773 *) Don't hard-code path to Perl interpreter on shebang line of Configure
9774 script. Instead use the usual Shell->Perl transition trick.
9775 [Ralf S. Engelschall]
9777 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
9778 (in addition to RSA certificates) to match the behaviour of `openssl dsa
9779 -noout -modulus' as it's already the case for `openssl rsa -noout
9780 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
9781 currently the public key is printed (a decision which was already done by
9782 `openssl dsa -modulus' in the past) which serves a similar purpose.
9783 Additionally the NO_RSA no longer completely removes the whole -modulus
9784 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
9786 [Ralf S. Engelschall]
9788 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
9789 BIO. See the source (crypto/evp/bio_ok.c) for more info.
9790 [Arne Ansper <arne@ats.cyber.ee>]
9792 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
9793 to be added. Now both 'req' and 'ca' can use new objects defined in the
9797 *) Add cool BIO that does syslog (or event log on NT).
9798 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
9800 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
9801 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
9802 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
9803 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
9806 *) Add preliminary config info for new extension code.
9809 *) Make RSA_NO_PADDING really use no padding.
9810 [Ulf Moeller <ulf@fitug.de>]
9812 *) Generate errors when private/public key check is done.
9815 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
9816 for some CRL extensions and new objects added.
9819 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
9820 key usage extension and fuller support for authority key id.
9823 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
9824 padding method for RSA, which is recommended for new applications in PKCS
9825 #1 v2.0 (RFC 2437, October 1998).
9826 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
9827 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
9828 against Bleichbacher's attack on RSA.
9829 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
9832 *) Updates to the new SSL compression code
9833 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9835 *) Fix so that the version number in the master secret, when passed
9836 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
9837 (because the server will not accept higher), that the version number
9838 is 0x03,0x01, not 0x03,0x00
9839 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9841 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
9842 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
9843 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
9846 *) Support for RAW extensions where an arbitrary extension can be
9847 created by including its DER encoding. See apps/openssl.cnf for
9851 *) Make sure latest Perl versions don't interpret some generated C array
9852 code as Perl array code in the crypto/err/err_genc.pl script.
9853 [Lars Weber <3weber@informatik.uni-hamburg.de>]
9855 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
9856 not many people have the assembler. Various Win32 compilation fixes and
9857 update to the INSTALL.W32 file with (hopefully) more accurate Win32
9861 *) Modify configure script 'Configure' to automatically create crypto/date.h
9862 file under Win32 and also build pem.h from pem.org. New script
9863 util/mkfiles.pl to create the MINFO file on environments that can't do a
9864 'make files': perl util/mkfiles.pl >MINFO should work.
9867 *) Major rework of DES function declarations, in the pursuit of correctness
9868 and purity. As a result, many evil casts evaporated, and some weirdness,
9869 too. You may find this causes warnings in your code. Zapping your evil
9870 casts will probably fix them. Mostly.
9873 *) Fix for a typo in asn1.h. Bug fix to object creation script
9874 obj_dat.pl. It considered a zero in an object definition to mean
9875 "end of object": none of the objects in objects.h have any zeros
9876 so it wasn't spotted.
9877 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
9879 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
9880 Masking (CBCM). In the absence of test vectors, the best I have been able
9881 to do is check that the decrypt undoes the encrypt, so far. Send me test
9882 vectors if you have them.
9885 *) Correct calculation of key length for export ciphers (too much space was
9886 allocated for null ciphers). This has not been tested!
9889 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
9890 message is now correct (it understands "crypto" and "ssl" on its
9891 command line). There is also now an "update" option. This will update
9892 the util/ssleay.num and util/libeay.num files with any new functions.
9894 perl util/mkdef.pl crypto ssl update
9895 it will update them.
9898 *) Overhauled the Perl interface (perl/*):
9899 - ported BN stuff to OpenSSL's different BN library
9900 - made the perl/ source tree CVS-aware
9901 - renamed the package from SSLeay to OpenSSL (the files still contain
9902 their history because I've copied them in the repository)
9903 - removed obsolete files (the test scripts will be replaced
9904 by better Test::Harness variants in the future)
9905 [Ralf S. Engelschall]
9907 *) First cut for a very conservative source tree cleanup:
9908 1. merge various obsolete readme texts into doc/ssleay.txt
9909 where we collect the old documents and readme texts.
9910 2. remove the first part of files where I'm already sure that we no
9911 longer need them because of three reasons: either they are just temporary
9912 files which were left by Eric or they are preserved original files where
9913 I've verified that the diff is also available in the CVS via "cvs diff
9914 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
9915 the crypto/md/ stuff).
9916 [Ralf S. Engelschall]
9918 *) More extension code. Incomplete support for subject and issuer alt
9919 name, issuer and authority key id. Change the i2v function parameters
9920 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
9921 what that's for :-) Fix to ASN1 macro which messed up
9922 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
9925 *) Preliminary support for ENUMERATED type. This is largely copied from the
9929 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
9930 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9932 *) Make sure `make rehash' target really finds the `openssl' program.
9933 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9935 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
9936 like to hear about it if this slows down other processors.
9939 *) Add CygWin32 platform information to Configure script.
9940 [Alan Batie <batie@aahz.jf.intel.com>]
9942 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
9943 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
9945 *) New program nseq to manipulate netscape certificate sequences
9948 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
9952 *) Fixes to BN code. Previously the default was to define BN_RECURSION
9953 but the BN code had some problems that would cause failures when
9954 doing certificate verification and some other functions.
9955 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9957 *) Add ASN1 and PEM code to support netscape certificate sequences.
9960 *) Add ASN1 and PEM code to support netscape certificate sequences.
9963 *) Add several PKIX and private extended key usage OIDs.
9966 *) Modify the 'ca' program to handle the new extension code. Modify
9967 openssl.cnf for new extension format, add comments.
9970 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
9971 and add a sample to openssl.cnf so req -x509 now adds appropriate
9975 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
9976 error code, add initial support to X509_print() and x509 application.
9979 *) Takes a deep breath and start addding X509 V3 extension support code. Add
9980 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
9981 stuff is currently isolated and isn't even compiled yet.
9984 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
9985 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
9986 Removed the versions check from X509 routines when loading extensions:
9987 this allows certain broken certificates that don't set the version
9988 properly to be processed.
9991 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
9992 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
9993 can still be regenerated with "make depend".
9996 *) Spelling mistake in C version of CAST-128.
9997 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
9999 *) Changes to the error generation code. The perl script err-code.pl
10000 now reads in the old error codes and retains the old numbers, only
10001 adding new ones if necessary. It also only changes the .err files if new
10002 codes are added. The makefiles have been modified to only insert errors
10003 when needed (to avoid needlessly modifying header files). This is done
10004 by only inserting errors if the .err file is newer than the auto generated
10005 C file. To rebuild all the error codes from scratch (the old behaviour)
10006 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
10007 or delete all the .err files.
10010 *) CAST-128 was incorrectly implemented for short keys. The C version has
10011 been fixed, but is untested. The assembler versions are also fixed, but
10012 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
10013 to regenerate it if needed.
10014 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
10015 Hagino <itojun@kame.net>]
10017 *) File was opened incorrectly in randfile.c.
10018 [Ulf Möller <ulf@fitug.de>]
10020 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
10021 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
10022 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
10023 al: it's just almost always a UTCTime. Note this patch adds new error
10024 codes so do a "make errors" if there are problems.
10027 *) Correct Linux 1 recognition in config.
10028 [Ulf Möller <ulf@fitug.de>]
10030 *) Remove pointless MD5 hash when using DSA keys in ca.
10031 [Anonymous <nobody@replay.com>]
10033 *) Generate an error if given an empty string as a cert directory. Also
10034 generate an error if handed NULL (previously returned 0 to indicate an
10035 error, but didn't set one).
10036 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
10038 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
10041 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
10042 parameters. This was causing a warning which killed off the Win32 compile.
10045 *) Remove C++ style comments from crypto/bn/bn_local.h.
10046 [Neil Costigan <neil.costigan@celocom.com>]
10048 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
10049 based on a text string, looking up short and long names and finally
10050 "dot" format. The "dot" format stuff didn't work. Added new function
10051 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
10052 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
10053 OID is not part of the table.
10056 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
10057 X509_LOOKUP_by_alias().
10060 *) Sort openssl functions by name.
10063 *) Get the gendsa program working (hopefully) and add it to app list. Remove
10064 encryption from sample DSA keys (in case anyone is interested the password
10068 *) Make _all_ *_free functions accept a NULL pointer.
10069 [Frans Heymans <fheymans@isaserver.be>]
10071 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
10073 [Anonymous <nobody@replay.com>]
10075 *) s_server should send the CAfile as acceptable CAs, not its own cert.
10076 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10078 *) Don't blow it for numeric -newkey arguments to apps/req.
10079 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10081 *) Temp key "for export" tests were wrong in s3_srvr.c.
10082 [Anonymous <nobody@replay.com>]
10084 *) Add prototype for temp key callback functions
10085 SSL_CTX_set_tmp_{rsa,dh}_callback().
10088 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
10089 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
10092 *) X509_name_add_entry() freed the wrong thing after an error.
10093 [Arne Ansper <arne@ats.cyber.ee>]
10095 *) rsa_eay.c would attempt to free a NULL context.
10096 [Arne Ansper <arne@ats.cyber.ee>]
10098 *) BIO_s_socket() had a broken should_retry() on Windoze.
10099 [Arne Ansper <arne@ats.cyber.ee>]
10101 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
10102 [Arne Ansper <arne@ats.cyber.ee>]
10104 *) Make sure the already existing X509_STORE->depth variable is initialized
10105 in X509_STORE_new(), but document the fact that this variable is still
10106 unused in the certificate verification process.
10107 [Ralf S. Engelschall]
10109 *) Fix the various library and apps files to free up pkeys obtained from
10110 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
10113 *) Fix reference counting in X509_PUBKEY_get(). This makes
10114 demos/maurice/example2.c work, amongst others, probably.
10115 [Steve Henson and Ben Laurie]
10117 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
10118 `openssl' and second, the shortcut symlinks for the `openssl <command>'
10119 are no longer created. This way we have a single and consistent command
10120 line interface `openssl <command>', similar to `cvs <command>'.
10121 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
10123 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
10124 BIT STRING wrapper always have zero unused bits.
10127 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
10130 *) Make the top-level INSTALL documentation easier to understand.
10133 *) Makefiles updated to exit if an error occurs in a sub-directory
10134 make (including if user presses ^C) [Paul Sutton]
10136 *) Make Montgomery context stuff explicit in RSA data structure.
10139 *) Fix build order of pem and err to allow for generated pem.h.
10142 *) Fix renumbering bug in X509_NAME_delete_entry().
10145 *) Enhanced the err-ins.pl script so it makes the error library number
10146 global and can add a library name. This is needed for external ASN1 and
10147 other error libraries.
10150 *) Fixed sk_insert which never worked properly.
10153 *) Fix ASN1 macros so they can handle indefinite length construted
10154 EXPLICIT tags. Some non standard certificates use these: they can now
10158 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
10159 into a single doc/ssleay.txt bundle. This way the information is still
10160 preserved but no longer messes up this directory. Now it's new room for
10161 the new set of documenation files.
10162 [Ralf S. Engelschall]
10164 *) SETs were incorrectly DER encoded. This was a major pain, because they
10165 shared code with SEQUENCEs, which aren't coded the same. This means that
10166 almost everything to do with SETs or SEQUENCEs has either changed name or
10167 number of arguments.
10168 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
10170 *) Fix test data to work with the above.
10173 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
10174 was already fixed by Eric for 0.9.1 it seems.
10175 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
10177 *) Autodetect FreeBSD3.
10180 *) Fix various bugs in Configure. This affects the following platforms:
10184 unixware-2.0-pentium
10188 *) Eliminate generated files from CVS. Reorder tests to regenerate files
10189 before they are needed.
10192 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
10196 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
10198 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
10199 changed SSLeay to OpenSSL in version strings.
10200 [Ralf S. Engelschall]
10202 *) Some fixups to the top-level documents.
10205 *) Fixed the nasty bug where rsaref.h was not found under compile-time
10206 because the symlink to include/ was missing.
10207 [Ralf S. Engelschall]
10209 *) Incorporated the popular no-RSA/DSA-only patches
10210 which allow to compile a RSA-free SSLeay.
10211 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
10213 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
10214 when "ssleay" is still not found.
10215 [Ralf S. Engelschall]
10217 *) Added more platforms to Configure: Cray T3E, HPUX 11,
10218 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
10220 *) Updated the README file.
10221 [Ralf S. Engelschall]
10223 *) Added various .cvsignore files in the CVS repository subdirs
10224 to make a "cvs update" really silent.
10225 [Ralf S. Engelschall]
10227 *) Recompiled the error-definition header files and added
10228 missing symbols to the Win32 linker tables.
10229 [Ralf S. Engelschall]
10231 *) Cleaned up the top-level documents;
10232 o new files: CHANGES and LICENSE
10233 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
10234 o merged COPYRIGHT into LICENSE
10235 o removed obsolete TODO file
10236 o renamed MICROSOFT to INSTALL.W32
10237 [Ralf S. Engelschall]
10239 *) Removed dummy files from the 0.9.1b source tree:
10240 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
10241 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
10242 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
10243 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
10244 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
10245 [Ralf S. Engelschall]
10247 *) Added various platform portability fixes.
10250 *) The Genesis of the OpenSSL rpject:
10251 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
10252 Young and Tim J. Hudson created while they were working for C2Net until
10254 [The OpenSSL Project]
10257 Changes between 0.9.0b and 0.9.1b [not released]
10259 *) Updated a few CA certificates under certs/
10262 *) Changed some BIGNUM api stuff.
10265 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
10266 DGUX x86, Linux Alpha, etc.
10269 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
10270 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
10274 *) Add -strparse option to asn1pars program which parses nested
10276 [Dr Stephen Henson <shenson@bigfoot.com>]
10278 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
10281 *) DSA fix for "ca" program.
10284 *) Added "-genkey" option to "dsaparam" program.
10287 *) Added RIPE MD160 (rmd160) message digest.
10290 *) Added -a (all) option to "ssleay version" command.
10293 *) Added PLATFORM define which is the id given to Configure.
10296 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
10299 *) Extended the ASN.1 parser routines.
10302 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
10305 *) Added a BN_CTX to the BN library.
10308 *) Fixed the weak key values in DES library
10311 *) Changed API in EVP library for cipher aliases.
10314 *) Added support for RC2/64bit cipher.
10317 *) Converted the lhash library to the crypto/mem.c functions.
10320 *) Added more recognized ASN.1 object ids.
10323 *) Added more RSA padding checks for SSL/TLS.
10326 *) Added BIO proxy/filter functionality.
10329 *) Added extra_certs to SSL_CTX which can be used
10330 send extra CA certificates to the client in the CA cert chain sending
10331 process. It can be configured with SSL_CTX_add_extra_chain_cert().
10334 *) Now Fortezza is denied in the authentication phase because
10335 this is key exchange mechanism is not supported by SSLeay at all.
10338 *) Additional PKCS1 checks.
10341 *) Support the string "TLSv1" for all TLS v1 ciphers.
10344 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
10345 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
10348 *) Fixed a few memory leaks.
10351 *) Fixed various code and comment typos.
10354 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
10355 bytes sent in the client random.
10356 [Edward Bishop <ebishop@spyglass.com>]
projects / dragonfly.git / blob